seasonal changes in arsenic concentrations and ... · school of science and engineering, university...

SEASONAL CHANGES IN ARSENIC CONCENTRATIONS ANDHYDROGEOCHEMISTRY OF CANADIAN CREEK, BALLARAT

(VICTORIA, AUSTRALIA)

KHAWAR SULTAN∗ and KIM DOWLINGSchool of Science and Engineering, University of Ballarat, Mount Helen, Victoria 3353, Australia

(∗author for correspondence, e-mail: [email protected])

(Received 13 April 2005; accepted 2 September 2005)

Abstract. A 10-month study of surface waters in Canadian Creek (Ballarat, Victoria, Australia)showed the significant influence of historic gold mining waste material. The investigation focussedon the hydrogeochemistry of the surface waters and soils in order to: (1) document the levels andseasonal trends in major, minor and trace elements in the creek, (2) identify the process by which Asis released from the soil/waste mining material to surface waters. For most dissolved major and traceelements (Na, Ca, Mg, K, and As) in surface waters, the concentrations decreased with the increasingrainfall and flow conditions except for Al and Fe. Two sites selected along the creek (<1 km apart)allowed evaluation of the possibility that mining waste material is contributing to the elevated Asconcentrations (up to 145 µg/l) in downstream surface water. Arsenic concentration varied more than28 fold seasonally and was highest in autumn and lowest in spring. Elevated concentrations of As(up to 1946 mg/kg) at the downstream site indicated the presence of a source of As concentrationin both surface and subsurface soils. Oxidation of arsenic sulphides under aerobic conditions withredox fluctuations (7 to 201 mV) could cause elevated As levels in the creek. Significant statisticalcorrelations among the major cations (Ca, Na and Mg) point to a common source(s) resulting inneutral to slightly alkaline (pH ∼ 6.5 to 7.8) surface water. Fe and Al secondary phases underoxidising conditions are a significant controlling mechanism for the mobilization of As in highlycontaminated soils (>1500 mg/kg) in the study area. The large As adsorption capacity of Fe and Alcould be limiting extreme mobilization into the water. Rainfall with relatively low pH is possiblycausing mobilisation of Al, Fe and As from highly alkaline soils (pH ≈ 9.0) into the nearby creek.

Keywords: arsenic, metal cations, mining, seasonality, surface water, Australia

1. Introduction

The contamination of the environment by metal/metalloids and the elucidation oftheir pathways through various environmental compartments (Cullen and Reimer,1989; Alloway, 1990; Smith et al., 1998) such as incorporation into surface water(Smedley and Kinniburgh, 2002 and references therein) and ultimately the foodchain, is of global significance. Metals can dissolve directly into river water fromabandoned waste material located along the river or indirectly by surface runoffand/or contaminated groundwater (Warner, 1998; Grabowski et al., 2001). Elevatedconcentrations of arsenic pose risks to plant (Baker et al., 1976), animal (Porcella,1987; Golub et al., 1998) and human health (Harrington et al., 1978; Hinwood et al.,

Water, Air, and Soil Pollution (2006) 169: 355–374 C© Springer 2006

356 K. SULTAN AND K. DOWLING

2003). Among the biggest sources of As are ore deposits with sulphide minerals,which may contain arsenic to a level of thousands of ppm. In mining activities,the processing of ores leads to the concentration of As-sulphide minerals whichweather geochemically to release As (Nriagu and Pacyna, 1988; Garcia and Alvarez,2003) and become local sources of contamination. Most of As is immobilized duethe presence of various minerals such as Fe and Al oxides. This oxide boundabundant source of arsenic can become mobile under the appropriate geochemicalconditions. A sudden pH change by acid rain (pH <4.5) and reduction in redox (Eh<100 mV) by flooding may cause mobility and the transport of a significant amountof arsenic, which is generally stable under prevailing environmental conditions.McGeehan et al. (1998) studied the effects of drying-flooding sequence underaerobic-anaerobic soil conditions and found that flooding resulted in a decrease insoil redox potential.

Regulatory authorities around the world (Commission for the European Com-munities, 1986; ATSDR, 1989; USEPA, 1996; WHO, 1996; ANZECC, 2000) haveestablished risk-based soil and water screening levels for As and other potentiallytoxic metals. For example, the drinking water standard is 7 µg/l (NHMRC, 1996)while the recreational water quality limit is 50 µg/l of As (ANZECC, 2000). The cur-rent health limit of As in soils is 100 mg/kg as per the guidelines of ANZECC/NHRC(1992) and there are many sites that exceed this limit (Hinwood et al., 1998).

Among the toxic elements, it is the presence of arsenic that has aroused concernin the Central Victorian region of Australia. Arsenic has a high natural backgroundsoil value and anthropogenic activities in the past such as gold mining have furtherconcentrated As levels (Smith et al., 1998). Hughes et al. (1997) have demonstratedthat gold mining industries have produced waste materials with high arsenic con-centrations, and this material is often built on for housing and has become part ofthe urban environment. This investigation focussed on the hydrogeochemistry ofthe surface waters to document the levels and seasonal trend in dissolved inorganicarsenic and other elements (Na, Ca, Mg, K, Al and Fe) in the Canadian creek.

2. Methods

2.1. STUDY AREA

The study area is located in Ballarat, approximately 120 km NW of Melbourne, Aus-tralia, near the suburb, Canadian (Figure 1). The sampling area is located near thefamous gold mining and now tourist site of Sovereign Hill in Victoria. Two samplinglocations were chosen along the Canadian Creek which carries discharges from up-stream catchments and urban run-off and drains into the Yarowee Creek locally.The two sampling locations were named ”Bridge” site and ”Trail” site and wereabout 1 km apart. The Bridge sampling site is located upstream and the Trail site islocated downstream near an historic mining waste dump (Figure 1). Gold mining

SEASONAL CHANGES IN ARSENIC CONCENTRATIONS 357

Figure 1. Map of the study area showing sampling locations on the Canadian Creek.

operations during and after the gold rush period had left the Trail sampling siteseverely degraded. The sampling sites were established with the aim of comparingthe hydrogeochemical changes along the creek near a point pollution source witha location which is relatively less disturbed by mining activities. The elevation ofthe two sampling sites is 423 m (Trail site) and 438 m (Bridge site) above sea level.

The climate of Ballarat is temperate with four seasons. The mean daily maxi-mum temperature for January is 24.9 ◦C in summer and the mean daily minimumtemperature for July is 3.2 ◦C during winter. The mean annual rainfall is 702 mm,59% of which is received from May to October.

Mining operations have produced waste piles of a heterogeneous nature that haveundergone crushing and occasional chemical treatment resulting in a mineralogical


complex material. These waste piles are difficult to locate because much of thearea now lies beneath the City of Ballarat due to urbanisation. This material isreacting over time with varying amounts of atmospheric input, seasonally producingcontamination to the Canadian Creek. In this paper, the hydrogeochemical featuresand seasonal changes in the concentration of As and other metals (Na, Ca, Mg, K,Al and Fe) from the two sites in the Canadian Creek are discussed.

2.2. GEOLOGY

The underlying geology is of Quaternary age (alluvium, silicious sand and gravels).Newer volcanic rocks of Quaternary to Tertiary age including basalt are also present.Other geological rock units of Ordovician age covering most of the study area, areslate, sandstone, sub-greywacke and mudstone (Clark et al., 1988; Ramsay et al.,1998; Taylor, 1998). Mesothermal quartz veins in the Ballarat goldfields are hostedby a thick sequence of Ordovician greywacke, which is interbedded with highlycarbonaceous, pyrite-bearing shale (Arne et al., 1998; Bierlein and McNaughton,1998). Ballarat gold deposits like other Central Victorian gold deposits (Gao andKwak, 1997) are typically sulphide-rich with pyrite (FeS) and arsenopyrite (FeAsS)up to several percent of total sulphides (Phillips and Hughes, 1996). Loellingite(FeAs2) is also present, but less widespread than arsenopyrite. Carbonate minerals,as reported by Phillips and Hughes (1996), include calcite (CaCO3), dolomite(CaMg (CO3)2), ankerite and siderite (FeCO3).

2.3. SAMPLING

Surface water samples (n = 40) were collected along the Canadian Creek duringa 10-month period from 16th March 2003 to 2nd January 2004 (Bridge site as B-type samples and Trail site as T-type samples, Tables I and II). Collecting samplesin those months facilitated a comparison among the seasons in the study area. Atboth sampling locations, surface water samples were collected at about two-weekintervals to determine the total dissolved inorganic concentrations of major (Na,Ca, Mg and K), minor (Al, Fe) and trace (As) elements. One domestic water samplewas also collected to analyse urban water influence on the surface water chemistryof the Canadian Creek. The samples were filtered through cellulose nitrate filters(0.45 µm pore size) and acidified with ultrapure HNO3 to pH ≤ 1.0 in the fieldto prevent precipitation/adsorption upon storage. To avoid contamination duringsampling and in subsequent handling of samples, trace metal-clean procedures wereemployed.

Water samples were stored in high-density polyethylene containers, which werepreviously washed with ultrapure 0.1 NHCl and rinsed with Milli-Q deionised water(≤18 M�-cm). Eight soil samples including both surface (0–10 cm) and subsurfacehorizons (10–20 and 20–30 cm) were collected to measure total concentrations ofeight elements from the two sampling sites. Also, six plant samples were collected


TAB

LE

I

Phys

ical

and

chem

ical

prop

ertie

sof

soils

Sam

ple

Dep

thC

lay

(%)

Silt

(%)

Sand

(%)

CE

CC

aM

gN

aK

As

FeA

lL

ocat

ion

ID(c

m)

(<2

µm

)(2

–20

µm

)(2

0–20

00µ

m)

pH1:

5(c

mol

c/kg

)(m

g/kg

)(m

g/kg

)(m

g/kg

)(m

g/kg

)(m

g/kg

)(g

/kg)

(g/k

g)

Bri

dge

site

B1S

0–10

12.1

48.4

39.5

6.9

6.4

367

1057

168

2166

3835

.016

.6

B2S

10–2

011

.642

.046

.46.

53.

421

650

114

719

5835

35.8

18.9

B3S

0–10

7.0

32.1

61.0

7.2

7.6

855

711

5329

154

11.8

3.6

B4S

10–2

08.

743

.647

.77.

07.

779

049

956

382

509.

64.

2

Tra

ilsi

teT

1S0–

106.

146

.947

.09.

158

.646

6363

7075

310

5916

1632

.24.

8

T2S

10–2

09.

958

.132

.09.

045

.940

9861

5325

462

019

4627

.72.

5

T3S

20–3

010

.159

.030

.99.

135

.042

9753

3723

383

518

1914

.18.

7

T4S

0–10

6.5

42.9

50.6

8.9

30.3

3488

3049

248

494

769

30.4

17.5

360 K. SULTAN AND K. DOWLINGTA

BL

EII

Che

mic

alco

mpo

sitio

n

Sam

ple

EC

Eh

Tem

p.C

aM

gN

aK

FeA

lA

sID

Dat

epH

(µS/

cm)

(mV

)(◦ C

)(m

g/l)

(mg/

l)(m

g/l)

(mg/

l)(m

g/l)

(mg/

l)(µ

g/l)

Surf

ace

wat

ers

from

Can

adia

nC

reek

(Bri

dge

site

)

B1

16M

AR

’03

7.0

642

112

20.6

29.8

15.7

756.

30.

070.

016.

1

B2

04A

PR’0

37.

149

315

416

.114

.313

.665

4.4

0.73

0.04

23.5

B3

17A

PR’0

37.

034

015

713

.111

.811

.050

3.7

0.98

0.32

8.2

B4

01M

AY

’03

7.2

518

135

12.1

21.2

14.8

703.

60.

240.

035.

1

B5

23M

AY

’03

7.4

280

143

9.0

7.9

9.2

392.

91.

050.

245.

1

B6

05JU

N’0

37.

538

316

010

.612

.89.

645

2.9

0.24

0.09

8.2

B7

23JU

N’0

37.

032

116

69.

511

.810

.343

2.8

0.85

0.23

16.3

B8

04JU

L’03

6.9

322

142

8.0

9.8

10.9

462.

81.

010.

256.

1

B9

21JU

L’03

7.6

366

172

9.0

10.3

12.3

492.

81.

170.

2014

.3

B10

03A

UG

’03

7.4

535

114

7.9

13.9

18.4

754.

01.

620.

377.

1

B11

19A

UG

’03

7.5

704

150

8.5

10.8

11.6

543.

00.

930.

3810

.2

B12

10SE

P’03

7.7

1040

9512

.125

.337

.114

04.

40.

910.

0726

.3

B13

17SE

P’03

7.8

691

9413

.114

.723

.293

3.7

0.96

0.31

8.2

B14

06O

CT

’03

7.8

694

6.8

12.6

12.5

22.7

904.

01.

110.

218.

2

B15

16O

CT

’03

7.8

1010

1812

.621

.135

.713

63.

51.

130.

079.

2

B16

01N

OV

’03

7.7

380

5511

.19.

412

.551

3.4

1.46

0.35

5.1

B17

12N

OV

’03

7.7

1110

1716

.623

.041

.114

83.

60.

940.

058.

3

B18

03D

EC

’03

7.8

880

1522

.122

.032

.910

92.

61.

300.

0313

.3

B19

17D

EC

’03

7.6

1290

1825

.639

.252

.117

44.

11.

490.

196.

1

B20

02JA

N’0

47.

699

068

25.1

38.4

48.7

132

4.2

1.40

0.16

7.1

(Con

tinu

edon

next

page

)

SEASONAL CHANGES IN ARSENIC CONCENTRATIONS 361TA

BL

EII

(Con

tinu

ed)

Sam

ple

EC

Eh

Tem

p.C

aM

gN

aK

FeA

lA

sID

Dat

epH

(µS/

cm)

(mV

)(◦ C

)(m

g/l)

(mg/

l)(m

g/l)

(mg/

l)(m

g/l)

(mg/

l)(µ

g/l)

Surf

ace

wat

ers

from

Tra

ilsi

tean

ddo

mes

ticw

ater

T1

16M

AR

’03

6.8

538

154

20.1

15.5

14.8

694.

20.

140.

0311

7.5

T2

04A

PR’0

36.

849

618

515

.114

.214

.464

5.1

1.04

0.26

145.

2

T3

17A

PR’0

36.

639

519

614

.68.

610

.656

3.8

0.69

0.62

112.

2

T4

01M

ay’0

36.

812

0520

111

.636

.849

.616

66.

80.

260.

0870

.0

T5

23M

AY

’03

7.0

323

175

8.5

9.0

10.3

422.

91.

530.

3662

.0

T6

05JU

N’0

37.

457

319

19.

016

.318

.681

3.8

0.99

0.08

104.

3

T7

23JU

N’0

36.

840

819

18.

512

.614

.358

3.4

1.59

0.18

84.5

T8

04JU

L’03

6.5

290

173

8.3

9.0

10.2

372.

71.

200.

3067

.3

T9

21JU

L’03

7.3

401

160

9.5

11.0

13.6

512.

81.

860.

1810

6.9

T10

03A

UG

’03

7.2

550

105

8.0

14.7

19.3

764.

01.

710.

2863

.4

T11

19A

UG

’03

7.2

708

163

9.5

10.7

11.0

463.

01.

120.

4485

.8

T12

10SE

P’03

7.4

980

135

13.1

23.1

34.0

134

4.9

1.04

0.17

100.

3

T13

17SE

P’03

7.7

681

120

13.7

14.8

21.8

873.

90.

920.

2138

.0

T14

06O

CT

’03

7.6

683

7815

.114

.322

.892

4.4

2.00

0.39

32.2

T15

16O

CT

’03

7.7

950

8013

.621

.232

.312

64.

40.

950.

1067

.3

T16

01N

OV

’03

7.5

390

6512

.110

.513

.049

4.3

1.33

0.56

44.5

T17

12N

OV

’03

7.6

1140

7819

.526

.342

.114

44.

70.

350.

2087

.1

T18

03D

EC

’03

7.7

820

6021

.124

.136

.311

02.

20.

410.

1412

8.6

T19

17D

EC

’03

7.5

800

6624

.224

.229

.397

4.0

0.57

0.34

103.

0

T20

02JA

N’0

47.

481

019

24.1

23.6

28.8

144

4.0

0.43

0.32

113.

3

Dom

estic

MA

R’0

38.

171

0–

–18

.324

.686

.03.

9–

––


(leaves of trees or grass with stems and leaves) to assess the bioavailability of con-taminants. There is a good ground cover of grasses and sedges along the creek bank.

Rainfall in the study area was also recorded during the sampling period tomonitor its influence on the hydrogeochemistry of the two sampling sites. A raingauge was installed (∼3 km south of the Bridge sampling site) in the School ofEngineering building at University of Ballarat, Mount Helen campus.

2.4. ANALYSIS

On site analysis included Eh (Pt electrode), pH and EC (electrical conductance)of surface water by using portable standard kits. Particle size distribution of soilsamples was determined using a Malvern instrument (model Mastersizer 2000) bylaser scattering at 10% sample obscuration. Soil pH (pH1:5) values were measuredat a soil:water ratio of 1:5 (w/v) following the method described by Rayment andHigginson (1992).

Preconcentration by evaporation of a 20 ml aliquot of each <0.45 µm filtrate wasdone to overcome the detection limit for the analysis of Al, Fe and As in surface wa-ters following the method described by Sherrell and Ross (1999). All other metals(Ca, Na, Mg and K) were analysed directly. Metals were analysed by graphite fur-nace atomic absorption spectroscopy (GFAAS) on a Varian Spectra AA-400 multielement instrument. Arsenic determinations were performed using Varian Spec-tra AA-20 and hydride generator VGA-76 Atomic Absorption Spectrophotometer(HGAAS) at the University of Ballarat. Prior to analysis by HGAAS, potassiumiodide (final concentration in the sample was 10% KI) was added to samples toreduce the arsenic to As(III) form.

About 4–5 g of soil sample was digested in about 15 ml aqua-regia (3:1,HCl:HNO3) for approximately 2 h, using a hot plate as a heating source. Thismethod is suitable for the analysis of elements like As and serves as an extractingmedia for mobile, mobilizable and pseudo-total fractions (Gupta et al., 1996). It isnot a total decomposition analysis, but released up to 90–95% of metal/metalloidinto the solution. Cation exchange capacity (CEC) of soils was analysed by extract-ing with an unbuffered NH4Cl (1 M, 12 h) solution following methods describedby Sumner and Miller (1996) and exchangeable cations were measured by graphitefurnace atomic absorption spectroscopy (GFAAS). About 0.5–1.0 g of plant samplewas weighed into 250 or 300 mm (diameter) Pyrex glass tubes, 6–8 ml nitric acidwas added and the sample tubes were placed in a programmable digestion blockcapable of ∼200 ◦C. After dissolution, samples were diluted to volume, mixed andfiltered prior to analysis.

From certified aqueous standards (BDH Spectorsol©R

), intermediate standardsof various concentrations were prepared. A control sample was analysed every15–20 samples for the QA/QC program. Precision and accuracy were estimatedat ±7% (evaluated from repeated measuring of standard solutions with known


concentrations) and ±10% (evaluated by analysing the certified reference materialGBW-07401), or better, at the mg/l and µg/l concentrations, respectively. Both fieldand laboratory blanks were below the detection limits for all elements in this study.Bulk soil samples were analysed with a scanning electron microscope (SEM) at theUniversity of Ballarat to examine the composition of certain minerals.

3. Results and Discussion

All physical and chemical data of soils are shown in Table I. Calcium contentsin soils ranged between 216 and 4663 mg/kg. Magnesium concentrations variedfrom 499 to 6370 mg/kg. Mg was dominant among alkali and alkaline earth metalsand results are in agreement with the work of Bierlein et al. (1999) who reportedMg-rich carbonates in the Ballarat region. Sodium concentrations in soils rangedfrom 53 to 753 mg/kg. Na registered the minimum concentration among the fourmajor elements and values of Na in the subsoil were about 3 fold lower comparedwith the topsoil. This is probably due to high leaching at the Trail site. Prevailingsoil conditions favour the solubilization of Na and leaving the soil horizons withpercolating water. Potassium levels varied from a minimum of 291 to a maximumof 2166 mg/kg. K in soils was positively correlated (r2 = 0.51) with the clay sizefraction (ϕ < 2 µm), possibly due to its fixation in secondary clay minerals. Anorder of abundance in soils for base cations followed as: Mg > Ca � K � Na.Generally, higher values were from the Trail site soils while lower values wereobserved for the Bridge site soils. Fe and Al contents varied from 9.6 to 35.8 g/kgand 2.5 to 18.9 g/kg respectively. Enrichment of Fe in soils has been observedwhich is consistent with natural background values. High values of Al in soils arecorrelated (r2 = 0.45) to K values because both form secondary silicate minerals.The highest As value measured was 1946 mg/kg (T2S) at the Trail sampling site andthe lowest value was 35.0 mg/kg (B2S) in soils from the Bridge site. Arsenic in soilsis greatly elevated at the Trail site (769–1946 mg/kg) as compared to the Bridgesite (35–54 mg/kg) (Table I). The maximum concentration was in mining wastematerial dumped at the Trail location near the Canadian Creek. When comparingthe As contents of soils at these two sites, all the values recorded in soils from theTrail site (n = 4) exceeded the Australian legal limits (ANZECC/NHRC, 1992).Average As concentrations (44.2 mg/kg) in soils from the Bridge site represent highbackground values, which suggests this site also has some mining influence. Morethan 34-fold higher As content in soils from the Trail site (average ∼1537 mg/kg)was found compared to the Bridge site which clearly indicates the presence ofAs-rich material (Table I).

Soil pH1:5 values were in contrast, neutral and alkaline, in soils from the twosampling locations. Both surface (0–10 cm) and subsurface (10–20 cm) soils wereneutral at the Bridge site (average pH1:5 ∼ 6.9) and alkaline (average pH1:5 ∼ 9.0)at the Trail site (Figure 2). Soils from the Trail site were higher by 2.0 pH units as


PH1:5

CE

C (

cmol

c/kg

) Trail site

Bridge site 0

10203040506070

6.0 7.0 8.0 9.0 10.0

Figure 2. Plot of CEC versus pH1:5 of soils from two sampling locations.

compared with soils from the Bridge site. The highly alkaline soils are probablyassociated with enrichment of Ca and Mg carbonates. The relatively constant pH(≈9.0) of top and subsurface soils at the Trail site seems to be regulated by Ca andMg carbonate buffering.

Cation exchange capacity (CEC) varied from 3.4 cmolc/kg (B2S) to58.6 cmolc/kg (T1S). Soils from the Trail site had more than six times higher CECvalues (average 42.5 cmolc/kg) than the Bridge site soils (average 6.3 cmolc/kg).Ca and Mg were the dominant exchangeable cations in soils and contributed morethan 75% to CEC of soils from the Trail site. CEC was greatest in topsoil sam-ple T1S (58.6 cmolc/kg) from the Trail site and it was substantially less in sampleB2S (3.4 cmolc/kg) soil from the Bridge site indicating a lack of exchangeable Cacontent (Table I and Figure 2). CEC generally decreased along the depth profile insoils from both locations. Low CEC values in subsurface soils are possibly due tothe low contents of clay size fraction and organic matter content.

The clay, silt and sand contents in soil samples varied as 5.7–12.1, 32.1–59.0,and 30.9–61.0%, respectively (Table I). While the silt size fraction was dominantin soils from the Trail site, the sand size fraction registered highest at the Bridgesite. The clay size fraction measured lowest in soils from both sampling locations.The silt fraction was positively correlated while the clay and sand fractions showeda negative correlation with soil CEC. This can be explained by the association oforganic matter content with the silt fraction as studied by Peinemmann et al. (2000).

Arsenic in plants varied from 0.4 to 2.1 mg/kg (dry weight) with higher valuesat the Trail site (Table III). Plants from the Bridge site showed low values (average∼0.59 mg/kg), which correspond to low As in soils (average ∼45.3 mg/kg) of thissite. The As concentration in plant samples clearly reflected its uptake (bioavail-ability), as the As levels in plants at the Trail site (average ∼1.57 mg/kg) werefound to be higher than the Bridge site. The variation in As concentrations in plantsfrom the two sites was attributed to different As levels in soils and to differencesin types/species. Similar findings have been reported by Baroni et al. (2004) andVillares et al. (2002) who described various factors affecting As uptake by plants in


TABLE III

Concentrations of total inorganic arsenic (dry weight) in plant samples from two locations

Location Sample ID Type of plant Common name As (mg/kg)

Bridge site B1P Grass Tussock grass 0.40

B2P Ground fern Austral bracken 0.73

B3P Shrub Myrtle Wattle 0.65

Trail site T1P Shrub Myrtle Wattle 2.10

T2P Shrub Myrtle Wattle 1.67

T3P Tree Eucalyptus 0.93

different soils. This finding agrees with the results from the Maldon area in Victoria,Australia (Sultan, 2005), where As levels in plants were related to As concentra-tions in soils. The myrtle wattle, a native shrub, showed particularly high levels ofAs (T1P sample, Table III) and can serve as a bioindicator of contamination in thearea. Various plant species have been known to concentrate metals from soils (Vil-lares et al., 2002; Haritonidis and Malea, 1995) under different soil environmentalconditions and, therefore, can be used as bioindicators. Low levels of As in plants(B1P and B3P) represent the background values in the area.

Total rainfall measured during the surface water sampling period was 602 mm,which was about 30 mm higher than the average rainfall from March to November(570 mm). Rainfall was more or less typical except for the two months, April andOctober, when it was recorded at 41 and 34% higher than the average, respectively.There were 99 rainy days during the sampling period. The rainfall amount wasnegatively correlated with the dissolved concentrations of major cations (Ca, Mg,Na and K) in surface water due to a dilution effect (Figure 3). Dissolved Fe and Alconcentrations in surface water showed an association with rainfall amount, which ispossibly due to their mobilization from soil into water by runoff and/or groundwaterflow. Rainfall, which has a much lower pH (4.1 to 7.7, average ∼5.9) than the soilpH1:5 (8.5 to 9.1), could in principle dissolve pH-sensitive elements like Al fromsoil. However, the high buffering capacity of soils increased their resistance topH changes and limits Al solubilization and mobilization. The hydraulic flushingallows already built up metals in the soil column to move to surface waters byinfiltration of low pH rainwater.

The surface waters are near neutral (pH ∼6.5–7.8) with lower pH values atthe Trail site (Table II). The lowering of pH (average decrease of 0.3 pH unitsalong the creek in less than a kilometre) suggested input at a downstream locationby some acid generation process. Electrical conductivities (EC) varied between280 to 1290 µS/cm depending upon the season (Figure 3). In general, creek waterchemistry showed an urban drainage signature (domestic water EC ≈ 710 µS/cm,Table II), which was diluted by rainfall and then affected by solute acquisitionbetween the two sites. The EC slightly increased from an average of 650 µS/cm


0

250

500

750

1000

1250

1500

6

6.5

7

7.5

80

0.2

0.4

0.6

0.8

0

10

20

30

40

50

0

50

100

150

a). EC (uS/cm)

b). pH

c).

d). Al (mg/l)

e). Ca (mg/l)

f). As (µg/l)

B ridge sit e T rail s ite

0

25

50

75

100

125

16M

ar'0

34-

Ap

r17

-Ap

r1-

May

23-M

ay5-

Jun

23-J

un

4-Ju

l21

-Jul

3-A

ug

19-A

ug

10-S

ep17

-Sep

6-O

ct16

-Oct

1-N

ov

12-N

ov

3-D

ec17

-Dec

02Ja

n'0

4

012345

678910

rain (mm) rainy days

Rai

n

Figure 3. Time series data of Canadian Creek area.

(n = 40) at the Bridge site to an average of 657 µS/cm at the Trail site. Theinconsistent variations in EC values and dissolved elements seem to be contributedby soil processes and/or irregular rainfall events (Figure 3). EC values showed aweak negative correlation with the rainfall amount, which was possibly due to thedilution effect. During a large rain event with more than 77 mm falling in two weeks,surface water EC increased due to the contribution of solutes from soils (Figure 3),but overall the area received more of its rain in the form of drizzly showers. Suchirregular storm events are the cause of a sudden increase in solute chemistry asweathering products are exposed to runoff and local flow through soil pores, asindicated by dissolved As and Al contents in surface water. The seasonal variationin EC suggests two possible main sources: the urban drainage chemistry duringlow rainfall period and the water passing through soil/waste material.

The Eh values of surface waters varied from slightly reducing, 6.8 mV, to oxidiz-ing, 201 mV. Higher Eh values, average 130 mV, at the Trail site than the Bridge site,average 100 mV, indicated oxidizing conditions. Eh changed seasonally with highervalues in autumn and lower values in summer, although data for the whole sum-mer period were not available. Eh showed negative correlation with pH in surface


0

50

100

150

200

250

6 6.5 7.5 8

Bridge site

Trail site

7

Eh

(mV

)

pH

Figure 4. Values of pH and Eh of Creek waters.

waters from the Trail site (r2 = 0.57, n = 20) and the Bridge site (r2 = 0.48,n = 20) (Figure 4). The variable redox conditions (10-fold variation) contributedto the interaction between water and redox sensitive minerals such as sulphides. Thesurface water temperature varied from 8.0 ◦C in winter to 25.5 ◦C in summer andgenerally increased downstream (Table II). The temperature was higher at theTrailsite throughout the sampling period except in autumn, during which it was slightlyhigher at the Bridge site.

Dissolved Ca concentrations in creek waters varied between 7.9 to 39.2 mg/land registered lowest during the winter/spring period (Figure 3). Ca concentrationdecreased slightly downstream from an average of 18.0 mg/l at the Bridge site to17.0 mg/l at the Trail site. Mg ranged between 9.2 and 52.1 mg/ with lowest val-ues in the winter season and did not register a significant change between the twosampling locations along the creek. Na, the concentration of which was low insoils, was dominant among major cations in all surface waters with average con-centrations of 84.2 and 86.5 mg/l at the Bridge and Trail sites respectively. A strongpositive correlation (r2 = 0.99) between Na concentrations and EC values showedsignificant addition of Na at both sampling locations. K concentrations in surfacewaters varied from 2.2 to 6.8 mg/l along the path, which indicated its leaching atthe downstream site. Concentrations of major cations followed more or less similartrends by decreasing during winter (June, July, August) and increasing in spring(September, October, November). This positive inter-relationship suggests the hy-drological control and dilution from a point source or groundwater contribution. Ingeneral, the order of abundance was Na � Mg > Ca � K in all water samples. Theinfluence of rainfall in the dilution of cations was observed, with rainfall showinga weak negative correlation with concentrations of Ca, Mg and Na.

Dissolved Fe and Al concentrations in water ranged from 0.07 to 2.0 mg/land 0.01 to 0.62 mg/l respectively. Al increased from an average concentrationof 0.18 mg/l to 0.26 mg/l along the creek path, which showed its addition at thedownstream location. Slightly higher Fe concentrations, average 1.01 mg/l, at the


Trail site as compared to Fe values, average 0.98 mg/l, at the Bridge site are dueto the presence of Fe-minerals in the mining waste material. Dissolved Fe and Alfollowed a similar pattern of seasonal variation with a few exceptions (Figure 3). Allsurface water data are shown in Table II. Lower dissolved Fe concentration mightbe due to its fixation by oxidation and subsequent precipitation as secondary phasesas confirmed by high soil Fe contents (up to 35.8 g/kg and average ∼24.6 g/kg).Lamb et al. (1996) also reported high soil Fe contents in the form of ferrihydrite(Fe(OH)3.nH2O) in the treatment of mining waste from the Ballarat area.

Arsenic concentrations in the surface waters varied from 5 to 26 µg/l at theBridge site and 32 to 145 µg/l at the Trail site. Dissolved As generally remainedconsistent with an average of 10.0 µg/l at the Bridge site throughout the samplingperiod with a few exceptions. Arsenic concentration showed more than a 4-foldseasonal variation at the Trail site. During winter and spring, the As levels insurface waters decreased although exceptions were also observed. It also showed atrend by increasing in the summer and autumn seasons (Figure 3).

The lowest values of As (∼5.0 µg/l) at the Bridge site point to the backgroundlevels of surface water in the area. More than an eight fold increase of As fromthe upstream Bridge site to the downstream Trail site within 1 km confirms thepollution source at a downstream location. Two likely mechanisms that could becontrolling the As release at the Trail site are: (1) oxidation of As bearing sulphideminerals, (2) the remobilisation of As from secondary phases of Fe and Al. The pHvalues of surface water at the Trail site are less (average ∼0.3 pH units) than thepH of the surface water at the Bridge site. It seems that acid being produced is wellneutralised due to the high buffering capacity of the soil material as shown by highCa and Mg concentrations in soils, as high as 4663 and 6370 mg/kg respectively.The second possible mechanism of As release under predominantly oxidising (upto 200 mV) and alkaline pH water conditions is its remobilisation from secondaryphases of Fe and Al. The negative correlation of As with Fe and Al indicates thatsome part of As is being desorbed while redox conditions are favouring Fe and Alfixation as secondary phases. A lack of significant correlation shows that it is notthe only source of this metalloid contamination.

The oxidation of arsenopyrite by exposure to oxygen and water produces acidwhich is being neutralized by a high content of carbonates of Ca present in soilsfrom mining waste dumps. The possible acid producing (Equations (1) and (2))reactions and subsequent consuming (Equation (3)) reaction are shown as:

2FeS2 + 7O2 + 2H2O → 2Fe3+ + 4SO2−4 + 4H+ (1)

Fe3+ + 3H2O → Fe(OH)3(s) + 3H+ (2)

H+ + CaCO3 → Ca2+ + HCO−3 (3)

Highly alkaline conditions in soils indicate that in addition to naturally occurringcarbonate minerals the soils might have been limed for AMD control, but the precise


Figure 5. Scanning electron microscopy (SEM) results showing: (a) Selected element x-ray mapsobtained from polished section of ferroan dolomite grains which are relatively abundant and can berecognised by inspection of the Cak and Mgk map panels, quartz is brighter in the Sik panel andarsenopyrite is bright in AsLα in the panel, illite is brighter in the Kk map. (b) EDX x-ray spectrumobtained from typical carbonate grain occurring in soil sample of ferroan dolomite. (c) SEM back-scattered image showing arsenopyrite (bright) oxidising to iron oxide hydrates (light grey rim onarsenopyrite) – other phases in the field include quartz, illite and chlorite. (d) AsLα element x-ray mapshowing dispersion of arsenic away from the iron phase.

history of the site is unknown. The presence of dolomite/ankerite in mining mate-rial is a typical composition of the abundant carbonate alteration found dispersedaround Ballarat style gold deposits as confirmed by SEM/EDS analysis in thisstudy (Figure 5). Neutralization processes produce enough alkalinity to raise soilpH (≈9.0) and hydrated iron oxides form as observed by a yellow to light orangecolour in soils. These secondary phases of iron act as sink for As depending onprevailing redox-pH conditions, which result in a complex As release and adsorp-tion process. SEM analysis shows arsenopyrite oxidising to iron oxide hydrateswith As dispersing away from the secondary iron phases (Figure 5). The elementx-ray maps (for As and Fe, Figure 5(a)) illustrate the concentration change of Asand Fe as indicated by the colour index. The lower concentration of As along the


rim reveals some retention of As, liberated by weathering processes. A slow releaseof As from soils by percolating water contributes to the signature of water chem-istry which enters the creek. Illite is identified as the second dominant clay mineralafter kaolinite in the central Victorian soils (Sultan, 2005). Overall, the absence ofmeaningful correlations indicates the likelihood of more than a single process ofAs release in the water-soil environment.

Seasonal changes in soil water content with depth cause the concentration andmobility of metal/metalloids to change significantly as the infiltrating water servesas the most important transport pathway into surface water. Considering that thesoils at the Trail site are contaminated with As, the overland flow is an importanttransport pathway during high rainfall periods.

The variations in As and Fe concentrations are most likely the result of interactionof water (rain water and/or groundwater) with sulphide rich mining waste underoxidising conditions. The seasonal influences suggest that incorporating water inthe contaminated soil probably did not have long residence times prior to enteringthe creek. The fixation of As in soils is largely a result of their high Fe content (upto 35.8 g/kg), which is serving as sink.

Dissolved As concentrations poorly correlated with dissolved Fe concentrations(Figure 6) in surface waters from the Trail site. This lack of relationship betweenAs and Fe implies that some other chemical reaction is taking place because usuallytheir solubility increases or decreases simultaneously (Gao and Burau, 1997). Underthe redox-pH conditions, oxidising and neutral respectively, encountered in thisstudy, As (V) will be the dominant species. In aerobic soils, 90% of dissolved Asis present as pentavalent arsenate (O’Neill, 1995) which strongly binds to Fe andAl hydroxides than As (III) (Cullen and Reimer, 1989). Most of the As, sorbedto Fe and Al in the soils, can become soluble if subjected to moderately reducingconditions.

A higher amount of sand and silt size fractions in the surface and subsurface soilhorizons increases infiltration and favours oxidising conditions. Arsenic solubility isperhaps also limited by the formation of insoluble As-sulphide minerals as reported

0

40

80

120

160

0.0 0.5 1.0 1.5 2.0 2.5

Fe (mg/L)

As

(µg/

L)

Figure 6. Scatter plot of dissolved as content versus Fe in waters of Trail site (n = 20).


by Ferguson and Gavis (1972). Arsenic concentrations are correlated (r2 = 0.60)to the silt size fraction (2–20 µm), which indicates the enrichment of adsorptionsites by the type and amount of clay minerals, and Fe and Al secondary phases. Thebehaviour of As in soils at the Trail site is complex as only a fraction of it is leaching,the high content of As in soils being stable under prevailing conditions. This limitedmobility of As is explained further by examining the amorphous oxides, clays andorganic matter present in the soils.

4. Conclusions

A systematic seasonal variation in surface water As and other measured elements(Ca, Mg, K, Na. Fe and Al) has been observed. A significant finding is that theconcentration of As changed markedly between seasons with the highest in autumnand the lowest in spring. The study showed the possibility that the As concentrationsin surface water are related to discharge and/or rainfall input in the area. Theinfluence of rainfall in dilution of creek water was observed with the rainfall showinga weak negative correlation with concentrations of Ca, Mg and Na.

Seasonally high As levels in urban drainage may affect water usage downstreamand limit its use. The more than eight-fold increase of As from the upstream Bridgesite to the downstream Trail site within 1 km confirms the contamination sourceat a downstream location. While upstream waters are influenced by urban waterdrainage, the surface water downstream carries a clear signature of mining materialchemistry. The primary oxidation of As-bearing sulphide phases generating acid iswell neutralized by Ca and Mg minerals in the soils. The high As retention capacityof soils and temporal trends of As may be used to understand the potential risksposed to the environment.

Due to the oxidising conditions, As solubility has been observed as being lowwhich could increase considerably under moderately reducing conditions. In highlycontaminated soils at the Trail site, the mobilization of As is controlled by oxidis-ing conditions and the presence of Fe and Al secondary phases. Also, the presenceof buffer materials such as carbonates and high CEC values are significantly con-tributing to the retention capacity of soils.

The results showed the relationship between the As contents in soils and plantsof the two sampling sites which confirmed the As bioavailability, particularly atthe Trail site. The results obtained can serve to increase an understanding of theseasonal changes of As concentrations in contaminated sites.

Acknowledgements

The assistance and guidance of Prof. Stafford McKnight was greatly appreciatedin the SEM analysis.


References

Alloway, B. J.: 1990, Heavy Metals in Soils, John Wiley & Sons, New York.ANZECC: 2000, Australian guidelines for water quality monitoring and reporting. Australian and

New Zealand Environment and Conservation Council, Canberra, Australia.Arne, D. C., Bierlein, F. P., McNaughton, N., Wilson, C. J. L. and Morand, V. J.: 1998, ‘Timing of gold

mineralisation in west and Central Victoria Australia: New constraints from SHRIMP II analysisof zircon grains from felsic intrusive rocks’, Ore Geology Reviews 13, 251–273.

ATSDR: 1989, ‘Toxicological profile for arsenic’, Agency for Toxic Substances and Disease Registry,U.S. Public Health Service, Atlanta, GA. ATSDR/TP-88/02.

Australian and New Zealand Environment Council/National Health and Medical Research Council:1992, Australian and New Zealand guidelines for the assessment and management of contaminatedsites. Australian Government, Canberra.

Baker, R. S., Barrentine, D. H., Bowman, W. L., Hawthorne, W. L. and Pettiet, J. V.: 1976, ‘Cropresponse and arsenic uptake following soil incorporation of MSMA’, Weed Science 24, 322–326.

Baroni, F., Boscagli, A., Di Lella, L. A., Protano, G. and Riccobono, F.: 2004, ‘Arsenic in soil andvegetation of contaminated areas in southern Tuscany (Italy),’ Journal of Geochemical Exploration81, 1–14.

Bierlein, F. P. and McNaughton, N. J.: 1998, ‘Pb isotope fingerprinting of mesothermal gold depositsfrom Central Victoria, Australia: Implications for ore genesis’, Mineralium Deposita 33, 633–638.

Bierlein, F. P., McKnight, S., Foster, D. A. and Arne, D. C.: 1999, Timing of gold mineralization inthe Ballarat goldfield, Central Victoria: constraints from 40Ar(39Ar results’, Australian Journalof Earth Sciences 46, 301–309.

Clark, I., Cook, B. and Cochrane, G. C.: 1988, ‘Victorian geology excursion guide. AustralianAcademy of Science in conjunction with the Geological Society of Australia’, Clay Minerals40, 365–368.

Commission for the European Communities :1986, Report on the protection of the environment andin particular of soil when sewage sludge issued in agriculture. Official Journal of the EuropeanCommunities, Report 86(278(EEC. CEC, Brussels, Belgium.

Cullen, W. R. and Reimer, K. J.:1989, ‘Arsenic speciation in the environment’, Chemical Reviews 89,713–764.

Ferguson, J. F. and Gavis, J.: 1972, ‘A review of the arsenic cycle in natural waters’, Water Research6, 1259–1274.

Gao, S. and Burau, R. G.: 1997, ‘Environmental factors affecting rates of arsenic evolution from andmineralisation of arsenicals in soil’, Journal of Environmental Quality 26, 753–763.

Gao, Z. L. and Kwak, T. A. P.: 1997, ‘The geochemistry of wall rock alteration in turbidites-hostedgold vein deposits, Central Victoria, Australia’, Journal of Geochemical Exploration 59, 259–274.

Garcia, S. A. and Alvarez, A. E.: 2003, ‘Arsenic in soils and waters and its relation to geology andmining activities (Salamanca Province, Spain)’, Journal of Geochemical Exploration 80, 69–79.

Golub, M., Macintosh, M. and Baumrind, N.: 1998, ‘Arsenic development toxicity: Animal studiesand human concerns’, Journal of Toxicology and Environmental Health B. Critical Reviews 1,199–241.

Grabowski, L. A., Houpis, J. L. J., Woods, W. I. and Johnson, K. A.: 2001, ‘Seasonal bioavailabilityof sediment-associated heavy metals along the Mississippi river floodplain’, Chemosphere 45,643–651.

Gupta, S. K., Vollmer, M. K. and Krebs, R.: 1996, ‘The importance of mobile, mobilizable and pseudototal heavy metal fractions in soil for three-level assessment and management’, The Science ofthe Total Environment 178, 11–20.


Haritonidis, S. and Malea, P.: 1995, ‘Seasonal and local variation of Cr, Ni and Co concentrationsin Ulva rigida C. Agardh and Entermosorpha linza (Linnaeus) from Thermaikos Gulf, Greece’,Environmental Pollution 89, 319–327.

Harrington, J. M., Middaugh, J. P., Morse, D. L. and Housworth. J.: 1978, ‘A survey of a populationexposed to high concentrations of arsenic in well water in Fairbanks, Alaska’, American Journalof Epidemiology 108, 377–385.

Hinwood, A., Bannister, R., Shugg, A. and Sim, M.: 1998, ‘Environmental arsenic in rural Victoria.An update’, Water 25, 34–36.

Hinwood, A. L., Sim, M. R., Jolley, D., de Klerk, N., Bastone, E. B., Gerostamoulos, J. andDrummer, O. H.:2003, ‘Hair and toenail arsenic concentrations of residents living in ar-eas with high environmental arsenic concentrations’, Environmental Health Perspectives 111,187–193.

Hughes, M. F.: 2002, ‘Arsenic toxicity and potential mechanisms of action’, Toxicology Letters 133,1–16.

Hughes, M. J., Phillip, G. N. and Gregory, L. M.: 1997, ‘Mineralogical domains in the Victorian goldprovinces, Maldon, and Carlin-Style potential. Proceedings of the AusIMM Annual Conference,Ballarat, Victoria, pp. 215–227.

Lamb, I. A., Hughes, M. J. and Hughes, C. E.: 1996, ‘Dispersion of arsenic in soil and water ofthe Ballarat goldfield, Victoria’, in 1st International Conference on Containments in the SoilEnvironment in the Australasian-Pacific Region, Feb., 18–23, Adelaide, Australia.

McGeehan, S. L., Fendorf, S. E. and Naylor, D. V.: 1998, ‘Alteration of arsenic sorption in flooded-dried soils,’ Soil Science Society of American Journal 62, 828–833.

NHMRC: 1996. Australian Drinking Water guidelines. National Water Quality Management Strategy,National Health and Medical Research Council and Agriculture and Resources managementCouncil of Australia and New Zealand, Canberra, Australia.

Nriagu, J. O. and Pacyna, J. M.: 1988, ‘Quantitative assessment of world wide contamination of air,water and soils by trace metals’, Nature 338, 134–139.

O’Neill, P.: 1995, ‘Arsenic’, in: Alloway, B.J. (ed.), Heavy Metals in Soils, Chapman and Hall,Glasgow, pp. 105–121.

Peinemmann, N., Amiotti, N. M., Zalba, P. and Villamill, M. B.: 2000, ‘Effect of Clay Minerals andOrganic Matter on the Cation Exchange Capacity of Silt Fractions’, Journal of Plant Nutritionand Soil Science 163, 47–52.

Phillips, G. N. and Hughes, M. J.: 1996, ‘The geology and gold deposits of the Victorian gold province’,Ore Geology Reviews 11, 255–302.

Porcella, D.: 1987, ‘Assessing potential toxicity in aquatic ecosystems’, EPRI Journal 12, 47–51.Ramsay, W. R. H., Bierlein, F. P., Arne, D. C. and VandenBerg, A. H. M.: 1998, ‘Turbidites-hosted

gold deposits of Central Victoria, Australia: Their regional setting, mineralising styles, and somegenetic constraints’, Ore Geology Reviews 13, 131–151.

Rayment, G. E. and Higginson, F. R.: 1992. Australian Laboratory Handbook of Soil and WaterChemical Methods, Inkata Press, Sydney.

Sherrell, R. M. and Ross, J. M.: 1999, ‘Temporal variability of trace metals in New Jersey Pinelandsstream: Relationships to discharge and pH’, Geochimica et Cosmochimica Acta 63, 3321–3336.

Smedley, P. L. and Kinniburgh, D. G.: 2002, ‘A review of the source, behaviour and distribution ofarsenic in natural waters’, Applied Geochemistry 17, 517–568.

Smith, E., Naidu, R. and Alston, A. M.: 1998, ‘Arsenic in the soil environment: A review’, Advancesin Agronomy 64, 149–195.

Sultan, K.: 2005, ‘Distribution of arsenic and heavy metals in soils and surface waters in Cen-tral Victoria (Ballarat, Creswick and Maldon)’, PhD Thesis, School of Science & Engineering,University of Ballarat, Australia.


Sumner, M. E. and Miller, W. P.: 1996, ‘Cation exchange capacity and exchange coefficients’, inD.L. Sparks (ed.), Methods of Soil Analysis, Part 3: Chemical Methods. Soil Science Society ofAmerica Books, Madison, WI.

Taylor, D. H.: 1998, ‘Ballarat gold deposits’, in D. A. Berkman and D. H. Mackenzie (eds.), Geologyof Australian and Papua New Guinea Mineral Deposits, The Australian Institute of Mining andMetallurgy, Melbourne, pp. 543–548.

USEPA: 1996, Soil screening guideline: User’s guidance. 540/R-96/018. U.S. Environmental Protec-tion Agency, Washington, DC.

Villares, R., Puente, X. and Carballeira, A.: 2002, ‘Seasonal variation and background levels of heavymetals in two green seaweeds,’ Environmental Pollution 119, 79–90.

Warner, K. L.: 1998, Water-Quality Assessment of the Lower Illinois River Basin: EnvironmentalSetting. Urbana, United States Geological Survey, United States Department of the Interior.

WHO: 1996, WHO Guidelines for Drinking-Water Quality, Vol. 2. Health criteria and other supportinginformation. World Health Organization, Geneva, Switzerland.

seasonal changes in arsenic concentrations and ... · school of science and engineering, university...

Documents