Longitudinal patterns of invertebrates in the hyporheiczone of a glacial river

F. MALARD*, D. GALASSI †, M. LAFONT ‡, SYLVAIN DOLEDEC* AND J.V. WARD §

*UMR CNRS 5023, Ecologie des Hydrosystemes Fluviaux, Universite Claude Bernard Lyon 1, Bat. Forel, Villeurbanne Cedex,

France

†Dipartimento di Scienze Ambientali, Universita di L’Aquila, Italy

‡UR Hydrobiologie, Cemagref, 3bis Quai Chauveau, Lyon Cedex, France

§Department of Limnology, EAWAG/ETH, Ueberlandstrasse, Duebendorf, Switzerland

SUMMARY

1. Longitudinal changes in physicochemical factors and the composition of the inverteb-

rate community were examined in the hyporheic zone of a glacial river (Val Roseg,

Switzerland) over a distance of 11 km from the glacier terminus. Multivariate analysis was

used to determine the habitat preferences of taxa along an upstream-downstream gradient

of increasing temperature and groundwater contribution to river flow.

2. The hyporheos conformed to the longitudinal distribution model described for

zoobenthic communities of glacial rivers in that taxonomic richness increased with

distance from the glacier terminus. Spatial variation in taxonomic richness was best

explained by temperature, the influence of groundwater, and the amount of organic

matter. The overriding importance of these variables on the distribution of taxa was

confirmed by the multivariate analysis.

3. The hyporheic zone contributed significantly to the overall biodiversity of the Roseg

River. Whereas insect larvae were predominant in the benthos, hyporheic invertebrates were

dominated by taxa belonging to the true groundwater fauna and the permanent hyporheos.

Several permanently aquatic taxa (e.g. Nematoda, Ostracoda, Cyclopoida, Harpacticoida,

Oligochaeta) appeared exclusively in the hyporheic zone or they extended farther upstream

in the hyporheic layer than in the benthic layer. Leuctridae, Nemouridae, and Heptageniidae

colonised hyporheic sediments where maximum water temperature was only 4 �C.

4. Despite strong seasonal changes in river discharge and physicochemistry in hyporheic

water, the density and distribution of the hyporheos varied little over time.

5. Taxonomic richness increased markedly in the downstream part of a floodplain reach

with an extensive upwelling zone. Upwelling groundwater not only maintained a

permanent flow of water but also created several species-rich habitats that added many

species to the community of the main channel.

Keywords: Alps, groundwater, hyporheos, nutrients, water temperature

Introduction

Global climate change and the rapid retreat of most

European glaciers have stimulated research on the

diversity and spatial distribution of glacial stream

invertebrates. Several studies have recently been con-

ducted in a variety of glacial streams across Europe

(Brittain et al., 2001; Burgherr & Ward, 2001; Fureder

et al., 2001; Gislason et al., 2001; Lods-Crozet et al.,

2001a; Maiolini & Lencioni, 2001; Robinson, Uehlinger

& Hieber, 2001; Snook & Milner, 2001), New Zealand

(Milner, Taylor & Winterbourn, 2001a) and Greenland

(Friberg et al., 2001), in order to identify the primary

Correspondence: Florian Malard, UMR CNRS 5023, Ecologie des

Hydrosystemes Fluviaux, Universite Claude Bernard - Lyon 1 -

Bat. Forel, 43 Bd 11 Novembre 1918, F-69 622 Villeurbanne

Cedex, France. E-mail: malard@univ-lyon1.fr

Freshwater Biology (2003) 48, 1709–1725

� 2003 Blackwell Publishing Ltd 1709

physical and chemical variables determining the dis-

tribution of macroinvertebrates in glacier-fed rivers.

Based on the comprehensive data set arising from

these studies, Milner et al. (2001b) validated and

refined an earlier conceptual model (Milner & Petts,

1994) that related longitudinal changes in benthic

macroinvertebrate communities in glacial-fed rivers to

downstream increases in water temperature and

channel stability. According to this model, the likely

first occurrence of macroinvertebrate taxa downstream

from the glacier terminus is determined primarily by

thresholds of temperature and channel stability. The

Diamesinae (Chironomidae) are the sole members of

the fauna where maximum water temperature (Tmax) is

below 2 �C, but Orthocladiinae, Tipulidae and Oligo-

chaeta are added to the community where Tmax ranges

between 2 and 4 �C. Perlodidae, Taeniopterygidae,

Baetidae, Simuliidae and Empididae appear where

Tmax exceeds 4 �C, whereas Nemouridae, Leuctridae,

Heptageniidae, Rhyacophilidae and Chironominae

colonise only when Tmax is above 8 �C.

The hyporheic zone beneath the stream bed of

glacial streams is physically more stable than the

benthic layer and may provide more favourable

thermal conditions (Malard et al., 2001a). Thus, this

hyporheic zone can contribute significantly to the

biodiversity of glacial streams. With few exceptions

(Tilzer, 1968; Husmann, 1975; Lafont & Malard, 2001;

Malard et al., 2001b), however, research has focussed

on the ecology of the surface benthos. Hyporheic

community gradients downstream of glacial margins

could depart from the conceptual model of Milner

et al. (2001b) for several reasons. First, the taxonomic

richness of hyporheic invertebrate communities does

not necessarily increase continuously downstream of

glacial margins because environmental conditions in

the bed sediments of glacial streams may be deter-

mined more by site-specific attributes, such as sedi-

ment porosity and permeability rather than by the

distance to the glacier (Malard et al., 2001b). Secondly,

permanent aquatic taxa, unable to tolerate the harsh

environmental conditions at the surface, may colonise

hyporheic sediments (Malard et al., 2001b). Thirdly,

the distribution of hyporheic taxa, especially those

belonging to the permanent hyporheos and true

groundwater fauna, may be determined by longitud-

inal changes in surface water/groundwater inter-

actions. Indeed, several conceptual models have

emphasised the role of surface water/groundwater

linkages on biodiversity along the river continuum

(Creuze des Chatelliers, 1991; Stanford & Ward, 1993).

This paper examines the composition and distribu-

tion of invertebrates in shallow hyporheic sediments of

a glacial river (Val Roseg, Switzerland) over a distance

of 11 km from the glacier terminus. The composition

and longitudinal pattern of benthic invertebrate com-

munities in this glacial river were reported by Burgh-

err & Ward (2001). Using a recent multivariate method

[the outlying mean index (OMI), Doledec, Chessel &

Gimaret-Carpentier, 2000], we determined the habitat

preferences of taxa along a gradient of increasing

temperature and groundwater contribution to stream

flow. We also examined whether the longitudinal

distribution of the hyporheos varied in response to the

distinct seasonal shift in discharge and environmental

conditions which characterised glacier-fed rivers. We

hypothesised that (i) the longitudinal pattern of

hyporheic invertebrate communities would follow an

additive model similar to that described for the

zoobenthic community and (ii) groundwater inputs

into the stream bed sediments would promote the

establishment of permanent aquatic taxa, thereby

increasing taxonomic richness in the hyporheic zone.

Methods

Study site

The Roseg River is a glacial tributary of the River Inn

(Danube catchment) in the Bernina Massif of the Swiss

Alps (Fig. 1). The catchment area is 66.5 km2, 30 % of

which is glaciated. Subalpine coniferous forest is

restricted to the valley sides (treeline: 2300 m a.s.l.),

with larch (Larix decidua Mill.), stone pine (Pinus

cembra L.) and mugo pine (P. mugo Turra) predomin-

ant. The main glaciers (Tschierva: 6.2 km2; Roseg:

8.5 km2) have been retreating at an average rate of

about 20 m year)1 since the middle of the 19th

century (end of the Little Ice Age) (Fig. 1).

Four different reaches were distinguished along the

11.5-km long corridor of the Roseg River (Fig. 1): an

unstable proglacial braided reach (length: 650 m), a

single thread channel incised in glacial till (length:

700 m), a complex flood plain (length: 2800 m; width:

130–510 m) the downstream part of which behaved

as a major groundwater upwelling zone, and a

long forested reach constrained by steep valley

slopes (length: 7200 m). Detailed information on the

1710 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

geomorphological, hydrological and hydrochemical

characteristics of this glacial river system have been

reported by several authors (Tockner et al., 1997;

Malard, Tockner & Ward, 1999; Ward et al., 1999;

Zah, Niederost & Uehlinger, 2000).

River discharge is gauged at the end of the

catchment by the Swiss National Hydrological and

Geological Survey (Fig. 2). Mean annual discharge for

the period 1955–97 was 2.8 m3 s)1. The river exhibits a

distinct annual flow pulse: daily mean discharge peaks

during the ice-melt season (from 6 to 10 m3 s)1 in July

and August) and is minimum in winter (ca. 0.2 m3 s)1

from December to March) when surface flow is

sustained solely by groundwater inputs.

Sampling

Sampling was carried out in June (river discharge:

Q ¼ 3.78 ± 0.51 m3 s)1), August (Q ¼ 7.85 ± 0.49 m3 s)1),

September (Q ¼ 3.19 ± 0.13 m3 s)1) and November

1997 (Q ¼ 0.65 ± 0.10 m3 s)1) (Fig. 2) at 11 sites

located over a distance of 11 km from the terminus

of the Tschierva Glacier (Fig. 1). Sites were more

concentrated where ground water was shown to

upwell in the downstream part of the floodplain

reach (Malard et al., 1999). All sites had permanent

surface and hyporheic flow, except site 4 which fell

dry from mid-December to the beginning of April.

During this period, the benthic and shallow hyporheic

sediments of most channels located in the upper flood

plain were also dry. Streambed sediments at all sites

were composed of an extremely heterogeneous mix-

ture of cobble, pebble, gravel, sand and silt, the

porosity of which varied between 8.6 and 22.9% (Zah,

2001). Surface water temperature was recorded at

hourly intervals using Minilog temperature loggers

(Vemco Ltd, Canada) deployed at sites 1–3, 5, 8 and

11. Physical characteristics of sampling sites are given

in Table 1.

Three replicate hyporheic samples were collected

at each site and date from randomly selected

locations in riffle/run habitats. However, only one

hyporheic replicate was collected at sites 6–11 in

November because of inclement weather. Hyporheic

invertebrates were collected by hammering a pipe

(0.025 m internal diameter, with 0.005-m holes at the

tip) to a depth of 30 cm below the streambed. Ten

litres of interstitial water and sediment were imme-

diately extracted using a Bou-Rouch pump (Bou &

Rouch, 1967) and filtered through a 100-lm-mesh

net. Samples were preserved in 4% formaldehyde.

Because the sampling method was based on pump-

ing water rather than removing an exact portion of

Fig. 2 Hydrograph of the Roseg River in 1997 (daily discharge).

Arrows indicate the sampling dates for hyporheic fauna.

Fig. 1 Map of the Val Roseg catchment showing the distribution

of geomorphological reaches and location of sampling sites

(1–11). Broken lines indicate the successive position of the

Tschierva Glacier since 1850.

Longitudinal pattern of glacial hyporheos 1711

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

habitat, the exact region of the streambed that was

sampled could not be guaranteed (but see Fraser &

Williams, 1997; Hunt & Stanley, 2000; Scarsbrook &

Halliday, 2002; Boulton, Dole-Olivier & Marmonier,

2003). A volume of 1 L of hyporheic water was

withdrawn for physicochemical analysis from the

third sampling pipe using a peristaltic pump (Willy

A. Bachofen AG, Basel, Switzerland). Water samples

were filtered (preashed Whatman GF/F filters;

0.7 lm) within 1–4 h, and the filtrate stored for

1–3 days at 4 �C prior to analysis.

Physicochemical analyses

Temperature, specific conductance (temperature ref-

erence: 20 �C), dissolved oxygen and pH were

measured in the field with portable meters (WTW

LF 323-B conductivity meter, WTW Oxi 330 oxygen

meter; Wissenschaftlich-Technische Werkstatten

GmbH & Co. KG, Weilheim, Germany. Orion 230A

pH meter; Thermo Orion, Beverly, MA, USA). Cal-

cium, magnesium and sodium were analysed with

an inductively coupled plasma-optical emission spec-

trometer (SPECTRO Analytical Instruments, Kleve,

Germany). Determination of total inorganic carbon

(TIC) was made by CO2 detection (Horiba IR-detec-

tor) after samples had been acidified and heated to

860 �C. Silica (SiO2) was analysed with the molyb-

date-heteropoly blue method (Clesceri, Greenberg &

Eaton, 1998). Specific conductance and the concentra-

tions of TIC, base cations (calcium, magnesium and

sodium) and silica were measured as indicators of

groundwater inputs into the stream because these

physicochemical variables displayed much higher

values in ground water than in glacial water (Malard

et al., 1999). Ammonium and nitrate were measured

with the indophenol-blue method (Clesceri et al.,

1998) and the automated hydrazine reduction

method (Downes, 1978), respectively. Soluble react-

ive phosphorous (SRP) and dissolved non-reactive

phosphorous (NRP) were measured with the molyb-

denum blue method (Vogler, 1965). Dissolved

organic carbon (DOC) was determined by wet

oxidation with subsequent acidification and CO2

IR-detection (Clesceri et al., 1998).

Identification of invertebrates

Invertebrates were sorted, counted and identified to

the lowest practicable level depending on their

development and condition. Ostracoda could not

be identified because their valves were decalcified

by formaldehyde. Most insect larvae were early

instars (body size approximately 3 mm) that were

identified to tribe, subfamily or family. Nematoda

and Hydracarina were not identified further. Fol-

lowing the classification by Gibert et al. (1994), taxa

were assigned to three ecological categories: occa-

sional hyporheos, permanent hyporheos, and stygo-

bites. The occasional hyporheos consisted mainly of

early instar larvae of aquatic insects that resided in

the bed sediment, whereas later life stages occurred

in the benthic layer. The permanent hyporheos was

composed of permanently aquatic taxa (i.e. Nematoda,

Table 1 Physical characteristics of surface water at the sampling sites. Means and ranges in parentheses are given for specific

conductance, turbidity (nephelometric turbidity units) and near-bed velocity. Annual degree days are expressed as Celsius tem-

perature units (CTU)

Site

Altitude

(m a. s. l.)

Distance

from glacier

terminus (m)

Annual mean

temperature

(�C)

Max. water

temperature

(�C)

Annual

degree days

(CTU)

Specific

conductance

(lS cm)1 at 20 �C)

Turbidity

(NTU)

Near-bed

velocity

(m s)1)

1 2122 414 0.48 2.10 176 36 (28–59) 143 (20–280) 0.67 (0.39–0.96)

2 2096 769 0.68 4.20 248 41 (30–75) 113 (9–226) 0.46 (0.19–0.69)

3 2064 1324 1.44 5.00 543 45 (30–76) 118 (22–280) 0.48 (0.05–0.76)

4 2045 1880 – – – 46 (31–79) 104 (16–255) –

5 2021 2579 2.36 7.60 806 47 (31–80) 94 (12–220) 0.40 (0.13–0.64)

6 2009 3218 – – – 49 (34–82) 80 (5–170) 0.45 (0.29–0.62)

7 1998 3668 – – – 49 (35–77) 75 (6–164) –

8 1981 4389 2.59 10.9 945 51 (35–82) 73 (4–150) 0.39 (0.24–0.59)

9 1901 6612 – – – – – 0.42 (0.15–0.68)

10 1862 7659 – – – 51 (38–80) 66 (6–148) 0.50 (0.13–0.80)

11 1773 10642 3.29 13.1 1302 51 (39–75) 53 (5–105) 0.42 (0.10–0.73)

1712 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

Oligochaeta, Ostracoda, Cyclopoida, Harpacticoida

and Hydracarina) that could spend their entire

lives either in the hyporheic layer or in the benthic

layer. Stygobites are true groundwater taxa that

complete their entire life cycle exclusively in subsur-

face water.

After removal of invertebrates, the remaining sedi-

ment from each sample was agitated and rinsed with

water to extract the organic matter that was only

loosely associated with the sediment (LOM: loosely

associated organic matter). The ash-free dry mass of

LOM was determined by drying and combustion and

expressed as milligram per 10-L hyporheic sample

(Pusch & Schwoerbel, 1994). Then, the ash-free dry

mass of the organic matter which remained on rinsed

sediments was also determined and expressed as

milligram strongly associated organic matter (SOM)

per gram of dry sediment.

Data analysis

Repeated measures analysis of variance (RMANOVARMANOVA)

was used to test for differences in taxonomic

richness and density of invertebrate assemblages

between sites and dates. The sampling date was

introduced as a repeated measures factor in the

analysis and its statistical significance was tested

using Wilks’ lambda multivariate test. Post hoc

Tukey’s honest significant difference (HSD) tests

were performed to determine pairwise differences

when significant differences among sites were

observed. Invertebrate densities were log10(x+1)

transformed prior to statistical analysis in order to

minimise differences among variances. Data were

tested for normality using the Chi-square test of

normality. Significance for all statistical analyses was

accepted at a ¼ 0.05; tests were performed with the

STATISTICA software package (STATSOFT, 1994).

The OMI analysis (Doledec et al., 2000) was used to

separate the habitat preferences of taxa along the

upstream–downstream environmental gradient. This

two-table ordination method decomposes variability

in the habitat of a taxon into three components. The

OMI, or taxon marginality, measures the distance

between the mean habitat conditions used by a taxon

(taxon centroid) and the mean habitat conditions in the

hyporheic zone of the stream. A Monte-Carlo permu-

tation test is used to check the statistical significance of

the marginality for each taxon. The tolerance which

corresponds to the dispersion of sampling sites con-

taining a taxon along the environmental gradient,

represents a measurement of habitat breadth. The

residual tolerance represents the proportion of vari-

ability in the habitat of a taxon that is not accounted for

by measured environmental variables.

The environmental data consisted of the average

values per site of 16 physicochemical variables meas-

ured in the hyporheic zone. The faunal data table

contained for each taxon the average number of

individuals collected at each site. Taxon densities

were log10(x+1) transformed to reduce the effect of

dominant taxa. The OMI analysis was used to position

taxa along the environmental gradient based on a

maximisation of their average marginality. The statis-

tical significance of the average marginality of all taxa

was tested using a global Monte-Carlo permutation

test. Multivariate analyses and graphical displays

were performed using ADE-4 software (Thioulouse

et al., 1997; software available on http://pbil.univ-

lyon1.fr/ADE-4/).

Results

Longitudinal changes in physicochemistry

The mean values (four dates) of physicochemical

variables measured at each site are given in Table 2.

From June to September, hyporheic water tempera-

ture increased from 1.5 �C in the proglacial reach to

6.4 �C at the downstream limit of the study (Fig. 3).

Hyporheic and surface water temperatures were

linearly correlated (P < 0.01) but the temperature of

hyporheic water temperature was on average 0.7 �Chigher than that of surface water (Fig. 4). Specific

conductance and the concentrations of TIC, base

cations and silica increased with distance from the

glacier terminus, although the increase was typically

more pronounced in the downstream part of the flood

plain (Fig. 3). The amount of LOM also increased

downstream and peaked in the constrained forested

reach (except site 10). In contrast, the highest concen-

trations of nitrate, ammonium and soluble reactive

phosphorus were measured immediately downstream

of the glacier terminus. The other physicochemical

variables did not exhibit a clear longitudinal pattern.

Dissolved oxygen was close to 100% saturation and

the concentrations of DOC and SOM remained

extremely low at all sites (Table 2).

Longitudinal pattern of glacial hyporheos 1713

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

Composition, richness and density of invertebrate

assemblages

More than 32 000 invertebrates were collected from

the hyporheic zone of the main channel of the Roseg

River. A total of 42 taxa were identified, 17 of which

belonged to the occasional hyporheos, 21 to the

permanent hyporheos and four were stygobites

(Table 3). Invertebrate assemblages were dominated

by taxa belonging to the stygobites and permanent

hyporheos, which collectively represented 50.2 and

31.2% of all individuals. Orthocladiinae (Chironom-

idae), Leuctridae (Plecoptera), and Heptageniidae

(Ephemeroptera) were by far the most abundant and

frequent insect taxa.

Taxonomic richness increased linearly with the

logarithm of density [richness ¼ 4.83 log10 (den-

sity) + 0.20, n ¼ 120, r2 ¼ 0.49, P < 0.0001). Taxo-

nomic richness and invertebrate density exhibited a

major discontinuity along the longitudinal dimension

(site effect: richness, F ¼ 39.81, P < 0.0001; density,

F ¼ 27.27, P < 0.0001, Fig. 5). Sites 1–4, located

upstream of the lower flood plain, contained signifi-

cantly fewer taxa and individuals than downstream

sites (Tukey’s tests, P < 0.05). Taxonomic richness and

invertebrate density varied significantly over time

(date effect: richness, F ¼ 28.41, P < 0.0001; density,

F ¼ 12.78, P < 0.0001, Fig. 5). Samples contained

significantly fewer taxa in August than in June,

September and November, and significantly more

taxa in November than in June and September

(Tukey’s tests, P < 0.01). Invertebrate density was

significantly lower in August and September than in

June and November. Cumulated taxonomic richness

(i.e. the total number of taxa collected at each site over

the four sampling dates) increased logarithmically

with distance from the glacier and was positively

related to the average hyporheic water temperature

and average concentrations of TIC and LOM (Fig. 6).

Distribution of taxa along the environmental gradient

The first axis of the OMI analysis arranged the

sampling sites along an upstream–downstream envi-

ronmental gradient of increasing temperature and

groundwater contribution to stream flow (Fig. 7a,b).

High concentrations of dissolved phosphorus (NRP

and SRP), ammonium and nitrates characterised

the most upstream sites (i.e. sites 1–3) whereas

Tab

le2

Av

erag

ev

alu

es(±

SD

,n¼

4d

ates

)o

fp

hy

sico

chem

ical

var

iab

les

mea

sure

din

the

hy

po

rhei

czo

ne

of

the

Ro

seg

Riv

er

Sam

pli

ng

site

s1

23

45

67

89

1011

Tem

per

atu

re(�

C)

1.4

±0.

21.

7±

0.6

3.0

±1.

13.

5±

1.3

3.8

±1.

33.

9±

1.6

4.1

±1.

64.

3±

1.8

4.8

±2.

05.

0±

2.1

5.2

±2.

3

Dis

solv

edo

xy

gen

(%)

97±

396

±9

99±

298

±2

96±

198

±3

98±

399

±0

99±

296

±5

97±

3

pH

6.8

±0.

06.

9±

0.1

6.7

±0.

16.

6±

0.1

6.6

±0.

26.

7±

0.2

6.7

±0.

26.

8±

0.1

6.7

±0.

26.

8±

0.1

6.8

±0.

1

Sp

ecifi

cco

nd

uct

ance

(lS

cm)

1)

43.7

±17

.346

.6±

19.2

46.7

±16

.645

.2±

15.3

48.2

±15

.748

.3±

15.7

51.6

±14

.053

.9±

15.9

57.6

±15

.555

.6±

15.0

55.8

±15

.7

TIC

(mg

L)

1)

3.8

±1.

04.

1±

1.2

4.2

±1.

24.

0±

1.1

4.3

±1.

24.

5±

1.4

5.0

±1.

35.

1±

1.5

5.3

±1.

25.

2±

1.2

5.2

±1.

0

Ca2

+(m

gL)

1)

8.5

±3.

19.

1±

3.5

9.4

±3.

08.

6±

2.9

9.2

±3.

39.

1±

3.1

10.0

±3.

010

.1±

2.9

10.8

±3.

110

.4±

2.9

10.5

±3.

1

Mg

2+

(mg

L)

1)

0.62

±0.

290.

59±

0.23

0.65

±0.

190.

60±

0.20

0.65

±0.

220.

60±

0.20

0.69

±0.

210.

82±

0.19

0.90

±0.

260.

91±

0.21

0.86

±0.

28

Na+

(mg

L)

1)

0.47

±0.

310.

44±

0.30

0.38

±0.

200.

31±

0.19

0.37

±0.

220.

41±

0.21

0.50

±0.

250.

56±

0.28

0.72

±0.

380.

69±

0.36

0.74

±0.

37

SiO

2(m

gL)

1)

1.81

±0.

501.

84±

0.59

1.48

±0.

501.

50±

0.48

1.62

±0.

451.

84±

0.47

2.07

±0.

532.

18±

0.61

2.89

±0.

732.

88±

0.92

3.00

±1.

01

NH

4+

-N(l

gL)

1)

6±

85

±9

6±

88

±5

5±

22

±2

4±

32

±1

3±

22

±1

4±

3

NO

3-N

(lg

L)

1)

307

±10

231

9±

112

260

±10

125

5±

8927

5±

108

227

±62

233

±69

249

±91

263

±86

261

±91

269

±10

0

SR

P(l

gL)

1)

8±

74

±4

4±

42

±1

3±

21

±1

1±

02

±1

2±

11

±1

2±

1

NR

P(l

gL)

1)

3±

21

±2

2±

22

±1

2±

11

±1

1±

11

±1

2±

11

±1

0±

0

DO

C(m

gL)

1)

0.3

±0.

20.

3±

0.2

0.4

±0.

20.

3±

0.2

0.5

±0.

50.

3±

0.0

0.5

±0.

20.

4±

0.2

0.4

±0.

20.

4±

0.1

0.4

±0.

1

LO

M(m

g10

L)

1)

15.8

±1.

612

.8±

2.0

15.1

±3.

114

.7±

4.2

16.3

±2.

715

.7±

2.4

16.4

±5.

022

.1±

5.3

26.1

±4.

017

.6±

4.9

31.7

±11

.4

SO

M(m

gg)

1d

w)

1.9

±0.

32.

2±

0.7

2.4

±0.

62.

3±

0.3

2.6

±0.

32.

7±

0.2

2.5

±0.

32.

4±

0.4

2.5

±0.

52.

5±

0.4

2.5

±0.

7

1714 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

temperature and the concentrations of groundwater

indicators (i.e. magnesium, sodium, silica, calcium,

specific conductance and TIC) and LOM increased

with increasing distance to the glacier terminus. The

average marginality of all taxa was highly significant

(P < 0.001, global Monte-Carlo permutation test).

Therefore, taxa were plotted along the first axis of

the OMI analysis, which represented 87.1% of the

explained variability (Fig. 7c). Twenty of 42 taxa,

among which were only five insect taxa, departed

significantly from a uniform distribution along the

environmental gradient (P < 0.1, Table 4). The tur-

bellarian Crenobia alpina and the harpacticoid Mara-

enobiotus insignipes insignipes were the only taxa that

occurred preferentially in the proglacial reach

(Fig. 7c). The other 17 taxa occurring in the proglacial

Fig. 3 Longitudinal patterns of temperature, total inorganic carbon (TIC), sodium, silica, ash-free dry mass of loosely associated

organic matter (LOM), nitrate, ammonium and soluble reactive phosphorus in shallow hyporheic water of the Roseg River. Filled

circles: 7 June 1997; open squares: 2 August; filled triangles: 27 September; open diamonds: 14 November.

Longitudinal pattern of glacial hyporheos 1715

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

reach were either uniformly distributed along the

environmental gradient (i.e. P > 0.10) or they had

about half of their habitat variability consisting of

residual tolerance (i.e. Troglochaetus cf. beranecki,

Corynoneura spp., Hydracarina, Orthocladiinae, Para-

stenocaris glacialis, Dorydrilus michaelseni (Table 4).

Twenty five of 42 taxa, among which eight taxa

showed a significantly non-uniform distribution, had

their centroids located in the downstream part of the

flood plain (i.e. between sites 5 and 8) (Fig. 7c). The

most downstream sites (i.e. sites 9–11) were colonised

preferentially by the oligochaetes Cernosvitoviella

atrata, C. carpatica, Nais communis, the copepods

Bryocamptus (A.) cuspidatus, Diacyclops bisetosus, the

ostracods, the empidid Rhabdomastix sp. and the

simuliids. The average tolerance of taxa to different

environmental conditions, i.e. the average of tolerance

values of all taxa present at a particular site, decreased

logarithmically with increasing distance from the

glacier margin (Fig. 8).

The initial data set (i.e. 4 dates · 11 sites · 42 taxa)

was reorganised in order to examine the temporal

stability of the distribution pattern of hyporheic

invertebrate assemblages (Fig. 9). Taxa were ordi-

nated according to their marginalities along axis 1 of

the OMI analysis. Taxon densities were graphically

displayed using squares, the sides of which are

proportional to the Napierian logarithm of the aver-

age number of individuals (n ¼ 3 replicates per site

and date). The longitudinal pattern of hyporheic

assemblages identified by the OMI analysis on the

average densities of taxa remained stable over time

(Fig. 9). Indeed, the position of taxa along the up-

stream–downstream environmental gradient from the

glacier margin varied little despite strong temporal

changes in river discharge (Fig. 2) and physicochem-

ical attributes of hyporheic water (Fig. 3). Several

insect taxa including Diamesinae, Orthocladiinae,

Chloroperlidae, Nemouridae, and Baetidae had signi-

ficantly higher densities in November than in June,

August, and September.

Discussion

Environmental gradient

Longitudinal changes in the physicochemistry of

shallow hyporheic water mirrored those described

for surface water in the Val Roseg (Tockner et al., 2002;

Uehlinger, Malard & Ward, 2003). The steep rise in

hyporheic water temperature at site 3 was caused by

the confluence of relatively warm water (i.e. average

summer temperature 4 �C) from Roseg Lake. The

glacier is the main source of particulate phosphorus

(glacial flour enriched in apatite) and nitrogen in the

Roseg catchment (Tockner et al., 2002), thereby result-

ing in high concentrations of SRP, ammonium and

nitrate in hyporheic water of the proglacial reach. The

increase in specific conductance and concentrations of

TIC, SiO2 and base cations in the downstream part

of the flood plain was caused by the upwelling of

ground water which occurred via either direct

groundwater inflow into the sediments of the main

channel or spring brooks on the floodplain surface

(Malard et al., 1999). Using a two-end member mixing

model based on the concentration of sodium, Malard

et al. (2001b, 2003) demonstrated that the contribution

of hillslope ground water to the flow of surface and

shallow hyporheic waters at the downstream end of

the flood plain varied seasonally from less than 10%

during the main ablation period in summer to more

than 60% in winter. Upwelling of ground water in the

lower part of the flood plain had no detectable effect

on the longitudinal pattern of temperature and nutri-

ent concentrations (except for SiO2) of shallow hypor-

heic water. This concurs with the results of previous

studies which demonstrated that the flood plain had a

negligible influence on the temperature and concen-

trations of dissolved nutrients in surface water of the

main channel (Tockner et al., 2002; Uehlinger et al.,

2003). The downstream increase in the amount of

LOM probably reflected the longitudinal change in

Fig. 4 Relationship between the temperature of shallow

hyporheic water and surface water. Broken line indicates 1 : 1

equivalence relationship.

1716 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

terrestrial vegetation and resulting inputs of allochth-

onous organic matter. Zah & Uehlinger (2001)

demonstrated that the input of allochthonous organic

matter to the Roseg River increased longitudinally as

the riparian zone changed from barren glacial sedi-

ments in the proglacial reach to subalpine coniferous

forests in the constrained reach.

Composition and distribution pattern of the hyporheos

Despite harsh environmental conditions, particularly

very low water temperature, the hyporheic zone of the

Roseg River harboured a relatively diverse inverte-

brate fauna. A total of 42 taxa were identified,

although most insect larvae could not be identified

Table 3 Mean density (±SD) and frequency of occurence of taxa collected in the hyporheic zone of the main channel (n ¼ 120

replicates). Abundance is expressed as number of individuals per 10 L of pumped water. OH ¼ occasional hyporheos, PH ¼ per-

permanent hyporheos, ST ¼ stygohite. Taxa were ordinated according to their position along the upstream-downstream environ-

mental gradient (see Table 4 and Fig. 7)

Code

Ecological

category

DensityFrequency

(%)Mean Range

Creanobia alphina Dana CRAL OH <0.1 ± 0.2 0–1 2.5

Maraenobiotus insignipes insignipes (Lilljeborg) MAIN PH 1.7 ± 5.1 0–46 37.3

Bryocamptus (A.) raethicus (Schmeil) BRRA PH <0.1 ± 0.3 0–3 2.5

Niphargus cf. tatrensis Wrzesniowsky NIPH ST <0.1 ± 0.1 0–1 0.8

Rhypholophus sp. RHYP OH 0.6 ± 1.2 0–9 29.7

Bathynellidae BATH ST 1.1 ± 3.1 0–22 21.2

Diamesinae DIAM OH 4.2 ± 8.0 0–45 66.1

Heptageniidae HEPT OH 9.9 ± 16.7 0–107 86.4

Tanytarsini TANY OH 0.3 ± 1.0 0–7 11.9

Nematoda NEMA PH 58.8 ± 109.0 0–667 89.0

Bryocamptus (R.) zschokkei (Schmeil) BRSZ PH <0.1 ± 0.2 0–2 2.5

Leuctridae LEUC OH 12.9 ± 18.0 0–120 88.1

Enchytraeidae sp. HENC PH <0.1 ± 0.2 0–1 3.4

Diacyclops languidus (G.O. Sars) DILA PH 0.2 ± 0.7 0–5 6.8

Limnephilidae LIMN OH 0.1 ± 0.4 0–3 4.2

Haplotaxis gordiodies (Hartman) HAPG PH 0.1 ± 0.2 0–1 5.1

Dicranota sp. DICR OH 0.4 ± 0.9 0–5 28.0

Wiedemannia sp. WIED OH 0.3 ± 0.9 0–6 17.8

Troglochaetus cf. beranecki Delachaux TROG ST 123.5 ± 212.2 0–1163 61.9

Corynoneura spp. CORY OH 1.4 ± 4.0 0–31 33.1

Hydracarina HYDR PH 0.9 ± 1.4 0–8 43.2

Other Orthocladiinae ORTH OH 18.6 ± 40.5 0–327 80.5

Diacyclops cf. disjunctus (Thallwitz) DIDI PH 2.0 ± 10.9 0–89 7.6

Parastenocaris glacialis Noodt PAGL ST 15.3 ± 43.5 0–272 54.2

Ceratopogonidae CERA OH 0.2 ± 1.1 0–8 10.2

Dorydrilus michaelseni Piguet DOMI PH 1.1 ± 3.1 0–27 39.8

Chloroperlidae CHLO OH 0.3 ± 0.7 0–4 21.2

Nemouridae NEMO OH 1.3 ± 10.6 0–115 21.2

Acanthocyclops robustus (G.O. Sars) ACRO PH 0.4 ± 2.2 0–20 10.2

Baetidae BAET OH 1.1 ± 8.6 0–93 11.0

Propappus volki Michaelsen PPVO PH 8.6 ± 17.6 0–157 46.6

Moraria (M.) alpina Stoch MOAL PH <0.1 ± 0.2 0–2 3.4

Bryocamptus (L.) echinatus (Marazek) BREC PH 0.3 ± 1.4 0–13 8.5

Cernosvitoviella atrata (Bretscher) CEAT PH 3.6 ± 14.3 0–113 26.3

Ostracoda OSTR PH 3.3 ± 9.1 0–60 30.5

Bryocamptus (A.) cuspidatus (Schmeil) BRCU PH 0.9 ± 6.1 0–66 15.3

Rhabdomastix sp. RHAB OH 0.1 ± 0.3 0–2 5.9

Cernosvitoviella carpatic Nielsen & Christensen CEOO PH 0.4 ± 1.9 0–16 9.3

Cognettia glandulosa Michaelsen COGL PH <0.1 ± 0.1 0–1 0.8

Simuliidae SIMU OH <0.1 ± 0.2 0–1 4.2

Diacyclops bisetosu (Rehberg) DIBI PH 4.2 ± 25.8 0–260 13.6

Nais communis Piguet NACO PH 0.4 ± 2.8 0–29 3.4

Longitudinal pattern of glacial hyporheos 1717

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

to species because of their small size. The longitudinal

pattern of hyporheic assemblages followed an addi-

tive model similar to that described by Milner et al.

(2001b) for the zoobenthic community. Taxa were

successively added to the community with increasing

distance from the glacier terminus. As species present

at the most upstream sites persisted downstream,

taxonomic richness increased with increasing distance

from the glacier margin. The harpacticoid M. insign-

ipes, which was the only abundant taxon to decline

Fig. 5 Longitudinal changes in the average taxonomic richness (filled circles) and density (open squares) of invertebrate assemblages

in the hyporheic zone of the Roseg River. Bars ¼ means ± SD (n ¼ 3 replicates per site and date). Only one replicate sample was

collected at sites 6–11 in November.

Fig. 6 Relationships between cumulated

taxonomic richness of hyporheic inver-

tebrate assemblages and distance to the

glacier, hyporheic water temperature,

concentration of total inorganic carbon

(TIC), and ash-free dry mass of loosely

associated organic matter (LOM). Open

dots correspond to outliers (site 4) which

were not considered in the calculations.

1718 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

progressively with increasing distance from the gla-

cier terminus, was classified by Husmann (1975) as a

‘Nordliche Gletscherrandarten’; i.e. a species that is

strictly bound to the border zone of glaciers. All other

taxa colonising the proglacial reach had broader

habitats (high tolerance) because they occurred all

along the river corridor. In line with the results

of previous studies on benthic macroinvertebrate

zonation downstream of glacier margins (Castella

et al., 2001), water temperature, various groundwater

indicators and the amount of LOM were the most

important variables explaining variations in the com-

position of hyporheic invertebrate assemblages along

the Roseg River. Indeed, these variables accounted for

74% of the appearance of the first axis of the OMI

analysis.

Fig. 7 Outlying mean index (OMI) analysis of hyporheic invertebrate assemblages. Only the first axis, which represented 87.1% of

the explained variability, was used for graphical presentation. (a) Canonical weights of environmental variables (DO: dissolved

oxygen; Temp: temperature; Cond: specific conductance). (b) Scores of sampling sites. (c) Distribution of taxa along the environmental

gradient as a function of their weighted average position along site scores. The sizes of black circles are proportional to the total

frequencies of taxa. Vertical lines correspond to standard deviations. Asterisks indicate taxa whose distribution deviates significantly

from a uniform distribution along the environmental gradient. See Table 3 for the codes of taxa.

Longitudinal pattern of glacial hyporheos 1719

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

The composition and distribution pattern of the

hyporheos in the Roseg River departed from Milner

et al. (2001b) conceptual model for zoobenthos in

several respects. Whereas zoobenthic communities

of glacier-fed rivers are dominated by insect larvae,

hyporheic invertebrate assemblages of the Roseg

River consisted mainly of true groundwater taxa

and taxa belonging to the permanent hyporheos.

The Diamesinae, which is usually the predominant

insect taxon of the zoobenthos in glacial streams

(Lods-Crozet et al., 2001b), represented about 50% of

all insect larvae collected in the benthic layer of the

Roseg River (Burgherr & Ward, 2001), but only 8%

of all insect larvae collected in the hyporheic zone.

Copepoda and Ostracoda occurred almost exclu-

sively in the hyporheic layer of the main channel.

Only 17 individual copepods and four ostracods

were found in 102 benthic samples collected in the

main channel (Burgherr & Ward, 2001), whereas

3206 Copepoda and 387 Ostracoda were recovered

from the 120 hyporheic samples. Several species of

oligochaetes, including D. michaelseni, Propappus volki

and C. atrata, were numerous further upstream in

the hyporheic zone of the Roseg River than at the

surface (Malard et al., 2001b). Copepoda (excluding

the stygobite P. glacialis) were mainly represented

by juvenile stages, representing 72% of all individ-

uals collected in the hyporheic zone of the main

channel. The harpacticoida, P. glacialis, is a cold-

stenotherm relict widely distributed in interstitial

habitats in Northern Europe (Enckell, 1969) that has,

however, been recorded recently from the Central

Apennines, Italy (Pesce, Galassi & Cottarelli, 1995).

Enckell (1969) who found a strikingly low percen-

tage of males among populations of P. glacialis in

Northern Europe, suggested that the species might

be at least temporarily parthenogenetic. This might

also hold true in the Val Roseg as we found only

six males of P. glacialis among a total of 1871

individuals.

Table 4 Percentages of total variability (Inertia) accounted for

by the outlying mean index (OMI), the tolerance index (Tol), and

the residual tolerance index (RTol). Taxa were ordinated

according to their position along the first factorial axis of the

OMI analysis. Species showing a significant OMI (P<0.1) are

indicated with an asterisk

Taxa Inertia OMI Tol Rtol

*Cral 20.5 66.1 13.5 20.4

*Main 22.6 36.6 32.0 31.4

Brra 6.9 62.2 11.7 26.1

Niph 11.9 100.0 0.0 0.0

Rhyp 21.3 6.9 40.8 52.3

Bath 12.1 32.3 16.2 51.6

Diam 16.2 1.7 44.8 53.5

Hept 14.7 1.6 39.2 59.2

Tany 10.7 7.0 18.9 74.1

Nema 15.8 0.4 14.9 84.7

Brsz 9.9 15.0 1.1 83.9

Leuc 15.0 1.8 46.3 51.9

Henc 20.6 8.1 6.3 85.6

Dila 16.1 11.3 13.1 75.5

Limn 12.0 16.1 23.4 60.5

Hapg 8.3 31.7 4.3 64.0

Dicr 16.0 6.7 51.3 42.0

Wied 14.7 10.4 29.6 60.0

*Trog 15.3 11.2 40.6 48.0

*Cory 13.2 14.6 29.9 55.6

*Hydr 14.0 14.0 37.7 48.3

*Orth 14.2 15.2 39.5 45.4

Didi 8.0 70.5 0.1 29.5

*Pagl 13.0 21.6 26.4 52.0

Cera 9.4 51.9 2.6 45.5

*Domi 16.1 20.4 41.3 38.3

*Chlo 14.5 28.7 21.1 50.2

Nemo 18.4 27.8 48.7 23.5

*Acro 11.7 43.4 17.3 39.2

*Baet 18.3 30.8 44.0 25.2

*Ppvo 13.5 42.0 13.8 44.3

Moal 14.5 59.7 16.1 24.1

Brec 12.3 63.8 9.6 26.6

*Ceat 14.1 57.2 12.0 30.8

*Ostr 14.4 60.1 10.5 29.4

*Brcu 16.9 55.9 23.2 20.9

*Rhab 16.7 71.6 9.1 19.4

*Ceoo 19.8 67.2 18.0 14.7

Cogl 18.6 100.0 0.0 0.0

*Simu 20.1 81.4 0.9 17.7

*Dibi 22.3 87.8 6.9 5.3

*Naco 23.4 100.0 0.0 0.0

Fig. 8 Relationship between the average tolerance values of sites

and distance from the glacier terminus. The open dot corres-

ponds to an outlier (site 4) which was not considered in the

calculation.

1720 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

Fig. 9 Longitudinal distribution of taxa in the hyporheic zone of the Roseg River. Taxa were ordinated according to their marginalities

along axis 1 of the OMI analysis (see Table 4). Asterisks indicate taxa whose distribution deviates significantly from a uniform

distribution along the upstream-downstream gradient. The side of squares is proportional to the Napierian logarithm of the average

number of individuals per site (n ¼ 3 replicates per date).

Longitudinal pattern of glacial hyporheos 1721

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

Several insect taxa colonised the hyporheic zone of

the Roseg River at a water temperature below that

predicted by the model of Milner et al. (2001b).

Leuctridae, Nemouridae, and Heptageniidae occurred

in hyporheic sediments of the proglacial reach

(Tmax � 4 �C), whereas they were only found to

colonise the benthic layer of European glacier-fed

rivers when Tmax exceeded 8 �C (Milner et al., 2001b).

In general, insect taxa exhibited similar distribution

patterns in the hyporheic and benthic sediments of the

Roseg River, although Leuctridae, Nemouridae,

Heptageniidae and Orthocladiinae were found

350 m further upstream in the hyporheic zone of the

proglacial reach (site 1) than in the benthos (site 2)

(Burgherr & Ward, 2001).

Several studies have demonstrated that the density

and composition of zoobenthic communities in

glacier-fed rivers shift seasonally in response to

changes in environmental conditions (Brittain et al.,

2001; Fureder et al., 2001; Robinson et al., 2001; Schutz

et al., 2001). In contrast, the density and distribution

pattern of hyporheic invertebrate assemblages in the

Roseg River differed little over time. Whereas the

average density (±SD) of benthic invertebrate assem-

blages varied from 1472 ± 1355 invertebrates per

square metre in August to 13 398 ± 6133 invertebrates

per square metre in October (Burgherr & Ward, 2001),

that of hyporheic invertebrate assemblages varied

only from 214 ± 233 invertebrates per 10-L sample in

August to 360 ± 311 invertebrates per 10-L sample in

June. Lower seasonal changes in the density of the

hyporheos might simply reflect the predominance of

taxa that show a continuous reproduction throughout

the year (i.e. true groundwater taxa). As observed in

the surface stream (Burgherr & Ward, 2001), the

density of several insect taxa in the hyporheic zone

peaked in autumn, a period of favourable environ-

mental conditions characterised by stable stream beds,

low shear stress and turbidity, and high food

availability because of the accrual of the chrysophyte

Hydrurus foetidus Kirch.

Ecological significance of the hyporheic zone

Several permanent aquatic taxa (e.g. Nematoda,

Ostracoda, Cyclopoida, Harpacticoida and, to a lesser

extent, Oligochaeta) that were unable to tolerate the

harsh environmental conditions prevailing in the

benthic zone of the glacial Roseg River were regularly

found in the hyporheos. Indeed, the hyporheic zone

was physically more stable, offered more favourable

thermal conditions and probably exhibited less

changes in area than the glacial surface environment.

Malard et al. (1999) demonstrated that the total length

of channels carrying water in the Roseg River flood

plain varied from 5.9 km in winter to 21.4 km in

summer. Moderate changes in volume of the hypor-

heic zone, as compared with the dramatic contraction

of the surface channel network in autumn and winter,

might also be responsible for the reduced variations in

the density of hyporheic invertebrate assemblages in

the Roseg River.

The hyporheic zone could also be used as a nursery

by some insect taxa (e.g. Leuctridae, Heptageniidae),

thereby allowing the maintenance of surface popula-

tions in the proglacial reach of the Roseg River, which

was characterised by low surface water temperature

(Tmax < 4 �C) and bed stability. Indeed, early instar

larvae of insects are likely to be more sensitive to high

flow velocities and bed movements than later instars.

Hyporheic sediments may also serve as a migration

pathway enabling the permanent hyporheos to

colonise suitable benthic habitats separated by inhos-

pitable stream segments. Recolonisation of suitable

benthic habitats via the hyporheic corridor can result

in discontinuous patterns of distribution within the

catchment. The surface stream at the outlet of Roseg

Lake, characterised by relatively high temperature

and bed stability, harboured six species of Oligochaeta

(N. communis, C. atrata, C. carpatica, Henlea sp.,

Hemienchytraeus sp. and Fridericia sp.), whereas Oligo-

chaeta were almost totally absent from the benthic

layer in the most upstream reaches of the Roseg River

(sites 1–5) (Malard et al., 2001b). Although hyporheic

invertebrate communities appeared to contribute sig-

nificantly to the overall biodiversity of the Roseg

River, we expect the ecological significance of the

hyporheic zone to vary markedly among glacial rivers

depending on the development of glacio-fluvial sedi-

ments and to the extent of permafrost and clogging of

the hyporheic zone by glacial flour.

Groundwater upwelling as modifiers of hyporheic

biodiversity pattern

Lakes and tributaries are important modifiers of

zoobenthic communities patterns in glacial rivers

(Burgherr & Ward, 2000; Brittain et al., 2001; Hieber

1722 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

et al., 2002). The interruption of a glacial river course

by a lake or the confluence with a non-glacial tributary

may result in the amelioration of physical conditions

(i.e. channel stability and temperature) within the

main glacial river, thereby allowing the establishment

of taxa that would otherwise occur further down-

stream. Non-glacial tributaries also constitute a source

of species that can colonise the main glacial river at

least temporarily via either drift or oviposition by

flying adults (Brittain et al., 2001; Saltveit, Haug &

Brittain, 2001; Knispel & Castella, 2003). Upwelling of

ground water within the bed of glacial streams may

also affect the longitudinal pattern of hyporheic

invertebrate communities by increasing hyporheic

water temperature (Malard et al., 2001a) and the

concentrations of nutrient and DOC. Subsurface

fluxes of heat, nutrients and DOC may stimulate

hyporheic microbial production, thereby increasing

food resources for the interstitial fauna.

Although groundwater upwelling had no influence

on temperature and nutrient concentrations in hypor-

heic sediments of the main channel of the Roseg River,

it ensured a permanent flow of hyporheic and surface

waters in the lower part of the flood plain and the

constrained reach. Malard et al. (2003) demonstrated

that water permanence was a key factor that deter-

mined differences in the diversity and density of

permanently aquatic taxa between hyporheic sedi-

ments of the upper and lower parts of the Roseg flood

plain. Moreover, Malard et al. (2003) and Burgherr,

Ward & Robinson (2002) demonstrated that the

emergence of ground water at the land surface in

the lower part of the flood plain generated several

habitats, the hyporheic and benthic layers of which

harboured species-rich communities comprising

numerous permanent aquatic taxa. These floodplain

habitats probably constituted a major source of per-

manent aquatic species to the hyporheic sediments of

the main channel. Thus, the upwelling of ground

water in the lower part of the floodplain reach

markedly increased the diversity and density of

hyporheic invertebrate assemblages within the hypor-

heic zone of the main channel by controlling not only

the degree of flow permanence but also the source of

colonisers. The retreat of the Tschierva and Roseg

glaciers, and the resulting increase in the relative

contribution of ground water to the river flow, would

modify the distribution range of species. Taxa that are

currently restricted to the lower reaches of the Roseg

River are expected to colonise more upstream sites if

the Roseg and Tschierva glaciers continue to retreat.

Acknowledgments

This study was supported by a research grant from

the Swiss National Science Foundation (SNF grant 21-

49243.96). We thank the large number of people who

contributed to the collection of water samples and

related data for the hyporheic investigation: P. Burgh-

err, C. Robinson, K. Tockner, U. Uehlinger, and R. Zah.

We are indebted to C. Boesch for processing the

hyporheic samples, B. Lods-Crozet for the identifica-

tion of Chironomidae and to Richard Illi and Bruno

Ribi for the chemical analyses. We are grateful to

Mr. Testa and his staff at the Roseg Hotel for their

hospitality and to the towns of Pontresina and

Samedan for providing access to the sampling area.

We also thank A.G. Hildrew for his editing work and

two anonymous reviewers for their comments that

improved the final version of this paper.

References

Bou C. & Rouch R. (1967) Un nouveau champ de

recherches sur la faune aquatique souterraine. Comptes

Rendus de l’Academie des Sciences de Paris, Sciences de la

Vie, 265, 369–370.

Boulton A.J., Dole-Olivier M.-J. & Marmonier P. (2003)

Optimizing a sampling strategy for assessing hypor-

heic invertebrate biodiversity using the Bou-Rouch

method: within-site replication and sample volume.

Archiv fur Hydrobiologie, 156, 431–456.

Brittain J.E., Saltveit S.J., Castella E., Bogen J., Bonsnes

T.E., Blakar I., Bremnes T., Haug I. & Velle G. (2001)

The macroinvertebrate communities of two contrasting

Norwegian glacial rivers in relation to environmental

variables. Freshwater Biology, 46, 1723–1736.

Burgherr P. & Ward J.V. (2000) Zoobenthos of kryal and

lake outlet biotopes in a glacial flood plain. Verhan-

dlungen der Internationalen Vereinigung fur Theoretische

und Angewandte Limnologie, 27, 1587–1590.

Burgherr P. & Ward J.V. (2001) Longitudinal and

seasonal distribution patterns of the benthic fauna of

an alpine glacial stream (Val Roseg, Swiss Alps).

Freshwater Biology, 46, 1705–1721.

Burgherr P., Ward J.V. & Robinson C.T. (2002) Seasonal

variation in zoobenthos across habitat gradients in an

alpine glacial flood plain (Val Roseg, Swiss Alps).

Journal of the North American Benthological Society, 21,

561–575.

Longitudinal pattern of glacial hyporheos 1723

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

Castella E., Adalsteinsson H., Brittain J.E. et al. (2001)

Macrobenthic invertebrate richness and composition

along a latitudinal gradient of European glacier-fed

streams. Freshwater Biology, 46, 1811–1831.

Clesceri L.S., Greenberg A.E. & Eaton A.D. (1998)

Standard Methods for the Examination of Water and

Wastewater. American Public Health Association,

NW, Washington, DC, USA.

Creuze des Chatelliers M. (1991) Geomorphological

processes and discontinuities in the macrodistribution

of the interstitial fauna. A working hypothesis.

Verhandlungen der Internationalen Vereinigung fur Theo-

retische und Angewandte Limnologie, 24, 1609–1612.

Doledec S., Chessel D. & Gimaret-Carpentier C. (2000)

Niche separation in community analysis: a new

method. Ecology, 81, 2914–2927.

Downes M.T. (1978) An improved hydrazine reduction

method for the automated determination of low nitrate

levels in freshwater. Water Research, 12, 673–675.

Enckell P.H. (1969) Distribution and dispersal of Para-

stenocarididae (Copepoda) in northern Europe. Oikos,

20, 493–507.

Fraser B.G. & Williams D.D. (1997) Accuracy and

precision in sampling hyporheic fauna. Canadian

Journal of Fisheries and Aquatic Sciences, 54, 1135–

1141.

Friberg N., Milner A.M., Svendsen L.M., Lindegaard C. &

Larsen S.E. (2001) Macroinvertebrate stream commu-

nities along regional and physico-chemical gradients

in Western Greenland. Freshwater Biology, 46, 1753–

1764.

Fureder L., Schutz C., Wallinger M. & Burger R. (2001)

Physico-chemistry and aquatic insects of a glacier-fed

and a spring-fed alpine stream. Freshwater Biology, 46,

1673–1690.

Gibert J., Stanford J.A., Dole-Olivier M.-J. & Ward J.V.

(1994) Basic attributes of groundwater ecosystems and

prospects for research. In: Groundwater ecology (Eds J.

Gibert, D.L. Danielopol & J.A.Stanford), pp. 7–40.

Academic Press, San Diego, USA.

Gislason G.M., Adalsteinsson H., Hansen I., Olafsson J.S.

& Svavarsdottir K. (2001) Longitudinal changes in

macroinvertebrate assemblages along a glacial river

system in central Iceland. Freshwater Biology, 46, 1737–

1751.

Hieber M., Robinson C.T., Uehlinger U. & Ward J.V.

(2002) Are alpine lakes outlets less harsh than

other alpine streams? Archiv fur Hydrobiologie, 154,

199–223.

Hunt G.W. & Stanley E.H. (2000) An evaluation of

alternative procedures using the Bou-Rouch method

for sampling hyporheic invertebrates. Canadian Journal

of Fisheries and Aquatic Sciences, 57, 1545–1550.

Husmann S. (1975) The boreoalpine distribution of

groundwater organisms in Europe. Verhandlungen der

Internationalen Vereinigung fur Theoretische und Ange-

wandte Limnologie, 19, 2983–2988.

Knispel S. & Castella E. (2003) Disruption of a long-

itudinal pattern in environmental factors and benthic

fauna by a glacial tributary. Freshwater Biology, 48, 604–

618.

Lafont M. & Malard F. (2001) Oligochaete communities

in the hyporheic zone of a glacial river, the Roseg

River, Switzerland. Hydrobiologia, 463, 75–81.

Lods-Crozet B., Castella E., Cambin D., Ilg C., Knispel S.

& Mayor-Simeant H. (2001a) Macroinvertebrate com-

munity structure in relation to environmental variables

in a Swiss glacial stream. Freshwater Biology, 46, 1641–

1661.

Lods-Crozet B., Lencioni V., Olafsson J.S., Snook D.L.,

Velle G., Brittain J.E., Castella E. & Rossaro B. (2001b)

Chironomid (Diptera: Chironomidae) communities in

six European glacier-fed streams. Freshwater Biology,

46, 1791–1809.

Maiolini B. & Lencioni V. (2001) Longitudinal distribu-

tion of macroinvertebrate assemblages in a glacially

influenced stream system in the Italian Alps. Fresh-

water Biology, 46, 1625–1639.

Malard F., Tockner K. & Ward J.V. (1999) Shifting

dominance of subcatchment water sources and flow

paths in a glacial floodplain, Val Roseg, Switzerland.

Arctic, Antarctic and Alpine Research, 31, 135–150.

Malard F., Mangin A., Uehlinger U. & Ward J.V. (2001a)

Thermal heterogeneity in the hyporheic zone of a

glacial floodplain. Canadian Journal of Fisheries and

Aquatic Sciences, 58, 1319–1335.

Malard F., Lafont M., Burgherr P. & Ward, J.V. (2001b) A

comparison of longitudinal patterns in hyporheic and

benthic oligochaete assemblages in a glacial river.

Arctic, Antarctic and Alpine Research, 33, 457–466.

Malard F., Ferreira D., Doledec S. & Ward J.V. (2003)

Influence of groundwater upwelling on the distribu-

tion of the hyporheos in a headwater river flood plain.

Archiv fur Hydrobiologie, 157, 89–116.

Milner A.M. & Petts G.E. (1994) Glacial rivers: physical

habitat and ecology. Freshwater Biology, 32, 295–307.

Milner A.M., Taylor R.C. & Winterbourn M.J. (2001a)

Longitudinal distribution of macroinvertebrates in two

glacier-fed New Zealand rivers. Freshwater Biology, 46,

1765–1775.

Milner A.M., Brittain J.E., Castella E. & Petts G.E. (2001b)

Trends of macroinvertebrate community structure in

glacier-fed rivers in relation to environmental condi-

tions: a synthesis. Freshwater Biology, 46, 1833–1847.

Pesce G.L., Galassi D.P. & Cottarelli V. (1995) Parasteno-

caris lorenzae n.sp., and first record of Parastenocaris

1724 F. Malard et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725

glacialis Noodt (Copepoda, Harpacticoida) from Italy.

Hydrobiologia, 302, 97–101.

Pusch M. & Schwoerbel J. (1994) Community respiration

in hyporheic sediments of a mountain stream (Steina,

Black Forest). Archiv fur Hydrobiologie, 130, 35–52.

Robinson C.T., Uehlinger U. & Hieber M. (2001) Spatio-

temporal variation in macroinvertebrate assemblages

of glacial streams in the Swiss Alps. Freshwater Biology,

46, 1663–1672.

Saltveit S.J., Haug I. & Brittain J.E. (2001) Invertebrate

drift in a glacial river and its non-glacial tributary.

Freshwater Biology, 46, 1777–1789.

Scarsbrook M.R. & Halliday J. (2002) Detecting patterns

in hyporheic community structure: does sampling

method alter the story? New Zealand Journal of Marine

and Freshwater Research, 36, 443–453.

Schutz C., Wallinger M., Burger R. & Fureder L. (2001)

Effects of snow cover on the benthic fauna in a glacier-

fed stream. Freshwater Biology, 46, 1691–1704.

Snook D. & Milner A.M. (2001) The influence of glacial

runoff on stream macroinvertebrate communities in

the Taillon catchment, French Pyrenees. Freshwater

Biology, 46, 1609–1623.

Stanford J.A. & Ward J.V. (1993) An ecosystem perspec-

tive of alluvial rivers: connectivity and the hyporheic

corridor. Journal of the North American Benthological

Society, 12, 48–60.

STATSOFT (1994) Statistica for the Macintosh. StatSoft,

Tulsa, Oklahoma.

Thioulouse J., Chessel D., Doledec S. & Olivier J.-M.

(1997) ADE-4: a multivariate analysis and graphical

display software. Statistics and Computing 7, 75–83.

Tilzer M. (1968) Zur Okologie und Besiedlung des

hochalpinen hyporheischen Interstitials im Arlbergge-

biet (Osterreich). Archiv fur Hydrobiologie, 65, 253–308.

Tockner K., Malard F., Burgherr P., Robinson C.T.,

Uehlinger U., Zah R. & Ward J.V. (1997) Physico-

chemical characterization of channel types in a glacial

floodplain ecosystem (Val Roseg, Switzerland). Archiv

fur Hydrobiologie, 140, 433–463.

Tockner K, Malard F., Uehlinger U. & Ward J.V. (2002)

Nutrients and organic matter in a glacial river flood-

plain system (Val Roseg, Switzerland). Limnology and

Oceanography, 47, 266–277.

Uehlinger U, Malard F. & Ward J.V. (2003) Thermal

patterns in the surface waters of a glacial river corridor

(Val Roseg, Switzerland). Freshwater Biology, 48, 284–

300.

Vogler P. (1965) Beitrage zur Phosphoranalytik in der

Limnologie. Fortschr. Wasserchemie und Grenzgebiete, 2,

109–119.

Ward J.V., Malard F., Tockner K. & Uehlinger U. (1999)

Influence of ground water on surface water conditions

in a glacial flood plain of the Swiss Alps. Hydrological

Processes, 13, 277–293.

Zah R. (2001) Patterns, pathways, and trophic transfer of

organic matter in a glacial stream ecosystem in the

Alps. PhD Thesis NRr. 13998. Swiss Federal Institute of

Technology, Zurich, Switzerland.

Zah R. & Uehlinger U. (2001) Particulate organic matter

inputs to a glacial stream ecosystem in the Swiss Alps.

Freshwater Biology, 46, 1597–1608.

Zah R., Niederost M. & Uehlinger U. (2000) Application

of photogrammetry in freshwater ecology: analysing

the morphology of a high Alpine floodplain. Interna-

tional Archive of Photogrammetry and Remote Sensing, 33,

1739–1746.

(Manuscript accepted 24 June 2003)

Longitudinal pattern of glacial hyporheos 1725

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 1709–1725