Arch Hydrobiol 158 4 485ndash516 Stuttgart December 2003

Comparative physical and chemical limnology oftwo Canadian High Arctic regions Alert (EllesmereIsland NU) and Mould Bay (Prince Patrick IslandNWT)

Dermot Antoniades1 Marianne S V Douglas1 and John P Smol2

With 6 figures 2 tables and 2 appendices

Abstract The physical and chemical limnological characteristics of 65 lakes andponds from two areas in the Canadian High Arctic were examined to determine differ-ences in regional limnology due to geological and vegetational characteristics as wellas other climate factors Sites in the Alert region of northern Ellesmere Island had rela-tively low concentrations of total phosphorus (median TP = 91 mg lndash1) and total N(median = 0465 mg lndash1) Dissolved organic carbon (DOC) concentrations were rela-tively low (median = 27 mg lndash1) reflecting the sparsity of vegetation in the regionWithin the Alert dataset there were pronounced differences in water chemistry be-tween small tundra ponds and larger deeper lakes The first axis of a principal com-ponents analysis of the Alert dataset reflected conductivity and nutrient gradients (l =028) while the second axis (l = 020) was related to metal concentrations Mould Baysites on Prince Patrick Island had relatively high concentrations of TP (mean =165 mg lndash1) total N (mean = 0616 mg lndash1) and DOC (mean = 67mg lndash1) Mean total Nand DOC were at the highest levels yet measured from any similar high arctic limno-logical survey while mean TP was the second highest high arctic value yet recorded inour surveys A principal components analysis of the Mould Bay data indicated that thetwo dominant gradients in the dataset were conductivity and related variables (l =030) and nutrients (l = 019) The differences in water chemistry variables betweenMould Bay and all previous high arctic surveys is attributable to the relatively densevegetation and deep soils present at Mould Bay relative to Alert and other high arcticregions

Key words climate factors polar freshwater systems high arctic limnology

1 Authorsrsquo addresses Department of Geology University of Toronto 22 Russell StToronto ON M5S 3B1 Canada2 PEARL Department of Biology Queenrsquos University Kingston ON K7L 3N6Canada Corresponding author

DOI 1011270003-913620030158-0485 0003-9136030158-0485 $ 800atilde 2003 E Schweizerbartrsquosche Verlagsbuchhandlung D-70176 Stuttgart

486 Dermot Antoniades Marianne S V Douglas and John P Smol

Introduction

The Canadian High Arctic is a region that has historically been underrepre-sented in limnological research (Schindler 2001) This can be attributed inlarge part to the logistical problems involved in sampling such remote loca-tions Until recently detailed limnological studies in the North American Arc-tic have been restricted to those undertaken in arctic Alaska (see Hobbie 1997)and a small number of lakes from Cornwallis Island (eg Schindler et al1974 Kalff amp Welch 1974) More recently a considerable volume of litera-ture has been developed on the limnology of lower arctic and subarctic regionsof northern Canada and Alaska (eg Gregory-Eaves et al 2000 Ruumlhland ampSmol 1998 Pienitz et al 1997 a b Kling et al 1992) arctic Fennoscandia(eg Korhola et al 2002 Weckstroumlm amp Korhola 2001 Blom et al 2000Weckstroumlm et al 1997) and Siberia (Duff et al 1999) However researchinto the freshwaters of high arctic regions remains sparse

With increasing awareness of the sensitivity of high arctic ecosystems andtheir vulnerability to ongoing climatic and environmental change (eg Rouse etal 1997) surveys have recently been initiated to document the baseline limno-logical conditions from around the High Arctic (eg Antoniades et al 2003Michelutti et al 2002 a b Hamilton et al 2001 Ellis-Evans et al 2001Lim et al 2001 Hamilton et al 2000 Douglas amp Smol 1994) as well as intoother aspects of high arctic limnology (eg Van Donk et al 2001 Villeneuveet al 2001 Markager et al 1999 Quesada et al 1999 Vezina amp Vincent1997) Nonetheless considerable additional study is required to understand thedynamics of these polar freshwater systems as the High Arctic represents a vastand varied region with many ecological climatic and geological gradients

Shallow tundra ponds dominate typical high arctic landscapes Theseponds (defined as being lt 2 m deep and freezing completely in winter) differmarkedly from lakes in their limnological characteristics Due to their smallvolumes tundra ponds have low capacity for the dilution of solutes and theirchemical limnological characteristics are thus highly sensitive to any changein external inputs (Douglas amp Smol 1999) Seasonal variations in arcticphysical and chemical limnology are largely dependent on the extent and dura-tion of ice within limnic systems which controls interactions with the atmos-phere and inputs of solar radiation (Smol 1988) High arctic lakes are gener-ally completely frozen for over nine months of the year and in some regionslake ice is perennial Ponds by comparison thaw and freeze earlier than lakesdue to their lower thermal capacities and are completely ice free for severalmonths each year As a result while water temperatures in high arctic lakesgenerally remain low (ie below 5 ƒC Schindler et al 1974) summer watertemperatures in high arctic ponds may even exceed ambient air temperature(Douglas amp Smol 1994)

Limnology of two Canadian High Arctic regions 487

Alert and Mould Bay are amongst the more studied regions in the Cana-dian High Arctic as logistical difficulties are eased by the presence of weatherstations However despite the volume of scientific investigation in each arealittle or no research has been undertaken on limnological questions The onlypublished study focusing on freshwaters near Alert was primarily concernedwith paleolimnology (Doubleday et al 1995) and no studies exist to date onthe limnology or paleolimnology of the Mould Bay region

Due to poor soil development and the inhospitable climate highly vege-tated regions are exceedingly rare in the High Arctic In the Mould Bay regionhowever the well-developed vegetation is strikingly different from the sparsevegetative cover found across most of the Canadian High Arctic Increases invegetation are expected to accompany predicted temperature and precipitationincreases in the High Arctic during the next century (Maxwell 1997) Assuch the effect of vegetation on the limnology of the Mould Bay region mayprovide insights into future conditions in other less vegetated high arctic re-gions

Consequently there were two primary objectives of this study Firstly toestablish the baseline limnological conditions in these previously uninvesti-gated regions and secondly to evaluate the effects of greater vegetation andsoil depth on the limnology of high arctic lakes and ponds by contrasting theregional water chemistry variability of our two study regions

This study forms part of a larger project by members of our research groupto document the limnological conditions of the Canadian High Arctic islands(eg Antoniades et al 2003 Michelutti et al 2002 a b Lim et al 2001Douglas amp Smol 1994) and will serve to augment our understanding of thefreshwaters of this little studied region In addition these data will be used indetermining diatom autecological characteristics that will form the basis forfuture paleoenvironmental reconstructions from these regions

Materials and methods

Site descript ion

Alert

Alert (80ƒ 30cent N 60ƒ 20cent W) the northernmost human settlement on the planet consistsof a Canadian military installation and a Meteorological Service of Canada (MSC) cli-mate station Alert is situated at an elevation of 62 m asl on the northeastern tip of El-lesmere Island in the Queen Elizabeth Islands (Fig 1) Local topography consists oflow rugged mountains that reach a maximum of 550 m asl The bedrock of the Alertarea is composed of Ordovician to Silurian sequences of carbonates and mudstones ofthe Hazen and Danish River formations (Christie 1964)

488 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 1 Location map of Alert and Mould Bay in the Canadian Arctic Archipelago

Alert has a cold dry climate with an annual mean temperature of ndash180 ƒC (Meteo-rological Service of Canada 2002) The growing season is very short averaging 28 de-gree days per year above 5 ƒC Alert is in the polar desert and receives an average an-nual precipitation of 154 mm (Meteorological Service of Canada 2002) Vegetation inthe area is sparse and low herbaceous shrubs and mosses are the only plants capableof surviving the harsh climate lack of moisture and poorly developed soils (Edlundamp Alt 1989) While tundra ponds dominate most regions in the High Arctic they arenot abundant near Alert However equally unusual among high arctic landscapes is theconcentration of large deep lakes (ie 1ndash3 km diam 10ndash 50 m deep) present within a30 km radius of the Alert base (Fig 2)

Mould Bay

Prince Patrick Island is situated along the western fringe of the Queen Elizabeth Is-lands (Fig 1) with its north coast directly in contact with the permanent polar pack ofthe Arctic Ocean The MSC Mould Bay weather station (76ƒ 14centN 119ƒ 20centW) was lo-

Limnology of two Canadian High Arctic regions 489

Fig 2 Map detail of Alert sampling sites

cated along the southern coast of Prince Patrick Island from 1948 until its closure in1995 and an automated station continues to record climate data at the site Mould Bayis situated in a region of low lying topography with rolling hills that reach a maximumelevation of 150 m It is underlain by Devonian clastic sediments of the Melville Islandgroup as well as fine to medium-grained sandstones of the Jurassic to CretaceousAwingak and Isachsen formations (Everett 1968)

Predominant winds on Prince Patrick Island are from the northwest bringing colddry air from the Arctic Ocean Mould Bay has an annual mean daily temperature ofndash175 ƒC and a July mean daily temperature of + 40 ƒC Mould Bay is amongst thedriest areas of the High Arctic with average annual precipitation of 111 mm per yearand a growing season of only 27 days per year (Meteorological Service of Canada2002)

The soils of the Mould Bay region are described as either tundra or polar desertsoils (Everett 1968) Drainage is typically poor with a shallow permafrost depth thatprevents the drainage of surface waters The soils are slightly acidic and extremely lowin organic matter particularly below a depth of 1ndash2 cm the typical depth of the upperorganic horizon (Everett 1968) Vegetation is extremely dense by high arctic stand-ards with a large percentage of lowland areas densely populated with mosses andgrasses Upland areas in the region are only sparsely vegetated due to lack of soil de-velopment caused by wind erosion and to the direct effects of wind (Everett 1968)

Sampling methods

Water samples were collected at 35 sites in the Mould Bay region between July 12 and21 1999 (Fig 3) and at 30 sites in the Alert area between July 24 and August 7 2000(Fig 2) Individual sites at Mould Bay were chosen in an effort to capture the max-imum range of physical and limnological variation present in the region At Alert vir-tually every accessible site within a 25 km radius of the base was included in the sam-ple set The largest site in the dataset Upper Dumbell Lake was sampled twice on the

490 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 3 Map detail of Mould Bay sampling sites

same day at locations approximately 15 km apart to investigate the homogeneity of thewater chemistry Two Alert sites (ie Lower Dumbell L and Kirk L) contained land-locked char while all other sites in the study are fishless

Sampling methods followed those of our other high arctic studies (eg Antonia-des et al 2003 Michelutti et al 2002 a b Lim et al 2001 Douglas amp Smol1994) and are summarized here For each site water temperatures were measuredusing a hand held thermometer held at approximately 03 m depth Conductivity(COND) and pH were measured in a field laboratory within hours of sampling using aYSI Model 33 conductivity meter and a handheld Hanna pH meter respectively Lati-tude and longitude were measured in the field with a handheld GPS receiver whileelevation (ELEV) was estimated from 1 50000 NTS topographic map sheets 89 B489 B5 (Mould Bay) and MCE 140 (Alert)

Water samples were collected for 37 water chemistry variables which were lateranalyzed at the National Water Research Institute (NWRI) in Burlington Ontario ac-

Limnology of two Canadian High Arctic regions 491

cording to standard protocols (Environment Canada 1994) Measured concentrationswere determined for silica (SiO2) metals (silver (Ag) aluminum (Al) barium (Ba)beryllium (Be) cadmium (Cd) cobalt (Co) chromium (Cr) copper (Cu) iron (Fe)lithium (Li) manganese (Mn) molybdenum (Mo) nickel (Ni) lead (Pb) strontium(Sr) vanadium (V) and zinc (Zn)) and major ions calcium (Ca) chloride (Cl) mag-nesium (Mg) potassium (K) sodium (Na) sulfate (SO4) and dissolved inorganic car-bon (DIC) Measurements of nutrients were made using water that was filtered through045 mm polytetrafluoroethylene (PTFE) filters and comprised total dissolved phos-phorus (TPF) soluble reactive phosphorus (SRP) total Kjeldahl nitrogen (TKN) am-monia (NH3) nitrite (NO2) nitrate-nitrite (NO3 +NO2) and dissolved organic carbon(DOC) Total phosphorus (TPU) was measured from unfiltered water For analysis ofparticulate organic carbon (POC) and nitrogen (PON) 150 ml of water was filteredthrough Whatman glass microfiber filters pre-ignited at 500 ƒC while 300 ml of waterwas filtered for chlorophyll-a (CHLA) analysis All filters were frozen immediatelyand stored in the dark between sampling and analysis CHLA was extracted by addi-tion of 90 acetone and the extract analyzed with a Beckman DU-62 spectrophoto-meter POC and PON were measured with a Perkin-Elmer 2400 CHN Elemental Ana-lyzer Due to a change in NWRI laboratory protocols Mould Bay samples were fil-tered using Whatman GFF filters (particle retention 07mm) while Alert samples usedWhatman GFC filters (particle retention = 12 mm) As such CHLA values are not di-rectly comparable between Alert and Mould Bay sites Total nitrogen (TN) was calcu-lated as the sum of TKN NO3 + NO2 and particulate organic nitrogen (PON) (see Ap-pendix 1)

Stat istical analyses

The datasets were screened prior to statistical analysis so that all variables that werebelow detection limits in at least half of the sites were removed (ie Ag Be Cd CoCr Mo Ni NO2 NO3+ NO2 Pb and V) from both datasets For statistical analysessingle measurements below detection limits were replaced with a value of half the de-tection limit At seven sites with no available CHLA measurement (ie MB-G MB-JEMB-O MB-X MB-Y MB-AD) the median CHLA value was substituted for statisticalanalysis Two sites in the Alert dataset (ie A-A A-B) had extremely high concentra-tions of most chemical limnological variables often two orders of magnitude greaterthan all other measured values Because chemical concentrations were generally verylow for most variables these values dramatically affected the means of many varia-bles Thus Alert median values are used in the discussion as they better typify sites inthe dataset

Covariation between limnological variables in the dataset was identified usingPearson correlation matrices with Bonferroni adjustment (Tables 1 and 2) The primarygradients controlling water chemistry were investigated by redundancy analysis (RDA)and standardized principal components analysis (PCA) using CANOCO version 4 (terBraak amp Šmilauer 1998) A preliminary PCA was used to identify outliers definedas any sites that exceeded the 95 confidence limit of the mean sample scores in thedataset (ie A-A A-B MB-O and MB-AF) (Birks et al 1990) However even when

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

486 Dermot Antoniades Marianne S V Douglas and John P Smol

Introduction

The Canadian High Arctic is a region that has historically been underrepre-sented in limnological research (Schindler 2001) This can be attributed inlarge part to the logistical problems involved in sampling such remote loca-tions Until recently detailed limnological studies in the North American Arc-tic have been restricted to those undertaken in arctic Alaska (see Hobbie 1997)and a small number of lakes from Cornwallis Island (eg Schindler et al1974 Kalff amp Welch 1974) More recently a considerable volume of litera-ture has been developed on the limnology of lower arctic and subarctic regionsof northern Canada and Alaska (eg Gregory-Eaves et al 2000 Ruumlhland ampSmol 1998 Pienitz et al 1997 a b Kling et al 1992) arctic Fennoscandia(eg Korhola et al 2002 Weckstroumlm amp Korhola 2001 Blom et al 2000Weckstroumlm et al 1997) and Siberia (Duff et al 1999) However researchinto the freshwaters of high arctic regions remains sparse

With increasing awareness of the sensitivity of high arctic ecosystems andtheir vulnerability to ongoing climatic and environmental change (eg Rouse etal 1997) surveys have recently been initiated to document the baseline limno-logical conditions from around the High Arctic (eg Antoniades et al 2003Michelutti et al 2002 a b Hamilton et al 2001 Ellis-Evans et al 2001Lim et al 2001 Hamilton et al 2000 Douglas amp Smol 1994) as well as intoother aspects of high arctic limnology (eg Van Donk et al 2001 Villeneuveet al 2001 Markager et al 1999 Quesada et al 1999 Vezina amp Vincent1997) Nonetheless considerable additional study is required to understand thedynamics of these polar freshwater systems as the High Arctic represents a vastand varied region with many ecological climatic and geological gradients

Shallow tundra ponds dominate typical high arctic landscapes Theseponds (defined as being lt 2 m deep and freezing completely in winter) differmarkedly from lakes in their limnological characteristics Due to their smallvolumes tundra ponds have low capacity for the dilution of solutes and theirchemical limnological characteristics are thus highly sensitive to any changein external inputs (Douglas amp Smol 1999) Seasonal variations in arcticphysical and chemical limnology are largely dependent on the extent and dura-tion of ice within limnic systems which controls interactions with the atmos-phere and inputs of solar radiation (Smol 1988) High arctic lakes are gener-ally completely frozen for over nine months of the year and in some regionslake ice is perennial Ponds by comparison thaw and freeze earlier than lakesdue to their lower thermal capacities and are completely ice free for severalmonths each year As a result while water temperatures in high arctic lakesgenerally remain low (ie below 5 ƒC Schindler et al 1974) summer watertemperatures in high arctic ponds may even exceed ambient air temperature(Douglas amp Smol 1994)

Limnology of two Canadian High Arctic regions 487

Alert and Mould Bay are amongst the more studied regions in the Cana-dian High Arctic as logistical difficulties are eased by the presence of weatherstations However despite the volume of scientific investigation in each arealittle or no research has been undertaken on limnological questions The onlypublished study focusing on freshwaters near Alert was primarily concernedwith paleolimnology (Doubleday et al 1995) and no studies exist to date onthe limnology or paleolimnology of the Mould Bay region

Due to poor soil development and the inhospitable climate highly vege-tated regions are exceedingly rare in the High Arctic In the Mould Bay regionhowever the well-developed vegetation is strikingly different from the sparsevegetative cover found across most of the Canadian High Arctic Increases invegetation are expected to accompany predicted temperature and precipitationincreases in the High Arctic during the next century (Maxwell 1997) Assuch the effect of vegetation on the limnology of the Mould Bay region mayprovide insights into future conditions in other less vegetated high arctic re-gions

Consequently there were two primary objectives of this study Firstly toestablish the baseline limnological conditions in these previously uninvesti-gated regions and secondly to evaluate the effects of greater vegetation andsoil depth on the limnology of high arctic lakes and ponds by contrasting theregional water chemistry variability of our two study regions

This study forms part of a larger project by members of our research groupto document the limnological conditions of the Canadian High Arctic islands(eg Antoniades et al 2003 Michelutti et al 2002 a b Lim et al 2001Douglas amp Smol 1994) and will serve to augment our understanding of thefreshwaters of this little studied region In addition these data will be used indetermining diatom autecological characteristics that will form the basis forfuture paleoenvironmental reconstructions from these regions

Materials and methods

Site descript ion

Alert

Alert (80ƒ 30cent N 60ƒ 20cent W) the northernmost human settlement on the planet consistsof a Canadian military installation and a Meteorological Service of Canada (MSC) cli-mate station Alert is situated at an elevation of 62 m asl on the northeastern tip of El-lesmere Island in the Queen Elizabeth Islands (Fig 1) Local topography consists oflow rugged mountains that reach a maximum of 550 m asl The bedrock of the Alertarea is composed of Ordovician to Silurian sequences of carbonates and mudstones ofthe Hazen and Danish River formations (Christie 1964)

488 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 1 Location map of Alert and Mould Bay in the Canadian Arctic Archipelago

Alert has a cold dry climate with an annual mean temperature of ndash180 ƒC (Meteo-rological Service of Canada 2002) The growing season is very short averaging 28 de-gree days per year above 5 ƒC Alert is in the polar desert and receives an average an-nual precipitation of 154 mm (Meteorological Service of Canada 2002) Vegetation inthe area is sparse and low herbaceous shrubs and mosses are the only plants capableof surviving the harsh climate lack of moisture and poorly developed soils (Edlundamp Alt 1989) While tundra ponds dominate most regions in the High Arctic they arenot abundant near Alert However equally unusual among high arctic landscapes is theconcentration of large deep lakes (ie 1ndash3 km diam 10ndash 50 m deep) present within a30 km radius of the Alert base (Fig 2)

Mould Bay

Prince Patrick Island is situated along the western fringe of the Queen Elizabeth Is-lands (Fig 1) with its north coast directly in contact with the permanent polar pack ofthe Arctic Ocean The MSC Mould Bay weather station (76ƒ 14centN 119ƒ 20centW) was lo-

Limnology of two Canadian High Arctic regions 489

Fig 2 Map detail of Alert sampling sites

cated along the southern coast of Prince Patrick Island from 1948 until its closure in1995 and an automated station continues to record climate data at the site Mould Bayis situated in a region of low lying topography with rolling hills that reach a maximumelevation of 150 m It is underlain by Devonian clastic sediments of the Melville Islandgroup as well as fine to medium-grained sandstones of the Jurassic to CretaceousAwingak and Isachsen formations (Everett 1968)

Predominant winds on Prince Patrick Island are from the northwest bringing colddry air from the Arctic Ocean Mould Bay has an annual mean daily temperature ofndash175 ƒC and a July mean daily temperature of + 40 ƒC Mould Bay is amongst thedriest areas of the High Arctic with average annual precipitation of 111 mm per yearand a growing season of only 27 days per year (Meteorological Service of Canada2002)

The soils of the Mould Bay region are described as either tundra or polar desertsoils (Everett 1968) Drainage is typically poor with a shallow permafrost depth thatprevents the drainage of surface waters The soils are slightly acidic and extremely lowin organic matter particularly below a depth of 1ndash2 cm the typical depth of the upperorganic horizon (Everett 1968) Vegetation is extremely dense by high arctic stand-ards with a large percentage of lowland areas densely populated with mosses andgrasses Upland areas in the region are only sparsely vegetated due to lack of soil de-velopment caused by wind erosion and to the direct effects of wind (Everett 1968)

Sampling methods

Water samples were collected at 35 sites in the Mould Bay region between July 12 and21 1999 (Fig 3) and at 30 sites in the Alert area between July 24 and August 7 2000(Fig 2) Individual sites at Mould Bay were chosen in an effort to capture the max-imum range of physical and limnological variation present in the region At Alert vir-tually every accessible site within a 25 km radius of the base was included in the sam-ple set The largest site in the dataset Upper Dumbell Lake was sampled twice on the

490 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 3 Map detail of Mould Bay sampling sites

same day at locations approximately 15 km apart to investigate the homogeneity of thewater chemistry Two Alert sites (ie Lower Dumbell L and Kirk L) contained land-locked char while all other sites in the study are fishless

Sampling methods followed those of our other high arctic studies (eg Antonia-des et al 2003 Michelutti et al 2002 a b Lim et al 2001 Douglas amp Smol1994) and are summarized here For each site water temperatures were measuredusing a hand held thermometer held at approximately 03 m depth Conductivity(COND) and pH were measured in a field laboratory within hours of sampling using aYSI Model 33 conductivity meter and a handheld Hanna pH meter respectively Lati-tude and longitude were measured in the field with a handheld GPS receiver whileelevation (ELEV) was estimated from 1 50000 NTS topographic map sheets 89 B489 B5 (Mould Bay) and MCE 140 (Alert)

Water samples were collected for 37 water chemistry variables which were lateranalyzed at the National Water Research Institute (NWRI) in Burlington Ontario ac-

Limnology of two Canadian High Arctic regions 491

cording to standard protocols (Environment Canada 1994) Measured concentrationswere determined for silica (SiO2) metals (silver (Ag) aluminum (Al) barium (Ba)beryllium (Be) cadmium (Cd) cobalt (Co) chromium (Cr) copper (Cu) iron (Fe)lithium (Li) manganese (Mn) molybdenum (Mo) nickel (Ni) lead (Pb) strontium(Sr) vanadium (V) and zinc (Zn)) and major ions calcium (Ca) chloride (Cl) mag-nesium (Mg) potassium (K) sodium (Na) sulfate (SO4) and dissolved inorganic car-bon (DIC) Measurements of nutrients were made using water that was filtered through045 mm polytetrafluoroethylene (PTFE) filters and comprised total dissolved phos-phorus (TPF) soluble reactive phosphorus (SRP) total Kjeldahl nitrogen (TKN) am-monia (NH3) nitrite (NO2) nitrate-nitrite (NO3 +NO2) and dissolved organic carbon(DOC) Total phosphorus (TPU) was measured from unfiltered water For analysis ofparticulate organic carbon (POC) and nitrogen (PON) 150 ml of water was filteredthrough Whatman glass microfiber filters pre-ignited at 500 ƒC while 300 ml of waterwas filtered for chlorophyll-a (CHLA) analysis All filters were frozen immediatelyand stored in the dark between sampling and analysis CHLA was extracted by addi-tion of 90 acetone and the extract analyzed with a Beckman DU-62 spectrophoto-meter POC and PON were measured with a Perkin-Elmer 2400 CHN Elemental Ana-lyzer Due to a change in NWRI laboratory protocols Mould Bay samples were fil-tered using Whatman GFF filters (particle retention 07mm) while Alert samples usedWhatman GFC filters (particle retention = 12 mm) As such CHLA values are not di-rectly comparable between Alert and Mould Bay sites Total nitrogen (TN) was calcu-lated as the sum of TKN NO3 + NO2 and particulate organic nitrogen (PON) (see Ap-pendix 1)

Stat istical analyses

The datasets were screened prior to statistical analysis so that all variables that werebelow detection limits in at least half of the sites were removed (ie Ag Be Cd CoCr Mo Ni NO2 NO3+ NO2 Pb and V) from both datasets For statistical analysessingle measurements below detection limits were replaced with a value of half the de-tection limit At seven sites with no available CHLA measurement (ie MB-G MB-JEMB-O MB-X MB-Y MB-AD) the median CHLA value was substituted for statisticalanalysis Two sites in the Alert dataset (ie A-A A-B) had extremely high concentra-tions of most chemical limnological variables often two orders of magnitude greaterthan all other measured values Because chemical concentrations were generally verylow for most variables these values dramatically affected the means of many varia-bles Thus Alert median values are used in the discussion as they better typify sites inthe dataset

Covariation between limnological variables in the dataset was identified usingPearson correlation matrices with Bonferroni adjustment (Tables 1 and 2) The primarygradients controlling water chemistry were investigated by redundancy analysis (RDA)and standardized principal components analysis (PCA) using CANOCO version 4 (terBraak amp Šmilauer 1998) A preliminary PCA was used to identify outliers definedas any sites that exceeded the 95 confidence limit of the mean sample scores in thedataset (ie A-A A-B MB-O and MB-AF) (Birks et al 1990) However even when

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 487

Alert and Mould Bay are amongst the more studied regions in the Cana-dian High Arctic as logistical difficulties are eased by the presence of weatherstations However despite the volume of scientific investigation in each arealittle or no research has been undertaken on limnological questions The onlypublished study focusing on freshwaters near Alert was primarily concernedwith paleolimnology (Doubleday et al 1995) and no studies exist to date onthe limnology or paleolimnology of the Mould Bay region

Due to poor soil development and the inhospitable climate highly vege-tated regions are exceedingly rare in the High Arctic In the Mould Bay regionhowever the well-developed vegetation is strikingly different from the sparsevegetative cover found across most of the Canadian High Arctic Increases invegetation are expected to accompany predicted temperature and precipitationincreases in the High Arctic during the next century (Maxwell 1997) Assuch the effect of vegetation on the limnology of the Mould Bay region mayprovide insights into future conditions in other less vegetated high arctic re-gions

Consequently there were two primary objectives of this study Firstly toestablish the baseline limnological conditions in these previously uninvesti-gated regions and secondly to evaluate the effects of greater vegetation andsoil depth on the limnology of high arctic lakes and ponds by contrasting theregional water chemistry variability of our two study regions

This study forms part of a larger project by members of our research groupto document the limnological conditions of the Canadian High Arctic islands(eg Antoniades et al 2003 Michelutti et al 2002 a b Lim et al 2001Douglas amp Smol 1994) and will serve to augment our understanding of thefreshwaters of this little studied region In addition these data will be used indetermining diatom autecological characteristics that will form the basis forfuture paleoenvironmental reconstructions from these regions

Materials and methods

Site descript ion

Alert

Alert (80ƒ 30cent N 60ƒ 20cent W) the northernmost human settlement on the planet consistsof a Canadian military installation and a Meteorological Service of Canada (MSC) cli-mate station Alert is situated at an elevation of 62 m asl on the northeastern tip of El-lesmere Island in the Queen Elizabeth Islands (Fig 1) Local topography consists oflow rugged mountains that reach a maximum of 550 m asl The bedrock of the Alertarea is composed of Ordovician to Silurian sequences of carbonates and mudstones ofthe Hazen and Danish River formations (Christie 1964)

488 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 1 Location map of Alert and Mould Bay in the Canadian Arctic Archipelago

Alert has a cold dry climate with an annual mean temperature of ndash180 ƒC (Meteo-rological Service of Canada 2002) The growing season is very short averaging 28 de-gree days per year above 5 ƒC Alert is in the polar desert and receives an average an-nual precipitation of 154 mm (Meteorological Service of Canada 2002) Vegetation inthe area is sparse and low herbaceous shrubs and mosses are the only plants capableof surviving the harsh climate lack of moisture and poorly developed soils (Edlundamp Alt 1989) While tundra ponds dominate most regions in the High Arctic they arenot abundant near Alert However equally unusual among high arctic landscapes is theconcentration of large deep lakes (ie 1ndash3 km diam 10ndash 50 m deep) present within a30 km radius of the Alert base (Fig 2)

Mould Bay

Prince Patrick Island is situated along the western fringe of the Queen Elizabeth Is-lands (Fig 1) with its north coast directly in contact with the permanent polar pack ofthe Arctic Ocean The MSC Mould Bay weather station (76ƒ 14centN 119ƒ 20centW) was lo-

Limnology of two Canadian High Arctic regions 489

Fig 2 Map detail of Alert sampling sites

cated along the southern coast of Prince Patrick Island from 1948 until its closure in1995 and an automated station continues to record climate data at the site Mould Bayis situated in a region of low lying topography with rolling hills that reach a maximumelevation of 150 m It is underlain by Devonian clastic sediments of the Melville Islandgroup as well as fine to medium-grained sandstones of the Jurassic to CretaceousAwingak and Isachsen formations (Everett 1968)

Predominant winds on Prince Patrick Island are from the northwest bringing colddry air from the Arctic Ocean Mould Bay has an annual mean daily temperature ofndash175 ƒC and a July mean daily temperature of + 40 ƒC Mould Bay is amongst thedriest areas of the High Arctic with average annual precipitation of 111 mm per yearand a growing season of only 27 days per year (Meteorological Service of Canada2002)

The soils of the Mould Bay region are described as either tundra or polar desertsoils (Everett 1968) Drainage is typically poor with a shallow permafrost depth thatprevents the drainage of surface waters The soils are slightly acidic and extremely lowin organic matter particularly below a depth of 1ndash2 cm the typical depth of the upperorganic horizon (Everett 1968) Vegetation is extremely dense by high arctic stand-ards with a large percentage of lowland areas densely populated with mosses andgrasses Upland areas in the region are only sparsely vegetated due to lack of soil de-velopment caused by wind erosion and to the direct effects of wind (Everett 1968)

Sampling methods

Water samples were collected at 35 sites in the Mould Bay region between July 12 and21 1999 (Fig 3) and at 30 sites in the Alert area between July 24 and August 7 2000(Fig 2) Individual sites at Mould Bay were chosen in an effort to capture the max-imum range of physical and limnological variation present in the region At Alert vir-tually every accessible site within a 25 km radius of the base was included in the sam-ple set The largest site in the dataset Upper Dumbell Lake was sampled twice on the

490 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 3 Map detail of Mould Bay sampling sites

same day at locations approximately 15 km apart to investigate the homogeneity of thewater chemistry Two Alert sites (ie Lower Dumbell L and Kirk L) contained land-locked char while all other sites in the study are fishless

Sampling methods followed those of our other high arctic studies (eg Antonia-des et al 2003 Michelutti et al 2002 a b Lim et al 2001 Douglas amp Smol1994) and are summarized here For each site water temperatures were measuredusing a hand held thermometer held at approximately 03 m depth Conductivity(COND) and pH were measured in a field laboratory within hours of sampling using aYSI Model 33 conductivity meter and a handheld Hanna pH meter respectively Lati-tude and longitude were measured in the field with a handheld GPS receiver whileelevation (ELEV) was estimated from 1 50000 NTS topographic map sheets 89 B489 B5 (Mould Bay) and MCE 140 (Alert)

Water samples were collected for 37 water chemistry variables which were lateranalyzed at the National Water Research Institute (NWRI) in Burlington Ontario ac-

Limnology of two Canadian High Arctic regions 491

cording to standard protocols (Environment Canada 1994) Measured concentrationswere determined for silica (SiO2) metals (silver (Ag) aluminum (Al) barium (Ba)beryllium (Be) cadmium (Cd) cobalt (Co) chromium (Cr) copper (Cu) iron (Fe)lithium (Li) manganese (Mn) molybdenum (Mo) nickel (Ni) lead (Pb) strontium(Sr) vanadium (V) and zinc (Zn)) and major ions calcium (Ca) chloride (Cl) mag-nesium (Mg) potassium (K) sodium (Na) sulfate (SO4) and dissolved inorganic car-bon (DIC) Measurements of nutrients were made using water that was filtered through045 mm polytetrafluoroethylene (PTFE) filters and comprised total dissolved phos-phorus (TPF) soluble reactive phosphorus (SRP) total Kjeldahl nitrogen (TKN) am-monia (NH3) nitrite (NO2) nitrate-nitrite (NO3 +NO2) and dissolved organic carbon(DOC) Total phosphorus (TPU) was measured from unfiltered water For analysis ofparticulate organic carbon (POC) and nitrogen (PON) 150 ml of water was filteredthrough Whatman glass microfiber filters pre-ignited at 500 ƒC while 300 ml of waterwas filtered for chlorophyll-a (CHLA) analysis All filters were frozen immediatelyand stored in the dark between sampling and analysis CHLA was extracted by addi-tion of 90 acetone and the extract analyzed with a Beckman DU-62 spectrophoto-meter POC and PON were measured with a Perkin-Elmer 2400 CHN Elemental Ana-lyzer Due to a change in NWRI laboratory protocols Mould Bay samples were fil-tered using Whatman GFF filters (particle retention 07mm) while Alert samples usedWhatman GFC filters (particle retention = 12 mm) As such CHLA values are not di-rectly comparable between Alert and Mould Bay sites Total nitrogen (TN) was calcu-lated as the sum of TKN NO3 + NO2 and particulate organic nitrogen (PON) (see Ap-pendix 1)

Stat istical analyses

The datasets were screened prior to statistical analysis so that all variables that werebelow detection limits in at least half of the sites were removed (ie Ag Be Cd CoCr Mo Ni NO2 NO3+ NO2 Pb and V) from both datasets For statistical analysessingle measurements below detection limits were replaced with a value of half the de-tection limit At seven sites with no available CHLA measurement (ie MB-G MB-JEMB-O MB-X MB-Y MB-AD) the median CHLA value was substituted for statisticalanalysis Two sites in the Alert dataset (ie A-A A-B) had extremely high concentra-tions of most chemical limnological variables often two orders of magnitude greaterthan all other measured values Because chemical concentrations were generally verylow for most variables these values dramatically affected the means of many varia-bles Thus Alert median values are used in the discussion as they better typify sites inthe dataset

Covariation between limnological variables in the dataset was identified usingPearson correlation matrices with Bonferroni adjustment (Tables 1 and 2) The primarygradients controlling water chemistry were investigated by redundancy analysis (RDA)and standardized principal components analysis (PCA) using CANOCO version 4 (terBraak amp Šmilauer 1998) A preliminary PCA was used to identify outliers definedas any sites that exceeded the 95 confidence limit of the mean sample scores in thedataset (ie A-A A-B MB-O and MB-AF) (Birks et al 1990) However even when

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

488 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 1 Location map of Alert and Mould Bay in the Canadian Arctic Archipelago

Alert has a cold dry climate with an annual mean temperature of ndash180 ƒC (Meteo-rological Service of Canada 2002) The growing season is very short averaging 28 de-gree days per year above 5 ƒC Alert is in the polar desert and receives an average an-nual precipitation of 154 mm (Meteorological Service of Canada 2002) Vegetation inthe area is sparse and low herbaceous shrubs and mosses are the only plants capableof surviving the harsh climate lack of moisture and poorly developed soils (Edlundamp Alt 1989) While tundra ponds dominate most regions in the High Arctic they arenot abundant near Alert However equally unusual among high arctic landscapes is theconcentration of large deep lakes (ie 1ndash3 km diam 10ndash 50 m deep) present within a30 km radius of the Alert base (Fig 2)

Mould Bay

Prince Patrick Island is situated along the western fringe of the Queen Elizabeth Is-lands (Fig 1) with its north coast directly in contact with the permanent polar pack ofthe Arctic Ocean The MSC Mould Bay weather station (76ƒ 14centN 119ƒ 20centW) was lo-

Limnology of two Canadian High Arctic regions 489

Fig 2 Map detail of Alert sampling sites

cated along the southern coast of Prince Patrick Island from 1948 until its closure in1995 and an automated station continues to record climate data at the site Mould Bayis situated in a region of low lying topography with rolling hills that reach a maximumelevation of 150 m It is underlain by Devonian clastic sediments of the Melville Islandgroup as well as fine to medium-grained sandstones of the Jurassic to CretaceousAwingak and Isachsen formations (Everett 1968)

Predominant winds on Prince Patrick Island are from the northwest bringing colddry air from the Arctic Ocean Mould Bay has an annual mean daily temperature ofndash175 ƒC and a July mean daily temperature of + 40 ƒC Mould Bay is amongst thedriest areas of the High Arctic with average annual precipitation of 111 mm per yearand a growing season of only 27 days per year (Meteorological Service of Canada2002)

The soils of the Mould Bay region are described as either tundra or polar desertsoils (Everett 1968) Drainage is typically poor with a shallow permafrost depth thatprevents the drainage of surface waters The soils are slightly acidic and extremely lowin organic matter particularly below a depth of 1ndash2 cm the typical depth of the upperorganic horizon (Everett 1968) Vegetation is extremely dense by high arctic stand-ards with a large percentage of lowland areas densely populated with mosses andgrasses Upland areas in the region are only sparsely vegetated due to lack of soil de-velopment caused by wind erosion and to the direct effects of wind (Everett 1968)

Sampling methods

Water samples were collected at 35 sites in the Mould Bay region between July 12 and21 1999 (Fig 3) and at 30 sites in the Alert area between July 24 and August 7 2000(Fig 2) Individual sites at Mould Bay were chosen in an effort to capture the max-imum range of physical and limnological variation present in the region At Alert vir-tually every accessible site within a 25 km radius of the base was included in the sam-ple set The largest site in the dataset Upper Dumbell Lake was sampled twice on the

490 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 3 Map detail of Mould Bay sampling sites

same day at locations approximately 15 km apart to investigate the homogeneity of thewater chemistry Two Alert sites (ie Lower Dumbell L and Kirk L) contained land-locked char while all other sites in the study are fishless

Sampling methods followed those of our other high arctic studies (eg Antonia-des et al 2003 Michelutti et al 2002 a b Lim et al 2001 Douglas amp Smol1994) and are summarized here For each site water temperatures were measuredusing a hand held thermometer held at approximately 03 m depth Conductivity(COND) and pH were measured in a field laboratory within hours of sampling using aYSI Model 33 conductivity meter and a handheld Hanna pH meter respectively Lati-tude and longitude were measured in the field with a handheld GPS receiver whileelevation (ELEV) was estimated from 1 50000 NTS topographic map sheets 89 B489 B5 (Mould Bay) and MCE 140 (Alert)

Water samples were collected for 37 water chemistry variables which were lateranalyzed at the National Water Research Institute (NWRI) in Burlington Ontario ac-

Limnology of two Canadian High Arctic regions 491

cording to standard protocols (Environment Canada 1994) Measured concentrationswere determined for silica (SiO2) metals (silver (Ag) aluminum (Al) barium (Ba)beryllium (Be) cadmium (Cd) cobalt (Co) chromium (Cr) copper (Cu) iron (Fe)lithium (Li) manganese (Mn) molybdenum (Mo) nickel (Ni) lead (Pb) strontium(Sr) vanadium (V) and zinc (Zn)) and major ions calcium (Ca) chloride (Cl) mag-nesium (Mg) potassium (K) sodium (Na) sulfate (SO4) and dissolved inorganic car-bon (DIC) Measurements of nutrients were made using water that was filtered through045 mm polytetrafluoroethylene (PTFE) filters and comprised total dissolved phos-phorus (TPF) soluble reactive phosphorus (SRP) total Kjeldahl nitrogen (TKN) am-monia (NH3) nitrite (NO2) nitrate-nitrite (NO3 +NO2) and dissolved organic carbon(DOC) Total phosphorus (TPU) was measured from unfiltered water For analysis ofparticulate organic carbon (POC) and nitrogen (PON) 150 ml of water was filteredthrough Whatman glass microfiber filters pre-ignited at 500 ƒC while 300 ml of waterwas filtered for chlorophyll-a (CHLA) analysis All filters were frozen immediatelyand stored in the dark between sampling and analysis CHLA was extracted by addi-tion of 90 acetone and the extract analyzed with a Beckman DU-62 spectrophoto-meter POC and PON were measured with a Perkin-Elmer 2400 CHN Elemental Ana-lyzer Due to a change in NWRI laboratory protocols Mould Bay samples were fil-tered using Whatman GFF filters (particle retention 07mm) while Alert samples usedWhatman GFC filters (particle retention = 12 mm) As such CHLA values are not di-rectly comparable between Alert and Mould Bay sites Total nitrogen (TN) was calcu-lated as the sum of TKN NO3 + NO2 and particulate organic nitrogen (PON) (see Ap-pendix 1)

Stat istical analyses

The datasets were screened prior to statistical analysis so that all variables that werebelow detection limits in at least half of the sites were removed (ie Ag Be Cd CoCr Mo Ni NO2 NO3+ NO2 Pb and V) from both datasets For statistical analysessingle measurements below detection limits were replaced with a value of half the de-tection limit At seven sites with no available CHLA measurement (ie MB-G MB-JEMB-O MB-X MB-Y MB-AD) the median CHLA value was substituted for statisticalanalysis Two sites in the Alert dataset (ie A-A A-B) had extremely high concentra-tions of most chemical limnological variables often two orders of magnitude greaterthan all other measured values Because chemical concentrations were generally verylow for most variables these values dramatically affected the means of many varia-bles Thus Alert median values are used in the discussion as they better typify sites inthe dataset

Covariation between limnological variables in the dataset was identified usingPearson correlation matrices with Bonferroni adjustment (Tables 1 and 2) The primarygradients controlling water chemistry were investigated by redundancy analysis (RDA)and standardized principal components analysis (PCA) using CANOCO version 4 (terBraak amp Šmilauer 1998) A preliminary PCA was used to identify outliers definedas any sites that exceeded the 95 confidence limit of the mean sample scores in thedataset (ie A-A A-B MB-O and MB-AF) (Birks et al 1990) However even when

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 489

Fig 2 Map detail of Alert sampling sites

cated along the southern coast of Prince Patrick Island from 1948 until its closure in1995 and an automated station continues to record climate data at the site Mould Bayis situated in a region of low lying topography with rolling hills that reach a maximumelevation of 150 m It is underlain by Devonian clastic sediments of the Melville Islandgroup as well as fine to medium-grained sandstones of the Jurassic to CretaceousAwingak and Isachsen formations (Everett 1968)

Predominant winds on Prince Patrick Island are from the northwest bringing colddry air from the Arctic Ocean Mould Bay has an annual mean daily temperature ofndash175 ƒC and a July mean daily temperature of + 40 ƒC Mould Bay is amongst thedriest areas of the High Arctic with average annual precipitation of 111 mm per yearand a growing season of only 27 days per year (Meteorological Service of Canada2002)

The soils of the Mould Bay region are described as either tundra or polar desertsoils (Everett 1968) Drainage is typically poor with a shallow permafrost depth thatprevents the drainage of surface waters The soils are slightly acidic and extremely lowin organic matter particularly below a depth of 1ndash2 cm the typical depth of the upperorganic horizon (Everett 1968) Vegetation is extremely dense by high arctic stand-ards with a large percentage of lowland areas densely populated with mosses andgrasses Upland areas in the region are only sparsely vegetated due to lack of soil de-velopment caused by wind erosion and to the direct effects of wind (Everett 1968)

Sampling methods

Water samples were collected at 35 sites in the Mould Bay region between July 12 and21 1999 (Fig 3) and at 30 sites in the Alert area between July 24 and August 7 2000(Fig 2) Individual sites at Mould Bay were chosen in an effort to capture the max-imum range of physical and limnological variation present in the region At Alert vir-tually every accessible site within a 25 km radius of the base was included in the sam-ple set The largest site in the dataset Upper Dumbell Lake was sampled twice on the

490 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 3 Map detail of Mould Bay sampling sites

same day at locations approximately 15 km apart to investigate the homogeneity of thewater chemistry Two Alert sites (ie Lower Dumbell L and Kirk L) contained land-locked char while all other sites in the study are fishless

Sampling methods followed those of our other high arctic studies (eg Antonia-des et al 2003 Michelutti et al 2002 a b Lim et al 2001 Douglas amp Smol1994) and are summarized here For each site water temperatures were measuredusing a hand held thermometer held at approximately 03 m depth Conductivity(COND) and pH were measured in a field laboratory within hours of sampling using aYSI Model 33 conductivity meter and a handheld Hanna pH meter respectively Lati-tude and longitude were measured in the field with a handheld GPS receiver whileelevation (ELEV) was estimated from 1 50000 NTS topographic map sheets 89 B489 B5 (Mould Bay) and MCE 140 (Alert)

Water samples were collected for 37 water chemistry variables which were lateranalyzed at the National Water Research Institute (NWRI) in Burlington Ontario ac-

Limnology of two Canadian High Arctic regions 491

cording to standard protocols (Environment Canada 1994) Measured concentrationswere determined for silica (SiO2) metals (silver (Ag) aluminum (Al) barium (Ba)beryllium (Be) cadmium (Cd) cobalt (Co) chromium (Cr) copper (Cu) iron (Fe)lithium (Li) manganese (Mn) molybdenum (Mo) nickel (Ni) lead (Pb) strontium(Sr) vanadium (V) and zinc (Zn)) and major ions calcium (Ca) chloride (Cl) mag-nesium (Mg) potassium (K) sodium (Na) sulfate (SO4) and dissolved inorganic car-bon (DIC) Measurements of nutrients were made using water that was filtered through045 mm polytetrafluoroethylene (PTFE) filters and comprised total dissolved phos-phorus (TPF) soluble reactive phosphorus (SRP) total Kjeldahl nitrogen (TKN) am-monia (NH3) nitrite (NO2) nitrate-nitrite (NO3 +NO2) and dissolved organic carbon(DOC) Total phosphorus (TPU) was measured from unfiltered water For analysis ofparticulate organic carbon (POC) and nitrogen (PON) 150 ml of water was filteredthrough Whatman glass microfiber filters pre-ignited at 500 ƒC while 300 ml of waterwas filtered for chlorophyll-a (CHLA) analysis All filters were frozen immediatelyand stored in the dark between sampling and analysis CHLA was extracted by addi-tion of 90 acetone and the extract analyzed with a Beckman DU-62 spectrophoto-meter POC and PON were measured with a Perkin-Elmer 2400 CHN Elemental Ana-lyzer Due to a change in NWRI laboratory protocols Mould Bay samples were fil-tered using Whatman GFF filters (particle retention 07mm) while Alert samples usedWhatman GFC filters (particle retention = 12 mm) As such CHLA values are not di-rectly comparable between Alert and Mould Bay sites Total nitrogen (TN) was calcu-lated as the sum of TKN NO3 + NO2 and particulate organic nitrogen (PON) (see Ap-pendix 1)

Stat istical analyses

The datasets were screened prior to statistical analysis so that all variables that werebelow detection limits in at least half of the sites were removed (ie Ag Be Cd CoCr Mo Ni NO2 NO3+ NO2 Pb and V) from both datasets For statistical analysessingle measurements below detection limits were replaced with a value of half the de-tection limit At seven sites with no available CHLA measurement (ie MB-G MB-JEMB-O MB-X MB-Y MB-AD) the median CHLA value was substituted for statisticalanalysis Two sites in the Alert dataset (ie A-A A-B) had extremely high concentra-tions of most chemical limnological variables often two orders of magnitude greaterthan all other measured values Because chemical concentrations were generally verylow for most variables these values dramatically affected the means of many varia-bles Thus Alert median values are used in the discussion as they better typify sites inthe dataset

Covariation between limnological variables in the dataset was identified usingPearson correlation matrices with Bonferroni adjustment (Tables 1 and 2) The primarygradients controlling water chemistry were investigated by redundancy analysis (RDA)and standardized principal components analysis (PCA) using CANOCO version 4 (terBraak amp Šmilauer 1998) A preliminary PCA was used to identify outliers definedas any sites that exceeded the 95 confidence limit of the mean sample scores in thedataset (ie A-A A-B MB-O and MB-AF) (Birks et al 1990) However even when

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

490 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 3 Map detail of Mould Bay sampling sites

same day at locations approximately 15 km apart to investigate the homogeneity of thewater chemistry Two Alert sites (ie Lower Dumbell L and Kirk L) contained land-locked char while all other sites in the study are fishless

Sampling methods followed those of our other high arctic studies (eg Antonia-des et al 2003 Michelutti et al 2002 a b Lim et al 2001 Douglas amp Smol1994) and are summarized here For each site water temperatures were measuredusing a hand held thermometer held at approximately 03 m depth Conductivity(COND) and pH were measured in a field laboratory within hours of sampling using aYSI Model 33 conductivity meter and a handheld Hanna pH meter respectively Lati-tude and longitude were measured in the field with a handheld GPS receiver whileelevation (ELEV) was estimated from 1 50000 NTS topographic map sheets 89 B489 B5 (Mould Bay) and MCE 140 (Alert)

Water samples were collected for 37 water chemistry variables which were lateranalyzed at the National Water Research Institute (NWRI) in Burlington Ontario ac-

Limnology of two Canadian High Arctic regions 491

cording to standard protocols (Environment Canada 1994) Measured concentrationswere determined for silica (SiO2) metals (silver (Ag) aluminum (Al) barium (Ba)beryllium (Be) cadmium (Cd) cobalt (Co) chromium (Cr) copper (Cu) iron (Fe)lithium (Li) manganese (Mn) molybdenum (Mo) nickel (Ni) lead (Pb) strontium(Sr) vanadium (V) and zinc (Zn)) and major ions calcium (Ca) chloride (Cl) mag-nesium (Mg) potassium (K) sodium (Na) sulfate (SO4) and dissolved inorganic car-bon (DIC) Measurements of nutrients were made using water that was filtered through045 mm polytetrafluoroethylene (PTFE) filters and comprised total dissolved phos-phorus (TPF) soluble reactive phosphorus (SRP) total Kjeldahl nitrogen (TKN) am-monia (NH3) nitrite (NO2) nitrate-nitrite (NO3 +NO2) and dissolved organic carbon(DOC) Total phosphorus (TPU) was measured from unfiltered water For analysis ofparticulate organic carbon (POC) and nitrogen (PON) 150 ml of water was filteredthrough Whatman glass microfiber filters pre-ignited at 500 ƒC while 300 ml of waterwas filtered for chlorophyll-a (CHLA) analysis All filters were frozen immediatelyand stored in the dark between sampling and analysis CHLA was extracted by addi-tion of 90 acetone and the extract analyzed with a Beckman DU-62 spectrophoto-meter POC and PON were measured with a Perkin-Elmer 2400 CHN Elemental Ana-lyzer Due to a change in NWRI laboratory protocols Mould Bay samples were fil-tered using Whatman GFF filters (particle retention 07mm) while Alert samples usedWhatman GFC filters (particle retention = 12 mm) As such CHLA values are not di-rectly comparable between Alert and Mould Bay sites Total nitrogen (TN) was calcu-lated as the sum of TKN NO3 + NO2 and particulate organic nitrogen (PON) (see Ap-pendix 1)

Stat istical analyses

The datasets were screened prior to statistical analysis so that all variables that werebelow detection limits in at least half of the sites were removed (ie Ag Be Cd CoCr Mo Ni NO2 NO3+ NO2 Pb and V) from both datasets For statistical analysessingle measurements below detection limits were replaced with a value of half the de-tection limit At seven sites with no available CHLA measurement (ie MB-G MB-JEMB-O MB-X MB-Y MB-AD) the median CHLA value was substituted for statisticalanalysis Two sites in the Alert dataset (ie A-A A-B) had extremely high concentra-tions of most chemical limnological variables often two orders of magnitude greaterthan all other measured values Because chemical concentrations were generally verylow for most variables these values dramatically affected the means of many varia-bles Thus Alert median values are used in the discussion as they better typify sites inthe dataset

Covariation between limnological variables in the dataset was identified usingPearson correlation matrices with Bonferroni adjustment (Tables 1 and 2) The primarygradients controlling water chemistry were investigated by redundancy analysis (RDA)and standardized principal components analysis (PCA) using CANOCO version 4 (terBraak amp Šmilauer 1998) A preliminary PCA was used to identify outliers definedas any sites that exceeded the 95 confidence limit of the mean sample scores in thedataset (ie A-A A-B MB-O and MB-AF) (Birks et al 1990) However even when

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 491

cording to standard protocols (Environment Canada 1994) Measured concentrationswere determined for silica (SiO2) metals (silver (Ag) aluminum (Al) barium (Ba)beryllium (Be) cadmium (Cd) cobalt (Co) chromium (Cr) copper (Cu) iron (Fe)lithium (Li) manganese (Mn) molybdenum (Mo) nickel (Ni) lead (Pb) strontium(Sr) vanadium (V) and zinc (Zn)) and major ions calcium (Ca) chloride (Cl) mag-nesium (Mg) potassium (K) sodium (Na) sulfate (SO4) and dissolved inorganic car-bon (DIC) Measurements of nutrients were made using water that was filtered through045 mm polytetrafluoroethylene (PTFE) filters and comprised total dissolved phos-phorus (TPF) soluble reactive phosphorus (SRP) total Kjeldahl nitrogen (TKN) am-monia (NH3) nitrite (NO2) nitrate-nitrite (NO3 +NO2) and dissolved organic carbon(DOC) Total phosphorus (TPU) was measured from unfiltered water For analysis ofparticulate organic carbon (POC) and nitrogen (PON) 150 ml of water was filteredthrough Whatman glass microfiber filters pre-ignited at 500 ƒC while 300 ml of waterwas filtered for chlorophyll-a (CHLA) analysis All filters were frozen immediatelyand stored in the dark between sampling and analysis CHLA was extracted by addi-tion of 90 acetone and the extract analyzed with a Beckman DU-62 spectrophoto-meter POC and PON were measured with a Perkin-Elmer 2400 CHN Elemental Ana-lyzer Due to a change in NWRI laboratory protocols Mould Bay samples were fil-tered using Whatman GFF filters (particle retention 07mm) while Alert samples usedWhatman GFC filters (particle retention = 12 mm) As such CHLA values are not di-rectly comparable between Alert and Mould Bay sites Total nitrogen (TN) was calcu-lated as the sum of TKN NO3 + NO2 and particulate organic nitrogen (PON) (see Ap-pendix 1)

Stat istical analyses

The datasets were screened prior to statistical analysis so that all variables that werebelow detection limits in at least half of the sites were removed (ie Ag Be Cd CoCr Mo Ni NO2 NO3+ NO2 Pb and V) from both datasets For statistical analysessingle measurements below detection limits were replaced with a value of half the de-tection limit At seven sites with no available CHLA measurement (ie MB-G MB-JEMB-O MB-X MB-Y MB-AD) the median CHLA value was substituted for statisticalanalysis Two sites in the Alert dataset (ie A-A A-B) had extremely high concentra-tions of most chemical limnological variables often two orders of magnitude greaterthan all other measured values Because chemical concentrations were generally verylow for most variables these values dramatically affected the means of many varia-bles Thus Alert median values are used in the discussion as they better typify sites inthe dataset

Covariation between limnological variables in the dataset was identified usingPearson correlation matrices with Bonferroni adjustment (Tables 1 and 2) The primarygradients controlling water chemistry were investigated by redundancy analysis (RDA)and standardized principal components analysis (PCA) using CANOCO version 4 (terBraak amp Šmilauer 1998) A preliminary PCA was used to identify outliers definedas any sites that exceeded the 95 confidence limit of the mean sample scores in thedataset (ie A-A A-B MB-O and MB-AF) (Birks et al 1990) However even when

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

492 Dermot Antoniades Marianne S V Douglas and John P Smol

Tab

le1

Ale

rtPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO4

089

100

SiO

2ndash

017

ndash0

251

00

POC

084

081

ndash0

231

00

PON

088

085

ndash0

210

981

00

DO

Cndash

002

006

060

002

006

100

DIC

020

014

022

019

017

055

100

SRP

081

073

ndash0

050

900

890

120

211

00

NH

3ndash

013

024

006

ndash0

010

020

490

00ndash

001

100

TK

N0

100

260

460

180

230

930

400

220

651

00

Al

074

063

ndash0

290

860

82ndash

021

009

083

ndash0

18ndash

010

100

Ba

093

089

ndash0

250

850

890

050

270

800

000

180

801

00

Cu

007

001

ndash0

190

120

11ndash

016

008

009

ndash0

22ndash

021

029

025

100

Fe0

700

60ndash

030

081

077

ndash0

240

060

79ndash

021

ndash0

140

990

790

411

00

Li

089

079

ndash0

200

880

890

110

330

86ndash

010

020

081

086

008

076

100

Mn

073

064

ndash0

280

800

77ndash

023

007

080

ndash0

20ndash

012

091

081

050

094

074

100

Sr0

840

81ndash

014

077

077

025

061

072

005

031

061

083

003

056

085

057

100

Zn

053

043

ndash0

340

640

59ndash

017

014

064

ndash0

22ndash

014

086

068

050

090

066

082

048

100

Ca

028

028

023

020

020

042

088

024

008

029

012

037

017

012

029

020

058

013

100

Mg

093

092

ndash0

200

850

870

160

370

790

070

290

690

900

030

640

900

640

940

520

361

00

Na

100

088

ndash0

170

850

89ndash

002

019

082

ndash0

140

100

750

920

070

710

900

730

840

540

270

921

00

K0

980

87ndash

023

090

092

ndash0

020

190

86ndash

013

011

081

093

008

077

094

078

084

061

022

093

099

100

TPU

071

064

ndash0

250

850

81ndash

019

007

079

ndash0

16ndash

007

079

067

009

075

075

079

059

060

013

066

072

077

100

TPF

048

050

025

047

053

061

057

039

020

063

017

050

011

013

048

023

066

005

050

058

048

047

029

100

Con

d0

990

90ndash

015

084

088

006

030

081

ndash0

080

170

720

930

060

670

900

700

890

520

380

950

990

970

690

541

00

pHndash

020

ndash0

010

24ndash

008

ndash0

080

52ndash

004

ndash0

080

440

60ndash

025

ndash0

14ndash

015

ndash0

27ndash

014

ndash0

32ndash

009

ndash0

26ndash

025

003

ndash0

20ndash

017

ndash0

200

26ndash

017

100

Ele

vndash

014

ndash0

030

21ndash

018

ndash0

120

16ndash

023

ndash0

020

220

23ndash

010

003

007

ndash0

06ndash

026

000

ndash0

24ndash

014

ndash0

09ndash

019

ndash0

14ndash

016

ndash0

30ndash

002

ndash0

140

241

00

Chl

-a0

390

61ndash

008

042

047

036

015

040

043

050

010

041

ndash0

240

040

410

090

460

000

140

550

390

390

240

440

430

410

111

00

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SR

PN

H3

TK

NA

lB

aC

uFe

Li

Mn

Sr

Zn

Ca

Mg

Na

KT

PU

TPF

Con

dpH

Ele

vC

hl-a

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 493

Tab

le2

Mou

ldB

ayPe

arso

nC

orre

latio

nM

atri

xw

ithB

onfe

rron

iAdj

uste

dPr

obab

ilitie

sB

old

deno

tes

plt

001

Ita

lics

deno

tep

lt0

05

Cl

100

SO

40

231

00

SiO

2ndash

007

005

100

PO

Cndash

004

ndash0

070

491

00

PO

Nndash

005

ndash0

060

511

001

00

DO

C0

020

220

430

150

171

00

DIC

001

054

016

ndash0

07ndash

004

061

100

SR

Pndash

009

013

063

048

049

068

022

100

NH

3ndash

014

ndash0

24ndash

025

ndash0

19ndash

018

035

019

017

100

TK

N0

060

040

220

070

080

910

580

520

561

00

Al

ndash0

07ndash

002

003

067

066

ndash0

26ndash

010

ndash0

02ndash

025

ndash0

211

00

Ba

031

043

008

052

052

019

045

013

ndash0

160

180

481

00

Cu

034

063

ndash0

11ndash

009

ndash0

080

110

420

01ndash

024

002

ndash0

050

451

00

Fe

ndash0

02ndash

006

064

094

094

024

ndash0

070

61ndash

018

011

054

048

ndash0

061

00

Li

042

028

047

033

035

043

040

019

ndash0

260

290

070

610

300

421

00

Mn

045

004

038

069

069

007

ndash0

090

31ndash

018

001

043

046

009

071

042

100

Sr

063

065

004

ndash0

10ndash

009

042

074

006

ndash0

010

41ndash

009

060

060

ndash0

080

580

251

00

Zn

000

ndash0

030

200

710

71ndash

011

ndash0

190

20ndash

021

ndash0

140

790

42ndash

007

064

012

052

ndash0

121

00

Ca

021

059

015

ndash0

12ndash

010

056

090

020

017

054

ndash0

120

470

43ndash

012

043

ndash0

020

83ndash

019

100

Mg

090

044

ndash0

05ndash

003

ndash0

030

170

31ndash

005

ndash0

140

17ndash

007

048

050

000

054

047

080

ndash0

040

391

00

Na

098

036

ndash0

06ndash

005

ndash0

060

070

11ndash

006

ndash0

160

07ndash

007

034

043

ndash0

040

420

400

69ndash

001

029

092

100

K0

740

250

250

200

210

310

400

11ndash

013

031

006

060

042

028

078

059

076

010

043

085

073

100

TP

U0

00ndash

003

040

098

098

011

ndash0

060

40ndash

020

005

075

055

ndash0

070

890

300

70ndash

005

076

ndash0

090

01ndash

001

020

100

TP

F0

040

120

610

550

560

610

120

820

120

470

060

17ndash

003

066

033

041

007

040

005

010

007

025

050

100

Con

d0

950

42ndash

003

ndash0

08ndash

008

019

030

ndash0

02ndash

010

021

ndash0

090

430

46ndash

006

051

039

083

ndash0

060

480

940

970

80ndash

003

006

100

pHndash

006

051

ndash0

12ndash

025

ndash0

250

360

860

010

200

39ndash

010

031

041

ndash0

300

15ndash

025

059

ndash0

350

780

230

040

16ndash

020

ndash0

170

211

00

Ele

vndash

029

ndash0

100

10ndash

018

ndash0

140

140

25ndash

010

007

017

003

000

ndash0

09ndash

014

007

ndash0

220

080

090

25ndash

026

ndash0

30ndash

003

ndash0

16ndash

013

ndash0

200

031

00

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

H3

TK

NA

lB

aC

uF

eL

iM

nS

rZ

nC

aM

gN

aK

TPU

TP

FC

ond

pHE

lev

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

494 Dermot Antoniades Marianne S V Douglas and John P Smol

Fig 4 Alert PCA biplot with outlier sites removed Inset indicates relative position ofoutlier sites when included passively in PCA analysis

included as passive sites in their respective analyses these sites influenced the ordina-tions to such a degree that they hampered the visualization of patterns among sites andwere thus removed from the biplots entirely Their positions on the PCA biplots rela-tive to the other sites are shown inset in Figs 4 and 5

A combined dataset from both study regions was analyzed with RDA in order todetermine to what degree water chemistry variation could be explained solely by geo-graphy Measured physical and chemical variables were analyzed using an explanatorydataset that consisted of a binary environmental variable representing either Alert orMould Bay

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 495

Fig 5 Mould Bay PCA biplot with outlier sites removed Inset as in Figure 4

Results

Conductivity pH and major ions

Thirty sites with diverse physical characteristics were sampled near AlertSites ranged in elevation from 13 to 147m asl (mean = 65 m asl) Seven ofthese sites were classified as lakes (ie gt2 m deep) that varied from ~450 to1400 m in diameter Water temperatures in the Alert sites ranged from 2 to15 ƒC and conductivity varied from 121 to 2000 mS cmndash1 Median conductivitywas 230 mS cmndash1 and except for two extreme sites (ie A-A 2000 mS cmndash1 andA-B 1650 mS cmndash1) Alert conductivity values fell within a more restrictedrange of (121ndash420 mS cmndash1) Alert sites were strongly alkaline ranging frompH 81 to 89 (mean = 84)

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

496 Dermot Antoniades Marianne S V Douglas and John P Smol

At Mould Bay 32 of 35 sites were classified as ponds which were typi-cally small (ie lt 50 m diam) and less than 1m deep The majority were lo-cated on low emergent grassy plains with 22 sites at or below 10 m in eleva-tion Mould Bay water temperatures fell within the restricted range of 4 to95 ƒC (mean = 64 ƒC) while conductivity values ranged from 25 to530 mS cmndash1 (mean = 115 mS cmndash1) Mould Bay mean pH was 79 with a rangefrom 70 to 86

Ca had the highest median concentration among major ions at Alert(378 mg lndash1 range = 188ndash675 mg lndash1) Mg Na and K had medians of 93 45and 05 mg lndash1 respectively and each had a range of at least one order of mag-nitude (see Appendix 1)

Median DIC was 263 mg lndash1 with a relatively restricted range from 121 to459 mg lndash1 Cl had a much broader range (06ndash703 mg lndash1) but a lower medianof 75 mg lndash1 SO4 was typically lower (median = 32 mg lndash1) although concen-trations ranged from 02 to 111 mg lndash1 Median concentrations at Alert wereCa gt Mg gtNa gtK and Ca was the dominant cation at 28 of 30 sites Two siteswere Na dominated (ie A-A A-B) due to greatly elevated Na and Mg Amonganions median concentrations were DIC gtCl gtSO4 and sites were generallyDIC dominated However at sites where Cl gtSO4 often both variables hadvery low concentrations and differences between the two were minimal

Mould Bay Ca and Na had similar means (132 mg lndash1 and 129 mg lndash1 re-spectively) Ca ranged from 26 to 293 mg lndash1 and Na from 16 to 988 mg lndash1Mean Mg was 52 mg lndash1 (range = 08ndash228 mg lndash1) while mean K was10 mg lndash1 (range lt 02 ndash40 mg lndash1) Cl had the highest mean (250 mg lndash1)among anions However concentrations ranged from 32 to 212 mg lndash1 and themedian value of 932 mglndash1 was more typical of Mould Bay sites DIC concen-trations reached a maximum of 225 mg lndash1 (mean = 94 mg lndash1) SO4 had thelowest mean concentration among major anions (77 mg lndash1 range = 06ndash384 mg lndash1) Mean concentrations were Ca gtNa gtMg gtK and Cl gt DIC gtSO4although five different relative cation and six different anion concentrationsexisted

Metals

Similar to most chemical variables metal concentrations were low in both theAlert and Mould Bay datasets with the exception of a small number of siteswith greatly elevated values Many of the metals analyzed were below detec-tion limits at the majority of sites and thus were removed from the datasetsprior to statistical analysis At Alert median Fe was 011 mg lndash1 with a rangefrom 0008 to 370 mg lndash1 Alert Al values ranged from below the detectionlimit (lt 01mglndash1) to 215 mglndash1 with a median value of 005 mg lndash1 while me-dian Mn was 00057mglndash1 (range = 00013 ndash00605 mglndash1)

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 497

Sites in the Mould Bay dataset had higher metal concentrations than thosefrom Alert Fe had a mean of 0724 mg lndash1 and a maximum concentration of720 mg lndash1 Al concentrations were generally low and although mean Al was012 mg lndash1 the median was 003 mg lndash1 Two Mould Bay sites were below theAl detection limit of 01 mg lndash1 while the maximum was 149 mg lndash1 Mnranged from 00011 to 01480 mglndash1 with a mean of 00185 mglndash1

Phosphorus nitrogen carbon and Chl-a

TPU at Alert ranged from 34 to 676 mg lndash1 with a median of 91 mg lndash1 Ac-cording to Wetzelrsquos (1983) classification scheme 17 of Alert sites were ul-tra-oligotrophic by TPU concentrations 37 were oligotrophic 33 weremesotrophic and 13 were eutrophic Mould Bay TPU ranged from 71 to1170 mg lndash1 with a mean of 165 mg lndash1 By TPU concentrations no Mould Baysites were ultra-oligotrophic 28 were oligotrophic 66 were mesotrophicand 6 were eutrophic

Five nitrogen measurements were made from each site (ie TKN NH3NO2 NO3+ NO2 and PON) NO2 and NO3 +NO2 were below their respectivedetection limit at the majority of sites in both study areas Alert TN rangedfrom 0107 to 1631 mg lndash1 (median = 0465 mg lndash1) TKN varied from 0086 to0736 mg lndash1 with a median of 0350 mg lndash1 Alert mean PON was 0032 mg lndash1

(range = 0015ndash0286 mg lndash1) and median NH3 was 0022 mg lndash1 althoughNH3 was below detection limits (ie 0005 mglndash1) at 14 sites

Mean Mould Bay TN was 0616 mg lndash1 (range = 0206 to 1430 mg lndash1) andTKN concentrations ranged from 0160 to 1360 mg lndash1 (mean = 0515 mg lndash1)PON concentrations were also highly variable ranging from 0030 to0684 mg lndash1 (mean = 0096 mg lndash1) while NH3 ranged from lt0005 to0122 mg lndash1 (mean = 0035 mg lndash1) TN TPU ratios ranged from 4 1 to 104 1at Alert sites with a mean of 46 1 while at Mould Bay the mean was 49 1with ratios ranging from 9 1 to 1011

Alert Chl-a concentrations were consistently low ranging from below thedetection limit of 01 mg lndash1 to a maximum of 26 mg lndash1 (median = 10 mg lndash1)POC varied from 0108 to 3090 mg lndash1 (median = 0328 mg lndash1) while DOCranged from 06 to 61mg lndash1 with a median of 27 mg lndash1 POC CHLA ratioswere calculated to assess the potential carbon sources as ratios greater than100 1 suggest allochthonous sources of POC (Eppley et al 1977) MedianPOC CHLA (348 1) implied that the majority of POC at Alert sites was al-lochthonously derived with values ranging from 106 1 to 1818 1

Due to analytical problems seven Mould Bay sites have no measured Chl-a values At a further 10 sites Chl-a was below the detection limit of 01mg lndash1At the 18 sites with available Chl-a measurements mean Chl-a was 08 mg lndash1POC and DOC values reflected higher catchment inputs of detrital matter

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

498 Dermot Antoniades Marianne S V Douglas and John P Smol

POC ranged from 0375 to 7090 mg lndash1 (mean = 0901 mg lndash1) while DOCranged from 11 to 137mg lndash1 (mean 67mg lndash1) POC CHLA ratios calculatedat the 18 sites in the Mould Bay dataset with available Chl-a values again sug-gested terrestrial sources for POC with a large range of values from 174 1 to8800 1 (mean = 1289 1)

Multivariate analyses

Several iterations of RDA and PCA were used to investigate patterns of varia-tion in the water chemistry datasets RDA was run to determine the degree towhich a strictly geographic factor could account for differences between theAlert and Mould Bay datasets and indicated that classification by island ex-plained 164 of the water chemistry variation The variables most highlycorrelated with the first RDA axis were as expected those for which one ofthe study regions had distinctively higher concentrations High DIC Ca andpH were most strongly associated with Alert High concentrations of TPF andDOC were most strongly linked with Mould Bay

PCA analysis indicated that there were two primary gradients in each envi-ronmental dataset The first two axes of the Alert PCA explained 476 of thewater chemistry variation with eigenvalues of 0276 and 0199 respectivelyConductivity-related variables and nutrients (ie Mg Sr TKN DOC TN andCond) most strongly controlled the first axis The second axis was moststrongly correlated with Al Fe Zn and Mn and thus represented a gradient ofmetal concentrations

The first two axes of the Mould Bay PCA had eigenvalues of 0302 and0188 respectively and collectively explained 490 of the variation in thedataset The first axis was again controlled by conductivity-related variables(ie Sr Ca Cond K Mg and DIC) while the second axis represented a nutri-ent gradient and was correlated with NH3 TN SRP and TPU

Discussion

Conductivity pH and major ions

The majority of the sites present in the study were shallow tundra ponds how-ever physical characteristics (Appendixes 1 2) differed between Alert andMould Bay Roughly one quarter of Alert sites were lakes compared with fe-wer than one in ten at Mould Bay Although the greater proportion of lakes inthe Alert dataset may partially account for differences in nutrient and soluteconcentrations between Alert and Mould Bay geology and vegetation appearto play the largest role in determining regional water chemistry Differences introphic status between Alert and Mould Bay sites can be attributed to vegeta-

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 499

tional regime while differences in ionic chemistry resulted largely from bed-rock differences High arctic water bodies are typically alkaline dilute andoligotrophic While Alert sites largely conformed to these trends the trophicstatus and ionic chemistry of Mould Bay sites were distinct among Canadianhigh arctic regions

Conductivity at both Alert and Mould Bay was within ranges found inother high arctic limnological studies (ie Antoniades et al 2003 Mi-chelutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Vezina ampVincent 1997 Douglas amp Smol 1994) No significant correlation was iden-tified between conductivity and elevation at either Alert or Mould Bay a cor-relation that previous high arctic studies have inferred represent distance fromthe ocean (Lim et al 2001 Michelutti et al 2002 b) However any relation-ship between conductivity and elevation at Alert may have been obscured bythe fact that the four largest most dilute lakes also had the lowest elevationsAt Mould Bay the eight sites closest to the Arctic Ocean had the eight highestconductivity values despite a range of elevations

The predominance of alkaline sites across the Canadian High Arctic resultsfrom the carbonate bedrock and glacial materials that dominate most of the re-gion In tundra ponds pH may also be elevated by photosynthesis during theextended daylight of the arctic summer pH typically ranges from 7 to 85 andvalues exceeding 9 have been reported (Vezina amp Vincent 1997 Hamiltonet al 2001) Alert sites were among the most alkaline reported from the HighArctic while Mould Bay sites which were situated on non-calcareous sand-stones were circumneutral to slightly alkaline due to the influence of carbon-ate surficial materials

Trends in major ion concentrations at Alert were broadly similar to thoseobserved in other high arctic surveys however Mould Bayrsquos major ions dif-fered from other high arctic sites Alert sites had the highest DIC and secondhighest Ca yet recorded in a limnological survey in the Canadian High Arcticresulting from the calcium carbonate bedrock underlying the Alert area Sim-ilar Ca concentrations were also recorded farther north on Ward Hunt Island(Villeneuve et al 2001) Conversely Mould Bay mean Ca and DIC were ator among the lowest found to date in a high arctic limnological survey re-flecting that Mould Bay is the lone study site not dominated by calcareousbedrock (Antoniades et al 2003 Michelutti et al 2002 a b Hamiltonet al 2001 Lim et al 2001) Na and Cl were at the upper end of typical higharctic ranges at both Mould Bay and Alert reflecting the greater proximity ofsites in this study to the ocean Concentrations of other major ions (Mg K andSO4) were within previously identified ranges in the Canadian High Arctic andnormal ranges for Canadian freshwaters (Antoniades et al 2003 Miche-lutti et al 2002 a b Hamilton et al 2001 Lim et al 2001 Douglas ampSmol 1994 McNeely et al 1979) Individual sites with elevated Na Cl Mg

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

500 Dermot Antoniades Marianne S V Douglas and John P Smol

and SO4 were either situated within very short distances of the ocean (ie MB-I MB-JE MB-JW MB-AG MB-AH) or may have been impacted by anthro-pogenic sources (ie A-A A-B)

K in tundra ponds may be increased by leaching from vascular plant litterparticularly during spring melt (Prentki et al 1980 Cornwell 1992) Highindividual K concentrations have been attributed to higher catchment vegeta-tion (Lim et al 2001 Michelutti et al 2002 a) however despite vegetationlevels that far exceed any observed in these studies Mould Bay K remainedlow and concentrations were similar to those from Alert

Na K in natural waters is typically between 2 1 and 3 1 (McNeely et al1979) However median Alert Na K was 9 1 while the ratio from Mould Baywas 4 1 The ratio of mean Na K in previous high arctic surveys has rangedfrom 8 1 to 13 1 and appears to increase with decreasing vegetation levelsThe highest Na K in the High Arctic was from Isachsen Ellef Ringnes Islanda region almost completely devoid of vegetation (Antoniades et al 2003)while the lowest value was from Mould Bay the most lushly vegetated regionamong these studies By comparison studies from the Canadian and AlaskanArctic mainland had Na K ratios ranging from 1 to 5 (Gregory-Eaves et al2000 Ruumlhland amp Smol 1998 Pienitz et al 1997 a b) and a survey fromVictoria Island the farthest south of the Canadian Arctic Islands had a ratio of2 1 High ratios from around the High Arctic relative to lower arctic sites maybe indicative of extremely low K contributions resulting from the depauperatevegetation regime at these sites

Metals

Similar to most high arctic datasets metals in both study areas were present inlow concentrations A notable exception was the high mean Fe at Mould Baywhich was above the range considered normal for Canadian waters(lt 05 mg lndash1 McNeely et al 1979) although a broad range of concentrationswas present (0083 to 7200 mg lndash1) While extreme sites in the dataset in-creased this mean (median Fe = 0267 mg lndash1) 12 sites still had Fe concentra-tions that exceeded 05 mg lndash1 Weathering of ferric minerals in the sandstonesof the Mould Bay area may be responsible for these elevated Fe levels In ad-dition while mean Fe at Alert was within typical high arctic and Canadianranges six sites in the dataset had values above 05 mg lndash1 Elevated Fe con-centrations have been observed at other high arctic sites in contact with pyrite-rich shale (Antoniades et al 2003 Hamilton et al 2001) At Alert crystal-lized pyrite was observed between bedrock joints and provided a potentialsource of Fe to Alert sites

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 501

Nutrients Chl-a and trophic status

Broad ranges of TPU concentrations were present in both regions however atAlert the majority of sites was oligotrophic to meso-oligotrophic while atMould Bay the majority of sites were mesotrophic Mean Mould Bay TPU ex-ceeded those from all other Canadian High Arctic studies (Michelutti et al2002 a b Hamilton et al 2001 Lim et al 2001 Douglas amp Smol 1994) ex-cept Isachsen Ellef Ringnes Island which was heavily influenced by phos-phatic shale (Antoniades et al 2003) TPU values similar to Mould Bay havealso been reported from lakes on Spitsbergen (van Donk et al 2001) The de-cay of organic matter from lush vegetation provides a potential source of in-creased phosphorus supply to Mould Bay sites not available in other less veg-etated high arctic locations

Alert mean TKN was higher than any previously reported in a high arcticsurvey but was further exceeded by mean TKN at Mould Bay Mould Baymean TN and NH3 were higher than those of all previous high arctic surveysand mean PON was only exceeded by that from Ellef Ringnes Island (Anto-niades et al 2003) TKN values greater than 05 mg lndash1 generally indicatehigh organic inputs (McNeely et al 1979) Although there was great between-site variability mean TKN and concentrations at 17 of 35 Mould Bay sites ex-ceeded this figure The higher TPU TKN and PON levels at Mould Bay are areflection of high vegetation levels which provide an abundant source of det-rital organic matter

Ninety-eight percent of the sites in this study were classified as oli-gotrophic by Chl-a concentrations which is typical of high arctic water bo-dies Alert Chl-a concentrations while low were somewhat higher than meas-urements from our other high arctic studies However recent methodologicalchanges may hamper direct comparisons of Chl-a concentrations with earlierpublished figures (see Methods) Regardless the consistently low Chl-a con-centrations in both study areas are indicative of low productivity in the watercolumn and were similar to those reported from the Canadian High Arctic andfrom Spitsbergen

TN TPU ratios were calculated in order to investigate the degree to whichphytoplankton productivity may be limited by nitrogen or phosphorus Higharctic lakes are typically dominated by benthic communities which can ac-count for over 95 of a sitersquos phototrophic productivity (Vezina amp Vincent1997 Villeneuve et al 2001) In addition nutrients and major ions withinarctic cyanobacterial mats can be concentrated by over an order of magnitudewhen compared with concentrations in the overlying water column (Ville-neuve et al 2001) As such estimates of nutrient limitation using samplestaken from the water column are applicable only to phytoplankton

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

502 Dermot Antoniades Marianne S V Douglas and John P Smol

TN TP ratios suggested P-limitation (ie sup317 1) at 93 of Alert sites and97 of Mould Bay sites Strong P-limitation was indicated at the majority ofsites in the dataset as 67 of Alert sites and 83 of Mould Bay sites ex-ceeded 30 1 Only three sites (A-E UDL MB-AF) had ratios suggestingN-limitation Of these two sites (ie UDL MB-AF) had anomalously highconcentrations of particulate matter and low TN TP ratios which likely resultfrom resuspended sediment and may not be reflective of N-limitation

Phosphorus and nitrogen were present in high concentrations at Mould Bayrelative to other high arctic sites yet Chl-a concentrations remained at the lowend of High Arctic ranges The importance of nitrogen as a limiting factor isoften underestimated in arctic and oligotrophic lakes (Levine amp Whalen2001 Elser et al 1990) In our study dissolved inorganic nitrogen specieswere at or below detection limits in many sites at both regions As such weperformed a series of linear regressions to investigate the response of Chl-a toincreases in TPU TN and TKN (Fig 6) and compared these to publishedequations for CHLA TP (ie Smith 1982 Dillon amp Rigler 1974) Despitethe suggestion of strong P-limitation by nutrient ratios Chl-a and TPU werenot significantly correlated (p pound005) and there was no significant relation be-tween Chl-a and TN or TKN The regression equations from the literature con-sistently overestimated Chl-a responses to increases in TPU in our sites(Fig 6 a b) This supports the findings of Flanagan et al (2003) who sug-gested that in arctic lakes smaller increases in algal biomass result from agiven increase in phosphorus than would occur in temperate sites Howeverthe lack of a relationship between Chl-a and N or P suggests that other factorsmay be controlling phytoplankton productivity at our sites Moreover as themajority of the production in these sites is from the periphyton open waterChl-a concentrations are not measuring periphytic production The precise roleof nutrients in regulating phytoplankton productivity in these and other higharctic sites is thus unclear and warrants future investigation through bioassaysand fertilization experiments

Alertrsquos median POC concentration was similar to those found at typicalsparsely vegetated high arctic sites while the Mould Bay median POC con-centration exceeded those reported from all high arctic regions except EllefRingnes Island where resuspension of sediments was a confounding factor(Antoniades et al 2003) POC CHLA ratios were significantly lower atAlert than those calculated in other high arctic surveys (Antoniades et al2003 Michelutti et al 2002 a b Lim et al 2001) and may reflect theslightly elevated Chl-a concentrations found at Alert Mean POC CHLA atboth study areas suggests that the majority of POC is derived from allochtho-nous and thus likely terrestrial sources a finding similar to all previous higharctic limnological surveys POC CHLA ratios were not calculated at MouldBay sites with no measured CHLA concentration however as ten of these

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 503

Fig 6 Individual regressions of Chl-a against TPU TN and TKN a) sites fail to segre-gate according to Smithrsquos (1982) regression equations for TN TP = 10 and 25 b)includes Chl-a TP regression of Dillon amp Rigler (1974)

sites were below the CHLA detection limit it is likely that our mean value un-der-represents the actual value for Mould Bay sites Despite this ratios fromMould Bay were much higher than those from Alert and elsewhere in the HighArctic further reinforcing the sensitivity of these sites to external detrital in-puts and the influence of the relatively lush vegetation in the Mould Bay arearelative to other more sparsely vegetated high arctic regions

Alert DOC concentrations were higher than those from Spitsbergen (Ellis-Evans et al 2001) and similar to those found elsewhere in the Canadian HighArctic with the exception of Ward Hunt Lake where DOC exceeded the max-imum found at Alert (Villeneuve et al 2001) This emphasizes the impor-tance of individual catchment characteristics in determining high arctic DOC

Mould Bay DOC greatly exceeded all values previously reported from ourhigh arctic surveys Concentrations were similar to those reported at or neartreeline from the subarctic Northwest Territories and arctic sites in northern

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

504 Dermot Antoniades Marianne S V Douglas and John P Smol

Russia (Pienitz et al 1997a Duff et al 1999) however DOC remained farlower than values reported from Alaska and western arctic Canada (Gregory-Eaves et al 2000 Pienitz et al 1997b) High arctic freshwater sites are typi-cally low in DOC and thus highly sensitive to ultraviolet radiation (Vincentamp Pienitz 1996) Predicted future increases in ultraviolet radiation are ex-pected to further stress high arctic lakes and ponds However high arctic veg-etation cover is also predicted to increase in response to higher temperaturesand greater precipitation High DOC at Mould Bay resulting from terrestrialorganic inputs suggest that the future effects of increased UV radiation onthese water bodies may be moderated by higher DOC concentrations

Outlier sites

The water chemistry of several sites was distinct from all others in the studyA-A and A-B had water chemistry characteristics that were distinctly differentfrom all other sites in the Alert dataset Site A-A had the highest concentrationin the dataset for 23 measured chemical variables including Cl SO4 Mg Naand K while A-B had the second highest concentration of these five variablesThese sites were among the closest to the Alert military base and were locatedshort distances from roads used to service outlying components of the installa-tion This proximity was the only distinguishing feature of these ponds assuch their elevated concentrations likely result from the construction andmaintenance of these roads

The TPU concentration of UDL (mean = 346 mg lndash1) was the highestamong non-impacted sites in the Alert dataset The TPF concentration at UDL(26 mg lndash1) was not comparably elevated and was in fact similar to that meas-ured at other lakes in the dataset Turbid waters resulting from high winds andlarge waves were observed while sampling UDL The adjacent LDL similarin size and general characteristics was sampled during calm conditions andhad a TPU concentration of only 41 mg lndash1 In addition the TN TP ratio ofUDL (5 1) closely approximates that of oligotrophic lake sediment (ie 3 1Downing amp McCauley 1992) As such high TPU at UDL is likely reflectiveof the contribution of resuspended particulate phosphorus

MB-AF TPU (1170 mg lndash1) was the highest amongst the Mould Bay sites Itwas the largest site sampled at Mould Bay and was sampled under conditionssimilar to UDL POC and PON were also elevated at MB-AF however TPFbarely exceeded the Mould Bay mean (93 vs 85 mg lndash1) MB-AF also had thelowest TN TP ratio among the Mould Bay sites Classifications of MB-AFand UDL as eutrophic (by TPU) are somewhat improbable given the trophicconditions of all other sites in the region The similarities between these sitesimply that measured TPU concentrations include resuspended particulate ma-terial and are not indicative of increased trophic levels

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 505

Tundra ponds are more sensitive than lakes to increasing concentration ofsolutes from external inputs and via evaporative concentration Because of thisdifference in physical characteristics concentrations of many water chemistryvariables differed greatly between lakes (n = 7) and ponds (n = 23) in the Alertdataset A series of simple t-tests indicated that 17 limnological variables dif-fered significantly (p pound005) between lakes and ponds (ie POC PON DOCDIC SiO2 NH3 TKN TN TPF SO4 Ca Mg Li Sr Cond T and pH) Con-centrations of these limnological variables were higher in ponds than in lakesas expected The difference in the chemical limnology between lakes andponds is indicative of the degree to which their concentrations are dependenton external inputs Because of their small volumes tundra ponds have muchlower capacities for dilution of allochthonous material and are thus highlysensitive to allochthonous deposition of solutes and nutrients Canonical Var-iates Analysis (CVA) indicated that the differences in concentrations of these17 variables accounted for 717 of water chemistry differences between lakesand ponds The variables that most strongly differentiated between lakes andponds were in descending order TKN DIC POC PON and Mg A secondCVA indicated that these five variables alone explained 637 of the waterchemistry differences between lakes and ponds

Mult ivariate analysis

Lakes and ponds were divided mainly into two groups along Alert PCA axis 1reflecting the difference in the evaporative concentration between the two sitetypes Accordingly the arrows representing site diameter and the conductivity-related variables are directly opposite to each other along the axis The higherdegree of variability in the chemical compositions of ponds was reflected intheir wide distribution along both PCA axes Lakes which were consistentlylow in solute concentrations more densely clustered than ponds and are sit-uated in the left-centre of the ordination

The Mould Bay PCA illustrates the effect of distance from the ocean onthe chemical composition of sites High conductivity sites situated near theocean are located in the right half of the ordination Proximity to the oceanappears to override site size as the primary determinant of solute concentra-tions however diameter is inversely related to nutrient and particulate con-centrations along PCA axis 2

Conclusions

The data presented in this study constitute the first documentation of baselinelimnological conditions for lakes and ponds in the Alert and Mould Bay re-

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

506 Dermot Antoniades Marianne S V Douglas and John P Smol

gion This information is crucial as these systems are likely to undergo dra-matic changes in the future due to changing climatic conditions The differentlimnology and higher DOC concentrations of the Mould Bay region relative toother less vegetated sites may provide examples of future limnological condi-tions that may be expected elsewhere in the High Arctic under the warmerwetter climate regimes and more densely vegetated ecosystems predictedwithin the next century

Acknowledgements

This research was completed with Natural Sciences and Engineering Research Councilof Canada and Polar Continental Shelf Project (PCSP) funding to MSVD and JPSand Northern Scientific Training Program and Ontario Graduate Scholarship fundingto DA We are extremely grateful to Derek Muir and Xiaowa Wang of the NWRI fortheir assistance with the water chemistry analyses This is PCSP contribution PCSPEacutePCP 01503

References

Antoniades D Douglas M S V amp Smol J P (2003) Physical and chemical lim-nology of 24 lakes and one pond from Isachsen Ellef Ringnes Island CanadianHigh Arctic ndash Internat Rev Hydrobiol 88 519 ndash539

Birks H J B Line J M Juggins S Stevenson A C amp ter Braak C J F(1990) Diatoms and pH reconstruction ndash Philosoph Transact Roy Soc LondonB 327 263 ndash278

Blom T Korhola A Weckstroumlm J Laing T Snyder J MacDonald G ampSmol J P (2000) Physical and chemical characterisation of small subarctic head-water lakes in Finnish Lapland and the Kola Peninsula ndash Internat Verein Limnol27 316 ndash320

Christie R L (1964) Geological reconnaissance of Northeastern Ellesmere IslandDistrict of Franklin ndash Geol Surv Can Mem 331 79 pp

Cornwell J C (1992) Cation export from Alaskan arctic watersheds ndash Hydrobiolo-gia 240 15ndash22

Dillon P J amp Rigler F H (1974) The phosphorus-chlorophyll relationship inlakes ndash Limnol Oceanogr 19 767ndash773

Doubleday N C Douglas M S V amp Smol J P (1995) Paleoenvironmentalstudies of black carbon deposition in the High Arctic a case study from NorthernEllesmere Island ndash Sci Total Environm 160161 661ndash668

Douglas M S V amp Smol J P (1999) Freshwater diatoms as indicators of environ-mental change in the High Arctic ndash In Stoermer E F amp Smol J P (eds) TheDiatoms Applications for the Environmental and Earth Sciences ndash CambridgeUniversity Press Cambridge pp 227ndash244

ndash ndash (1994) Limnology of high arctic ponds (Cape Herschel Ellesmere IslandNWT) ndash Arch Hydrobiol 131 401ndash 434

Downing J A amp McCauley E (1992) The nitrogenphosphorus relationship inlakes ndash Limnol Oceanogr 37 936 ndash945

Limnology of two Canadian High Arctic regions 507

Duff K E Laing T E Smol J P amp Lean D R S (1999) Limnological charac-teristics of lakes located across arctic treeline in northern Russia ndash Hydrobiologia391 205 ndash222

Edlund S A amp Alt B T (1989) Regional Congruence of Vegetation and SummerClimate Patterns in the Queen Elizabeth Islands Northwest Territories Canada ndashArctic 42 3ndash 23

Ellis-Evans J C Galchenko V Laybourn-Parry J Mylnikov A P amp PetzW (2001) Environmental characteristics and microbial plankton activity of fresh-water environments at Kongsfjorden Spitsbergen (Svalbard) ndash Arch Hydrobiol151 609 ndash632

Elser J J Marzolf E R amp Goldman C R (1990) Phosphorus and nitrogen lim-itation of phytoplankton growth in the freshwaters of North America a review andcritique of experimental enrichments ndash Can J Fish Aquat Sci 47 1468ndash1477

Environment Canada (1994) Manual of Analytical Methods ndash National Laboratory forEnvironmental Testing Canadian Centre for Inland Waters Burlington Ontario 2vols 1097 pp

Eppley R W Harrison W G Chisolm S W amp Stuart E (1977) Particulate or-ganic matter in surface waters off California and its relation to photosynthesis ndashJ Mar Res 35 671ndash 696

Everett K R (1968) Soil Development in the Isachsen and Mould Bay AreasQueen Elizabeth Islands Northwest Territories Canada ndash Institute of Polar Stud-ies Report No 24 Columbus Ohio State University Research Foundation

Flanagan K M McCauley E Wrona F amp Prowse T (2003) Climate changethe potential for latitudinal effects on algal biomass in aquatic ecosystems ndash CanJ Fish Aquat Sci 60 635 ndash639

Gregory-Eaves I Smol J P Finney B P Lean D R S amp Edwards M(2000) Characteristics and variation in lakes along a north-south transect inAlaska ndash Arch Hydrobiol 147 193 ndash223

Hamilton P B Gajewski K Atkinson D E amp Lean D R S (2001) Physicaland chemical limnology of 204 lakes from the Canadian Arctic Archipelago ndash Hy-drobiologia 457 133 ndash148

Hamilton P B Gajewski K McNeely R N amp Lean D R S (2000) Physicalchemical and biological characteristics of lakes from the Slidre River basin onFosheim Peninsula Ellesmere Island Nunavut ndash In Garneau M amp Alt B T(eds) Environmental response to climate change in the Canadian High Arctic ndashGeol Surv Can Bull 539 235 ndash248

Hobbie J E (1997) History of Limnology in Alaska Expeditions and Major Projectsndash In Milner A M amp Oswood M W (eds) Freshwaters of Alaska ndash EcologicalSyntheses ndash Ecological Studies 119 Springer New York pp 45ndash60

Kalff J amp Welch H E (1974) Phytoplankton production in Char Lake a naturalpolar lake and in Meretta Lake a polluted polar lake Cornwallis Island North-west Territories ndash Can J Fish Aquat Sci 31 621ndash 636

Kling G W OrsquoBrien W J Miller M C amp Hershey A E (1992) The biogeo-chemistry and zoogeography of lakes and rivers in arctic Alaska ndash Hydrobiologia240 1ndash14

Korhola A Weckstroumlm J amp Blom T (2002) Relationships between lake andland-cover features along latitudinal vegetation ecotones in arctic Fennoscandia ndashArch Hydrobiol Suppl 1392 203 ndash 235

508 Dermot Antoniades Marianne S V Douglas and John P Smol

Levine M A amp Whalen S C (2001) Nutrient limitation of phytoplankton produc-tion in Alaskan Arctic foothill lakes ndash Hydrobiologia 455 189 ndash201

Lim D S S Douglas M S V Smol J P amp Lean D R S (2001) Physical andchemical limnological characteristics of 38 lakes and ponds on Bathurst IslandNunavut Canadian High Arctic ndash Internat Rev Hydrobiol 86 1ndash22

Markager S Vincent W F amp Tang E P Y (1999) Carbon fixation by phyto-plankton in high Arctic lakes Implications of low temperature for photosynthesisndash Limnol Oceanogr 44 597ndash607

Maxwell J B (1997) Responding to Global Climate Change in Canadarsquos Arctic ndashCanada Country Study Climate Impacts and Adaptation vol 2 EnvironmentCanada Downsview Ontario 82 pp

McNeely R N Neimanis V P amp Dwyer L (1979) Water Quality Sourcebook Aguide to water quality parameters ndash Minister of Supply and Services Canada Ot-tawa Ontario 88 pp

Meteorological Service of Canada (2002) Canadian Climate Normals 1971ndash 2000 ndashEnvironment Canada Ottawa Ontario

Michelutti N Douglas M S V Lean D R S amp Smol J P (2002 a) Physicaland chemical limnology of 34 ultra-oligotrophic lakes and ponds near WynniattBay Victoria Island Arctic Canada ndash Hydrobiologia 482 1ndash13

ndash ndash ndash ndash (2002 b) Limnological characteristics of 38 lakes and ponds on AxelHeiberg Island High Arctic Canada ndash Internat Rev Hydrobiol 87 385 ndash399

Pienitz R Smol J P amp Lean D R S (1997a) Physical and chemical limnologyof 59 lakes located between the southern Yukon and the Tuktoyaktuk PeninsulaNorthwest Territories (Canada) ndash Can J Fish Aquat Sci 54 330 ndash346

ndash ndash ndash (1997b) Physical and chemical limnology of 24 lakes located between Yel-lowknife and Contwoyto Lake Northwest Territories (Canada) ndash Can J FishAquat Sci 54 347ndash 358

Prentki R T Miller M C Barsdate R J Alexander V Kelley J ampCoyne P (1980) Chemistry ndash In Hobbie J E (ed) Limnology of TundraPonds Barrow Alaska ndash Dowden Hutschinson and Ross Inc Stroudsburg PApp 76ndash178

Quesada A Vincent W F amp Lean D R S (1999) Community and pigmentstructure of Arctic cyanobacterial assemblages the occurrence and distribution ofUV-absorbing compounds ndash FEMS Microbiol Ecol 28 315 ndash323

Rouse W R Douglas M S V Hecky R E Hershey A E Kling G W Le-sack L Marsh P McDonald M Nicholson B J Roulet N T amp SmolJ P (1997) Effects of climate change on the waters of Arctic and Subarctic NorthAmerica ndash Hydrological Processes 11 973 ndash902

Ruumlhland K amp Smol J P (1998) Limnological Characteristics of 70 Lakes Spann-ing Arctic Treeline from Coronation Gulf to Great Slave Lake in the CentralNorthwest Territories Canada ndash Internat Rev Hydrobiol 83 183ndash203

Schindler D W (2001) The cumulative effects of climate warming and other hu-man stresses on Canadian freshwaters in the new millennium ndash Can J FishAquat Sci 58 18ndash29

Schindler D W Welch H E Kalff J Brunskill G J amp Kritsch N (1974)Physical and chemical limnology of Char Lake Cornwallis Island (75 ƒN Lat) ndash JFish Resour Bd Can 31 585 ndash 607

Limnology of two Canadian High Arctic regions 509

Smith V H (1982) The nitrogen and phosphorus dependence of algal biomass inlakes An empirical and theoretical analysisndash Limnol Oceanogr 27 1101ndash1112

Smol J P (1988) Paleoclimate proxy data from freshwater arctic diatoms ndash Verh In-ternat Verein Limnol 23 837ndash844

ter Braak C J F amp Šmilauer P (1998) CANOCO and CANOPost Version 4 forWindows ndash Microcomputer Power Ithaca NY USA

van Donk E Faafeng B A de Lange H J amp Hessen D O (2001) Differentialsensitivity to natural ultraviolet radiation among phytoplankton species in Arcticlakes (Spitspergen Norway) ndash Plant Ecology 154 249 ndash 259

Vezina S amp Vincent W F (1997) Arctic cyanobacteria and limnological propertiesof their environment Bylot Island Northwest Territories Canada (73 ƒN 80 ƒW) ndashPolar Biology 17 523 ndash534

Villeneuve V Vincent W F amp Komaacuterek J (2001) Community structure and mi-crohabitat characteristics of cyanobacterial mats in an extreme high Arctic envi-ronment Ward Hunt Lake ndash Nova Hedwigia Beiheft 123 199 ndash224

Vincent W F amp Pienitz R (1996) Sensitivity of high-latitude freshwater ecosys-tems to global change temperature and solar ultraviolet radiation ndash GeoscienceCanada 23 231ndash 236

Weckstroumlm J amp Korhola A (2001) Patterns in the distribution composition anddiversity of diatom assemblages in relation to ecoclimatic factors in Arctic Lap-land ndash J Biogeogr 28 31ndash 45

Weckstroumlm J Korhola A amp Blom T (1997) Diatoms as quantitative indicatorsof pH and water temperature in subarctic Fennoscandian lakes ndash Hydrobiologia347 171ndash184

Wetzel R G (1983) Limnology 2nd ed ndash Saunders Publishing Philadelphia PA767 pp

Submitted 17 March 2003 accepted 29 August 2003

510 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x1

Ale

rtW

ater

Che

mis

try

DL

=B

elow

dete

ctio

nlim

itsN

A=

valu

eno

tava

ilabl

e

Site

Cl

SO4

SiO

2PO

CPO

ND

OC

DIC

SRP

NO

2N

O3+

NO

2N

H3

TK

NA

lB

aB

em

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

A-A

703

011

10

044

309

00

286

24

347

48

000

10

888

lt0

005

045

72

150

0205

lt0

0002

A-B

553

092

30

511

210

014

22

733

22

0lt

000

10

220

lt0

005

037

60

510

0152

lt0

0002

A-C

22

35

023

024

70

029

11

250

06

lt0

001

lt0

010

lt0

005

015

30

060

0022

lt0

0002

Hilg

ard

L

06

18

040

011

60

017

08

212

08

lt0

001

003

3lt

000

50

086

084

000

77lt

000

02A

-E7

85

80

041

060

007

20

625

12

00

076

007

3lt

000

50

106

107

000

43lt

000

02A

-F10

712

30

241

080

010

02

733

20

90

114

011

7lt

000

50

327

024

000

52lt

000

02A

-G6

21

53

120

387

005

26

145

91

30

004

lt0

010

004

90

669

012

000

43lt

000

02A

-H10

61

40

600

519

005

94

934

41

4lt

000

1lt

001

00

077

065

00

090

0052

lt0

0002

A-I

33

09

411

033

30

043

60

225

13

lt0

001

lt0

010

007

70

667

001

000

16lt

000

02A

-J43

511

90

550

542

004

73

841

21

5lt

000

1lt

001

00

078

041

80

310

0048

lt0

0002

A-K

219

85

149

019

00

023

39

240

11

lt0

001

lt0

010

008

30

602

002

000

32lt

000

02M

oss

P51

66

10

830

108

001

60

719

80

8lt

000

1lt

001

00

013

009

10

040

0025

lt0

0002

A-M

196

28

275

030

10

028

35

416

13

000

1lt

001

00

037

035

30

010

0034

lt0

0002

A-N

130

85

056

038

70

037

48

291

12

000

1lt

001

00

020

055

60

040

0032

lt0

0002

A-O

378

92

033

027

80

028

18

324

09

000

10

099

lt0

005

020

40

210

0032

lt0

0002

A-P

12

02

429

034

80

040

37

241

13

lt0

001

lt0

010

000

60

471

004

000

19lt

000

02A

-Q1

01

01

810

220

002

84

035

21

2lt

000

1lt

001

00

012

039

30

020

0038

lt0

0002

A-R

127

710

015

073

70

080

41

215

12

000

20

016

022

90

736

004

000

66lt

000

02A

-S1

30

70

370

407

004

72

236

30

7lt

000

10

084

lt0

005

034

90

030

0030

lt0

0002

Whi

teP

13

32

069

017

30

023

19

280

09

lt0

001

030

4lt

000

50

235

008

000

54lt

000

02K

irk

L

104

26

060

019

60

024

09

214

05

000

20

031

003

00

165

001

000

20lt

000

02S

elf

P0

62

90

340

339

004

50

912

11

8lt

000

10

025

lt0

005

013

20

040

0022

lt0

0002

A-W

25

105

180

021

20

027

28

263

07

000

10

048

002

20

399

lt0

010

0033

lt0

0002

A-X

09

30

199

032

80

031

32

250

09

lt0

001

lt0

010

001

60

435

lt0

010

0043

lt0

0002

A-Y

243

146

116

015

90

015

27

378

10

000

20

036

001

40

264

001

000

43lt

000

02A

-Z46

417

11

610

260

002

53

340

91

0lt

000

1lt

001

00

022

035

00

080

0037

lt0

0002

A-A

A2

01

00

220

329

004

31

016

40

6lt

000

10

021

lt0

005

020

30

140

0025

lt0

0002

LD

L2

51

30

590

224

003

20

716

60

50

001

008

30

016

014

10

020

0016

lt0

0002

A-A

E7

53

52

050

291

002

74

237

30

90

083

008

6lt

000

50

448

001

000

34lt

000

02U

DL

-W6

92

00

920

263

002

30

723

11

1lt

000

10

031

lt0

005

010

20

050

0019

lt0

0002

UD

L-E

70

21

093

040

20

035

07

228

10

000

10

039

lt0

005

012

00

240

0025

lt0

0002

Mea

n52

013

41

150

475

004

92

728

61

20

009

007

20

026

034

40

210

0045

lt0

0002

Med

ian

75

32

060

032

80

032

27

263

10

000

20

061

002

20

350

005

000

34lt

000

02M

axim

um70

30

111

04

293

090

028

66

145

94

80

114

088

80

229

073

62

150

0205

lt0

0002

Min

imum

06

02

004

010

80

015

06

121

05

000

10

016

000

60

086

lt0

010

0016

lt0

0002

Limnology of two Canadian High Arctic regions 511

App

endi

x1

Con

tinue

d

Site

Cd

Co

Cr

Cu

FeL

iM

nM

oN

iPb

SrV

Zn

Ca

Mg

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

A-A

lt0

001

000

30

005

000

83

700

001

50

0605

000

30

007

lt0

005

027

000

003

001

050

256

5A

-Blt

000

10

001

000

10

007

090

50

009

002

660

002

000

2lt

000

50

2230

lt0

001

000

353

443

0A

-Clt

000

10

001

lt0

001

000

40

082

000

20

0013

lt0

001

lt0

002

lt0

005

007

62lt

000

1lt

000

128

97

3H

ilgar

dL

lt

000

10

001

000

20

013

185

00

001

002

46lt

000

10

004

lt0

005

005

170

003

000

831

64

0A

-Elt

000

10

002

000

30

009

189

00

005

003

350

001

000

4lt

000

50

0932

000

20

006

310

105

A-F

lt0

001

lt0

001

000

10

012

048

00

002

000

90lt

000

10

002

lt0

005

010

400

001

000

337

814

5A

-Glt

000

10

001

lt0

001

000

60

185

000

40

0104

000

1lt

000

2lt

000

50

1130

lt0

001

000

267

57

2A

-Hlt

000

1lt

000

10

001

000

60

184

000

40

0055

000

20

002

lt0

005

009

99lt

000

10

003

402

114

A-I

lt0

001

lt0

001

lt0

001

000

60

083

000

20

0022

000

1lt

000

2lt

000

50

0543

lt0

001

000

128

85

9A

-Jlt

000

10

001

000

10

005

045

40

002

000

58lt

000

10

002

lt0

005

013

20lt

000

10

003

549

166

A-K

lt0

001

lt0

001

lt0

001

000

80

105

000

10

0032

lt0

001

lt0

002

lt0

005

008

04lt

000

10

001

254

134

Mos

sP

lt

000

1lt

000

1lt

000

10

002

008

8lt

000

10

0031

lt0

001

lt0

002

lt0

005

008

92lt

000

1lt

000

127

89

3A

-Mlt

000

1lt

000

1lt

000

10

008

004

6lt

000

10

0071

lt0

001

lt0

002

lt0

005

014

10lt

000

10

002

618

95

A-N

lt0

001

000

10

002

000

40

073

000

40

0033

000

20

003

lt0

005

011

50lt

000

10

004

209

166

A-O

lt0

001

000

10

001

000

50

330

000

40

0055

000

10

013

lt0

005

012

40lt

000

10

002

411

139

A-P

lt0

001

lt0

001

lt0

001

000

30

077

lt0

001

000

55lt

000

1lt

000

2lt

000

50

0566

lt0

001

lt0

001

290

69

A-Q

lt0

001

lt0

001

lt0

001

000

70

050

lt0

001

000

92lt

000

1lt

000

2lt

000

50

0734

lt0

001

lt0

001

538

55

A-R

lt0

001

lt0

001

lt0

001

000

30

086

000

10

0087

lt0

001

000

2lt

000

50

1140

lt0

001

lt0

001

398

194

A-S

lt0

001

lt0

001

lt0

001

000

50

082

000

10

0059

lt0

001

lt0

002

lt0

005

010

90lt

000

10

001

402

100

Whi

teP

lt0

001

000

1lt

000

10

018

067

7lt

000

10

0270

lt0

001

000

2lt

000

50

0713

lt0

001

000

346

35

1K

irk

L

lt0

001

000

1lt

000

10

004

001

40

001

000

130

001

lt0

002

lt0

005

005

74lt

000

10

001

305

50

Self

Plt

000

1lt

000

10

001

000

60

109

lt0

001

000

95lt

000

1lt

000

2lt

000

50

0263

lt0

001

000

118

81

4A

-Wlt

000

1lt

000

1lt

000

10

004

000

8lt

000

10

0016

lt0

001

lt0

002

lt0

005

006

43lt

000

1lt

000

140

76

5A

-Xlt

000

1lt

000

1lt

000

10

004

001

5lt

000

10

0015

lt0

001

lt0

002

lt0

005

007

21lt

000

10

001

325

74

A-Y

lt0

001

000

10

002

000

80

151

000

20

0033

000

10

003

lt0

005

010

60lt

000

10

004

550

133

A-Z

lt0

001

000

10

001

000

80

238

000

30

0057

000

1lt

000

2lt

000

50

1220

lt0

001

000

255

619

5A

-AA

lt0

001

lt0

001

lt0

001

000

60

270

lt0

001

000

65lt

000

10

024

lt0

005

003

62lt

000

10

003

230

22

LD

Llt

000

1lt

000

1lt

000

10

008

008

4lt

000

10

0047

lt0

001

lt0

002

lt0

005

003

99lt

000

1lt

000

125

12

9A

-AE

lt0

001

lt0

001

lt0

001

000

30

029

000

10

0028

lt0

001

lt0

002

lt0

005

010

60lt

000

1lt

000

151

19

9U

DL

-Wlt

000

10

001

lt0

001

000

20

107

000

10

0055

000

1lt

000

2lt

000

50

0599

lt0

001

000

133

14

8U

DL

-Elt

000

10

001

000

10

009

061

5lt

000

10

0206

lt0

001

000

2lt

000

50

0644

lt0

001

000

436

35

4M

ean

lt0

001

000

10

001

000

60

422

000

20

0104

000

10

002

lt0

005

009

500

000

000

239

111

8M

edia

nlt

000

10

001

000

10

006

010

70

002

000

570

001

000

3lt

000

50

0892

000

30

003

378

93

Max

imum

lt0

001

000

30

005

001

83

700

001

50

0605

000

30

024

lt0

005

027

000

003

001

067

556

5M

inim

umlt

000

10

001

000

10

002

000

80

001

000

130

001

000

2lt

000

50

0263

000

10

001

188

14

512 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x1

Con

tinue

d

Site

Na

KA

gT

PU

TPF

TN

Chl

-aP

OC

T

NT

PC

ond

TpH

EL

EV

Dia

mL

atitu

deL

ongi

tude

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

m gL

ndash1

CH

L-a

m Scm

ndash1

(ƒC

)(m

asl)

(m)

A-A

404

018

40

002

676

77

163

11

718

181

241

2000

78

245

6582

ƒ 29

608cent

N62

ƒ 25

505cent

WA

-B30

90

121

000

233

99

70

738

23

526

122

116

507

83

4530

82ƒ 2

960

5centN

62ƒ 2

693

3centW

A-C

40

05

000

34

83

20

182

06

412

138

115

22

82

4033

82ƒ 2

917

1centN

62ƒ 2

554

6centW

Hil

gard

L

04

04

000

24

81

50

136

lt0

1N

A28

114

09

825

143

750

82ƒ 2

618

0centN

63ƒ 0

663

5centW

A-E

86

26

000

239

82

10

251

09

1178

16

119

015

835

2315

82ƒ 2

713

7centN

63ƒ 0

115

2centW

A-F

69

12

000

121

18

70

544

16

675

126

126

015

865

2510

082

ƒ 27

137cent

N63

ƒ 01

152cent

WA

-G2

20

20

003

125

79

072

11

135

21

581

311

118

1595

782

ƒ 30

196cent

N62

ƒ 30

121cent

WA

-H6

11

70

002

91

53

070

91

632

41

781

259

138

572

982

ƒ 31

208cent

N62

ƒ 30

634cent

WA

-I3

1lt

02

000

213

06

30

710

10

333

155

116

815

865

2580

82ƒ 3

115

2centN

62ƒ 1

772

1centW

A-J

246

12

000

210

83

80

465

09

602

143

142

013

845

3165

82ƒ 2

960

2centN

62ƒ 5

088

7centW

A-K

122

08

000

26

06

20

625

NA

NA

104

125

112

89

120

1282

ƒ 29

025cent

N62

ƒ 54

214cent

WM

oss

P

266

14

000

24

83

00

107

06

180

122

130

26

82

8565

082

ƒ 28

674cent

N62

ƒ 54

242cent

WA

-M9

90

50

002

84

65

038

10

650

21

451

282

88

157

4582

ƒ 28

020cent

N62

ƒ 56

418cent

WA

-N12

52

40

002

87

49

059

31

624

21

681

238

148

740

1882

ƒ 27

790cent

N62

ƒ 55

040cent

WA

-O21

11

20

002

156

48

033

11

027

81

211

282

138

426

300

82ƒ 2

730

0centN

63ƒ 5

605

0centW

A-P

08

lt0

20

002

76

46

051

11

034

81

671

165

68

712

522

582

ƒ 28

334cent

N62

ƒ 32

387cent

WA

-Q0

70

20

002

88

61

042

11

022

01

481

230

68

411

645

82ƒ 2

828

6centN

62ƒ 2

881

6centW

A-R

37

06

000

215

05

70

832

26

283

155

127

87

87

122

2582

ƒ 28

492cent

N62

ƒ 26

528cent

WA

-S3

50

5lt

000

115

65

80

480

10

407

131

121

56

84

4530

82ƒ 2

922

6centN

62ƒ 2

429

1centW

Whi

teP

0

60

2lt

000

18

74

00

562

lt0

1N

A65

118

56

58

311

545

082

ƒ 27

093cent

N62

ƒ 52

388cent

WK

irk

L

52

04

000

13

40

70

220

09

218

165

117

04

82

1390

082

ƒ 27

891cent

N62

ƒ 49

190cent

WS

elf

P0

4lt

02

lt0

001

120

17

020

21

917

81

171

135

48

313

760

082

ƒ 26

552cent

N62

ƒ 01

618cent

WA

-W0

80

2lt

000

16

73

20

474

14

151

171

120

26

85

147

5882

ƒ 27

646cent

N62

ƒ 12

358cent

WA

-X0

90

20

001

57

28

046

61

521

91

821

183

78

611

852

82ƒ 2

856

8centN

62ƒ 1

189

4centW

A-Y

102

04

lt0

001

51

30

031

51

510

61

621

329

78

417

4082

ƒ 29

123cent

N62

ƒ 11

956cent

WA

-Z21

10

7lt

000

16

03

50

375

14

186

163

138

88

84

1823

82ƒ 2

911

1centN

62ƒ 1

310

7centW

A-A

A1

00

2lt

000

113

81

30

267

lt0

1N

A19

112

14

81

3780

82ƒ 2

738

9centN

61ƒ 3

189

9centW

LD

L1

60

2lt

000

14

13

40

256

05

448

162

113

26

83

2011

5082

ƒ 29

066cent

N62

ƒ 38

411cent

WA

-AE

50

02

lt0

001

101

54

056

10

558

21

561

268

98

4535

1482

ƒ 29

855cent

N62

ƒ 23

444cent

WU

DL

-W4

40

40

001

376

21

015

60

552

61

41

179

45

835

2114

0082

ƒ 29

354cent

N62

ƒ 33

954cent

WU

DL

-E4

50

5lt

000

131

63

10

194

09

447

16

117

14

58

221

1400

82ƒ 2

930

1centN

62ƒ 2

913

8centW

Mea

n29

51

60

001

146

45

046

51

137

91

461

331

82

84

6524

2M

edia

n4

50

50

002

91

40

046

51

034

81

481

230

78

445

55M

axim

um40

40

184

000

367

69

71

631

26

1818

110

41

2000

158

914

714

00M

inim

um0

4lt

02

000

13

40

70

107

05

106

14

112

12

81

137

Limnology of two Canadian High Arctic regions 513

App

endi

x2

Mou

ldB

ayW

ater

Che

mis

try

DL

=be

low

dete

ctio

nlim

itsN

A=

valu

eno

tava

ilabl

e

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-A5

924

00

310

880

008

37

510

71

1lt

000

1lt

001

00

122

071

20

010

0143

MB

-B9

323

80

570

552

005

48

016

81

20

002

001

90

054

057

20

090

0154

MB

-C14

10

38

025

065

00

053

68

70

13

000

1lt

001

00

033

049

0lt

001

000

64M

B-D

118

05

50

300

394

004

35

12

91

00

001

lt0

010

002

70

347

008

000

57M

B-E

134

023

10

360

440

004

35

33

50

6lt

000

1lt

001

00

026

037

20

080

0147

MB

-F17

40

75

059

066

50

062

69

49

09

000

1lt

001

00

025

042

70

010

0083

MB

-G17

70

32

017

062

10

068

72

76

09

000

1lt

001

00

028

052

30

010

0074

MB

-H8

220

90

080

669

007

013

714

51

60

001

lt0

010

010

51

360

001

001

28M

B-I

575

025

51

400

683

008

010

121

01

30

001

lt0

010

001

40

645

003

000

33M

B-J

E60

80

381

013

070

80

083

93

203

13

000

1lt

001

00

044

061

10

020

0358

MB

-JW

615

038

40

130

725

007

69

020

31

20

001

lt0

010

001

00

561

004

003

45M

B-K

569

28

006

057

90

072

39

37

06

000

1lt

001

00

050

028

00

030

0035

MB

-L7

035

20

280

568

007

03

04

70

90

001

014

60

025

021

20

580

0113

MB

-M7

8412

50

040

375

004

84

511

21

00

001

lt0

010

003

50

353

031

002

47M

B-N

938

07

005

107

00

111

93

48

13

000

1lt

001

00

075

077

40

260

0064

MB

-O12

70

28

354

614

00

620

135

73

30

000

1lt

001

00

010

077

40

100

0260

MB

-P6

475

10

110

412

004

38

24

90

4lt

000

1lt

001

00

028

047

8lt

001

000

75M

B-Q

607

14

006

049

90

067

61

76

07

000

1lt

001

00

046

045

60

010

0208

MB

-R9

120

60

090

560

005

25

58

10

50

001

lt0

010

001

80

427

003

001

11M

B-S

148

00

60

130

760

010

22

65

40

40

001

lt0

010

005

60

253

001

000

79M

B-T

315

40

036

043

30

060

43

15

08

000

2lt

001

0lt

000

50

221

024

000

91M

B-U

119

011

51

760

430

006

56

917

40

4lt

000

1lt

001

0lt

000

50

555

037

002

18M

B-V

128

014

80

320

465

007

09

922

10

40

001

lt0

010

001

60

743

002

003

22M

B-W

365

14

011

058

70

086

112

225

10

000

1lt

001

00

104

097

20

010

0250

MB

-X5

566

40

460

488

004

87

810

21

00

001

lt0

010

007

10

568

002

001

16M

B-Y

577

98

020

049

50

065

69

81

12

000

1lt

001

00

016

054

70

010

0127

MB

-Z5

882

10

510

601

007

58

08

30

7lt

000

1lt

001

00

058

071

70

020

0073

MB

-AA

747

65

127

043

70

059

78

121

13

000

1lt

001

00

021

062

40

040

0134

MB

-AB

109

04

80

140

382

005

45

48

20

6lt

000

1lt

001

00

010

039

90

020

0144

MB

-AC

624

18

005

044

60

043

26

52

06

lt0

001

lt0

010

002

30

263

001

001

07

514 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-AD

395

12

008

041

30

030

19

51

02

lt0

001

lt0

010

lt0

005

020

00

030

0119

MB

-AE

380

15

013

046

50

046

11

39

02

000

2lt

001

00

026

016

00

080

0091

MB

-AF

233

06

50

227

090

068

43

57

30

8lt

000

1lt

001

0lt

000

50

347

149

004

81M

B-A

G20

100

29

016

042

30

046

78

70

06

lt0

001

lt0

010

001

80

711

003

002

63M

B-A

H21

200

89

010

043

80

046

41

41

05

lt0

001

lt0

010

003

30

376

002

002

23M

ean

249

87

70

410

901

009

66

79

40

90

001

000

50

035

051

50

120

0158

Med

ian

932

40

017

055

20

065

69

76

09

000

1lt

001

00

028

049

00

030

0127

Max

imum

212

0038

43

547

090

068

413

722

53

00

002

014

60

122

136

01

490

0481

Min

imum

315

06

004

037

50

030

11

15

02

000

1D

L

DL

0

160

DL

0

0033

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Alt

000

02lt

000

1lt

000

10

001

000

50

623

000

40

0023

000

1lt

000

2lt

000

50

0373

lt0

001

000

416

0M

B-B

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

305

000

50

0056

lt0

001

lt0

002

lt0

005

005

68lt

000

10

002

255

MB

-Clt

000

02lt

000

1lt

000

1lt

000

10

005

027

80

002

000

12lt

000

1lt

000

2lt

000

50

0287

lt0

001

000

112

1M

B-D

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

262

000

20

0051

lt0

001

lt0

002

lt0

005

001

81lt

000

10

001

59

MB

-Elt

000

02lt

000

1lt

000

1lt

000

10

003

026

40

003

001

51lt

000

1lt

000

2lt

000

50

0323

lt0

001

000

310

9M

B-F

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

746

000

40

0080

lt0

001

lt0

002

lt0

005

003

01lt

000

10

002

104

MB

-Glt

000

02lt

000

1lt

000

10

001

000

70

235

000

30

0181

lt0

001

lt0

002

lt0

005

003

34lt

000

10

001

125

MB

-Hlt

000

02lt

000

1lt

000

1lt

000

10

004

056

00

001

001

23lt

000

10

003

lt0

005

006

44lt

000

10

002

165

MB

-Ilt

000

02lt

000

1lt

000

1lt

000

10

007

014

80

003

003

05lt

000

1lt

000

2lt

000

50

0942

lt0

001

000

223

6M

B-J

Elt

000

02lt

000

1lt

000

1lt

000

10

009

034

80

006

001

19lt

000

1lt

000

2lt

000

50

1250

lt0

001

000

227

9M

B-J

Wlt

000

02lt

000

1lt

000

1lt

000

10

025

076

40

006

001

31lt

000

1lt

000

2lt

000

50

1260

000

10

002

278

MB

-Klt

000

02lt

000

1lt

000

1lt

000

10

006

020

30

001

000

45lt

000

1lt

000

2lt

000

50

0186

lt0

001

000

14

3M

B-L

lt0

0002

lt0

001

lt0

001

000

10

007

115

00

001

001

70lt

000

1lt

000

2lt

000

50

0268

000

10

005

54

MB

-Mlt

000

02lt

000

1lt

000

10

001

000

81

140

000

20

0141

lt0

001

000

2lt

000

50

0557

000

10

003

153

MB

-Nlt

000

02lt

000

1lt

000

10

001

000

41

150

000

40

0169

lt0

001

000

2lt

000

50

0104

lt0

001

000

44

3M

B-O

lt0

0002

000

20

002

000

10

004

720

00

011

011

90lt

000

10

006

lt0

005

002

570

001

000

76

9M

B-P

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

174

000

40

0014

lt0

001

lt0

002

lt0

005

002

62lt

000

10

001

57

MB

-Qlt

000

02lt

000

1lt

000

1lt

000

10

007

024

90

003

000

44lt

000

1lt

000

2lt

000

50

0273

lt0

001

000

16

6

Limnology of two Canadian High Arctic regions 515

App

endi

x2

Con

tinue

d

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Rlt

000

02lt

000

1lt

000

1lt

000

10

008

013

00

004

000

53lt

000

1lt

000

2lt

000

50

0227

lt0

001

000

15

1M

B-S

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

216

000

10

0030

lt0

001

lt0

002

lt0

005

002

04lt

000

10

002

66

MB

-Tlt

000

02lt

000

1lt

000

10

001

000

50

236

000

10

0038

lt0

001

000

4lt

000

50

0093

000

10

012

26

MB

-Ult

000

02lt

000

10

001

000

10

005

085

80

010

001

05lt

000

10

002

lt0

005

008

640

001

000

324

9M

B-V

lt0

0002

lt0

001

lt0

001

000

10

011

015

00

012

001

18lt

000

1lt

000

2lt

000

50

1160

lt0

001

000

129

3M

B-W

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

519

000

30

0175

lt0

001

lt0

002

lt0

005

009

01lt

000

10

002

267

MB

-Xlt

000

02lt

000

1lt

000

1lt

000

10

002

025

50

002

000

30lt

000

1lt

000

2lt

000

50

0400

lt0

001

lt0

001

177

MB

-Ylt

000

02lt

000

1lt

000

1lt

000

10

005

023

40

002

000

18lt

000

1lt

000

2lt

000

50

0404

lt0

001

000

117

1M

B-Z

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

144

000

10

0026

lt0

001

lt0

002

lt0

005

002

78lt

000

10

001

145

MB

-AA

lt0

0002

lt0

001

lt0

001

lt0

001

000

80

125

000

20

0011

lt0

001

lt0

002

lt0

005

005

59lt

000

10

001

218

MB

-AB

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

267

000

40

0050

lt0

001

lt0

002

lt0

005

002

53lt

000

10

001

57

MB

-AC

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

090

000

20

0031

lt0

001

lt0

002

lt0

005

001

55lt

000

10

001

32

MB

-AD

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

083

000

10

0034

lt0

001

lt0

002

lt0

005

001

80lt

000

10

001

44

MB

-AE

lt0

0002

lt0

001

lt0

001

lt0

001

000

60

402

000

10

0114

lt0

001

lt0

002

lt0

005

001

52lt

000

10

002

34

MB

-AF

lt0

0002

000

10

002

000

30

005

490

00

005

011

30lt

000

10

005

000

50

0398

000

50

015

101

MB

-AG

lt0

0002

lt0

001

lt0

001

lt0

001

000

70

351

000

90

0024

lt0

001

000

2lt

000

50

1040

lt0

001

000

215

1M

B-A

Hlt

000

02lt

000

10

001

lt0

001

001

00

596

000

70

1480

lt0

001

000

2lt

000

50

1280

lt0

001

000

315

2M

ean

lt0

0002

000

10

001

000

10

006

072

40

004

001

85lt

000

10

001

lt0

001

004

80lt

000

10

003

132

Med

ian

lt0

0002

000

20

002

000

10

005

026

70

003

000

560

001

000

20

005

003

200

001

000

212

1M

axim

umlt

000

020

002

000

20

003

002

57

200

001

20

1480

000

10

006

000

50

1280

000

50

015

293

Min

imum

DL

D

L

DL

D

L

000

20

083

000

10

0011

DL

D

L

DL

0

0090

DL

D

L

26

Site

Mg

Na

KA

gT

PU

TPF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-A2

62

60

7lt

000

115

511

30

795

01

8800

151

175

70

82

1020

76ƒ1

334

5cent11

9ƒ17

677

centM

B-B

28

43

10

000

19

89

10

645

04

1380

166

110

97

58

48

2076

ƒ12

915cent

119ƒ

169

20cent

MB

-C2

56

10

3lt

000

116

18

80

543

05

1300

134

178

80

80

815

76ƒ1

291

5cent11

9ƒ17

267

centM

B-D

23

48

lt0

2lt

000

19

46

50

390

lt0

1N

A41

153

80

77

812

76ƒ1

291

2cent11

9ƒ17

381

centM

B-E

43

51

02

lt0

001

111

85

041

50

948

91

371

738

07

66

1076

ƒ13

205cent

119ƒ

193

30cent

MB

-F2

97

4lt

02

lt0

001

123

84

048

90

513

301

401

838

07

75

876

ƒ13

351cent

119ƒ

193

47cent

MB

-G3

18

90

6lt

000

111

48

20

591

NA

NA

521

898

07

95

1876

ƒ13

330cent

119ƒ

192

14cent

516 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Mg

Na

KA

gT

PUT

PF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-H5

83

21

1lt

000

117

915

81

430

09

743

180

197

60

81

3112

76ƒ1

494

3cent11

9ƒ23

129

centM

B-I

114

425

16

lt0

001

169

125

072

50

513

661

431

256

60

85

325

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

E11

942

21

8lt

000

113

311

30

694

NA

NA

521

279

60

85

328

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

W12

143

11

8lt

000

113

99

50

637

05

1450

146

127

86

08

63

3576

ƒ14

530cent

119ƒ

220

11cent

MB

-K1

73

10

3lt

000

110

68

60

352

04

1448

133

132

50

74

5025

076

ƒ15

001cent

119ƒ

225

17cent

MB

-L2

05

80

7lt

000

114

67

20

428

lt0

1N

A29

147

60

75

4310

076

ƒ14

792cent

119ƒ

221

40cent

MB

-M6

23

41

1lt

000

110

17

90

401

lt0

1N

A40

189

60

81

135

76ƒ1

436

7cent11

9ƒ17

185

centM

B-N

31

41

13

lt0

001

143

104

088

5lt

01

NA

621

476

07

58

1276

ƒ14

417cent

119ƒ

171

05cent

MB

-O4

24

02

4lt

000

177

923

91

394

NA

NA

181

606

07

08

376

ƒ14

453cent

119ƒ

169

70cent

MB

-P3

33

40

2lt

000

110

47

10

521

lt0

1N

A50

147

60

76

780

76ƒ1

453

2cent11

9ƒ16

884

centM

B-Q

45

28

10

lt0

001

112

74

052

3lt

01

NA

471

536

07

96

1076

ƒ14

862cent

119ƒ

170

98cent

MB

-R6

24

10

9lt

000

18

17

00

479

lt0

1N

A59

160

60

80

625

76ƒ1

486

2cent11

9ƒ17

098

centM

B-S

23

60

05

lt0

001

105

61

035

51

550

71

341

574

07

37

3076

ƒ13

364cent

119ƒ

192

83cent

MB

-T0

81

60

2lt

000

113

811

60

281

NA

NA

201

255

07

078

1076

ƒ15

958cent

119ƒ

261

47cent

MB

-U6

26

02

2lt

000

111

27

70

620

lt0

1N

A55

113

85

08

375

125

76ƒ1

664

1cent11

9ƒ25

841

centM

B-V

75

70

27

lt0

001

118

56

081

32

221

11

691

156

60

83

7714

076

ƒ16

778cent

119ƒ

261

15cent

MB

-W5

81

91

5lt

000

110

56

81

058

24

245

110

11

124

60

84

6212

76ƒ1

670

4cent11

9ƒ25

163

centM

B-X

11

28

lt0

2lt

000

18

56

00

616

NA

NA

721

758

07

962

3076

ƒ16

421cent

119ƒ

260

98cent

MB

-Y1

12

60

2lt

000

19

47

20

612

NA

NA

651

777

57

737

2576

ƒ16

418cent

119ƒ

229

82cent

MB

-Z1

02

60

4lt

000

114

16

40

792

04

1503

156

165

90

78

3618

76ƒ1

618

1cent11

9ƒ21

726

centM

B-A

A1

73

30

7lt

000

17

14

60

683

14

312

196

196

95

81

3338

76ƒ1

613

0cent11

9ƒ21

512

centM

B-A

B6

34

41

0lt

000

17

24

00

453

22

174

163

167

60

80

3230

76ƒ1

476

0cent11

9ƒ17

655

centM

B-A

C3

92

70

8lt

000

17

86

80

306

09

496

139

142

60

77

327

76ƒ1

482

0cent11

9ƒ17

705

centM

B-A

D2

72

10

6lt

000

18

54

40

230

NA

NA

271

386

07

73

7576

ƒ14

758cent

119ƒ

177

34cent

MB

-AE

19

24

05

lt0

001

81

43

020

61

142

31

251

294

07

74

3076

ƒ14

685cent

119ƒ

177

37cent

MB

-AF

50

117

13

lt0

001

117

09

31

031

69

1028

19

199

55

78

470

076

ƒ14

484cent

119ƒ

190

88cent

MB

-AG

173

988

31

lt0

001

114

91

075

7lt

01

NA

661

510

60

75

260

76ƒ1

368

8cent11

9ƒ20

500

centM

B-A

H22

896

24

0lt

000

114

16

70

422

lt0

1N

A30

153

06

07

52

110

76ƒ1

366

5cent11

9ƒ20

222

centM

ean

52

129

10

lt0

001

165

85

061

60

866

31

491

115

64

79

2161

Med

ian

33

41

10

000

111

27

70

591

09

885

147

175

60

78

825

Max

imum

228

988

40

000

111

70

239

143

06

988

001

101

153

09

58

678

700

Min

imum

08

16

DL

D

L

71

40

020

60

117

41

91

254

07

02

3

Limnology of two Canadian High Arctic regions 511

App

endi

x1

Con

tinue

d

Site

Cd

Co

Cr

Cu

FeL

iM

nM

oN

iPb

SrV

Zn

Ca

Mg

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

A-A

lt0

001

000

30

005

000

83

700

001

50

0605

000

30

007

lt0

005

027

000

003

001

050

256

5A

-Blt

000

10

001

000

10

007

090

50

009

002

660

002

000

2lt

000

50

2230

lt0

001

000

353

443

0A

-Clt

000

10

001

lt0

001

000

40

082

000

20

0013

lt0

001

lt0

002

lt0

005

007

62lt

000

1lt

000

128

97

3H

ilgar

dL

lt

000

10

001

000

20

013

185

00

001

002

46lt

000

10

004

lt0

005

005

170

003

000

831

64

0A

-Elt

000

10

002

000

30

009

189

00

005

003

350

001

000

4lt

000

50

0932

000

20

006

310

105

A-F

lt0

001

lt0

001

000

10

012

048

00

002

000

90lt

000

10

002

lt0

005

010

400

001

000

337

814

5A

-Glt

000

10

001

lt0

001

000

60

185

000

40

0104

000

1lt

000

2lt

000

50

1130

lt0

001

000

267

57

2A

-Hlt

000

1lt

000

10

001

000

60

184

000

40

0055

000

20

002

lt0

005

009

99lt

000

10

003

402

114

A-I

lt0

001

lt0

001

lt0

001

000

60

083

000

20

0022

000

1lt

000

2lt

000

50

0543

lt0

001

000

128

85

9A

-Jlt

000

10

001

000

10

005

045

40

002

000

58lt

000

10

002

lt0

005

013

20lt

000

10

003

549

166

A-K

lt0

001

lt0

001

lt0

001

000

80

105

000

10

0032

lt0

001

lt0

002

lt0

005

008

04lt

000

10

001

254

134

Mos

sP

lt

000

1lt

000

1lt

000

10

002

008

8lt

000

10

0031

lt0

001

lt0

002

lt0

005

008

92lt

000

1lt

000

127

89

3A

-Mlt

000

1lt

000

1lt

000

10

008

004

6lt

000

10

0071

lt0

001

lt0

002

lt0

005

014

10lt

000

10

002

618

95

A-N

lt0

001

000

10

002

000

40

073

000

40

0033

000

20

003

lt0

005

011

50lt

000

10

004

209

166

A-O

lt0

001

000

10

001

000

50

330

000

40

0055

000

10

013

lt0

005

012

40lt

000

10

002

411

139

A-P

lt0

001

lt0

001

lt0

001

000

30

077

lt0

001

000

55lt

000

1lt

000

2lt

000

50

0566

lt0

001

lt0

001

290

69

A-Q

lt0

001

lt0

001

lt0

001

000

70

050

lt0

001

000

92lt

000

1lt

000

2lt

000

50

0734

lt0

001

lt0

001

538

55

A-R

lt0

001

lt0

001

lt0

001

000

30

086

000

10

0087

lt0

001

000

2lt

000

50

1140

lt0

001

lt0

001

398

194

A-S

lt0

001

lt0

001

lt0

001

000

50

082

000

10

0059

lt0

001

lt0

002

lt0

005

010

90lt

000

10

001

402

100

Whi

teP

lt0

001

000

1lt

000

10

018

067

7lt

000

10

0270

lt0

001

000

2lt

000

50

0713

lt0

001

000

346

35

1K

irk

L

lt0

001

000

1lt

000

10

004

001

40

001

000

130

001

lt0

002

lt0

005

005

74lt

000

10

001

305

50

Self

Plt

000

1lt

000

10

001

000

60

109

lt0

001

000

95lt

000

1lt

000

2lt

000

50

0263

lt0

001

000

118

81

4A

-Wlt

000

1lt

000

1lt

000

10

004

000

8lt

000

10

0016

lt0

001

lt0

002

lt0

005

006

43lt

000

1lt

000

140

76

5A

-Xlt

000

1lt

000

1lt

000

10

004

001

5lt

000

10

0015

lt0

001

lt0

002

lt0

005

007

21lt

000

10

001

325

74

A-Y

lt0

001

000

10

002

000

80

151

000

20

0033

000

10

003

lt0

005

010

60lt

000

10

004

550

133

A-Z

lt0

001

000

10

001

000

80

238

000

30

0057

000

1lt

000

2lt

000

50

1220

lt0

001

000

255

619

5A

-AA

lt0

001

lt0

001

lt0

001

000

60

270

lt0

001

000

65lt

000

10

024

lt0

005

003

62lt

000

10

003

230

22

LD

Llt

000

1lt

000

1lt

000

10

008

008

4lt

000

10

0047

lt0

001

lt0

002

lt0

005

003

99lt

000

1lt

000

125

12

9A

-AE

lt0

001

lt0

001

lt0

001

000

30

029

000

10

0028

lt0

001

lt0

002

lt0

005

010

60lt

000

1lt

000

151

19

9U

DL

-Wlt

000

10

001

lt0

001

000

20

107

000

10

0055

000

1lt

000

2lt

000

50

0599

lt0

001

000

133

14

8U

DL

-Elt

000

10

001

000

10

009

061

5lt

000

10

0206

lt0

001

000

2lt

000

50

0644

lt0

001

000

436

35

4M

ean

lt0

001

000

10

001

000

60

422

000

20

0104

000

10

002

lt0

005

009

500

000

000

239

111

8M

edia

nlt

000

10

001

000

10

006

010

70

002

000

570

001

000

3lt

000

50

0892

000

30

003

378

93

Max

imum

lt0

001

000

30

005

001

83

700

001

50

0605

000

30

024

lt0

005

027

000

003

001

067

556

5M

inim

umlt

000

10

001

000

10

002

000

80

001

000

130

001

000

2lt

000

50

0263

000

10

001

188

14

512 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x1

Con

tinue

d

Site

Na

KA

gT

PU

TPF

TN

Chl

-aP

OC

T

NT

PC

ond

TpH

EL

EV

Dia

mL

atitu

deL

ongi

tude

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

m gL

ndash1

CH

L-a

m Scm

ndash1

(ƒC

)(m

asl)

(m)

A-A

404

018

40

002

676

77

163

11

718

181

241

2000

78

245

6582

ƒ 29

608cent

N62

ƒ 25

505cent

WA

-B30

90

121

000

233

99

70

738

23

526

122

116

507

83

4530

82ƒ 2

960

5centN

62ƒ 2

693

3centW

A-C

40

05

000

34

83

20

182

06

412

138

115

22

82

4033

82ƒ 2

917

1centN

62ƒ 2

554

6centW

Hil

gard

L

04

04

000

24

81

50

136

lt0

1N

A28

114

09

825

143

750

82ƒ 2

618

0centN

63ƒ 0

663

5centW

A-E

86

26

000

239

82

10

251

09

1178

16

119

015

835

2315

82ƒ 2

713

7centN

63ƒ 0

115

2centW

A-F

69

12

000

121

18

70

544

16

675

126

126

015

865

2510

082

ƒ 27

137cent

N63

ƒ 01

152cent

WA

-G2

20

20

003

125

79

072

11

135

21

581

311

118

1595

782

ƒ 30

196cent

N62

ƒ 30

121cent

WA

-H6

11

70

002

91

53

070

91

632

41

781

259

138

572

982

ƒ 31

208cent

N62

ƒ 30

634cent

WA

-I3

1lt

02

000

213

06

30

710

10

333

155

116

815

865

2580

82ƒ 3

115

2centN

62ƒ 1

772

1centW

A-J

246

12

000

210

83

80

465

09

602

143

142

013

845

3165

82ƒ 2

960

2centN

62ƒ 5

088

7centW

A-K

122

08

000

26

06

20

625

NA

NA

104

125

112

89

120

1282

ƒ 29

025cent

N62

ƒ 54

214cent

WM

oss

P

266

14

000

24

83

00

107

06

180

122

130

26

82

8565

082

ƒ 28

674cent

N62

ƒ 54

242cent

WA

-M9

90

50

002

84

65

038

10

650

21

451

282

88

157

4582

ƒ 28

020cent

N62

ƒ 56

418cent

WA

-N12

52

40

002

87

49

059

31

624

21

681

238

148

740

1882

ƒ 27

790cent

N62

ƒ 55

040cent

WA

-O21

11

20

002

156

48

033

11

027

81

211

282

138

426

300

82ƒ 2

730

0centN

63ƒ 5

605

0centW

A-P

08

lt0

20

002

76

46

051

11

034

81

671

165

68

712

522

582

ƒ 28

334cent

N62

ƒ 32

387cent

WA

-Q0

70

20

002

88

61

042

11

022

01

481

230

68

411

645

82ƒ 2

828

6centN

62ƒ 2

881

6centW

A-R

37

06

000

215

05

70

832

26

283

155

127

87

87

122

2582

ƒ 28

492cent

N62

ƒ 26

528cent

WA

-S3

50

5lt

000

115

65

80

480

10

407

131

121

56

84

4530

82ƒ 2

922

6centN

62ƒ 2

429

1centW

Whi

teP

0

60

2lt

000

18

74

00

562

lt0

1N

A65

118

56

58

311

545

082

ƒ 27

093cent

N62

ƒ 52

388cent

WK

irk

L

52

04

000

13

40

70

220

09

218

165

117

04

82

1390

082

ƒ 27

891cent

N62

ƒ 49

190cent

WS

elf

P0

4lt

02

lt0

001

120

17

020

21

917

81

171

135

48

313

760

082

ƒ 26

552cent

N62

ƒ 01

618cent

WA

-W0

80

2lt

000

16

73

20

474

14

151

171

120

26

85

147

5882

ƒ 27

646cent

N62

ƒ 12

358cent

WA

-X0

90

20

001

57

28

046

61

521

91

821

183

78

611

852

82ƒ 2

856

8centN

62ƒ 1

189

4centW

A-Y

102

04

lt0

001

51

30

031

51

510

61

621

329

78

417

4082

ƒ 29

123cent

N62

ƒ 11

956cent

WA

-Z21

10

7lt

000

16

03

50

375

14

186

163

138

88

84

1823

82ƒ 2

911

1centN

62ƒ 1

310

7centW

A-A

A1

00

2lt

000

113

81

30

267

lt0

1N

A19

112

14

81

3780

82ƒ 2

738

9centN

61ƒ 3

189

9centW

LD

L1

60

2lt

000

14

13

40

256

05

448

162

113

26

83

2011

5082

ƒ 29

066cent

N62

ƒ 38

411cent

WA

-AE

50

02

lt0

001

101

54

056

10

558

21

561

268

98

4535

1482

ƒ 29

855cent

N62

ƒ 23

444cent

WU

DL

-W4

40

40

001

376

21

015

60

552

61

41

179

45

835

2114

0082

ƒ 29

354cent

N62

ƒ 33

954cent

WU

DL

-E4

50

5lt

000

131

63

10

194

09

447

16

117

14

58

221

1400

82ƒ 2

930

1centN

62ƒ 2

913

8centW

Mea

n29

51

60

001

146

45

046

51

137

91

461

331

82

84

6524

2M

edia

n4

50

50

002

91

40

046

51

034

81

481

230

78

445

55M

axim

um40

40

184

000

367

69

71

631

26

1818

110

41

2000

158

914

714

00M

inim

um0

4lt

02

000

13

40

70

107

05

106

14

112

12

81

137

Limnology of two Canadian High Arctic regions 513

App

endi

x2

Mou

ldB

ayW

ater

Che

mis

try

DL

=be

low

dete

ctio

nlim

itsN

A=

valu

eno

tava

ilabl

e

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-A5

924

00

310

880

008

37

510

71

1lt

000

1lt

001

00

122

071

20

010

0143

MB

-B9

323

80

570

552

005

48

016

81

20

002

001

90

054

057

20

090

0154

MB

-C14

10

38

025

065

00

053

68

70

13

000

1lt

001

00

033

049

0lt

001

000

64M

B-D

118

05

50

300

394

004

35

12

91

00

001

lt0

010

002

70

347

008

000

57M

B-E

134

023

10

360

440

004

35

33

50

6lt

000

1lt

001

00

026

037

20

080

0147

MB

-F17

40

75

059

066

50

062

69

49

09

000

1lt

001

00

025

042

70

010

0083

MB

-G17

70

32

017

062

10

068

72

76

09

000

1lt

001

00

028

052

30

010

0074

MB

-H8

220

90

080

669

007

013

714

51

60

001

lt0

010

010

51

360

001

001

28M

B-I

575

025

51

400

683

008

010

121

01

30

001

lt0

010

001

40

645

003

000

33M

B-J

E60

80

381

013

070

80

083

93

203

13

000

1lt

001

00

044

061

10

020

0358

MB

-JW

615

038

40

130

725

007

69

020

31

20

001

lt0

010

001

00

561

004

003

45M

B-K

569

28

006

057

90

072

39

37

06

000

1lt

001

00

050

028

00

030

0035

MB

-L7

035

20

280

568

007

03

04

70

90

001

014

60

025

021

20

580

0113

MB

-M7

8412

50

040

375

004

84

511

21

00

001

lt0

010

003

50

353

031

002

47M

B-N

938

07

005

107

00

111

93

48

13

000

1lt

001

00

075

077

40

260

0064

MB

-O12

70

28

354

614

00

620

135

73

30

000

1lt

001

00

010

077

40

100

0260

MB

-P6

475

10

110

412

004

38

24

90

4lt

000

1lt

001

00

028

047

8lt

001

000

75M

B-Q

607

14

006

049

90

067

61

76

07

000

1lt

001

00

046

045

60

010

0208

MB

-R9

120

60

090

560

005

25

58

10

50

001

lt0

010

001

80

427

003

001

11M

B-S

148

00

60

130

760

010

22

65

40

40

001

lt0

010

005

60

253

001

000

79M

B-T

315

40

036

043

30

060

43

15

08

000

2lt

001

0lt

000

50

221

024

000

91M

B-U

119

011

51

760

430

006

56

917

40

4lt

000

1lt

001

0lt

000

50

555

037

002

18M

B-V

128

014

80

320

465

007

09

922

10

40

001

lt0

010

001

60

743

002

003

22M

B-W

365

14

011

058

70

086

112

225

10

000

1lt

001

00

104

097

20

010

0250

MB

-X5

566

40

460

488

004

87

810

21

00

001

lt0

010

007

10

568

002

001

16M

B-Y

577

98

020

049

50

065

69

81

12

000

1lt

001

00

016

054

70

010

0127

MB

-Z5

882

10

510

601

007

58

08

30

7lt

000

1lt

001

00

058

071

70

020

0073

MB

-AA

747

65

127

043

70

059

78

121

13

000

1lt

001

00

021

062

40

040

0134

MB

-AB

109

04

80

140

382

005

45

48

20

6lt

000

1lt

001

00

010

039

90

020

0144

MB

-AC

624

18

005

044

60

043

26

52

06

lt0

001

lt0

010

002

30

263

001

001

07

514 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-AD

395

12

008

041

30

030

19

51

02

lt0

001

lt0

010

lt0

005

020

00

030

0119

MB

-AE

380

15

013

046

50

046

11

39

02

000

2lt

001

00

026

016

00

080

0091

MB

-AF

233

06

50

227

090

068

43

57

30

8lt

000

1lt

001

0lt

000

50

347

149

004

81M

B-A

G20

100

29

016

042

30

046

78

70

06

lt0

001

lt0

010

001

80

711

003

002

63M

B-A

H21

200

89

010

043

80

046

41

41

05

lt0

001

lt0

010

003

30

376

002

002

23M

ean

249

87

70

410

901

009

66

79

40

90

001

000

50

035

051

50

120

0158

Med

ian

932

40

017

055

20

065

69

76

09

000

1lt

001

00

028

049

00

030

0127

Max

imum

212

0038

43

547

090

068

413

722

53

00

002

014

60

122

136

01

490

0481

Min

imum

315

06

004

037

50

030

11

15

02

000

1D

L

DL

0

160

DL

0

0033

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Alt

000

02lt

000

1lt

000

10

001

000

50

623

000

40

0023

000

1lt

000

2lt

000

50

0373

lt0

001

000

416

0M

B-B

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

305

000

50

0056

lt0

001

lt0

002

lt0

005

005

68lt

000

10

002

255

MB

-Clt

000

02lt

000

1lt

000

1lt

000

10

005

027

80

002

000

12lt

000

1lt

000

2lt

000

50

0287

lt0

001

000

112

1M

B-D

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

262

000

20

0051

lt0

001

lt0

002

lt0

005

001

81lt

000

10

001

59

MB

-Elt

000

02lt

000

1lt

000

1lt

000

10

003

026

40

003

001

51lt

000

1lt

000

2lt

000

50

0323

lt0

001

000

310

9M

B-F

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

746

000

40

0080

lt0

001

lt0

002

lt0

005

003

01lt

000

10

002

104

MB

-Glt

000

02lt

000

1lt

000

10

001

000

70

235

000

30

0181

lt0

001

lt0

002

lt0

005

003

34lt

000

10

001

125

MB

-Hlt

000

02lt

000

1lt

000

1lt

000

10

004

056

00

001

001

23lt

000

10

003

lt0

005

006

44lt

000

10

002

165

MB

-Ilt

000

02lt

000

1lt

000

1lt

000

10

007

014

80

003

003

05lt

000

1lt

000

2lt

000

50

0942

lt0

001

000

223

6M

B-J

Elt

000

02lt

000

1lt

000

1lt

000

10

009

034

80

006

001

19lt

000

1lt

000

2lt

000

50

1250

lt0

001

000

227

9M

B-J

Wlt

000

02lt

000

1lt

000

1lt

000

10

025

076

40

006

001

31lt

000

1lt

000

2lt

000

50

1260

000

10

002

278

MB

-Klt

000

02lt

000

1lt

000

1lt

000

10

006

020

30

001

000

45lt

000

1lt

000

2lt

000

50

0186

lt0

001

000

14

3M

B-L

lt0

0002

lt0

001

lt0

001

000

10

007

115

00

001

001

70lt

000

1lt

000

2lt

000

50

0268

000

10

005

54

MB

-Mlt

000

02lt

000

1lt

000

10

001

000

81

140

000

20

0141

lt0

001

000

2lt

000

50

0557

000

10

003

153

MB

-Nlt

000

02lt

000

1lt

000

10

001

000

41

150

000

40

0169

lt0

001

000

2lt

000

50

0104

lt0

001

000

44

3M

B-O

lt0

0002

000

20

002

000

10

004

720

00

011

011

90lt

000

10

006

lt0

005

002

570

001

000

76

9M

B-P

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

174

000

40

0014

lt0

001

lt0

002

lt0

005

002

62lt

000

10

001

57

MB

-Qlt

000

02lt

000

1lt

000

1lt

000

10

007

024

90

003

000

44lt

000

1lt

000

2lt

000

50

0273

lt0

001

000

16

6

Limnology of two Canadian High Arctic regions 515

App

endi

x2

Con

tinue

d

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Rlt

000

02lt

000

1lt

000

1lt

000

10

008

013

00

004

000

53lt

000

1lt

000

2lt

000

50

0227

lt0

001

000

15

1M

B-S

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

216

000

10

0030

lt0

001

lt0

002

lt0

005

002

04lt

000

10

002

66

MB

-Tlt

000

02lt

000

1lt

000

10

001

000

50

236

000

10

0038

lt0

001

000

4lt

000

50

0093

000

10

012

26

MB

-Ult

000

02lt

000

10

001

000

10

005

085

80

010

001

05lt

000

10

002

lt0

005

008

640

001

000

324

9M

B-V

lt0

0002

lt0

001

lt0

001

000

10

011

015

00

012

001

18lt

000

1lt

000

2lt

000

50

1160

lt0

001

000

129

3M

B-W

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

519

000

30

0175

lt0

001

lt0

002

lt0

005

009

01lt

000

10

002

267

MB

-Xlt

000

02lt

000

1lt

000

1lt

000

10

002

025

50

002

000

30lt

000

1lt

000

2lt

000

50

0400

lt0

001

lt0

001

177

MB

-Ylt

000

02lt

000

1lt

000

1lt

000

10

005

023

40

002

000

18lt

000

1lt

000

2lt

000

50

0404

lt0

001

000

117

1M

B-Z

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

144

000

10

0026

lt0

001

lt0

002

lt0

005

002

78lt

000

10

001

145

MB

-AA

lt0

0002

lt0

001

lt0

001

lt0

001

000

80

125

000

20

0011

lt0

001

lt0

002

lt0

005

005

59lt

000

10

001

218

MB

-AB

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

267

000

40

0050

lt0

001

lt0

002

lt0

005

002

53lt

000

10

001

57

MB

-AC

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

090

000

20

0031

lt0

001

lt0

002

lt0

005

001

55lt

000

10

001

32

MB

-AD

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

083

000

10

0034

lt0

001

lt0

002

lt0

005

001

80lt

000

10

001

44

MB

-AE

lt0

0002

lt0

001

lt0

001

lt0

001

000

60

402

000

10

0114

lt0

001

lt0

002

lt0

005

001

52lt

000

10

002

34

MB

-AF

lt0

0002

000

10

002

000

30

005

490

00

005

011

30lt

000

10

005

000

50

0398

000

50

015

101

MB

-AG

lt0

0002

lt0

001

lt0

001

lt0

001

000

70

351

000

90

0024

lt0

001

000

2lt

000

50

1040

lt0

001

000

215

1M

B-A

Hlt

000

02lt

000

10

001

lt0

001

001

00

596

000

70

1480

lt0

001

000

2lt

000

50

1280

lt0

001

000

315

2M

ean

lt0

0002

000

10

001

000

10

006

072

40

004

001

85lt

000

10

001

lt0

001

004

80lt

000

10

003

132

Med

ian

lt0

0002

000

20

002

000

10

005

026

70

003

000

560

001

000

20

005

003

200

001

000

212

1M

axim

umlt

000

020

002

000

20

003

002

57

200

001

20

1480

000

10

006

000

50

1280

000

50

015

293

Min

imum

DL

D

L

DL

D

L

000

20

083

000

10

0011

DL

D

L

DL

0

0090

DL

D

L

26

Site

Mg

Na

KA

gT

PU

TPF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-A2

62

60

7lt

000

115

511

30

795

01

8800

151

175

70

82

1020

76ƒ1

334

5cent11

9ƒ17

677

centM

B-B

28

43

10

000

19

89

10

645

04

1380

166

110

97

58

48

2076

ƒ12

915cent

119ƒ

169

20cent

MB

-C2

56

10

3lt

000

116

18

80

543

05

1300

134

178

80

80

815

76ƒ1

291

5cent11

9ƒ17

267

centM

B-D

23

48

lt0

2lt

000

19

46

50

390

lt0

1N

A41

153

80

77

812

76ƒ1

291

2cent11

9ƒ17

381

centM

B-E

43

51

02

lt0

001

111

85

041

50

948

91

371

738

07

66

1076

ƒ13

205cent

119ƒ

193

30cent

MB

-F2

97

4lt

02

lt0

001

123

84

048

90

513

301

401

838

07

75

876

ƒ13

351cent

119ƒ

193

47cent

MB

-G3

18

90

6lt

000

111

48

20

591

NA

NA

521

898

07

95

1876

ƒ13

330cent

119ƒ

192

14cent

516 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Mg

Na

KA

gT

PUT

PF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-H5

83

21

1lt

000

117

915

81

430

09

743

180

197

60

81

3112

76ƒ1

494

3cent11

9ƒ23

129

centM

B-I

114

425

16

lt0

001

169

125

072

50

513

661

431

256

60

85

325

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

E11

942

21

8lt

000

113

311

30

694

NA

NA

521

279

60

85

328

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

W12

143

11

8lt

000

113

99

50

637

05

1450

146

127

86

08

63

3576

ƒ14

530cent

119ƒ

220

11cent

MB

-K1

73

10

3lt

000

110

68

60

352

04

1448

133

132

50

74

5025

076

ƒ15

001cent

119ƒ

225

17cent

MB

-L2

05

80

7lt

000

114

67

20

428

lt0

1N

A29

147

60

75

4310

076

ƒ14

792cent

119ƒ

221

40cent

MB

-M6

23

41

1lt

000

110

17

90

401

lt0

1N

A40

189

60

81

135

76ƒ1

436

7cent11

9ƒ17

185

centM

B-N

31

41

13

lt0

001

143

104

088

5lt

01

NA

621

476

07

58

1276

ƒ14

417cent

119ƒ

171

05cent

MB

-O4

24

02

4lt

000

177

923

91

394

NA

NA

181

606

07

08

376

ƒ14

453cent

119ƒ

169

70cent

MB

-P3

33

40

2lt

000

110

47

10

521

lt0

1N

A50

147

60

76

780

76ƒ1

453

2cent11

9ƒ16

884

centM

B-Q

45

28

10

lt0

001

112

74

052

3lt

01

NA

471

536

07

96

1076

ƒ14

862cent

119ƒ

170

98cent

MB

-R6

24

10

9lt

000

18

17

00

479

lt0

1N

A59

160

60

80

625

76ƒ1

486

2cent11

9ƒ17

098

centM

B-S

23

60

05

lt0

001

105

61

035

51

550

71

341

574

07

37

3076

ƒ13

364cent

119ƒ

192

83cent

MB

-T0

81

60

2lt

000

113

811

60

281

NA

NA

201

255

07

078

1076

ƒ15

958cent

119ƒ

261

47cent

MB

-U6

26

02

2lt

000

111

27

70

620

lt0

1N

A55

113

85

08

375

125

76ƒ1

664

1cent11

9ƒ25

841

centM

B-V

75

70

27

lt0

001

118

56

081

32

221

11

691

156

60

83

7714

076

ƒ16

778cent

119ƒ

261

15cent

MB

-W5

81

91

5lt

000

110

56

81

058

24

245

110

11

124

60

84

6212

76ƒ1

670

4cent11

9ƒ25

163

centM

B-X

11

28

lt0

2lt

000

18

56

00

616

NA

NA

721

758

07

962

3076

ƒ16

421cent

119ƒ

260

98cent

MB

-Y1

12

60

2lt

000

19

47

20

612

NA

NA

651

777

57

737

2576

ƒ16

418cent

119ƒ

229

82cent

MB

-Z1

02

60

4lt

000

114

16

40

792

04

1503

156

165

90

78

3618

76ƒ1

618

1cent11

9ƒ21

726

centM

B-A

A1

73

30

7lt

000

17

14

60

683

14

312

196

196

95

81

3338

76ƒ1

613

0cent11

9ƒ21

512

centM

B-A

B6

34

41

0lt

000

17

24

00

453

22

174

163

167

60

80

3230

76ƒ1

476

0cent11

9ƒ17

655

centM

B-A

C3

92

70

8lt

000

17

86

80

306

09

496

139

142

60

77

327

76ƒ1

482

0cent11

9ƒ17

705

centM

B-A

D2

72

10

6lt

000

18

54

40

230

NA

NA

271

386

07

73

7576

ƒ14

758cent

119ƒ

177

34cent

MB

-AE

19

24

05

lt0

001

81

43

020

61

142

31

251

294

07

74

3076

ƒ14

685cent

119ƒ

177

37cent

MB

-AF

50

117

13

lt0

001

117

09

31

031

69

1028

19

199

55

78

470

076

ƒ14

484cent

119ƒ

190

88cent

MB

-AG

173

988

31

lt0

001

114

91

075

7lt

01

NA

661

510

60

75

260

76ƒ1

368

8cent11

9ƒ20

500

centM

B-A

H22

896

24

0lt

000

114

16

70

422

lt0

1N

A30

153

06

07

52

110

76ƒ1

366

5cent11

9ƒ20

222

centM

ean

52

129

10

lt0

001

165

85

061

60

866

31

491

115

64

79

2161

Med

ian

33

41

10

000

111

27

70

591

09

885

147

175

60

78

825

Max

imum

228

988

40

000

111

70

239

143

06

988

001

101

153

09

58

678

700

Min

imum

08

16

DL

D

L

71

40

020

60

117

41

91

254

07

02

3

512 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x1

Con

tinue

d

Site

Na

KA

gT

PU

TPF

TN

Chl

-aP

OC

T

NT

PC

ond

TpH

EL

EV

Dia

mL

atitu

deL

ongi

tude

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

m gL

ndash1

CH

L-a

m Scm

ndash1

(ƒC

)(m

asl)

(m)

A-A

404

018

40

002

676

77

163

11

718

181

241

2000

78

245

6582

ƒ 29

608cent

N62

ƒ 25

505cent

WA

-B30

90

121

000

233

99

70

738

23

526

122

116

507

83

4530

82ƒ 2

960

5centN

62ƒ 2

693

3centW

A-C

40

05

000

34

83

20

182

06

412

138

115

22

82

4033

82ƒ 2

917

1centN

62ƒ 2

554

6centW

Hil

gard

L

04

04

000

24

81

50

136

lt0

1N

A28

114

09

825

143

750

82ƒ 2

618

0centN

63ƒ 0

663

5centW

A-E

86

26

000

239

82

10

251

09

1178

16

119

015

835

2315

82ƒ 2

713

7centN

63ƒ 0

115

2centW

A-F

69

12

000

121

18

70

544

16

675

126

126

015

865

2510

082

ƒ 27

137cent

N63

ƒ 01

152cent

WA

-G2

20

20

003

125

79

072

11

135

21

581

311

118

1595

782

ƒ 30

196cent

N62

ƒ 30

121cent

WA

-H6

11

70

002

91

53

070

91

632

41

781

259

138

572

982

ƒ 31

208cent

N62

ƒ 30

634cent

WA

-I3

1lt

02

000

213

06

30

710

10

333

155

116

815

865

2580

82ƒ 3

115

2centN

62ƒ 1

772

1centW

A-J

246

12

000

210

83

80

465

09

602

143

142

013

845

3165

82ƒ 2

960

2centN

62ƒ 5

088

7centW

A-K

122

08

000

26

06

20

625

NA

NA

104

125

112

89

120

1282

ƒ 29

025cent

N62

ƒ 54

214cent

WM

oss

P

266

14

000

24

83

00

107

06

180

122

130

26

82

8565

082

ƒ 28

674cent

N62

ƒ 54

242cent

WA

-M9

90

50

002

84

65

038

10

650

21

451

282

88

157

4582

ƒ 28

020cent

N62

ƒ 56

418cent

WA

-N12

52

40

002

87

49

059

31

624

21

681

238

148

740

1882

ƒ 27

790cent

N62

ƒ 55

040cent

WA

-O21

11

20

002

156

48

033

11

027

81

211

282

138

426

300

82ƒ 2

730

0centN

63ƒ 5

605

0centW

A-P

08

lt0

20

002

76

46

051

11

034

81

671

165

68

712

522

582

ƒ 28

334cent

N62

ƒ 32

387cent

WA

-Q0

70

20

002

88

61

042

11

022

01

481

230

68

411

645

82ƒ 2

828

6centN

62ƒ 2

881

6centW

A-R

37

06

000

215

05

70

832

26

283

155

127

87

87

122

2582

ƒ 28

492cent

N62

ƒ 26

528cent

WA

-S3

50

5lt

000

115

65

80

480

10

407

131

121

56

84

4530

82ƒ 2

922

6centN

62ƒ 2

429

1centW

Whi

teP

0

60

2lt

000

18

74

00

562

lt0

1N

A65

118

56

58

311

545

082

ƒ 27

093cent

N62

ƒ 52

388cent

WK

irk

L

52

04

000

13

40

70

220

09

218

165

117

04

82

1390

082

ƒ 27

891cent

N62

ƒ 49

190cent

WS

elf

P0

4lt

02

lt0

001

120

17

020

21

917

81

171

135

48

313

760

082

ƒ 26

552cent

N62

ƒ 01

618cent

WA

-W0

80

2lt

000

16

73

20

474

14

151

171

120

26

85

147

5882

ƒ 27

646cent

N62

ƒ 12

358cent

WA

-X0

90

20

001

57

28

046

61

521

91

821

183

78

611

852

82ƒ 2

856

8centN

62ƒ 1

189

4centW

A-Y

102

04

lt0

001

51

30

031

51

510

61

621

329

78

417

4082

ƒ 29

123cent

N62

ƒ 11

956cent

WA

-Z21

10

7lt

000

16

03

50

375

14

186

163

138

88

84

1823

82ƒ 2

911

1centN

62ƒ 1

310

7centW

A-A

A1

00

2lt

000

113

81

30

267

lt0

1N

A19

112

14

81

3780

82ƒ 2

738

9centN

61ƒ 3

189

9centW

LD

L1

60

2lt

000

14

13

40

256

05

448

162

113

26

83

2011

5082

ƒ 29

066cent

N62

ƒ 38

411cent

WA

-AE

50

02

lt0

001

101

54

056

10

558

21

561

268

98

4535

1482

ƒ 29

855cent

N62

ƒ 23

444cent

WU

DL

-W4

40

40

001

376

21

015

60

552

61

41

179

45

835

2114

0082

ƒ 29

354cent

N62

ƒ 33

954cent

WU

DL

-E4

50

5lt

000

131

63

10

194

09

447

16

117

14

58

221

1400

82ƒ 2

930

1centN

62ƒ 2

913

8centW

Mea

n29

51

60

001

146

45

046

51

137

91

461

331

82

84

6524

2M

edia

n4

50

50

002

91

40

046

51

034

81

481

230

78

445

55M

axim

um40

40

184

000

367

69

71

631

26

1818

110

41

2000

158

914

714

00M

inim

um0

4lt

02

000

13

40

70

107

05

106

14

112

12

81

137

Limnology of two Canadian High Arctic regions 513

App

endi

x2

Mou

ldB

ayW

ater

Che

mis

try

DL

=be

low

dete

ctio

nlim

itsN

A=

valu

eno

tava

ilabl

e

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-A5

924

00

310

880

008

37

510

71

1lt

000

1lt

001

00

122

071

20

010

0143

MB

-B9

323

80

570

552

005

48

016

81

20

002

001

90

054

057

20

090

0154

MB

-C14

10

38

025

065

00

053

68

70

13

000

1lt

001

00

033

049

0lt

001

000

64M

B-D

118

05

50

300

394

004

35

12

91

00

001

lt0

010

002

70

347

008

000

57M

B-E

134

023

10

360

440

004

35

33

50

6lt

000

1lt

001

00

026

037

20

080

0147

MB

-F17

40

75

059

066

50

062

69

49

09

000

1lt

001

00

025

042

70

010

0083

MB

-G17

70

32

017

062

10

068

72

76

09

000

1lt

001

00

028

052

30

010

0074

MB

-H8

220

90

080

669

007

013

714

51

60

001

lt0

010

010

51

360

001

001

28M

B-I

575

025

51

400

683

008

010

121

01

30

001

lt0

010

001

40

645

003

000

33M

B-J

E60

80

381

013

070

80

083

93

203

13

000

1lt

001

00

044

061

10

020

0358

MB

-JW

615

038

40

130

725

007

69

020

31

20

001

lt0

010

001

00

561

004

003

45M

B-K

569

28

006

057

90

072

39

37

06

000

1lt

001

00

050

028

00

030

0035

MB

-L7

035

20

280

568

007

03

04

70

90

001

014

60

025

021

20

580

0113

MB

-M7

8412

50

040

375

004

84

511

21

00

001

lt0

010

003

50

353

031

002

47M

B-N

938

07

005

107

00

111

93

48

13

000

1lt

001

00

075

077

40

260

0064

MB

-O12

70

28

354

614

00

620

135

73

30

000

1lt

001

00

010

077

40

100

0260

MB

-P6

475

10

110

412

004

38

24

90

4lt

000

1lt

001

00

028

047

8lt

001

000

75M

B-Q

607

14

006

049

90

067

61

76

07

000

1lt

001

00

046

045

60

010

0208

MB

-R9

120

60

090

560

005

25

58

10

50

001

lt0

010

001

80

427

003

001

11M

B-S

148

00

60

130

760

010

22

65

40

40

001

lt0

010

005

60

253

001

000

79M

B-T

315

40

036

043

30

060

43

15

08

000

2lt

001

0lt

000

50

221

024

000

91M

B-U

119

011

51

760

430

006

56

917

40

4lt

000

1lt

001

0lt

000

50

555

037

002

18M

B-V

128

014

80

320

465

007

09

922

10

40

001

lt0

010

001

60

743

002

003

22M

B-W

365

14

011

058

70

086

112

225

10

000

1lt

001

00

104

097

20

010

0250

MB

-X5

566

40

460

488

004

87

810

21

00

001

lt0

010

007

10

568

002

001

16M

B-Y

577

98

020

049

50

065

69

81

12

000

1lt

001

00

016

054

70

010

0127

MB

-Z5

882

10

510

601

007

58

08

30

7lt

000

1lt

001

00

058

071

70

020

0073

MB

-AA

747

65

127

043

70

059

78

121

13

000

1lt

001

00

021

062

40

040

0134

MB

-AB

109

04

80

140

382

005

45

48

20

6lt

000

1lt

001

00

010

039

90

020

0144

MB

-AC

624

18

005

044

60

043

26

52

06

lt0

001

lt0

010

002

30

263

001

001

07

514 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-AD

395

12

008

041

30

030

19

51

02

lt0

001

lt0

010

lt0

005

020

00

030

0119

MB

-AE

380

15

013

046

50

046

11

39

02

000

2lt

001

00

026

016

00

080

0091

MB

-AF

233

06

50

227

090

068

43

57

30

8lt

000

1lt

001

0lt

000

50

347

149

004

81M

B-A

G20

100

29

016

042

30

046

78

70

06

lt0

001

lt0

010

001

80

711

003

002

63M

B-A

H21

200

89

010

043

80

046

41

41

05

lt0

001

lt0

010

003

30

376

002

002

23M

ean

249

87

70

410

901

009

66

79

40

90

001

000

50

035

051

50

120

0158

Med

ian

932

40

017

055

20

065

69

76

09

000

1lt

001

00

028

049

00

030

0127

Max

imum

212

0038

43

547

090

068

413

722

53

00

002

014

60

122

136

01

490

0481

Min

imum

315

06

004

037

50

030

11

15

02

000

1D

L

DL

0

160

DL

0

0033

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Alt

000

02lt

000

1lt

000

10

001

000

50

623

000

40

0023

000

1lt

000

2lt

000

50

0373

lt0

001

000

416

0M

B-B

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

305

000

50

0056

lt0

001

lt0

002

lt0

005

005

68lt

000

10

002

255

MB

-Clt

000

02lt

000

1lt

000

1lt

000

10

005

027

80

002

000

12lt

000

1lt

000

2lt

000

50

0287

lt0

001

000

112

1M

B-D

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

262

000

20

0051

lt0

001

lt0

002

lt0

005

001

81lt

000

10

001

59

MB

-Elt

000

02lt

000

1lt

000

1lt

000

10

003

026

40

003

001

51lt

000

1lt

000

2lt

000

50

0323

lt0

001

000

310

9M

B-F

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

746

000

40

0080

lt0

001

lt0

002

lt0

005

003

01lt

000

10

002

104

MB

-Glt

000

02lt

000

1lt

000

10

001

000

70

235

000

30

0181

lt0

001

lt0

002

lt0

005

003

34lt

000

10

001

125

MB

-Hlt

000

02lt

000

1lt

000

1lt

000

10

004

056

00

001

001

23lt

000

10

003

lt0

005

006

44lt

000

10

002

165

MB

-Ilt

000

02lt

000

1lt

000

1lt

000

10

007

014

80

003

003

05lt

000

1lt

000

2lt

000

50

0942

lt0

001

000

223

6M

B-J

Elt

000

02lt

000

1lt

000

1lt

000

10

009

034

80

006

001

19lt

000

1lt

000

2lt

000

50

1250

lt0

001

000

227

9M

B-J

Wlt

000

02lt

000

1lt

000

1lt

000

10

025

076

40

006

001

31lt

000

1lt

000

2lt

000

50

1260

000

10

002

278

MB

-Klt

000

02lt

000

1lt

000

1lt

000

10

006

020

30

001

000

45lt

000

1lt

000

2lt

000

50

0186

lt0

001

000

14

3M

B-L

lt0

0002

lt0

001

lt0

001

000

10

007

115

00

001

001

70lt

000

1lt

000

2lt

000

50

0268

000

10

005

54

MB

-Mlt

000

02lt

000

1lt

000

10

001

000

81

140

000

20

0141

lt0

001

000

2lt

000

50

0557

000

10

003

153

MB

-Nlt

000

02lt

000

1lt

000

10

001

000

41

150

000

40

0169

lt0

001

000

2lt

000

50

0104

lt0

001

000

44

3M

B-O

lt0

0002

000

20

002

000

10

004

720

00

011

011

90lt

000

10

006

lt0

005

002

570

001

000

76

9M

B-P

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

174

000

40

0014

lt0

001

lt0

002

lt0

005

002

62lt

000

10

001

57

MB

-Qlt

000

02lt

000

1lt

000

1lt

000

10

007

024

90

003

000

44lt

000

1lt

000

2lt

000

50

0273

lt0

001

000

16

6

Limnology of two Canadian High Arctic regions 515

App

endi

x2

Con

tinue

d

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Rlt

000

02lt

000

1lt

000

1lt

000

10

008

013

00

004

000

53lt

000

1lt

000

2lt

000

50

0227

lt0

001

000

15

1M

B-S

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

216

000

10

0030

lt0

001

lt0

002

lt0

005

002

04lt

000

10

002

66

MB

-Tlt

000

02lt

000

1lt

000

10

001

000

50

236

000

10

0038

lt0

001

000

4lt

000

50

0093

000

10

012

26

MB

-Ult

000

02lt

000

10

001

000

10

005

085

80

010

001

05lt

000

10

002

lt0

005

008

640

001

000

324

9M

B-V

lt0

0002

lt0

001

lt0

001

000

10

011

015

00

012

001

18lt

000

1lt

000

2lt

000

50

1160

lt0

001

000

129

3M

B-W

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

519

000

30

0175

lt0

001

lt0

002

lt0

005

009

01lt

000

10

002

267

MB

-Xlt

000

02lt

000

1lt

000

1lt

000

10

002

025

50

002

000

30lt

000

1lt

000

2lt

000

50

0400

lt0

001

lt0

001

177

MB

-Ylt

000

02lt

000

1lt

000

1lt

000

10

005

023

40

002

000

18lt

000

1lt

000

2lt

000

50

0404

lt0

001

000

117

1M

B-Z

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

144

000

10

0026

lt0

001

lt0

002

lt0

005

002

78lt

000

10

001

145

MB

-AA

lt0

0002

lt0

001

lt0

001

lt0

001

000

80

125

000

20

0011

lt0

001

lt0

002

lt0

005

005

59lt

000

10

001

218

MB

-AB

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

267

000

40

0050

lt0

001

lt0

002

lt0

005

002

53lt

000

10

001

57

MB

-AC

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

090

000

20

0031

lt0

001

lt0

002

lt0

005

001

55lt

000

10

001

32

MB

-AD

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

083

000

10

0034

lt0

001

lt0

002

lt0

005

001

80lt

000

10

001

44

MB

-AE

lt0

0002

lt0

001

lt0

001

lt0

001

000

60

402

000

10

0114

lt0

001

lt0

002

lt0

005

001

52lt

000

10

002

34

MB

-AF

lt0

0002

000

10

002

000

30

005

490

00

005

011

30lt

000

10

005

000

50

0398

000

50

015

101

MB

-AG

lt0

0002

lt0

001

lt0

001

lt0

001

000

70

351

000

90

0024

lt0

001

000

2lt

000

50

1040

lt0

001

000

215

1M

B-A

Hlt

000

02lt

000

10

001

lt0

001

001

00

596

000

70

1480

lt0

001

000

2lt

000

50

1280

lt0

001

000

315

2M

ean

lt0

0002

000

10

001

000

10

006

072

40

004

001

85lt

000

10

001

lt0

001

004

80lt

000

10

003

132

Med

ian

lt0

0002

000

20

002

000

10

005

026

70

003

000

560

001

000

20

005

003

200

001

000

212

1M

axim

umlt

000

020

002

000

20

003

002

57

200

001

20

1480

000

10

006

000

50

1280

000

50

015

293

Min

imum

DL

D

L

DL

D

L

000

20

083

000

10

0011

DL

D

L

DL

0

0090

DL

D

L

26

Site

Mg

Na

KA

gT

PU

TPF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-A2

62

60

7lt

000

115

511

30

795

01

8800

151

175

70

82

1020

76ƒ1

334

5cent11

9ƒ17

677

centM

B-B

28

43

10

000

19

89

10

645

04

1380

166

110

97

58

48

2076

ƒ12

915cent

119ƒ

169

20cent

MB

-C2

56

10

3lt

000

116

18

80

543

05

1300

134

178

80

80

815

76ƒ1

291

5cent11

9ƒ17

267

centM

B-D

23

48

lt0

2lt

000

19

46

50

390

lt0

1N

A41

153

80

77

812

76ƒ1

291

2cent11

9ƒ17

381

centM

B-E

43

51

02

lt0

001

111

85

041

50

948

91

371

738

07

66

1076

ƒ13

205cent

119ƒ

193

30cent

MB

-F2

97

4lt

02

lt0

001

123

84

048

90

513

301

401

838

07

75

876

ƒ13

351cent

119ƒ

193

47cent

MB

-G3

18

90

6lt

000

111

48

20

591

NA

NA

521

898

07

95

1876

ƒ13

330cent

119ƒ

192

14cent

516 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Mg

Na

KA

gT

PUT

PF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-H5

83

21

1lt

000

117

915

81

430

09

743

180

197

60

81

3112

76ƒ1

494

3cent11

9ƒ23

129

centM

B-I

114

425

16

lt0

001

169

125

072

50

513

661

431

256

60

85

325

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

E11

942

21

8lt

000

113

311

30

694

NA

NA

521

279

60

85

328

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

W12

143

11

8lt

000

113

99

50

637

05

1450

146

127

86

08

63

3576

ƒ14

530cent

119ƒ

220

11cent

MB

-K1

73

10

3lt

000

110

68

60

352

04

1448

133

132

50

74

5025

076

ƒ15

001cent

119ƒ

225

17cent

MB

-L2

05

80

7lt

000

114

67

20

428

lt0

1N

A29

147

60

75

4310

076

ƒ14

792cent

119ƒ

221

40cent

MB

-M6

23

41

1lt

000

110

17

90

401

lt0

1N

A40

189

60

81

135

76ƒ1

436

7cent11

9ƒ17

185

centM

B-N

31

41

13

lt0

001

143

104

088

5lt

01

NA

621

476

07

58

1276

ƒ14

417cent

119ƒ

171

05cent

MB

-O4

24

02

4lt

000

177

923

91

394

NA

NA

181

606

07

08

376

ƒ14

453cent

119ƒ

169

70cent

MB

-P3

33

40

2lt

000

110

47

10

521

lt0

1N

A50

147

60

76

780

76ƒ1

453

2cent11

9ƒ16

884

centM

B-Q

45

28

10

lt0

001

112

74

052

3lt

01

NA

471

536

07

96

1076

ƒ14

862cent

119ƒ

170

98cent

MB

-R6

24

10

9lt

000

18

17

00

479

lt0

1N

A59

160

60

80

625

76ƒ1

486

2cent11

9ƒ17

098

centM

B-S

23

60

05

lt0

001

105

61

035

51

550

71

341

574

07

37

3076

ƒ13

364cent

119ƒ

192

83cent

MB

-T0

81

60

2lt

000

113

811

60

281

NA

NA

201

255

07

078

1076

ƒ15

958cent

119ƒ

261

47cent

MB

-U6

26

02

2lt

000

111

27

70

620

lt0

1N

A55

113

85

08

375

125

76ƒ1

664

1cent11

9ƒ25

841

centM

B-V

75

70

27

lt0

001

118

56

081

32

221

11

691

156

60

83

7714

076

ƒ16

778cent

119ƒ

261

15cent

MB

-W5

81

91

5lt

000

110

56

81

058

24

245

110

11

124

60

84

6212

76ƒ1

670

4cent11

9ƒ25

163

centM

B-X

11

28

lt0

2lt

000

18

56

00

616

NA

NA

721

758

07

962

3076

ƒ16

421cent

119ƒ

260

98cent

MB

-Y1

12

60

2lt

000

19

47

20

612

NA

NA

651

777

57

737

2576

ƒ16

418cent

119ƒ

229

82cent

MB

-Z1

02

60

4lt

000

114

16

40

792

04

1503

156

165

90

78

3618

76ƒ1

618

1cent11

9ƒ21

726

centM

B-A

A1

73

30

7lt

000

17

14

60

683

14

312

196

196

95

81

3338

76ƒ1

613

0cent11

9ƒ21

512

centM

B-A

B6

34

41

0lt

000

17

24

00

453

22

174

163

167

60

80

3230

76ƒ1

476

0cent11

9ƒ17

655

centM

B-A

C3

92

70

8lt

000

17

86

80

306

09

496

139

142

60

77

327

76ƒ1

482

0cent11

9ƒ17

705

centM

B-A

D2

72

10

6lt

000

18

54

40

230

NA

NA

271

386

07

73

7576

ƒ14

758cent

119ƒ

177

34cent

MB

-AE

19

24

05

lt0

001

81

43

020

61

142

31

251

294

07

74

3076

ƒ14

685cent

119ƒ

177

37cent

MB

-AF

50

117

13

lt0

001

117

09

31

031

69

1028

19

199

55

78

470

076

ƒ14

484cent

119ƒ

190

88cent

MB

-AG

173

988

31

lt0

001

114

91

075

7lt

01

NA

661

510

60

75

260

76ƒ1

368

8cent11

9ƒ20

500

centM

B-A

H22

896

24

0lt

000

114

16

70

422

lt0

1N

A30

153

06

07

52

110

76ƒ1

366

5cent11

9ƒ20

222

centM

ean

52

129

10

lt0

001

165

85

061

60

866

31

491

115

64

79

2161

Med

ian

33

41

10

000

111

27

70

591

09

885

147

175

60

78

825

Max

imum

228

988

40

000

111

70

239

143

06

988

001

101

153

09

58

678

700

Min

imum

08

16

DL

D

L

71

40

020

60

117

41

91

254

07

02

3

Limnology of two Canadian High Arctic regions 513

App

endi

x2

Mou

ldB

ayW

ater

Che

mis

try

DL

=be

low

dete

ctio

nlim

itsN

A=

valu

eno

tava

ilabl

e

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-A5

924

00

310

880

008

37

510

71

1lt

000

1lt

001

00

122

071

20

010

0143

MB

-B9

323

80

570

552

005

48

016

81

20

002

001

90

054

057

20

090

0154

MB

-C14

10

38

025

065

00

053

68

70

13

000

1lt

001

00

033

049

0lt

001

000

64M

B-D

118

05

50

300

394

004

35

12

91

00

001

lt0

010

002

70

347

008

000

57M

B-E

134

023

10

360

440

004

35

33

50

6lt

000

1lt

001

00

026

037

20

080

0147

MB

-F17

40

75

059

066

50

062

69

49

09

000

1lt

001

00

025

042

70

010

0083

MB

-G17

70

32

017

062

10

068

72

76

09

000

1lt

001

00

028

052

30

010

0074

MB

-H8

220

90

080

669

007

013

714

51

60

001

lt0

010

010

51

360

001

001

28M

B-I

575

025

51

400

683

008

010

121

01

30

001

lt0

010

001

40

645

003

000

33M

B-J

E60

80

381

013

070

80

083

93

203

13

000

1lt

001

00

044

061

10

020

0358

MB

-JW

615

038

40

130

725

007

69

020

31

20

001

lt0

010

001

00

561

004

003

45M

B-K

569

28

006

057

90

072

39

37

06

000

1lt

001

00

050

028

00

030

0035

MB

-L7

035

20

280

568

007

03

04

70

90

001

014

60

025

021

20

580

0113

MB

-M7

8412

50

040

375

004

84

511

21

00

001

lt0

010

003

50

353

031

002

47M

B-N

938

07

005

107

00

111

93

48

13

000

1lt

001

00

075

077

40

260

0064

MB

-O12

70

28

354

614

00

620

135

73

30

000

1lt

001

00

010

077

40

100

0260

MB

-P6

475

10

110

412

004

38

24

90

4lt

000

1lt

001

00

028

047

8lt

001

000

75M

B-Q

607

14

006

049

90

067

61

76

07

000

1lt

001

00

046

045

60

010

0208

MB

-R9

120

60

090

560

005

25

58

10

50

001

lt0

010

001

80

427

003

001

11M

B-S

148

00

60

130

760

010

22

65

40

40

001

lt0

010

005

60

253

001

000

79M

B-T

315

40

036

043

30

060

43

15

08

000

2lt

001

0lt

000

50

221

024

000

91M

B-U

119

011

51

760

430

006

56

917

40

4lt

000

1lt

001

0lt

000

50

555

037

002

18M

B-V

128

014

80

320

465

007

09

922

10

40

001

lt0

010

001

60

743

002

003

22M

B-W

365

14

011

058

70

086

112

225

10

000

1lt

001

00

104

097

20

010

0250

MB

-X5

566

40

460

488

004

87

810

21

00

001

lt0

010

007

10

568

002

001

16M

B-Y

577

98

020

049

50

065

69

81

12

000

1lt

001

00

016

054

70

010

0127

MB

-Z5

882

10

510

601

007

58

08

30

7lt

000

1lt

001

00

058

071

70

020

0073

MB

-AA

747

65

127

043

70

059

78

121

13

000

1lt

001

00

021

062

40

040

0134

MB

-AB

109

04

80

140

382

005

45

48

20

6lt

000

1lt

001

00

010

039

90

020

0144

MB

-AC

624

18

005

044

60

043

26

52

06

lt0

001

lt0

010

002

30

263

001

001

07

514 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-AD

395

12

008

041

30

030

19

51

02

lt0

001

lt0

010

lt0

005

020

00

030

0119

MB

-AE

380

15

013

046

50

046

11

39

02

000

2lt

001

00

026

016

00

080

0091

MB

-AF

233

06

50

227

090

068

43

57

30

8lt

000

1lt

001

0lt

000

50

347

149

004

81M

B-A

G20

100

29

016

042

30

046

78

70

06

lt0

001

lt0

010

001

80

711

003

002

63M

B-A

H21

200

89

010

043

80

046

41

41

05

lt0

001

lt0

010

003

30

376

002

002

23M

ean

249

87

70

410

901

009

66

79

40

90

001

000

50

035

051

50

120

0158

Med

ian

932

40

017

055

20

065

69

76

09

000

1lt

001

00

028

049

00

030

0127

Max

imum

212

0038

43

547

090

068

413

722

53

00

002

014

60

122

136

01

490

0481

Min

imum

315

06

004

037

50

030

11

15

02

000

1D

L

DL

0

160

DL

0

0033

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Alt

000

02lt

000

1lt

000

10

001

000

50

623

000

40

0023

000

1lt

000

2lt

000

50

0373

lt0

001

000

416

0M

B-B

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

305

000

50

0056

lt0

001

lt0

002

lt0

005

005

68lt

000

10

002

255

MB

-Clt

000

02lt

000

1lt

000

1lt

000

10

005

027

80

002

000

12lt

000

1lt

000

2lt

000

50

0287

lt0

001

000

112

1M

B-D

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

262

000

20

0051

lt0

001

lt0

002

lt0

005

001

81lt

000

10

001

59

MB

-Elt

000

02lt

000

1lt

000

1lt

000

10

003

026

40

003

001

51lt

000

1lt

000

2lt

000

50

0323

lt0

001

000

310

9M

B-F

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

746

000

40

0080

lt0

001

lt0

002

lt0

005

003

01lt

000

10

002

104

MB

-Glt

000

02lt

000

1lt

000

10

001

000

70

235

000

30

0181

lt0

001

lt0

002

lt0

005

003

34lt

000

10

001

125

MB

-Hlt

000

02lt

000

1lt

000

1lt

000

10

004

056

00

001

001

23lt

000

10

003

lt0

005

006

44lt

000

10

002

165

MB

-Ilt

000

02lt

000

1lt

000

1lt

000

10

007

014

80

003

003

05lt

000

1lt

000

2lt

000

50

0942

lt0

001

000

223

6M

B-J

Elt

000

02lt

000

1lt

000

1lt

000

10

009

034

80

006

001

19lt

000

1lt

000

2lt

000

50

1250

lt0

001

000

227

9M

B-J

Wlt

000

02lt

000

1lt

000

1lt

000

10

025

076

40

006

001

31lt

000

1lt

000

2lt

000

50

1260

000

10

002

278

MB

-Klt

000

02lt

000

1lt

000

1lt

000

10

006

020

30

001

000

45lt

000

1lt

000

2lt

000

50

0186

lt0

001

000

14

3M

B-L

lt0

0002

lt0

001

lt0

001

000

10

007

115

00

001

001

70lt

000

1lt

000

2lt

000

50

0268

000

10

005

54

MB

-Mlt

000

02lt

000

1lt

000

10

001

000

81

140

000

20

0141

lt0

001

000

2lt

000

50

0557

000

10

003

153

MB

-Nlt

000

02lt

000

1lt

000

10

001

000

41

150

000

40

0169

lt0

001

000

2lt

000

50

0104

lt0

001

000

44

3M

B-O

lt0

0002

000

20

002

000

10

004

720

00

011

011

90lt

000

10

006

lt0

005

002

570

001

000

76

9M

B-P

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

174

000

40

0014

lt0

001

lt0

002

lt0

005

002

62lt

000

10

001

57

MB

-Qlt

000

02lt

000

1lt

000

1lt

000

10

007

024

90

003

000

44lt

000

1lt

000

2lt

000

50

0273

lt0

001

000

16

6

Limnology of two Canadian High Arctic regions 515

App

endi

x2

Con

tinue

d

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Rlt

000

02lt

000

1lt

000

1lt

000

10

008

013

00

004

000

53lt

000

1lt

000

2lt

000

50

0227

lt0

001

000

15

1M

B-S

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

216

000

10

0030

lt0

001

lt0

002

lt0

005

002

04lt

000

10

002

66

MB

-Tlt

000

02lt

000

1lt

000

10

001

000

50

236

000

10

0038

lt0

001

000

4lt

000

50

0093

000

10

012

26

MB

-Ult

000

02lt

000

10

001

000

10

005

085

80

010

001

05lt

000

10

002

lt0

005

008

640

001

000

324

9M

B-V

lt0

0002

lt0

001

lt0

001

000

10

011

015

00

012

001

18lt

000

1lt

000

2lt

000

50

1160

lt0

001

000

129

3M

B-W

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

519

000

30

0175

lt0

001

lt0

002

lt0

005

009

01lt

000

10

002

267

MB

-Xlt

000

02lt

000

1lt

000

1lt

000

10

002

025

50

002

000

30lt

000

1lt

000

2lt

000

50

0400

lt0

001

lt0

001

177

MB

-Ylt

000

02lt

000

1lt

000

1lt

000

10

005

023

40

002

000

18lt

000

1lt

000

2lt

000

50

0404

lt0

001

000

117

1M

B-Z

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

144

000

10

0026

lt0

001

lt0

002

lt0

005

002

78lt

000

10

001

145

MB

-AA

lt0

0002

lt0

001

lt0

001

lt0

001

000

80

125

000

20

0011

lt0

001

lt0

002

lt0

005

005

59lt

000

10

001

218

MB

-AB

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

267

000

40

0050

lt0

001

lt0

002

lt0

005

002

53lt

000

10

001

57

MB

-AC

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

090

000

20

0031

lt0

001

lt0

002

lt0

005

001

55lt

000

10

001

32

MB

-AD

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

083

000

10

0034

lt0

001

lt0

002

lt0

005

001

80lt

000

10

001

44

MB

-AE

lt0

0002

lt0

001

lt0

001

lt0

001

000

60

402

000

10

0114

lt0

001

lt0

002

lt0

005

001

52lt

000

10

002

34

MB

-AF

lt0

0002

000

10

002

000

30

005

490

00

005

011

30lt

000

10

005

000

50

0398

000

50

015

101

MB

-AG

lt0

0002

lt0

001

lt0

001

lt0

001

000

70

351

000

90

0024

lt0

001

000

2lt

000

50

1040

lt0

001

000

215

1M

B-A

Hlt

000

02lt

000

10

001

lt0

001

001

00

596

000

70

1480

lt0

001

000

2lt

000

50

1280

lt0

001

000

315

2M

ean

lt0

0002

000

10

001

000

10

006

072

40

004

001

85lt

000

10

001

lt0

001

004

80lt

000

10

003

132

Med

ian

lt0

0002

000

20

002

000

10

005

026

70

003

000

560

001

000

20

005

003

200

001

000

212

1M

axim

umlt

000

020

002

000

20

003

002

57

200

001

20

1480

000

10

006

000

50

1280

000

50

015

293

Min

imum

DL

D

L

DL

D

L

000

20

083

000

10

0011

DL

D

L

DL

0

0090

DL

D

L

26

Site

Mg

Na

KA

gT

PU

TPF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-A2

62

60

7lt

000

115

511

30

795

01

8800

151

175

70

82

1020

76ƒ1

334

5cent11

9ƒ17

677

centM

B-B

28

43

10

000

19

89

10

645

04

1380

166

110

97

58

48

2076

ƒ12

915cent

119ƒ

169

20cent

MB

-C2

56

10

3lt

000

116

18

80

543

05

1300

134

178

80

80

815

76ƒ1

291

5cent11

9ƒ17

267

centM

B-D

23

48

lt0

2lt

000

19

46

50

390

lt0

1N

A41

153

80

77

812

76ƒ1

291

2cent11

9ƒ17

381

centM

B-E

43

51

02

lt0

001

111

85

041

50

948

91

371

738

07

66

1076

ƒ13

205cent

119ƒ

193

30cent

MB

-F2

97

4lt

02

lt0

001

123

84

048

90

513

301

401

838

07

75

876

ƒ13

351cent

119ƒ

193

47cent

MB

-G3

18

90

6lt

000

111

48

20

591

NA

NA

521

898

07

95

1876

ƒ13

330cent

119ƒ

192

14cent

516 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Mg

Na

KA

gT

PUT

PF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-H5

83

21

1lt

000

117

915

81

430

09

743

180

197

60

81

3112

76ƒ1

494

3cent11

9ƒ23

129

centM

B-I

114

425

16

lt0

001

169

125

072

50

513

661

431

256

60

85

325

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

E11

942

21

8lt

000

113

311

30

694

NA

NA

521

279

60

85

328

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

W12

143

11

8lt

000

113

99

50

637

05

1450

146

127

86

08

63

3576

ƒ14

530cent

119ƒ

220

11cent

MB

-K1

73

10

3lt

000

110

68

60

352

04

1448

133

132

50

74

5025

076

ƒ15

001cent

119ƒ

225

17cent

MB

-L2

05

80

7lt

000

114

67

20

428

lt0

1N

A29

147

60

75

4310

076

ƒ14

792cent

119ƒ

221

40cent

MB

-M6

23

41

1lt

000

110

17

90

401

lt0

1N

A40

189

60

81

135

76ƒ1

436

7cent11

9ƒ17

185

centM

B-N

31

41

13

lt0

001

143

104

088

5lt

01

NA

621

476

07

58

1276

ƒ14

417cent

119ƒ

171

05cent

MB

-O4

24

02

4lt

000

177

923

91

394

NA

NA

181

606

07

08

376

ƒ14

453cent

119ƒ

169

70cent

MB

-P3

33

40

2lt

000

110

47

10

521

lt0

1N

A50

147

60

76

780

76ƒ1

453

2cent11

9ƒ16

884

centM

B-Q

45

28

10

lt0

001

112

74

052

3lt

01

NA

471

536

07

96

1076

ƒ14

862cent

119ƒ

170

98cent

MB

-R6

24

10

9lt

000

18

17

00

479

lt0

1N

A59

160

60

80

625

76ƒ1

486

2cent11

9ƒ17

098

centM

B-S

23

60

05

lt0

001

105

61

035

51

550

71

341

574

07

37

3076

ƒ13

364cent

119ƒ

192

83cent

MB

-T0

81

60

2lt

000

113

811

60

281

NA

NA

201

255

07

078

1076

ƒ15

958cent

119ƒ

261

47cent

MB

-U6

26

02

2lt

000

111

27

70

620

lt0

1N

A55

113

85

08

375

125

76ƒ1

664

1cent11

9ƒ25

841

centM

B-V

75

70

27

lt0

001

118

56

081

32

221

11

691

156

60

83

7714

076

ƒ16

778cent

119ƒ

261

15cent

MB

-W5

81

91

5lt

000

110

56

81

058

24

245

110

11

124

60

84

6212

76ƒ1

670

4cent11

9ƒ25

163

centM

B-X

11

28

lt0

2lt

000

18

56

00

616

NA

NA

721

758

07

962

3076

ƒ16

421cent

119ƒ

260

98cent

MB

-Y1

12

60

2lt

000

19

47

20

612

NA

NA

651

777

57

737

2576

ƒ16

418cent

119ƒ

229

82cent

MB

-Z1

02

60

4lt

000

114

16

40

792

04

1503

156

165

90

78

3618

76ƒ1

618

1cent11

9ƒ21

726

centM

B-A

A1

73

30

7lt

000

17

14

60

683

14

312

196

196

95

81

3338

76ƒ1

613

0cent11

9ƒ21

512

centM

B-A

B6

34

41

0lt

000

17

24

00

453

22

174

163

167

60

80

3230

76ƒ1

476

0cent11

9ƒ17

655

centM

B-A

C3

92

70

8lt

000

17

86

80

306

09

496

139

142

60

77

327

76ƒ1

482

0cent11

9ƒ17

705

centM

B-A

D2

72

10

6lt

000

18

54

40

230

NA

NA

271

386

07

73

7576

ƒ14

758cent

119ƒ

177

34cent

MB

-AE

19

24

05

lt0

001

81

43

020

61

142

31

251

294

07

74

3076

ƒ14

685cent

119ƒ

177

37cent

MB

-AF

50

117

13

lt0

001

117

09

31

031

69

1028

19

199

55

78

470

076

ƒ14

484cent

119ƒ

190

88cent

MB

-AG

173

988

31

lt0

001

114

91

075

7lt

01

NA

661

510

60

75

260

76ƒ1

368

8cent11

9ƒ20

500

centM

B-A

H22

896

24

0lt

000

114

16

70

422

lt0

1N

A30

153

06

07

52

110

76ƒ1

366

5cent11

9ƒ20

222

centM

ean

52

129

10

lt0

001

165

85

061

60

866

31

491

115

64

79

2161

Med

ian

33

41

10

000

111

27

70

591

09

885

147

175

60

78

825

Max

imum

228

988

40

000

111

70

239

143

06

988

001

101

153

09

58

678

700

Min

imum

08

16

DL

D

L

71

40

020

60

117

41

91

254

07

02

3

514 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Cl

SO

4S

iO2

PO

CP

ON

DO

CD

ICSR

PN

O2

NO

3+

NO

2N

H3

TK

NA

lB

am

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

m gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

MB

-AD

395

12

008

041

30

030

19

51

02

lt0

001

lt0

010

lt0

005

020

00

030

0119

MB

-AE

380

15

013

046

50

046

11

39

02

000

2lt

001

00

026

016

00

080

0091

MB

-AF

233

06

50

227

090

068

43

57

30

8lt

000

1lt

001

0lt

000

50

347

149

004

81M

B-A

G20

100

29

016

042

30

046

78

70

06

lt0

001

lt0

010

001

80

711

003

002

63M

B-A

H21

200

89

010

043

80

046

41

41

05

lt0

001

lt0

010

003

30

376

002

002

23M

ean

249

87

70

410

901

009

66

79

40

90

001

000

50

035

051

50

120

0158

Med

ian

932

40

017

055

20

065

69

76

09

000

1lt

001

00

028

049

00

030

0127

Max

imum

212

0038

43

547

090

068

413

722

53

00

002

014

60

122

136

01

490

0481

Min

imum

315

06

004

037

50

030

11

15

02

000

1D

L

DL

0

160

DL

0

0033

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Alt

000

02lt

000

1lt

000

10

001

000

50

623

000

40

0023

000

1lt

000

2lt

000

50

0373

lt0

001

000

416

0M

B-B

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

305

000

50

0056

lt0

001

lt0

002

lt0

005

005

68lt

000

10

002

255

MB

-Clt

000

02lt

000

1lt

000

1lt

000

10

005

027

80

002

000

12lt

000

1lt

000

2lt

000

50

0287

lt0

001

000

112

1M

B-D

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

262

000

20

0051

lt0

001

lt0

002

lt0

005

001

81lt

000

10

001

59

MB

-Elt

000

02lt

000

1lt

000

1lt

000

10

003

026

40

003

001

51lt

000

1lt

000

2lt

000

50

0323

lt0

001

000

310

9M

B-F

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

746

000

40

0080

lt0

001

lt0

002

lt0

005

003

01lt

000

10

002

104

MB

-Glt

000

02lt

000

1lt

000

10

001

000

70

235

000

30

0181

lt0

001

lt0

002

lt0

005

003

34lt

000

10

001

125

MB

-Hlt

000

02lt

000

1lt

000

1lt

000

10

004

056

00

001

001

23lt

000

10

003

lt0

005

006

44lt

000

10

002

165

MB

-Ilt

000

02lt

000

1lt

000

1lt

000

10

007

014

80

003

003

05lt

000

1lt

000

2lt

000

50

0942

lt0

001

000

223

6M

B-J

Elt

000

02lt

000

1lt

000

1lt

000

10

009

034

80

006

001

19lt

000

1lt

000

2lt

000

50

1250

lt0

001

000

227

9M

B-J

Wlt

000

02lt

000

1lt

000

1lt

000

10

025

076

40

006

001

31lt

000

1lt

000

2lt

000

50

1260

000

10

002

278

MB

-Klt

000

02lt

000

1lt

000

1lt

000

10

006

020

30

001

000

45lt

000

1lt

000

2lt

000

50

0186

lt0

001

000

14

3M

B-L

lt0

0002

lt0

001

lt0

001

000

10

007

115

00

001

001

70lt

000

1lt

000

2lt

000

50

0268

000

10

005

54

MB

-Mlt

000

02lt

000

1lt

000

10

001

000

81

140

000

20

0141

lt0

001

000

2lt

000

50

0557

000

10

003

153

MB

-Nlt

000

02lt

000

1lt

000

10

001

000

41

150

000

40

0169

lt0

001

000

2lt

000

50

0104

lt0

001

000

44

3M

B-O

lt0

0002

000

20

002

000

10

004

720

00

011

011

90lt

000

10

006

lt0

005

002

570

001

000

76

9M

B-P

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

174

000

40

0014

lt0

001

lt0

002

lt0

005

002

62lt

000

10

001

57

MB

-Qlt

000

02lt

000

1lt

000

1lt

000

10

007

024

90

003

000

44lt

000

1lt

000

2lt

000

50

0273

lt0

001

000

16

6

Limnology of two Canadian High Arctic regions 515

App

endi

x2

Con

tinue

d

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Rlt

000

02lt

000

1lt

000

1lt

000

10

008

013

00

004

000

53lt

000

1lt

000

2lt

000

50

0227

lt0

001

000

15

1M

B-S

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

216

000

10

0030

lt0

001

lt0

002

lt0

005

002

04lt

000

10

002

66

MB

-Tlt

000

02lt

000

1lt

000

10

001

000

50

236

000

10

0038

lt0

001

000

4lt

000

50

0093

000

10

012

26

MB

-Ult

000

02lt

000

10

001

000

10

005

085

80

010

001

05lt

000

10

002

lt0

005

008

640

001

000

324

9M

B-V

lt0

0002

lt0

001

lt0

001

000

10

011

015

00

012

001

18lt

000

1lt

000

2lt

000

50

1160

lt0

001

000

129

3M

B-W

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

519

000

30

0175

lt0

001

lt0

002

lt0

005

009

01lt

000

10

002

267

MB

-Xlt

000

02lt

000

1lt

000

1lt

000

10

002

025

50

002

000

30lt

000

1lt

000

2lt

000

50

0400

lt0

001

lt0

001

177

MB

-Ylt

000

02lt

000

1lt

000

1lt

000

10

005

023

40

002

000

18lt

000

1lt

000

2lt

000

50

0404

lt0

001

000

117

1M

B-Z

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

144

000

10

0026

lt0

001

lt0

002

lt0

005

002

78lt

000

10

001

145

MB

-AA

lt0

0002

lt0

001

lt0

001

lt0

001

000

80

125

000

20

0011

lt0

001

lt0

002

lt0

005

005

59lt

000

10

001

218

MB

-AB

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

267

000

40

0050

lt0

001

lt0

002

lt0

005

002

53lt

000

10

001

57

MB

-AC

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

090

000

20

0031

lt0

001

lt0

002

lt0

005

001

55lt

000

10

001

32

MB

-AD

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

083

000

10

0034

lt0

001

lt0

002

lt0

005

001

80lt

000

10

001

44

MB

-AE

lt0

0002

lt0

001

lt0

001

lt0

001

000

60

402

000

10

0114

lt0

001

lt0

002

lt0

005

001

52lt

000

10

002

34

MB

-AF

lt0

0002

000

10

002

000

30

005

490

00

005

011

30lt

000

10

005

000

50

0398

000

50

015

101

MB

-AG

lt0

0002

lt0

001

lt0

001

lt0

001

000

70

351

000

90

0024

lt0

001

000

2lt

000

50

1040

lt0

001

000

215

1M

B-A

Hlt

000

02lt

000

10

001

lt0

001

001

00

596

000

70

1480

lt0

001

000

2lt

000

50

1280

lt0

001

000

315

2M

ean

lt0

0002

000

10

001

000

10

006

072

40

004

001

85lt

000

10

001

lt0

001

004

80lt

000

10

003

132

Med

ian

lt0

0002

000

20

002

000

10

005

026

70

003

000

560

001

000

20

005

003

200

001

000

212

1M

axim

umlt

000

020

002

000

20

003

002

57

200

001

20

1480

000

10

006

000

50

1280

000

50

015

293

Min

imum

DL

D

L

DL

D

L

000

20

083

000

10

0011

DL

D

L

DL

0

0090

DL

D

L

26

Site

Mg

Na

KA

gT

PU

TPF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-A2

62

60

7lt

000

115

511

30

795

01

8800

151

175

70

82

1020

76ƒ1

334

5cent11

9ƒ17

677

centM

B-B

28

43

10

000

19

89

10

645

04

1380

166

110

97

58

48

2076

ƒ12

915cent

119ƒ

169

20cent

MB

-C2

56

10

3lt

000

116

18

80

543

05

1300

134

178

80

80

815

76ƒ1

291

5cent11

9ƒ17

267

centM

B-D

23

48

lt0

2lt

000

19

46

50

390

lt0

1N

A41

153

80

77

812

76ƒ1

291

2cent11

9ƒ17

381

centM

B-E

43

51

02

lt0

001

111

85

041

50

948

91

371

738

07

66

1076

ƒ13

205cent

119ƒ

193

30cent

MB

-F2

97

4lt

02

lt0

001

123

84

048

90

513

301

401

838

07

75

876

ƒ13

351cent

119ƒ

193

47cent

MB

-G3

18

90

6lt

000

111

48

20

591

NA

NA

521

898

07

95

1876

ƒ13

330cent

119ƒ

192

14cent

516 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Mg

Na

KA

gT

PUT

PF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-H5

83

21

1lt

000

117

915

81

430

09

743

180

197

60

81

3112

76ƒ1

494

3cent11

9ƒ23

129

centM

B-I

114

425

16

lt0

001

169

125

072

50

513

661

431

256

60

85

325

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

E11

942

21

8lt

000

113

311

30

694

NA

NA

521

279

60

85

328

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

W12

143

11

8lt

000

113

99

50

637

05

1450

146

127

86

08

63

3576

ƒ14

530cent

119ƒ

220

11cent

MB

-K1

73

10

3lt

000

110

68

60

352

04

1448

133

132

50

74

5025

076

ƒ15

001cent

119ƒ

225

17cent

MB

-L2

05

80

7lt

000

114

67

20

428

lt0

1N

A29

147

60

75

4310

076

ƒ14

792cent

119ƒ

221

40cent

MB

-M6

23

41

1lt

000

110

17

90

401

lt0

1N

A40

189

60

81

135

76ƒ1

436

7cent11

9ƒ17

185

centM

B-N

31

41

13

lt0

001

143

104

088

5lt

01

NA

621

476

07

58

1276

ƒ14

417cent

119ƒ

171

05cent

MB

-O4

24

02

4lt

000

177

923

91

394

NA

NA

181

606

07

08

376

ƒ14

453cent

119ƒ

169

70cent

MB

-P3

33

40

2lt

000

110

47

10

521

lt0

1N

A50

147

60

76

780

76ƒ1

453

2cent11

9ƒ16

884

centM

B-Q

45

28

10

lt0

001

112

74

052

3lt

01

NA

471

536

07

96

1076

ƒ14

862cent

119ƒ

170

98cent

MB

-R6

24

10

9lt

000

18

17

00

479

lt0

1N

A59

160

60

80

625

76ƒ1

486

2cent11

9ƒ17

098

centM

B-S

23

60

05

lt0

001

105

61

035

51

550

71

341

574

07

37

3076

ƒ13

364cent

119ƒ

192

83cent

MB

-T0

81

60

2lt

000

113

811

60

281

NA

NA

201

255

07

078

1076

ƒ15

958cent

119ƒ

261

47cent

MB

-U6

26

02

2lt

000

111

27

70

620

lt0

1N

A55

113

85

08

375

125

76ƒ1

664

1cent11

9ƒ25

841

centM

B-V

75

70

27

lt0

001

118

56

081

32

221

11

691

156

60

83

7714

076

ƒ16

778cent

119ƒ

261

15cent

MB

-W5

81

91

5lt

000

110

56

81

058

24

245

110

11

124

60

84

6212

76ƒ1

670

4cent11

9ƒ25

163

centM

B-X

11

28

lt0

2lt

000

18

56

00

616

NA

NA

721

758

07

962

3076

ƒ16

421cent

119ƒ

260

98cent

MB

-Y1

12

60

2lt

000

19

47

20

612

NA

NA

651

777

57

737

2576

ƒ16

418cent

119ƒ

229

82cent

MB

-Z1

02

60

4lt

000

114

16

40

792

04

1503

156

165

90

78

3618

76ƒ1

618

1cent11

9ƒ21

726

centM

B-A

A1

73

30

7lt

000

17

14

60

683

14

312

196

196

95

81

3338

76ƒ1

613

0cent11

9ƒ21

512

centM

B-A

B6

34

41

0lt

000

17

24

00

453

22

174

163

167

60

80

3230

76ƒ1

476

0cent11

9ƒ17

655

centM

B-A

C3

92

70

8lt

000

17

86

80

306

09

496

139

142

60

77

327

76ƒ1

482

0cent11

9ƒ17

705

centM

B-A

D2

72

10

6lt

000

18

54

40

230

NA

NA

271

386

07

73

7576

ƒ14

758cent

119ƒ

177

34cent

MB

-AE

19

24

05

lt0

001

81

43

020

61

142

31

251

294

07

74

3076

ƒ14

685cent

119ƒ

177

37cent

MB

-AF

50

117

13

lt0

001

117

09

31

031

69

1028

19

199

55

78

470

076

ƒ14

484cent

119ƒ

190

88cent

MB

-AG

173

988

31

lt0

001

114

91

075

7lt

01

NA

661

510

60

75

260

76ƒ1

368

8cent11

9ƒ20

500

centM

B-A

H22

896

24

0lt

000

114

16

70

422

lt0

1N

A30

153

06

07

52

110

76ƒ1

366

5cent11

9ƒ20

222

centM

ean

52

129

10

lt0

001

165

85

061

60

866

31

491

115

64

79

2161

Med

ian

33

41

10

000

111

27

70

591

09

885

147

175

60

78

825

Max

imum

228

988

40

000

111

70

239

143

06

988

001

101

153

09

58

678

700

Min

imum

08

16

DL

D

L

71

40

020

60

117

41

91

254

07

02

3

Limnology of two Canadian High Arctic regions 515

App

endi

x2

Con

tinue

d

Site

Be

Cd

Co

Cr

Cu

Fe

Li

Mn

Mo

Ni

Pb

SrV

Zn

Ca

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1

MB

-Rlt

000

02lt

000

1lt

000

1lt

000

10

008

013

00

004

000

53lt

000

1lt

000

2lt

000

50

0227

lt0

001

000

15

1M

B-S

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

216

000

10

0030

lt0

001

lt0

002

lt0

005

002

04lt

000

10

002

66

MB

-Tlt

000

02lt

000

1lt

000

10

001

000

50

236

000

10

0038

lt0

001

000

4lt

000

50

0093

000

10

012

26

MB

-Ult

000

02lt

000

10

001

000

10

005

085

80

010

001

05lt

000

10

002

lt0

005

008

640

001

000

324

9M

B-V

lt0

0002

lt0

001

lt0

001

000

10

011

015

00

012

001

18lt

000

1lt

000

2lt

000

50

1160

lt0

001

000

129

3M

B-W

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

519

000

30

0175

lt0

001

lt0

002

lt0

005

009

01lt

000

10

002

267

MB

-Xlt

000

02lt

000

1lt

000

1lt

000

10

002

025

50

002

000

30lt

000

1lt

000

2lt

000

50

0400

lt0

001

lt0

001

177

MB

-Ylt

000

02lt

000

1lt

000

1lt

000

10

005

023

40

002

000

18lt

000

1lt

000

2lt

000

50

0404

lt0

001

000

117

1M

B-Z

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

144

000

10

0026

lt0

001

lt0

002

lt0

005

002

78lt

000

10

001

145

MB

-AA

lt0

0002

lt0

001

lt0

001

lt0

001

000

80

125

000

20

0011

lt0

001

lt0

002

lt0

005

005

59lt

000

10

001

218

MB

-AB

lt0

0002

lt0

001

lt0

001

lt0

001

000

50

267

000

40

0050

lt0

001

lt0

002

lt0

005

002

53lt

000

10

001

57

MB

-AC

lt0

0002

lt0

001

lt0

001

lt0

001

000

30

090

000

20

0031

lt0

001

lt0

002

lt0

005

001

55lt

000

10

001

32

MB

-AD

lt0

0002

lt0

001

lt0

001

lt0

001

000

40

083

000

10

0034

lt0

001

lt0

002

lt0

005

001

80lt

000

10

001

44

MB

-AE

lt0

0002

lt0

001

lt0

001

lt0

001

000

60

402

000

10

0114

lt0

001

lt0

002

lt0

005

001

52lt

000

10

002

34

MB

-AF

lt0

0002

000

10

002

000

30

005

490

00

005

011

30lt

000

10

005

000

50

0398

000

50

015

101

MB

-AG

lt0

0002

lt0

001

lt0

001

lt0

001

000

70

351

000

90

0024

lt0

001

000

2lt

000

50

1040

lt0

001

000

215

1M

B-A

Hlt

000

02lt

000

10

001

lt0

001

001

00

596

000

70

1480

lt0

001

000

2lt

000

50

1280

lt0

001

000

315

2M

ean

lt0

0002

000

10

001

000

10

006

072

40

004

001

85lt

000

10

001

lt0

001

004

80lt

000

10

003

132

Med

ian

lt0

0002

000

20

002

000

10

005

026

70

003

000

560

001

000

20

005

003

200

001

000

212

1M

axim

umlt

000

020

002

000

20

003

002

57

200

001

20

1480

000

10

006

000

50

1280

000

50

015

293

Min

imum

DL

D

L

DL

D

L

000

20

083

000

10

0011

DL

D

L

DL

0

0090

DL

D

L

26

Site

Mg

Na

KA

gT

PU

TPF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-A2

62

60

7lt

000

115

511

30

795

01

8800

151

175

70

82

1020

76ƒ1

334

5cent11

9ƒ17

677

centM

B-B

28

43

10

000

19

89

10

645

04

1380

166

110

97

58

48

2076

ƒ12

915cent

119ƒ

169

20cent

MB

-C2

56

10

3lt

000

116

18

80

543

05

1300

134

178

80

80

815

76ƒ1

291

5cent11

9ƒ17

267

centM

B-D

23

48

lt0

2lt

000

19

46

50

390

lt0

1N

A41

153

80

77

812

76ƒ1

291

2cent11

9ƒ17

381

centM

B-E

43

51

02

lt0

001

111

85

041

50

948

91

371

738

07

66

1076

ƒ13

205cent

119ƒ

193

30cent

MB

-F2

97

4lt

02

lt0

001

123

84

048

90

513

301

401

838

07

75

876

ƒ13

351cent

119ƒ

193

47cent

MB

-G3

18

90

6lt

000

111

48

20

591

NA

NA

521

898

07

95

1876

ƒ13

330cent

119ƒ

192

14cent

516 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Mg

Na

KA

gT

PUT

PF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-H5

83

21

1lt

000

117

915

81

430

09

743

180

197

60

81

3112

76ƒ1

494

3cent11

9ƒ23

129

centM

B-I

114

425

16

lt0

001

169

125

072

50

513

661

431

256

60

85

325

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

E11

942

21

8lt

000

113

311

30

694

NA

NA

521

279

60

85

328

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

W12

143

11

8lt

000

113

99

50

637

05

1450

146

127

86

08

63

3576

ƒ14

530cent

119ƒ

220

11cent

MB

-K1

73

10

3lt

000

110

68

60

352

04

1448

133

132

50

74

5025

076

ƒ15

001cent

119ƒ

225

17cent

MB

-L2

05

80

7lt

000

114

67

20

428

lt0

1N

A29

147

60

75

4310

076

ƒ14

792cent

119ƒ

221

40cent

MB

-M6

23

41

1lt

000

110

17

90

401

lt0

1N

A40

189

60

81

135

76ƒ1

436

7cent11

9ƒ17

185

centM

B-N

31

41

13

lt0

001

143

104

088

5lt

01

NA

621

476

07

58

1276

ƒ14

417cent

119ƒ

171

05cent

MB

-O4

24

02

4lt

000

177

923

91

394

NA

NA

181

606

07

08

376

ƒ14

453cent

119ƒ

169

70cent

MB

-P3

33

40

2lt

000

110

47

10

521

lt0

1N

A50

147

60

76

780

76ƒ1

453

2cent11

9ƒ16

884

centM

B-Q

45

28

10

lt0

001

112

74

052

3lt

01

NA

471

536

07

96

1076

ƒ14

862cent

119ƒ

170

98cent

MB

-R6

24

10

9lt

000

18

17

00

479

lt0

1N

A59

160

60

80

625

76ƒ1

486

2cent11

9ƒ17

098

centM

B-S

23

60

05

lt0

001

105

61

035

51

550

71

341

574

07

37

3076

ƒ13

364cent

119ƒ

192

83cent

MB

-T0

81

60

2lt

000

113

811

60

281

NA

NA

201

255

07

078

1076

ƒ15

958cent

119ƒ

261

47cent

MB

-U6

26

02

2lt

000

111

27

70

620

lt0

1N

A55

113

85

08

375

125

76ƒ1

664

1cent11

9ƒ25

841

centM

B-V

75

70

27

lt0

001

118

56

081

32

221

11

691

156

60

83

7714

076

ƒ16

778cent

119ƒ

261

15cent

MB

-W5

81

91

5lt

000

110

56

81

058

24

245

110

11

124

60

84

6212

76ƒ1

670

4cent11

9ƒ25

163

centM

B-X

11

28

lt0

2lt

000

18

56

00

616

NA

NA

721

758

07

962

3076

ƒ16

421cent

119ƒ

260

98cent

MB

-Y1

12

60

2lt

000

19

47

20

612

NA

NA

651

777

57

737

2576

ƒ16

418cent

119ƒ

229

82cent

MB

-Z1

02

60

4lt

000

114

16

40

792

04

1503

156

165

90

78

3618

76ƒ1

618

1cent11

9ƒ21

726

centM

B-A

A1

73

30

7lt

000

17

14

60

683

14

312

196

196

95

81

3338

76ƒ1

613

0cent11

9ƒ21

512

centM

B-A

B6

34

41

0lt

000

17

24

00

453

22

174

163

167

60

80

3230

76ƒ1

476

0cent11

9ƒ17

655

centM

B-A

C3

92

70

8lt

000

17

86

80

306

09

496

139

142

60

77

327

76ƒ1

482

0cent11

9ƒ17

705

centM

B-A

D2

72

10

6lt

000

18

54

40

230

NA

NA

271

386

07

73

7576

ƒ14

758cent

119ƒ

177

34cent

MB

-AE

19

24

05

lt0

001

81

43

020

61

142

31

251

294

07

74

3076

ƒ14

685cent

119ƒ

177

37cent

MB

-AF

50

117

13

lt0

001

117

09

31

031

69

1028

19

199

55

78

470

076

ƒ14

484cent

119ƒ

190

88cent

MB

-AG

173

988

31

lt0

001

114

91

075

7lt

01

NA

661

510

60

75

260

76ƒ1

368

8cent11

9ƒ20

500

centM

B-A

H22

896

24

0lt

000

114

16

70

422

lt0

1N

A30

153

06

07

52

110

76ƒ1

366

5cent11

9ƒ20

222

centM

ean

52

129

10

lt0

001

165

85

061

60

866

31

491

115

64

79

2161

Med

ian

33

41

10

000

111

27

70

591

09

885

147

175

60

78

825

Max

imum

228

988

40

000

111

70

239

143

06

988

001

101

153

09

58

678

700

Min

imum

08

16

DL

D

L

71

40

020

60

117

41

91

254

07

02

3

516 Dermot Antoniades Marianne S V Douglas and John P Smol

App

endi

x2

Con

tinue

d

Site

Mg

Na

KA

gT

PUT

PF

TN

Chl

-aPO

C

TN

C

ond

TpH

Ele

vD

iam

Lat

itude

Lon

gitu

dem

gL

ndash1

mg

Lndash

1m

gL

ndash1

mg

Lndash

1m g

Lndash

1m g

Lndash

1m

gL

ndash1

mg

Lndash

1C

HL

-aT

Pm S

cmndash

1(ƒ

C)

(mas

l)(m

)(ƒ

N)

(ƒW

)

MB

-H5

83

21

1lt

000

117

915

81

430

09

743

180

197

60

81

3112

76ƒ1

494

3cent11

9ƒ23

129

centM

B-I

114

425

16

lt0

001

169

125

072

50

513

661

431

256

60

85

325

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

E11

942

21

8lt

000

113

311

30

694

NA

NA

521

279

60

85

328

76ƒ1

453

0cent11

9ƒ22

011

centM

B-J

W12

143

11

8lt

000

113

99

50

637

05

1450

146

127

86

08

63

3576

ƒ14

530cent

119ƒ

220

11cent

MB

-K1

73

10

3lt

000

110

68

60

352

04

1448

133

132

50

74

5025

076

ƒ15

001cent

119ƒ

225

17cent

MB

-L2

05

80

7lt

000

114

67

20

428

lt0

1N

A29

147

60

75

4310

076

ƒ14

792cent

119ƒ

221

40cent

MB

-M6

23

41

1lt

000

110

17

90

401

lt0

1N

A40

189

60

81

135

76ƒ1

436

7cent11

9ƒ17

185

centM

B-N

31

41

13

lt0

001

143

104

088

5lt

01

NA

621

476

07

58

1276

ƒ14

417cent

119ƒ

171

05cent

MB

-O4

24

02

4lt

000

177

923

91

394

NA

NA

181

606

07

08

376

ƒ14

453cent

119ƒ

169

70cent

MB

-P3

33

40

2lt

000

110

47

10

521

lt0

1N

A50

147

60

76

780

76ƒ1

453

2cent11

9ƒ16

884

centM

B-Q

45

28

10

lt0

001

112

74

052

3lt

01

NA

471

536

07

96

1076

ƒ14

862cent

119ƒ

170

98cent

MB

-R6

24

10

9lt

000

18

17

00

479

lt0

1N

A59

160

60

80

625

76ƒ1

486

2cent11

9ƒ17

098

centM

B-S

23

60

05

lt0

001

105

61

035

51

550

71

341

574

07

37

3076

ƒ13

364cent

119ƒ

192

83cent

MB

-T0

81

60

2lt

000

113

811

60

281

NA

NA

201

255

07

078

1076

ƒ15

958cent

119ƒ

261

47cent

MB

-U6

26

02

2lt

000

111

27

70

620

lt0

1N

A55

113

85

08

375

125

76ƒ1

664

1cent11

9ƒ25

841

centM

B-V

75

70

27

lt0

001

118

56

081

32

221

11

691

156

60

83

7714

076

ƒ16

778cent

119ƒ

261

15cent

MB

-W5

81

91

5lt

000

110

56

81

058

24

245

110

11

124

60

84

6212

76ƒ1

670

4cent11

9ƒ25

163

centM

B-X

11

28

lt0

2lt

000

18

56

00

616

NA

NA

721

758

07

962

3076

ƒ16

421cent

119ƒ

260

98cent

MB

-Y1

12

60

2lt

000

19

47

20

612

NA

NA

651

777

57

737

2576

ƒ16

418cent

119ƒ

229

82cent

MB

-Z1

02

60

4lt

000

114

16

40

792

04

1503

156

165

90

78

3618

76ƒ1

618

1cent11

9ƒ21

726

centM

B-A

A1

73

30

7lt

000

17

14

60

683

14

312

196

196

95

81

3338

76ƒ1

613

0cent11

9ƒ21

512

centM

B-A

B6

34

41

0lt

000

17

24

00

453

22

174

163

167

60

80

3230

76ƒ1

476

0cent11

9ƒ17

655

centM

B-A

C3

92

70

8lt

000

17

86

80

306

09

496

139

142

60

77

327

76ƒ1

482

0cent11

9ƒ17

705

centM

B-A

D2

72

10

6lt

000

18

54

40

230

NA

NA

271

386

07

73

7576

ƒ14

758cent

119ƒ

177

34cent

MB

-AE

19

24

05

lt0

001

81

43

020

61

142

31

251

294

07

74

3076

ƒ14

685cent

119ƒ

177

37cent

MB

-AF

50

117

13

lt0

001

117

09

31

031

69

1028

19

199

55

78

470

076

ƒ14

484cent

119ƒ

190

88cent

MB

-AG

173

988

31

lt0

001

114

91

075

7lt

01

NA

661

510

60

75

260

76ƒ1

368

8cent11

9ƒ20

500

centM

B-A

H22

896

24

0lt

000

114

16

70

422

lt0

1N

A30

153

06

07

52

110

76ƒ1

366

5cent11

9ƒ20

222

centM

ean

52

129

10

lt0

001

165

85

061

60

866

31

491

115

64

79

2161

Med

ian

33

41

10

000

111

27

70

591

09

885

147

175

60

78

825

Max

imum

228

988

40

000

111

70

239

143

06

988

001

101

153

09

58

678

700

Min

imum

08

16

DL

D

L

71

40

020

60

117

41

91

254

07

02

3