benthic macroinvertebrate assemblages and macrohabitat connectivity in mediterranean-climate streams...

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors nonprofit publishers academic institutions researchlibraries and research funders in the common goal of maximizing access to critical research

Benthic macroinvertebrate assemblages and macrohabitat connectivity inMediterranean-climate streams of northern CaliforniaAuthor(s) Nuacuteria Bonada Maria Rieradevall Narciacutes Prat and Vincent H ReshSource Journal of the North American Benthological Society 25(1)32-43 2006Published By The Society for Freshwater ScienceDOI httpdxdoiorg1018990887-3593(2006)25[32BMAAMC]20CO2URL httpwwwbiooneorgdoifull1018990887-359328200629255B323ABMAAMC5D20CO3B2

BioOne (wwwbiooneorg) is a nonprofit online aggregation of core research in the biological ecological andenvironmental sciences BioOne provides a sustainable online platform for over 170 journals and books publishedby nonprofit societies associations museums institutions and presses

Your use of this PDF the BioOne Web site and all posted and associated content indicates your acceptance ofBioOnersquos Terms of Use available at wwwbiooneorgpageterms_of_use

Usage of BioOne content is strictly limited to personal educational and non-commercial use Commercial inquiriesor rights and permissions requests should be directed to the individual publisher as copyright holder

J N Am Benthol Soc 2006 25(1)32ndash43 2006 by The North American Benthological Society

Benthic macroinvertebrate assemblages and macrohabitatconnectivity in Mediterranean-climate streams of northern California

Nuria Bonada1 Maria Rieradevall2 AND Narcıs Prat3

Department of Ecology University of Barcelona E-08028 Barcelona Spain

Vincent H Resh4

Department of Environmental Science Policy and Management University of CaliforniaBerkeley California 96704-3112 USA

Abstract Drought leads to a loss of longitudinal and lateral hydrologic connectivity which causes director indirect changes in stream ecosystem properties Changes in macrohabitat availability from a rifflendashpoolsequence to isolated pools are among the most conspicuous consequences of connectivity lossMacroinvertebrate assemblages were compared among 3 distinct stream macrohabitats (riffles [R] poolsconnected to riffles [Pc] disconnected pools [Pd]) of 19 Mediterranean-climate sites in northern California toexamine the influence of loss of habitat resulting from drought disturbance At the time of sampling 10 siteswere perennial and included R and Pc macrohabitats whereas 9 sites were intermittent and included onlyPd macrohabitats Taxa richness was more variable in Pd and taxa richness was significantly lower in Pdthan in Pc but not R These results suggested a decline in richness between Pc and Pd that might beassociated with loss of connectivity Lower Ephemeroptera Plecoptera and Trichoptera (EPT) richnessrelative to Odonata Coleoptera and Heteroptera (OCH) richness was observed for Pd than R and Pcmacrohabitats Family composition was more similar between R and Pc than between R or Pc and Pdmacrohabitats This similarity may be associated with greater connectivity between R and Pcmacrohabitats Correspondence analysis indicated that macroinvertebrate composition changed along agradient from R to Pc and Pd that was related to a perennialndashintermittent gradient across sites Highvariability among macroinvertebrate assemblages in Pd could have been related to variability in theduration of intermittency In cluster analysis macroinvertebrate assemblages were grouped by macro-habitat first and then by site suggesting that the macrohabitat filter had a greater influence onmacroinvertebrate assemblages than did local site characteristics Few taxa were found exclusively in Pcand this macrohabitat shared numerous taxa with R and Pd indicating that Pc may act as a bridge betweenR and Pd during drought Drought is regarded as a ramp disturbance but our results suggest that theresponse of macroinvertebrate assemblages to the loss of hydrological connectivity among macrohabitats isgradual at least in Mediterranean-climate streams where drying is gradual However the changes may bemore dramatic in arid and semiarid streams or in Mediterranean-climate streams if drying is rapid

Key words macrohabitats riffles pools connectivity drought intermittent streams macroinvertebra-tes Mediterranean-climate streams

Environmental factors affecting organisms in stream

ecosystems can be thought of as hierarchically

organized (Frissell et al 1986 Church 1996) Factors

operating at each level of the hierarchy act as

mechanistic filters that influence the presence and

abundance of biota across a range of scales (Poff 1997)

Spatial and temporal heterogeneity of habitat (the

lsquolsquohabitat templetrsquorsquo) are 2 important components in the

organization of stream biotic communities (South-

wood 1977 1988 Townsend and Hildrew 1994) Of the

static (substrate) and dynamic (flow) features included

in the habitat-templet approach mean velocity and

complex hydraulic characteristics best explain the

distribution of benthic macroinvertebrates (Statzner

et al 1988) and riffles and pools are the most obvious

1 Present address CNRS-Ecologie des HydrosystemesFluviaux Universite Claude Bernard Lyon 1 F-69622Villeurbanne Cedex FranceE-mail nuriabonadauniv-lyon1fr

2 E-mail adresses mrieradevallubedu3 npratubedu4 vreshnatureberkeleyedu

32

macrohabitat features that can be distinguished in astream reach (Frissell et al 1986)

Stream ecologists traditionally have addressed howdifferences in physical factors between riffles andconnected pools affect macroinvertebrates (Logan andBrooker 1983 Cooper et al 1986 Brown and Brussock1991 Boulton and Lake 1992 Angradi 1996 Riberaand Vogler 2000 Carter and Fend 2001) periphyton(Keithan and Lowe 1985 Rosenfeld and Hudson 1997Whitledge and Rabeni 2000) and fish (Young 2001Inoue and Nunokawa 2002) The physical differencesbetween riffle and pool macrohabitats (eg velocityparticle size depth) are apparent but macrohabitat-related differences in macroinvertebrate compositionare not always clear For example some studies havereported similar numbers of taxa in riffles and pools(Egglishaw and Mackay 1967 Armitage et al 1974Logan and Brooker 1983) whereas others havereported that riffles are richer (Brown and Brussock1991 Carter and Fend 2001) or poorer (McCulloch1986 Boulton and Lake 1992) in taxa than pools Somestudies have reported that macroinvertebrate taxo-nomic composition differs between riffles and pools(ie composition is unique in each macrohabitatScullion et al 1982 McCulloch 1986) but significantoverlap also has been reported suggesting thatmacroinvertebrate taxonomic composition in thesemacrohabitats is not as discrete as often is presumed(Rabeni et al 2002) Moreover few of the above studieswere done in intermittent streams (Brown andBrussock 1991 Boulton and Lake 1992) Thus thedisconnected pool has been largely neglected as aseparate benthic macrohabitat from the connectedrifflendashpool sequence

In streams hydrologic connectivity refers to longi-tudinal (upstreamndashdownstream) lateral (riparianndashriv-er channel) and vertical (surface waterndashgroundwater)transfers of water (Ward et al 1999 Lake 2003)Droughts and floods are major disturbances affectingconnectivity with droughts decreasing connectivityand floods restoring it The riffle portions of rifflendashpoolsequences are gradually lost as drought progressesand the remaining pools often become disconnectedSuch pools may persist until the following wet seasonwhen floods restore rifflendashpool sequences (Boulton andLake 1992 Williams 1996 Gasith and Resh 1999 Lake2000 2003 Boulton 2003) Decreases in hydrologicconnectivity may cause changes in existing benthicmacrohabitats while also creating new macrohabitatseg riffles (R) pools connected to riffles (Pc) andisolated or disconnected pools (Pd) Change in macro-habitat patchiness over time is associated withvariability in natural annual discharge and can beconsiderable depending on stream size substrate

conditions and other features (Lake 2000 Bonada2003) Thus streams may simultaneously have R Pcand Pd macrohabitats that may persist for days ormonths depending on annual climate geomorphol-ogy or substrate (Bonada 2003)

Mediterranean-climate streams are characterized bysequential floods and droughts that are seasonallypredictable but show variable intensity (McElravy et al1989 Molina et al 1994 Gasith and Resh 1999)Seasonal drought is viewed as a ramp disturbance(ie the strength of the disturbance increases steadilyover time Lake 2000 2003) but ecological responses todrought may be stepped (ie responses occurring asirregular steps as disturbance increases) rather thanramped Steps are especially common when changescaused by drought cross critical thresholds such asisolation of the wetted bed from littoral vegetation lossof flow or loss of surface and hyporheic water (Boulton2003) However responses to critical thresholds mayvary with the intensity and timing of drought and maydepend on local habitat characteristics Thus moreempirical information is needed before ecologicalresponses to drought disturbance can be characterized

We examined assemblages of benthic macroinverte-brates at the macrohabitat (ie R Pc and Pd) scale atsites in Mediterranean-climate streams that rangedalong a gradient from perennial to intermittentHabitat is a key filter for macroinvertebrate assemb-lages (Poff 1997) so we hypothesized that assemblageswould differ among macrohabitats and that taxonomiccomposition would be more similar within the samemacrohabitat type at different sites than amongdifferent macrohabitats at a single site We alsohypothesized that reduced connectivity during dryingwould produce gradual rather than rapid changes inassemblages at the macrohabitat scale We predictedthat changes in assemblages would occur along agradient from R through Pc to Pd with Pc serving as abridge macrohabitat between connected and discon-nected conditions Last we predicted that assemblageswould be most dissimilar between Pd and PcRmacrohabitats and that Pc and R assemblages wouldbe most similar because of their high connectivity

Methods

Study area

Macroinvertebrates were quantified from 19 sitesalong 18 streams in the Mediterranean-climate area ofnorthern California during spring 2002 At the time ofsampling 9 sites were from 8 intermittent streamscontaining only Pd macrohabitats whereas the other10 sites were from 10 perennial streams containing Rand Pc macrohabitats Sites were distributed in the

2006] 33MACROHABITAT CONNECTIVITY AND MACROINVERTEBRATES

inland mountains of the Sierra Nevada and the coastalranges north and south of San Francisco California

Sites in the Sierra Nevada were fast-flowing reachesin forested basins with medium slopes and substratescomposed mainly of boulders and cobbles Instreamautotrophic biomass consisted mostly of diatom matsencrusting green algae and cyanobacteria and somemacrophytes Riparian vegetation was mostly AlnusSalix and Populus with some Pinus lambertiana andPseudotsuga menziesii intruding from the upland con-iferous forest

Sites in the northern coastal ranges had steepchannels containing coarse substrate except for theSan Geronimo River site where gravels sand andbedrock predominated Filamentous green algaediatoms and some macrophytes occurred at somesites Riparian vegetation was mostly Quercus lobataAlnus Corylus cornuta Sequoia sempervirens andUmbellularia californica

Sites in the southern coastal ranges had narrowsteep channels with substrates and instream vegeta-tion similar to the northern coastal sites Riparianvegetation was mostly Quercus lobata Platanus race-mosa Juglans hindsii Populus Salix Alnus Coryluscornuta and Umbellularia californica

Sampling

Sites were sampled following the GUADALMEDProject methods (Jaimez-Cuellar et al 2002) used inMediterranean-climate regions of Spain At each sitestreamwater pH temperature dissolved O2 (DO)conductivity and discharge were measured Macro-algae and macrophytes were identified visually in thefield and the number of taxa recorded Macroalgaltaxa were divided into those with filamentous andthose with encrusted growth forms Habitat diversitywas quantified using the fluvial habitat index (IHFPardo et al 2002 wwwguadalmedorg) developed forMediterranean-climate streams IHF includes 7 metricsand scores vary from 0 to 100 high values indicatehigh habitat diversity

Macroinvertebrates were sampled from R Pc and Pdmacrohabitats using a 250-lm-mesh kicknet Sampleswere examined sequentially in the field and additionalsamples were taken until 2 successive samples fromeach macrohabitat yielded no additional macroinverte-brate families Samples were preserved in 70 ethanolsorted to family in the laboratory and counted Taxawere quantified by rank abundance because of thesemiquantitative nature of the samples Rank abun-dances were 1frac141 to 3 individualssample 2frac14 4 to 10individualssample 3frac14 11 to 100 individualssampleand 4frac14100 individualssample

Data analysis

Analysis of variance (ANOVA) or KruskalndashWallistests were used to compare physical and chemicalvariables macroalgal and macrophyte richness andgeomorphologic characteristics of the IHF (IHF 1ndash7)between perennial (R and Pc) and intermittent sites(Pd) Total taxa richness and Ephemeroptera Plecop-tera and Trichoptera (EPT) richness relative to EPT thornOdonata Coleoptera and Heteroptera (OCH) richness(EPT[EPTthornOCH]) were analyzed among macrohabi-tats (R Pc and Pd) Data were checked for normalityand homogeneity of variances using KolmogorovndashSmirnov and Levenersquos tests respectively and trans-formed when appropriate STATISTICA (1999 Stat-Soft Tulsa Oklahoma) was used for all analyses

The Multi-Response Permutation Procedure (MRPP)computed with PC-ORD (version 420 MjM SoftwareGleneden Beach Oregon) was used to identify among-macrohabitat differences in macroinvertebrate compo-sition MRPP is a multivariate test for differencesamong predefined groups and provides the statistic Aand a p-value by permutation MRPP is nonparamet-ric so it is more appropriate than multiple analysis ofvariance for comparing relative abundances of speciesin data matrices containing many 0 counts Corre-spondence Analysis (CA) was done with CANOCO(version 4 Centre for Biometry Wageningen TheNetherlands) to describe among-macrohabitat differ-ences in macroinvertebrate assemblages CA uses v2

distance and is recommended over other ordinationmethods (eg Principal Components Analysis) whenusing rank abundances or when the data havenumerous 0 values (Legendre and Legendre 1998)

Cluster analysis based on BrayndashCurtis similarity wasdone using a flexible method (b frac14 025) to assesssimilarity in macroinvertebrate assemblages among RPc and Pd macrohabitats and to ascertain if assemb-lages were more similar within the same macrohabitatat different sites than among different macrohabitats ata single site Cluster analysis allows more detailedassessment of pairwise distances among macroinver-tebrate assemblages in macrohabitats at different sitesthan ordination techniques which include the wholeassociation matrix (Legendre and Legendre 1998)

The Indicator Value (IndVal Dufrene and Legendre1997) method was used to determine the mostrepresentative macroinvertebrate taxa in each macro-habitat IndVal identifies characteristic taxa from apredefined group (ie R Pc and Pd) based on therelative frequency of occurrence of a taxon in samplesof one group and the mean relative abundance of thattaxon in samples of one group compared with allgroups An indicator value (IV) and a p-value obtained

34 [Volume 25N BONADA ET AL

by Monte Carlo permutations (9999 permutations)

were associated with each taxon Only taxa with an IV

25 (ie species present in 50 of the samples from

one group and with a relative abundance in that group

of 50 Dufrene and Legendre 1997) were retained

Results

Environmental factors

Perennial sites (R and Pc macrohabitats) were

characterized by significantly higher DO saturation

than intermittent (Pd macrohabitat) sites (Table 1)

Streamwater pH temperature and conductivity were

similar among sites Discharge was highly variable in

perennial sites ranging from 32 to 6000 Ls Taxa

richness of filamentous and encrusted algae and

macrophytes did not differ between perennial and

intermittent sites (Table 1)

IHF scores were significantly higher in perennial

than intermittent sites (Table 1) Riffle embeddedness

and pool sedimentation (IHF-1) differed significantly

between perennial and intermittent sites (Table 1)

Riffle frequency (IHF-2) and velocitydepth regime

(IHF-4) also differed between perennial and intermit-

tent sites because riffles were absent from intermittent

sites No other habitat characteristics differed between

perennial and intermittent sites (Table 1)

Macroinvertebrate taxa richness and EPT(EPTthornOCH) ratio

Macroinvertebrate taxa richness was more variable

in Pd than in R or Pc (Fig 1A) Taxa richness was

higher in Pc than Pd (F frac14 427 p frac14 002 Fig 1A) The

EPT(EPTthornOCH) ratio differed significantly among

macrohabitats (H frac14 2224 p 0001) the EPT

(EPTthornOCH) ratio was highest in R and lowest in Pd

(Fig 1B)

Macroinvertebrate composition

Macroinvertebrate composition differed signifi-

cantly between R and Pd (MRPP A frac14 02714 p

0001) R and Pc (Afrac14 01182 p 0001) and Pc and Pd

(A frac14 01022 p 0001) The first 2 CA axes explained

284 of total variability among assemblages in the 3

macrohabitats Assemblages were ordered from R to

Pc and Pd along Axis 1 (Fig 2A) Assemblages in Pd

were separated along Axis 2 indicating high varia-

bility of macroinvertebrate composition among sites of

this macrohabitat Pd assemblages near the top of Axis

2 had high abundances of Chaoboridae Lymnaeidae

and Hydraenidae (Appendix Fig 2B) Pc assemblages

were largely intermediate between R and Pd Some Pc

assemblages were similar to Pd assemblages sharing

Corixidae Naucoridae and Hydrophilidae whereas

other Pc assemblages were more similar to R assemb-

TABLE 1 Mean (61 SD) values for environmental variables in perennial (riffle [R] and connected pool [Pc] macrohabitats) andintermittent (disconnected pool [Pd] macrohabitat) sites Taxa richness values for filamentous and encrusted macroalgae andmacrophytes were based on taxonomic identifications that were made visually in the field The total value of the Fluvial HabitatIndex (IHF) and the individual values of each habitat variable are presented IHF-1 frac14 riffle embeddedness (for perennial sites) orsedimentation in pools (for intermittent sites) IHF-2 frac14 riffle frequency IHF-3 frac14 substrate composition IHF-4 frac14 velocitydepthregime IHF-5 frac14 shading of stream bed IHF-6 frac14 elements of heterogeneity (including litter deposition) and IHF-7 frac14 aquaticvegetation cover

Variable Perennial Intermittent

ANOVA KruskalndashWallis test

F p H p

pH 76 (018) 77 (019) 2130 014Dissolved O2 ( saturation) 922 (1853) 496 (2439) 1862 001Temperature (8C) 167 (298) 183 (303) 142 025Conductivity (lScm) 2602 (18199) 3644 (15942) 1745 020Discharge (Ls) 13438 (200314) 00 (000) 1508 001Filamentous algae richness 07 (067) 06 (053) 017 068Encrusted algae richness 11 (120) 10 (050) 001 079Macrophyte richness 10 (115) 08 (083) 009 076IHF 690 (506) 517 (714) 37921 001IHF-1 100 (000) 50 (354) 1117 001IHF-2 86 (250) 20 (000) 1332 001IHF-3 151 (242) 167 (235) 118 028IHF-4 62 (063) 44 (088) 1140 001IHF-5 88 (257) 73 (218) 228 013IHF-6 78 (239) 73 (173) 053 046IHF-7 125 (425) 89 (417) 3484 008


lages sharing Helicopsychidae Odontoceridae andHydroptilidae (Fig 2B)

BrayndashCurtis cluster analysis showed a high degree ofseparation among assemblages in the 3 macrohabitats(Fig 3) The 1st split separated assemblages in Pd fromthose in R and Pc and the 2nd split separatedassemblages in R and Pc further Within macrohabitattypes assemblages in R were most similar to each otherwhereas assemblages in Pd were the most dissimilarconsistent with the CA results (Fig 3) However a fewsites were exceptional The assemblages in R in CoyoteCreek (gR in Fig 3) were more similar to assemblages inPc than to other assemblages in R whereas theassemblage in Pc in Schneider Creek (jPc in Fig 3)grouped more closely with assemblages in R than withother assemblages in Pc Last the assemblage in Pc inSausal Burns Creek (aPc in Fig 3) was more similar toassemblages in Pd than to assemblages in R or Pc

Of the taxa that occurred in R only 46 wereexclusive to the R macrohabitat Seventeen families(mostly EPT taxa) had an IV value significantly 25(characteristic taxa) in R (Table 2) The stonefliesChloroperlidae Perlidae Nemouridae and Peltoperli-dae were characteristic of R Nemouridae and Chlor-operlidae also had IVs 25 in Pc but these IV valueswere nonsignificant (Table 2) The caddisflies Hydro-psychidae Rhyacophilidae Polycentropodidae Philo-potamidae and Glossosomatidae and the mayfliesHeptageniidae and Ephemerellidae were characteristicof R Limnephilidae occurred in R and Pc but wascharacteristic of Pc (Table 2) The triclad familyDugesiidae and the dipteran families SimuliidaeTipulidae and Empididae were characteristic of R butEmpididae had a nonsignificant IV value in Pc (Table 2)

Pc shared many taxa with R and Pd Only 38 oftaxa were exclusive to Pc macrohabitat and only 6families (Ceratopogonidae Leptophlebiidae Limne-philidae Lepidostomatidae Calamoceratidae andGomphidae) were characteristic of Pc (Table 2)Leptophlebiidae also had a high but nonsignificantIV value in Pd Several taxa in Pc had IV values 25but were characteristic of R or Pd For exampleElmidae Baetidae Nemouridae EphemerellidaeChloroperlidae and Heptageniidae all had relativelyhigh IV values in Pc but were characteristic of RGerridae Dytiscidae and Sialidae had high IV valuesin Pc but were characteristic of Pd

Pd had the highest of taxa that were exclusive(113) to the Pd macrohabitat and 14 taxa werecharacteristic of this macrohabitat (Table 2) Hetero-ptera families especially Gerridae and Veliidae werecharacteristic of Pd as were the Coleoptera familiesDytiscidae Gyrinidae and Hydraenidae the Molluscfamilies Planorbidae Physidae and Lymnaeidae andthe crustacean groups Copepoda and Cladocera Othercharacteristic taxa were Culicidae Oribatidae Gam-maridae and Sialidae Oligochaeta was the only taxonwith IV 25 in R and Pd but Oligochaeta was notcharacteristic of either R or Pd

Discussion

Uniqueness of benthic assemblages in macrohabitats

Earlier studies comparing riffle and pool macro-invertebrate faunas reported distinct assemblages ineach macrohabitat (Scullion et al 1982 Logan andBrooker 1983 McCulloch 1986) However the numberof exclusive taxa reported in these macrohabitats wasvariable McCulloch (1986) found more unique taxa inpools (probably Pc) than in R whereas Scullion et al(1982) found the opposite pattern We found a slightlyhigher number of exclusive taxa in R than in Pc but a

FIG 1 Box-and-whisker plots showing the number ofmacroinvertebrate taxa (A) and ratio of EphemeropteraPlecoptera and Trichoptera (EPT) to EPT thorn OdonataColeoptera and Heteroptera (OCH) family richness [EPT(EPTthornOCH)] (B) in riffle (R) connected pool (Pc) anddisconnected pool (Pd) macrohabitats Plots show medians(horizontal bars) 25th and 75th percentiles (boxes) and 10th

and 90th percentiles (whiskers extending from boxes)


lower number of exclusive taxa in R than in PdCharacteristic taxa in R Pc and Pd were generallyconsistent with those of other studies (Armitage et al1974 Rabeni and Minshall 1977 Scullion et al 1982McCulloch 1986 Malmqvist et al 1993)

As in other studies (Scullion et al 1982 Logan andBrooker 1983 McCulloch 1986) R had a predominanceof EPT taxa whereas OCH taxa were more abundantin Pc and Pd than in R Pd macrohabitats had thelowest EPT(EPTthornOCH) ratios of the 3 macrohabitats

FIG 2 Correspondence Analysis (CA) plots of sites (A) and macroinvertebrate taxa (B) The of total variance explained byeach axis is shown next to the axis R frac14 riffles Pc frac14 connected pools Pd frac14 disconnected pools Codes for taxa as in Appendix


Pd could be a refuge from desiccation for some

tolerant lentic (ie OCH) but probably not for

rheophilic (ie EPT) macroinvertebrates (Brown and

Brussock 1991) This pattern would explain the low

similarity between R and Pd in our sites and it

underscores the value of using the EPT(EPTthornOCH)

ratio to characterize the assemblages in each macro-

habitat Pd also had several characteristic noninsect

taxa including Crustacea and Mollusca Except for

Physidae Mollusca are rarely recorded in intermittent

sites (Williams 1987) The presence of some Mollusca

in Pd could be explained by morphological (Eckblad

1973) or life-cycle (Brown 1982) adaptations along

with high variability in hydrology among years

(McElravy et al 1989) which could yield permanencein one year and intermittence the next

Does macrohabitat connectivity influence taxa richness

Differences in taxa richness between riffles andpools have been associated with habitat stability(McCulloch 1986 Boulton and Lake 1992) peakannual discharge and reach gradient (Carter and Fend2001) but sampling method and taxonomic resolutionalso are important factors (Logan and Brooker 1983)The degree of habitat connectivity also may influenceassemblage richness and diversity Drought causes lossof hydrologic connectivity among stream habitatsLow connectivity is hypothesized to yield low richness

FIG 3 BrayndashCurtis dendrogram of macroinvertebrate composition Rfrac14 riffles Pcfrac14 connected pools Pdfrac14disconnected pools Rand Pc assemblages with the same initial letter are from the same sampling site and river afrac14Sausal Burns Creek bfrac14San GeronimoRiver cfrac14Cronan Creek dfrac14Webb Creek efrac14Redwood Creek ffrac14 Saratoga Creek gfrac14Coyote Creek hfrac14 Spanish Creek ifrac14 SlateCreek j frac14 Schneider Creek


because exchanges of matter energy and organismsare constrained between patches (Ward et al 1999)Therefore low taxa richness should be expected indisconnected habitats Sheldon et al (2002) reportedlower taxa richness in ephemeral and temporary lakesthan semipermanent river channels in an Australiandryland river In our study taxa richness was lower inPd than in Pc presumably because of decreasedhydrologic connectivity in Pd In contrast taxa rich-ness was similar between R and Pc presumablybecause of relatively high connectivity between R Pcand other parts of the whole fluvial ecosystem

Boulton (2003) suggested that decreases in taxarichness at the stream-reach scale follow a stepped (vsramped) pattern during drought with faunal re-sponses occurring as irregular steps instead of gradualchanges as drying disturbance progresses At themacrohabitat scale one would expect R to dry firstTaxa richness should increase slightly in Pc as R driesbecause Pc may shelter some R taxa from desiccationbefore they emerge or die (ie concentration ofindividuals in Pc as drying progresses) Pc macro-habitat becomes Pd macrohabitat after cessation offlow so Pd could have similar taxa richness to Pc

before environmental conditions become harsh andtaxa richness declines because of desiccation Thussimilarity of taxa richness between Pc and Pd shouldbe a function of the time since cessation of flow In ourstudy we did not know how long Pd macrohabitatshad been disconnected from R macrohabitats but thehigh variability of richness among Pd macrohabitatmay have been related to the duration of hydrologicseparation The coefficient of variation (CV) of taxarichness was lower in Pc (115) than in R (153) orPd (261) suggesting that Pc may be the leastvariable macrohabitat during drought These patternssuggest that changes in taxa richness at the macro-habitat scale follow a ramped pattern during droughtie the loss of R is gradual and Pc provides aconsistent refuge before drying converts Pc to Pd

Does connectivity influence macroinvertebrate composition

Sheldon et al (2002) reported that temporary lakeassemblages were distinct from assemblages in morepermanent sites and we observed significant differ-ences in macroinvertebrate assemblage structureamong 3 stream macrohabitats The greatest similarity

TABLE 2 Indicator Value (IV) scores and associated p-values obtained by Monte Carlo permutations (9999 runs) formacroinvertebrate taxa in riffle (R) connected pool (Pc) and disconnected pool (Pd) macrohabitats Taxa in each macrohabitat arelisted in order of decreasing IV and only taxa with IV 25 are shown Characteristic taxa for a macrohabitat type have p 005

R Pc Pd

Taxon IV p Taxon IV p Taxon IV p

Hydropsychidae 964 001 Ceratopogonidae 626 005 Planorbidae 788 001Simuliidae 932 001 Leptophlebiidae 594 005 Gerridae 725 001Rhyacophilidae 795 001 Limnephilidae 563 005 Culicidae 717 001Chloroperlidae 792 001 Lepidostomatidae 55 005 Physidae 698 001Tipulidae 772 001 Elmidae 542 015 Copepoda 648 001Perlidae 77 001 Baetidae 536 014 Oribatidae 631 001Heptageniidae 712 001 Ostracoda 536 010 Dystiscidae 601 005Ephemerellidae 686 001 Gerridae 529 008 Gammaridae 556 001Nemouridae 663 001 Dytiscidae 525 010 Oligochaeta 521 044Baetidae 643 001 Empididae 432 020 Sialidae 49 005Elmidae 64 001 Calamoceratidae 40 001 Veliidae 475 001Empididae 623 001 Nemouridae 395 032 Leptophlebiidae 457 077Oligochaeta 544 026 Gomphidae 362 005 Cladocera 454 005Hydracarina 536 032 Sialidae 362 020 Dixidae 409 044Polycentropodidae 475 005 Ephemerellidae 348 063 Ostracoda 404 072Philopotamidae 475 001 Chloroperlidae 341 067 Gyrinidae 40 005Glossosomatidae 449 005 Heptageniidae 341 054 Hydraenidae 40 005Hydroptilidae 448 013 Nematoda 306 043 Hydrophilidae 342 012Psephenidae 393 018 Sericostomatidae 274 025 Lymnaeidae 333 005Peltoperlidae 368 005 Oribatidae 27 083 Corixidae 289 034Dugesiidae 353 005 Corduliidae 265 008 Haliplidae 262 010Brachycentridae 317 009 Psephenidae 257 068Uenoidae 308 020Limnephilidae 283 073Psychodidae 262 019Perlodidae 255 018


in taxa composition occurred between hydrologicallyconnected R and Pc macrohabitats Moreover thegreater variability in taxonomic composition observedin Pd (cf R and Pc) may have been a consequence ofdifferences in the duration of disconnection of Pd fromR at different sites (Stanley et al 1997 Lake 2003)

In temporary ponds predator abundance mayincrease as the duration of inundation increases(Schneider and Frost 1996) In our study macro-invertebrate composition of some Pd assemblageswas similar to composition of Pc assemblages whereasother Pd assemblages had a fauna distinctly differentfrom Pc assemblages and also had a high abundance ofpredators (eg Chaoboridae) This pattern suggeststhat some Pd macrohabitats were disconnected from Rfor a longer period than other Pd macrohabitats Onthe other hand differential colonization and predationin Pd may produce distinctive macroinvertebrateassemblages even among Pd sites that have beendisconnected for the same length of time

We propose that the high variability among macro-invertebrate assemblages in Pd macrohabitat at differ-ent sites may be explained by 1) the duration of thehydrologic disconnection of Pd from nearby R macro-habitats 2) differential macroinvertebrate colonization3) differential predation by colonizing predators and 4)other environmental factors that differ naturally amongPd macrohabitats For example changes in physico-chemical variables sedimentation rates and associatedchanges in algal and macrophyte biomass may occurafter loss of hydrologic connectivity during droughtand macroinvertebrate responses may be associatedwith these changes (Boulton 2003) We did not findsignificant differences in macrophyte and macroalgalrichness aquatic vegetation cover litter deposition orsubstrate composition between perennial and inter-mittent sites at the reach scale (Table 1) however someof these factors may be important at the microhabitatscale (eg within a single Pd) and could explain inpart the faunal uniqueness of a given site

Overall our results indicate that differences betweenR and Pc macrohabitats were greater than differencesamong perennial sites throughout the study area Ourresults are consistent with those reported from 2 Texasstreams (McCulloch 1986) and 3 Appalachian streams(Angradi 1996) but they differ from several UK streams(Logan and Brooker 1983) where riffles and pools at thesame site were more similar to each other than theywere to similar macrohabitats at different sites Angradi(1996) proposed that the scale of observation accountsfor whether one or the other conclusion is reached atlarge spatial scales high similarity is expected betweenassemblages in different macrohabitats within a sitebecause other large-scale filters (eg basin character-

istics zoogeography) may be more influential on faunalassemblages than local habitat characteristics In astudy comparing common taxa in riffles and pools inseveral Mediterranean-climate regions Bonada (2003)reported that differences among regions were greaterthan differences between macrohabitats Bonada (2003)attributed this result to differential influence of local vsregional environmental and historical processes How-ever relatively small-scale studies such as ours suggestthat macrohabitat type has a greater influence onmacroinvertebrate assemblages than local character-istics of a site

In summary our data suggest that at least for theseMediterranean-climate streams changes in taxa rich-ness and macroinvertebrate assemblages duringdrought follow a ramp model (but see Acuna 2004)Pc macrohabitat represents an intermediate conditionbetween RPc and Pd macrohabitats The rate ofdrying and other abiotic or biotic factors may becrucial in determining assemblage responses (Boulton2003) and rate of drying (ramp slope) may determinewhether faunal responses are stepped or ramped Rateof drying depends on climatic conditions channelmorphology and discharge during the wet season(Bonada 2003) Drought in Mediterranean-climatestreams may have a milder effect on macroinvertebrateassemblages than drought in more arid climates wherechannel drying is faster and may be similar tointermittence in temperate-deciduous streams wherestreambed drying may occur more slowly duringdrought On the other hand even in Mediterranean-climate streams ramped or stepped models couldapply depending on the habitat characteristics of a siteA ramped pattern could develop when site geo-morphology and hydrology allow isolated pools toform gradually during drought A stepped responsecould occur in situations where drying occurs sud-denly (eg gravel substrate very low winter flow)Natural or anthropogenic alterations of Mediterraneanstreams that increase aridity or modify habitatcharacteristics (eg climate change riparian vegeta-tion alterations water abstraction) may alter the wayassemblages respond to drought emphasizing thevulnerability of these ecosystems

Acknowledgements

We thank Patina Mendez Deborah Rudnick LeahBeche Faith Kearns and Raphael Mazor at theUniversity of California Berkeley for field andlaboratory assistance Sergi Vives at the University ofBarcelona Spain for statistical advice and AndrewBoulton and 2 anonymous referees for helpful com-ments on the manuscript This research was supported


by the GUADALMED Project (HID98ndash0323-C05 andREN2001ndash3438-C07) and a grant to NB from theMinisterio de Ciencia y Tecnologıa Spain

Literature Cited

ACUNA V 2004 Riparian forest and hydrology influences onthe ecosystem structure and function of a Mediterraneanstream PhD Thesis University of Barcelona BarcelonaSpain

ANGRADI T R 1996 Inter-habitat variation in benthiccommunity structure function and organic matterstorage in 3 Appalachian headwater streams Journal ofthe North American Benthological Society 1542ndash63

ARMITAGE P D A M MACHALE AND D C CRISP 1974 Asurvey of stream invertebrates in the Cow Green basin(Upper Teesdale) before inundation Freshwater Biology4369ndash398

BONADA N 2003 Ecology of the macroinvertebrate com-munities in Mediterranean rivers at different scales andorganization levels PhD Thesis University of BarcelonaBarcelona Spain (Available from httpwwwtdxcbuces)

BOULTON A J 2003 Parallels and contrasts in the effect ofdrought on stream macroinvertebrate assemblagesFreshwater Biology 481173ndash1185

BOULTON A J AND P S LAKE 1992 The ecology of twointermittent streams in Victoria Australia II Compar-isons of faunal composition between habitats rivers andyears Freshwater Biology 2799ndash121

BROWN A V AND P P BRUSSOCK 1991 Comparisons ofbenthic invertebrates between riffles and pools Hydro-biologia 22099ndash108

BROWN K M 1982 Resource overlap and competition inpond snails an experimental analysis Ecology 63412ndash422

CARTER J L AND S V FEND 2001 Inter-annual changes in thebenthic community structure of riffles and pools inreaches of contrasting gradient Hydrobiologia 459187ndash200

CHURCH M 1996 Channel morphology and typology Pages185ndash202 in G Petts and P Calow (editors) River flowsand channel forms Blackwell Science Oxford UK

COOPER S D T L DUDLEY AND N HEMPHILL 1986 Thebiology of chaparral streams in southern CaliforniaPages 139ndash151 in J DeVries (editor) Proceedings of theChaparral Ecosystem Research Conference Report No62 California Water Resources Center Davis California

DUFRENE M AND P LEGENDRE 1997 Species assemblages andindicator species the need for a flexible asymmetricalapproach Ecological Monographs 67345ndash366

ECKBLAD J W 1973 Population studies of three aquaticgastropods in an intermittent backwater Hydrobiologia41199ndash219

EGGLISHAW H J AND D W MACKAY 1967 A survey of thebottom fauna of streams in the Scottish Highlands IIISeasonal changes in the fauna of three streams Hydro-biologia 30305ndash334

FRISSELL C A W J LISS C E WARREN AND M D HURLEY

1986 A hierarchical framework for stream habitatclassification viewing streams in a watershed contextEnvironmental Management 10199ndash214

GASITH A AND V H RESH 1999 Streams in Mediterraneanclimate regions abiotic influences and biotic responses topredictable seasonal events Annual Review of Ecologyand Systematics 3051ndash81

INOUE M AND M NUNOKAWA 2002 Effects of longitudinalvariations in stream habitat structure on fish abundancean analysis based on subunit-scale habitat classificationFreshwater Biology 471594ndash1607

JAIMEZ-CUELLAR P S VIVAS N BONADA S ROBLES AMELLADO M ALVAREZ J AVILES J CASAS M ORTEGA IPARDO N PRAT M RIERADEVALL C E SAINZ-CANTERO ASANCHEZ-ORTEGA M L SUAREZ M TORO M R VIDAL-ABARCA C ZAMORA-MUNOZ AND J ALBA-TERCEDOR 2002Protocolo Guadalmed (PRECE) Limnetica 21187ndash204

KEITHAN E D AND R L LOWE 1985 Primary production andspatial structure of phytolithic growth in streams in theGreat Smoky Mountains National Park TennesseeHydrobiologia 12359ndash67

LAKE P S 2000 Disturbance patchiness and diversity instreams Journal of the North American BenthologicalSociety 19573ndash592

LAKE P S 2003 Ecological effects of perturbation by droughtin flowing waters Freshwater Biology 481161ndash1172

LEGENDRE P AND L LEGENDRE 1998 Numerical ecology 2nd

edition Elsevier Amsterdam The Netherlands

LOGAN P AND M P BROOKER 1983 The macroinvertebratefaunas of riffles and pools Water Research 17263ndash270

MALMQVIST B A N NILSSON M BAEZ P D ARMITAGE AND JBLACKBURN 1993 Stream macroinvertebrate communitiesin the island of Tenerife Archiv fur Hydrobiologie 128209ndash235

MCCULLOCH D L 1986 Benthic macroinvertebrate distribu-tions in the riffle-pool communities of two east Texasstreams Hydrobiologia 13561ndash70

MCELRAVY E P G A LAMBERTI AND V H RESH 1989 Year-to-year variation in the aquatic macroinvertebrate fauna ofa northern California stream Journal of the NorthAmerican Benthological Society 851ndash63

MOLINA C M R VIDAL-ABARCA AND M L SUAREZ 1994Floods in arid south-east Spanish areas a historical andenvironmental review Pages 271ndash278 in G Rossi (editor)Coping with floods Kluwer Academic PublishersDordrecht The Netherlands

PARDO I M ALVAREZ J CASAS J L MORENO S VIVAS NBONADA J ALBA-TERCEDOR P JAIMEZ-CUELLAR G MOYA NPRAT S ROBLES M L SUAREZ M TORO AND M R VIDAL-ABARCA 2002 El habitat de los rıos mediterraneosDiseno de un ındice de diversidad de habitat Limnetica21115ndash133

POFF N L 1997 Landscape filters and species traits towardsmechanistic understanding and prediction in streamecology Journal of the North American BenthologicalSociety 6391ndash409

RABENI C F K E DOISY AND D L GALAT 2002 Testing thebiological basis of a stream habitat classification using


benthic invertebrates Ecological Applications 12782ndash796

RABENI C F AND G W MINSHALL 1977 Factors affecting themicrodistribution of stream benthic insects Oikos 293ndash43

RIBERA I AND A P VOGLER 2000 Habitat types as adeterminant of species range sizes the example oflotic-lentic differences in aquatic Coleoptera BiologicalJournal of the Linnean Society 7133ndash52

ROSENFELD J S AND J J HUDSON 1997 Primary productionbacterial production and invertebrate biomass in poolsand riffles in southern Ontario streams Archiv furHydrobiologie 139301ndash316

SCHNEIDER D W AND T M FROST 1996 Habitat duration andcommunity structure in temporary ponds Journal of theNorth American Benthological Society 1564ndash86

SCULLION J C A PARISH N MORGAN AND R W EDWARDS1982 Comparison of benthic macroinvertebrate faunaand substratum composition in riffles and pools in theimpounded River Elan and the unregulated River Wyemid-Wales Freshwater Biology 12579ndash595

SHELDON F A J BOULTON AND J T PUCKRIDGE 2002Conservation value of variable connectivity aquaticinvertebrate assemblages of channel and floodplainhabitats of a central Australian arid-zone river CooperCreek Biological Conservation 10313ndash31

SOUTHWOOD T R E 1977 Habitat the templet for ecologicalstrategies Journal of Animal Ecology 46337ndash365

SOUTHWOOD T R E 1988 Tactics strategies and templetsOikos 523ndash18

STANLEY E H S G FISHER AND N B GRIMM 1997 Ecosystem

expansion and contraction in streams BioScience 47427ndash

435

STATZNER B J A GORE AND V H RESH 1988 Hydraulic

stream ecology observed patterns and potential appli-

cations Journal of the North American Benthological

Society 7307ndash360

TOWNSEND C R AND A G HILDREW 1994 Species traits in

relation to a habitat templet for river systems Fresh-

water Biology 31265ndash275

WARD J V K TOCKNER AND F SCHIEMER 1999 Biodiversity of

floodplain river ecosystems ecotones and connectivity

Regulated Rivers Research and Management 15125ndash

139

WHITLEDGE G W AND C F RABENI 2000 Benthic community

metabolism in three habitats in an Ozark stream

Hydrobiologia 437165ndash170

WILLIAMS D D 1987 The ecology of temporary waters The

Blackburn Press Caldwell New Jersey

WILLIAMS D D 1996 Environmental constraints in tempo-

rary waters and their consequences for the insect fauna

Journal of the North American Benthological Society 15

634ndash650

YOUNG K A 2001 Habitat diversity and species diversity

testing the competition hypothesis with juvenile salmo-

nids Oikos 9587ndash93

Received 21 August 2003

Accepted 27 November 2005


APPENDIX Codes for macroinvertebrate taxa plotted in Fig 2B

Taxon Code Taxon Code

Aeshnidae 1 Hydroptilidae 45Amelitidae 2 Lepidostomatidae 46Ancylidae 3 Leptoceridae 47Asellidae 4 Leptophlebiidae 48Baetidae 5 Lestidae 49Belostomatidae 6 Leuctridae 50Brachycentridae 7 Limnephilidae 51Caenidae 8 Lymnaeidae 52Calamoceratidae 9 Muscidae 53Calopterygidae 10 Naucoridae 54Cambaridae 11 Nematoda 55Ceratopogonidae 12 Nemouridae 56Chaoboridae 13 Notonectidae 57Chironomidae 14 Odontoceridae 58Chloroperlidae 15 Oligochaeta 59Cladocera 16 Oribatidae 60Coenagrionidae 17 Ostracoda 61Copepoda 18 Peltoperlidae 62Cordulegasteridae 19 Perlidae 63Corixidae 20 Perlodidae 64Corydalidae 21 Philopotamidae 65Culicidae 22 Physidae 66Dixidae 23 Planariidae 67Dugesiidae 24 Planorbidae 68Dytiscidae 25 Polycentropodidae 69Elmidae 26 Psephenidae 70Empididae 27 Psychodidae 71Ephemerellidae 28 Psychomyiidae 72Ephydridae 29 Pteronarcyiidae 73Erpobdellidae 30 Pyralidae 74Gammaridae 31 Rhyacophilidae 75Gerridae 32 Sciomyzidae 76Glossosomatidae 33 Sericostomatidae 77Gomphidae 34 Sialidae 78Gyrinidae 35 Simuliidae 79Haliplidae 36 Siphlonuridae 80Helicopsychidae 37 Sphaeridae 81Helophoridae 38 Stratiomyidae 82Heptageniidae 39 Tabanidae 83Hydracarina 40 Tipulidae 84Hydraenidae 41 Tricorythidae 85Hydrobiidae 42 Uenoidae 86Hydrophilidae 43 Veliidae 87Hydropsychidae 44


J N Am Benthol Soc 2006 25(1)32ndash43 2006 by The North American Benthological Society

Benthic macroinvertebrate assemblages and macrohabitatconnectivity in Mediterranean-climate streams of northern California

Nuria Bonada1 Maria Rieradevall2 AND Narcıs Prat3

Department of Ecology University of Barcelona E-08028 Barcelona Spain

Vincent H Resh4

Department of Environmental Science Policy and Management University of CaliforniaBerkeley California 96704-3112 USA

Abstract Drought leads to a loss of longitudinal and lateral hydrologic connectivity which causes director indirect changes in stream ecosystem properties Changes in macrohabitat availability from a rifflendashpoolsequence to isolated pools are among the most conspicuous consequences of connectivity lossMacroinvertebrate assemblages were compared among 3 distinct stream macrohabitats (riffles [R] poolsconnected to riffles [Pc] disconnected pools [Pd]) of 19 Mediterranean-climate sites in northern California toexamine the influence of loss of habitat resulting from drought disturbance At the time of sampling 10 siteswere perennial and included R and Pc macrohabitats whereas 9 sites were intermittent and included onlyPd macrohabitats Taxa richness was more variable in Pd and taxa richness was significantly lower in Pdthan in Pc but not R These results suggested a decline in richness between Pc and Pd that might beassociated with loss of connectivity Lower Ephemeroptera Plecoptera and Trichoptera (EPT) richnessrelative to Odonata Coleoptera and Heteroptera (OCH) richness was observed for Pd than R and Pcmacrohabitats Family composition was more similar between R and Pc than between R or Pc and Pdmacrohabitats This similarity may be associated with greater connectivity between R and Pcmacrohabitats Correspondence analysis indicated that macroinvertebrate composition changed along agradient from R to Pc and Pd that was related to a perennialndashintermittent gradient across sites Highvariability among macroinvertebrate assemblages in Pd could have been related to variability in theduration of intermittency In cluster analysis macroinvertebrate assemblages were grouped by macro-habitat first and then by site suggesting that the macrohabitat filter had a greater influence onmacroinvertebrate assemblages than did local site characteristics Few taxa were found exclusively in Pcand this macrohabitat shared numerous taxa with R and Pd indicating that Pc may act as a bridge betweenR and Pd during drought Drought is regarded as a ramp disturbance but our results suggest that theresponse of macroinvertebrate assemblages to the loss of hydrological connectivity among macrohabitats isgradual at least in Mediterranean-climate streams where drying is gradual However the changes may bemore dramatic in arid and semiarid streams or in Mediterranean-climate streams if drying is rapid

Key words macrohabitats riffles pools connectivity drought intermittent streams macroinvertebra-tes Mediterranean-climate streams

Environmental factors affecting organisms in stream

ecosystems can be thought of as hierarchically

organized (Frissell et al 1986 Church 1996) Factors

operating at each level of the hierarchy act as

mechanistic filters that influence the presence and

abundance of biota across a range of scales (Poff 1997)

Spatial and temporal heterogeneity of habitat (the

lsquolsquohabitat templetrsquorsquo) are 2 important components in the

organization of stream biotic communities (South-

wood 1977 1988 Townsend and Hildrew 1994) Of the

static (substrate) and dynamic (flow) features included

in the habitat-templet approach mean velocity and

complex hydraulic characteristics best explain the

distribution of benthic macroinvertebrates (Statzner

et al 1988) and riffles and pools are the most obvious

1 Present address CNRS-Ecologie des HydrosystemesFluviaux Universite Claude Bernard Lyon 1 F-69622Villeurbanne Cedex FranceE-mail nuriabonadauniv-lyon1fr

2 E-mail adresses mrieradevallubedu3 npratubedu4 vreshnatureberkeleyedu

32







Methods

Study area







Sampling



Data analysis












Results



























(Fig 1B)





















F p H p








Discussion



































R Pc Pd










Acknowledgements




Literature Cited













































435











139









634ndash650

















Methods

Study area







Sampling



Data analysis












Results



























(Fig 1B)





















F p H p








Discussion



































R Pc Pd










Acknowledgements




Literature Cited













































435











139









634ndash650















Sampling



Data analysis












Results



























(Fig 1B)





















F p H p








Discussion



































R Pc Pd










Acknowledgements




Literature Cited













































435











139









634ndash650
















Results



























(Fig 1B)





















F p H p








Discussion



































R Pc Pd










Acknowledgements




Literature Cited













































435











139









634ndash650
















Discussion



































R Pc Pd










Acknowledgements




Literature Cited













































435











139









634ndash650








































R Pc Pd










Acknowledgements




Literature Cited













































435











139









634ndash650

















Acknowledgements




Literature Cited













































435











139









634ndash650












Literature Cited













































435











139









634ndash650






















435











139









634ndash650











benthic macroinvertebrate assemblages and macrohabitat connectivity in mediterranean-climate streams...

Documents