Research Collection
Doctoral Thesis
Spatio-temporal community patterns of lotic zoobenthos across habitat gradients in an alpine glacial stream ecosystem
Author(s): Burgherr, Peter
Publication Date: 2000
Permanent Link: https://doi.org/10.3929/ethz-a-004105273
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.
ETH Library Diss. ETH No. 13829
Spatio-temporal community patterns of lotie zoobenthos across habitat gradients in an alpine glacial stream ecosystem
A dissertation submitted to the
SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH
for the degree of
Doctor of Natural Sciences (Dr. sc. nat.)
presented by
Peter Burgherr
Dipl. Natw. ETH (ETH Zurich)
born August 19,1969
citizen of Triengen (LU)
accepted on the recommendation of
Prof. Dr. J.V. Ward, examiner
Prof. Dr. J.E. Brittain, co-examiner
Dr. C.T. Robinson, co-examiner
Zürich, 2000 Chapters 2, 3 and 4 are in press
Burgherr, P. &c Ward, J.V. (in press) Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie 27.
Burgherr, P. &C Ward, J.V. (in press) Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps). Fresh¬ water Biology.
Burgherr, P., Ward, J.V. & Glatthaar, R. (in press) Diversity, distribution and seasonality of the Simuliidae fauna in a glacial stream system in the Swiss Alps. Archiv für Hydrobiologie.
Chapters 5 and 6 are submitted
Burgherr, P., Ward, J.V. & Robinson, C.T. Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps). Sub¬ mitted to Journal of the North American Benthological Society.
Burgherr, P., Robinson, C.T. & Zah, R. Genetic structure and feeding plasticity of Acrophylax zerberus (Trichoptera: Limnephilidae) in a Swiss alpine flood plain. Submitted to International Review of Hydrobiology. Contents
Summary 1
Zusammenfassung 5
Chapter 1 Introduction 9
Chapter 2 Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain 17
Chapter 3 Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps) 25
Chapter 4 Diversity, distribution and seasonality of the Simuliidae fauna in a glacial stream system in the Swiss Alps 53
Chapter 5 Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps) 71
Chapter 6 Genetic structure and feeding plasticity of Acrophylax zerberus (Trichoptera: Limnephilidae) in a Swiss alpine flood plain 97
Chapter 7 Synopsis 121
Verdankung 131
Curriculum Vitae 135 I ff** I 1 1^ 1 Summary
Summary
The research conducted for this thesis investigated the zoobenthos in a stream of this work was dedicated ecosystem of an alpine glacial flood plain. The major part to the examination of spatio-temporal patterns in macroinvertebrate distributions, and to the identification of relationships between faunal and environmental gradi¬ from these de¬ ents. The final chapter aimed to test specific hypotheses generated scriptive gradient analyses. The first study compared the benthic macroinvertebrate assemblages between Non-metric multidimensional a kryal stream and an adjacent proglacial lake outlet. scaling in combination with analysis of similarity and a similarity percentages proce¬ total dure was used to examine differences between sites. Community structure and densities were less variable in the lake outlet than in the kryal channel. Furthermore, > 95% the kryal stream zoobenthos was clearly dominated by Chironomidae (i.e., Diamesa spp.), whereas Ephemeroptera, Plecoptera and Trichoptera, as well as Chironomidae, contributed substantially to the zoobenthic community in the lake outlet. In contrast to the predominance of filter-feeding organisms reported for low- lake out¬ elevation lake outlets, a scraper predominance was found in the proglacial let. Temperature, substrate stability, turbidity, and flow regime constituted major environmental factors responsible for the observed differences in zoobenthic assem¬ blage structure between the kryal and proglacial lake outlet stream. The second study addressed the temporal variability in longitudinal patterns of macroinvertebrate distributions in relation to environmental gradients along the 11.3 km long Roseg River. Co-inertia analysis clearly revealed a strong relationship between environmental conditions and benthic community structure. In contrast to Milner & Petts' conceptual model from 1994, macroinvertebrate communities fol¬ lowed predicted zonation patterns for alpine glacial streams only in summer. In ad¬ dition to this temporal variability in longitudinal response patterns, a similarity in found. temporal patterns among individual sites along the longitudinal gradient was In summary, the results suggest that longitudinal zoobenthic gradients were not solely related to temperature and channel stability. Seasonal shifts in sources and pathways of water (i.e., extent of glacial influence), and periods of favourable environmental conditions (in spring and late autumn/early winter) also had a strong influence on zoobenthic distributions.
Glacier-fed streams are thought to provide a poor habitat for Simuliidae, yet these animals are often the first non-chironomid taxa occurring in glacier-fed streams. These 2 contradictory statements were the motivation for the simuliid study pre¬ sented in chapter 4. In a first step, an assessment of the species present in the main glacial channel was conducted. Samples were collected along the longitudinal gradi¬ ent from sites near the glacier terminus down to the river mouth. Different channel types in the glacial flood plain also were surveyed for additional species. In a second 2 Summary
step, longitudinal patterns in simuliid associations during different seasons were in¬ vestigated. Last, relationships between simuliid distribution patterns and environ¬ mental factors were examined using partial least squares regression. About 20% of all Simuliidae species currently known in Switzerland were recorded in the Val Roseg flood plain, indicating glacial streams provide more adequate habitats for these in¬ sects than generally perceived. Temperature and channel stability constrained Simuliidae colonization close to the glacier. Although suspended particular organic matter represent the most important food source for Simuliidae, results suggest that feeding plasticity may be a key factor for persistence in this harsh environment. In contrast to temperate and tropical river-flood plain systems, glacial flood plains and their ecological dynamics have received little attention, although they comprise predominant landscape features in recent and former proglacial areas. Therefore, the primary goal of the fourth study was to gain insight into how habitat heterogeneity in the glacial flood plain of Val Roseg affects the structure and compo¬ sition of benthic macroinvertebrate communities. For this purpose a study design combining the longitudinal perspective with a lateral habitat gradient covering 3 distinct habitat types was applied. These investigations demonstrated that connectiv¬ ity and spatio-temporal heterogeneity were linked to variations in ecosystem size (i.e., seasonal changes in the length of floodplain channels). Connectivity and spatio- temporal heterogeneity appear to be the primary structuring agents that set the habi¬ tat templet for the community structure in each habitat (channel) type in the glacial flood plain of Val Roseg. The fifth study was initiated to explore how the complex mosaic of different stream types constrains dispersal and influences other life history traits of an aquatic macroinvertebrate species commonly found in alpine glacial flood plains. The ge¬ netic structure, energy flow and food preferences of a limnephilid caddisfly (Acrophylax zerberus Brauer) in 3 spatially distinct stream types was analyzed. Little migration between habitats was found within a single generation. Furthermore, re¬ sults of allozyme electrophoresis suggested that dispersal most likely occurs through flying adults in this glacial flood plain. Carbon and nitrogen stable isotope analyses and food preference experiments were indicative of a generalist feeding strategy by A. zerberus, suggesting feeding plasticity may be a key factor for the successful dis¬ persal and persistence of this patchily distributed species. In conclusion, alpine glacial flood plains are highly dynamic systems with respect to both their environmental conditions and zoobenthic communities. Sea¬ sonal changes in the balance between glacial meltwater and groundwater contribu¬ tions result in shifting patterns of spatial heterogeneity. Spatial heterogeneity re¬ duces the negative effects of summer high flows (e.g., high turbidity, unstable substrates) by providing numerous réfugia for the biota, thus enhancing overall eco¬ system stability and biodiversity. From an organismic view, plasticity in life history traits (e.g., feeding behavior) may play an important role in sustaining aquatic insect populations in alpine glacial streams.
£kt sink Isaf Zusammenfassung 5
Zusammenfassung
In der vorliegenden Dissertation wurde das Zoobenthos verschiedener Fliessgewässertypen einer alpinen glazialen Schwemmebene untersucht. Der Haupt¬ teil der Arbeit befasst sich mit der Analyse räumlich-zeitlicher Muster in der Verbrei¬ tung der Makroinvertebraten sowie den Wechselwirkungen zwischen den faunisti- schen Gradienten und den verschiedenen Umweltfaktoren. Schliesslich wurden die
aus den deskriptiven Gradientenanalysen hervorgegangenen Hypothesen an einem spezifischen Beispiel getestet. Die erste Studie vergleicht die benthischen Makroinvertebraten-Gemeinschaf- ten eines Gletscherbachs mit denen eines angrenzenden proglazialen Seeausflusses. Mit Hilfe einer nicht-metrischen, multidimensionalen Skalierung in Kombination mit einer Ähnlichkeitsanalyse und dem sogenannten „similarity percentages"-Ver- fahren wurden die Unterschiede zwischen den beiden Stellen analysiert. Die Struktur der Makroinvertebratengemeinschaft und die Gesamtabundanzen waren im Seeaus- fluss weniger variabel als im Gletscherbach. Des weiteren dominierten die Chironomidae das Zoobenthos im Gletscherbach (d.h. > 95% Diamesa spp.), wo¬ hingegen Ephemeroptera, Plecoptera und Trichoptera, zusammen mit den
Chironomidae beträchtlich zur Diversität der benthischen Gemeinschaft des Seeaus¬ flusses beitrugen. Während in Seeausflüssen tieferer Lagen filtrierende Makro¬ invertebraten dominieren, wurde im proglazialen Seeausfluss ein Vorherrschen von Weidegängern festgestellt. Temperatur, Substratstabilität, Turbidität und Abfluss¬ regime waren die wesentlichen Umweltfaktoren, die die unterschiedlichen Struktu¬ ren des Zoobenthos im Gletscherbach und im proglazialen Seeausfluss bewirkten. Die zweite Studie befasst sich mit der zeitlichen Variabilität der longitudina¬ len Verteilungsmuster der Makroinvertebraten in Beziehung zu den Gradienten ver¬ schiedener Umweltfaktoren entlang des 11.3 km langen Rosegbaches. Eine Co-inertia Analyse zeigte einen signifikanten Zusammenhang zwischen den Umweltbedingungen und der Struktur der benthischen Gemeinschaften. Die longitudinale Zonierung der Makroinvertebraten-Gemeinschaften stimmte mit dem konzeptionellen Modell von Milner &c Petts von 1994 nur im Sommer überein, da das Modell die saisonalen Änderungen in der Gemeinschaftsstruktur nicht vollständig berücksichtigt. Zusätz¬ lich zu dieser saisonalen Variabilität in den longitudinalen Verteilungsmustern, wie¬ sen die einzelnen Probestellen übereinstimmende saisonale Veränderungen entlang des longitudinalen Gradienten auf. Zusammenfassend deuten die Resultate darauf hin, dass die longitudinalen Gradienten des Zoobenthos nicht nur durch die Wasser¬ temperatur und die Gerinnestabilität bestimmt werden, sondern dass Zeiträume vor¬ teilhafter Umweltbedingungen (im Frühling und Spätherbst/frühen Winter), die durch saisonale Verschiebungen im Ursprung und den Fliesswegen des Wassers (d.h. des Anteils Gletscherschmelzwasser und Grundwasser) beeinflusst werden, ebenfalls ei¬ nen starken Einfluss auf die zoobenthischen Verteilungsmuster haben. 6 Zusammenfassung
Gletschergespeiste Bäche werden als ungeeignete Habitate für Simuliidae er¬ achtet, obwohl diese Insektenlarven nach den Chironomidae oft zu den ersten Besiedlern von Gletscherbächen gehören. Diese sich widersprechenden Aussagen waren die Motiviation für die Simuliidae-Studie (Kapitel 4). Für eine Bestandsauf¬ nahme der im Hauptgerinne des Gletscherbaches vorkommenden Arten wurden Pro¬ ben entlang des longitudinalen Gradienten von Stellen nahe der Gletscherzunge bis zur Mündung genommen. In der glazialen Schwemmebene wurden verschiedene Gerinnetypen bezüglich zusätzlicher Simuliidae-Arten untersucht. Zusätzlich wur¬ den die longitudinalen Verteilungsmuster der Simuliidae-Gemeinschaften in verschie¬ denen Jahreszeiten untersucht. Mit Hilfe der „partial least squares" Regression wur¬ den die Wechselwirkungen zwischen den Verteilungsmustern der Simuliidae und ausgewählten Umweltfaktoren analysiert. Ungefähr 20% aller zur Zeit in der Schweiz nachgewiesenen Simuliidae-Arten wurden in der glazialen Schwemmebene des Val Rosegs gefunden, was darauf hinweist, dass Gletscherbäche ein geeigneteres Habitat für diese Insektenlarven darstellen als generell angenommen. Nahe dem Gletscher wurde die Besiedlung durch Simuliidae durch Temperatur und Gerinnestabilität beeinträchtigt. Obwohl partikuläre organische Schwebstoffe die wichtigste Nahrungs¬ quelle für die Simuliidae in diesem Ökosystem darstellten, deuten die Resultate dar¬ auf hin, dass eine flexible Nahrungswahl ein Schlüsselfaktor für das Überleben bei diesen harschen Umweltbedingungen sein kann. Im Gegensatz zu Schwemmebenen der gemässigten und tropischen Zone, wurden glaziale Schwemmebenen und ihre ökologische Dynamik bislang kaum er¬ forscht, obwohl sie vorherrschende Landschaftselemente in proglazialen Gebieten darstellen. Deshalb war das Hauptziel der vierten Studie, den Einfluss der Habitats- Heterogenität in der glazialen Schwemmebene des Val Roseg auf die Struktur und Zusammensetzung der benthischen Makroinvertebraten-Gemeinschaften zu unter¬ suchen. Die gewählte Versuchsanordnung kombinierte eine longitudinale Perspekti¬ ve mit einem lateralen Habitatsgradienten, so dass 3 verschiedene Habitatstypen erfasst wurden. Diese Untersuchungen zeigten, dass saisonale Änderungen in der Gesamtgerinnelänge in der Schwemmebene den Vernetzungsgrad der Gerinne und dadurch die räumlich-zeitliche Habitatsheterogenität und -Stabilität beeinflussten. Diese Faktoren definieren im wesentlichen die Habitatsbedingungen für das Zoo¬ benthos eines jeden Gerinnetyps in der glazialen Schwemmebene des Val Roseg. In der fünften Studie wurde am konkreten Beispiel einer in alpinen glazialen Schwemmebenen häufig gefundenen aquatischen Makroinvertebraten-Art der Ein¬ fluss des komplexen Mosaiks verschiedener Fliessgewässertypen auf die Verbreitung und das Ernährungsverhalten untersucht. Die genetische Struktur, der Energiefluss, sowie die Ernährungspräferenzen einer Köcherfliege der Familie Limnephilidae (Acrophylax zerberus Brauer) wurde in 3 räumlich deutlich verschiedenen Fliessgewässertypen untersucht. Innerhalb einer Generation wurde nur ein geringer Austausch zwischen den Populationen in den 3 verschiedenen Habitaten gefunden. Darüberhinaus deuten die Resultate der Allozym-Elektrophorese darauf hin, dass die Ausbreitung in dieser glazialen Schwemmebene sehr wahrscheinlich durch geflü- Zusammenfassung 7
gelte Adulttiere erfolgt. Stabile Isotopen-Analysen von Kohlenstoff und Stickstoff, sowie Experimente bezüglich der Ernährungspräferenzen ergaben eine opportunisti¬ sche Ernährungsstrategie vonA. zerberus. Dieses Ergebnis deutet daraufhin, dass ein flexibles Ernährungsverhalten einen Schlüsselfaktor für eine erfolgreiche Ausbrei¬ tung und ein dauerhaftes Vorkommen dieser lückenhaft verteilten Art darstellt. Zusammenfassend lässt sich festhalten, dass glaziale Schwemmebenen hoch¬ dynamische Ökosysteme darstellen, sowohl bezüglich ihrer Umweltbedingungen, als auch hinsichtlich ihrer zoobenthischen Lebensgemeinschaften. Saisonale Änderun¬ gen des Anteils von Gletscherschmelzwasser und Grundwasser resultieren in sich verändernden Mustern räumlicher Heterogenität und bestimmen die Zeiträume vor¬ teilhafter Umweltbedingungen. Die räumliche Heterogenität reduziert die negativen Effekte des hohen Abflusses im Sommer (zum Beispiel Turbidität, Geschiebetrieb) durch die Bereitstellung zahlreicher Refugialräume für die Biota, wodurch die Stabi¬ lität und Biodiversität des gesamten Ökosystems erhöht wird. Aus einer organismi¬ schen Sichtweise scheinen flexible Ausprägungen verschiedener Merkmale im Lebens¬ zyklus (zum Beispiel im Ernährungsverhalten) für eine erfolgreiche Besiedlung alpi¬ ner Gletscherbäche eine wichtige Rolle zu spielen.
Introduction 9
Chapter 1
Introduction
Give me space and motion, and I will give you a world. Descartes (1596 -1650)
1.1 The alpine environment
Areas covered by alpine and arctic vegetation, ranging from 80° N to 67° S and reaching elevations of more than 6000 m a.s.l. in the subtropics, amount to roughly 11 million km2 or 8% (5% arctic, 3% alpine) of the terrestrial surface of the earth (Chapin III & Körner, 1995). Another 15.9 million km2 worldwide are cov¬ ered by ice; 2909 km2 of which form the 4500 glaciers in the Alps (Maisch, 1992). Streams and rivers are integral features of alpine landscapes (Füreder, 1999). For example, 31% of the discharge of the Rhine River originates from the Alps, and make up 11% of its catchment. Alpine stream ecosystems have recently received increased attention, as it has become widely recognized that these high elevation headwater streams play an important role in the dynamics of major river systems throughout the world (Füreder, 1999). Although high mountain regions have long been developed and cultivated by man (Franz, 1979), today's human activities are influencing and altering the func¬ tioning of ecosystems at a rate and magnitude much faster than „normal evolution" (Chapin III & Körner, 1994). Alpine stream ecosystems have been proposed to be particularly sensitive to climate change and human impacts (i.e., water resource de¬ velopments, tourism) (McGregor et al., 1995), yet relatively little is known of their ecology (Ward, 1994). Whether global warming would lead to the recession of gla¬ ciers through increased ablation or their expansion through increased precipitation is still under debate. Irrespective of which scenario takes place, changes in hydrol¬ ogy, temperature, and channel stability should cause predictable changes in the struc¬ ture of zoobenthic communities in alpine stream systems (Milner &C Petts, 1994; McGregor et al, 1995). However, such predictions require a much more compre¬ hensive understanding of the interplay between abiotic factors and biotic communi¬ ties. 10 Introduction
1.2 Alpine stream types
Based on their local water sources and temperature regimes, Ward (1994) distinguished 3 major types of alpine streams: (1) kryal stream segments fed by gla¬ cial meltwater, (2) krenal segments fed by ground water, and (3) rhithral segments dominated by snowmelt/rainfall. Harsh environmental conditions (i.e., cold tem¬ peratures and low food resources) are a general feature of alpine streams. Alpine stream networks often form a complex mosaic of kryal, krenal and rhithral segments because of differences in climate, geomorphology and hydrology of alpine environ¬ ments (Ward, 1994; McGregor et al, 1995). Kryal stream segments are characterized by (1) year-round low temperatures, (2) large diel flow fluctuations in summer with peaks in late afternoon from glacial melting, (3) high loads of suspended sediments („glacial flour") from glacial scour¬ ing, and (4) a gradient of channel stability with sites close to the glacier being highly unstable and stability increasing downstream (Milner &C Petts, 1994; Ward, 1994). The kryal can be further subdivided into eukryal, the temporary aquatic habitat on or within the glacier, metakryal adjacent to the glacier (Tmax < 2°C), and hypokryal where diel and seasonal variability in temperature is greater, but maximum tempera¬ ture does not exceed 4°C (Ward, 1992; Füreder, 1999). Downstream of the kryal segment, a glacio-rhithral zone fed by a mixture of water sources can extend for a considerable distance, with temporal changes in discharge and temperature reflect¬ ing the relative proportion of glacial influence (Füreder, 1999). In contrast, groundwater-fed streams or krenal biotopes typically exhibit (1) constant flow re¬ gimes, (2) low temperature amplitudes close to the source that can increase consider¬ ably downstream, (3) physico-chemical constancy, and (4) high water clarity (Ward, 1994). Temperatures of non-thermal springs are close to the mean annual air tem¬ perature at their sources, and exhibit a summer-cool/winter-warm pattern (Ward, 1985).
1.3 Overview on research in alpine streams
Investigations on the fauna of glacier-fed streams started over 100 years ago (Nordenskjöld, 1870 in Franz, 1979). At the beginning of the 20th century the works of Steinmann Thienemann Steinbock Dorier and (1907), (1912), (1934) , (1937) Léger (1937) established the ecological study of high mountain streams in Europe. Later studies of alpine streams have dealt mostly with hydrology, glaciology, geomorphology and physico-chemical attributes (Gurnell & Fenn, 1985; Röthlisberger & Lang, 1987; Prowse, 1994; Gurnell et al., 1999). In contrast, comprehensive studies examining longitudinal distribution patterns of alpine benthic macroinvertebrate communities (e.g., Elgmork & Saether, 1970; Kownacka & Kownacki, 1972; Kownacki & Kownacka, 1973; Allan, 1975; Kownacki, 1991) or the autecology of individual species or higher taxonomic groups (e.g., Saether, 1968; Introduction 11
Steffan, 1971; Irons III et al, 1993) are rather scarce, and often of limited duration and intensity. Comparative studies of biota in different alpine stream types have received little attention (e.g., Kownacki, 1991; Füreder et al, 1998), and virtually nothing is known regarding the aquatic biota of glacial flood plains.
1.4 Thesis goals
The present thesis aimed to enhance our limited knowledge of alpine glacial streams in general and to contribute to the understanding of the ecology of glacial floodplain systems in particular; systems that constitute predominant landscape ele¬ ments in recent and former proglacial areas. The overall objective was to determine how spatio-temporal habitat heterogeneity contributes to distribution patterns of lotie zoobenthos. For this purpose, investigations were conducted in the glacial floodplain stream ecosystem of Val Roseg in the Swiss Alps. In a synthesis of Alaskan and European literature reporting on the ecology of glacial-fed streams, Milner & Petts (1994) proposed a qualitative model relating gradients of zoobenthic communities to temperature and channel stability; both be¬ ing a function of distance from the glacier terminus and time since déglaciation. They suggest that close to the glacier, where summer temperatures typically do not exceed 2°C, chironomids of the genus Diamesa are generally the dominant, if not the only, animals. Where maximum temperatures are between 2 and 4°C, other Diamesinae, Orthocladiinae and Simuliidae may be present. Further downstream where water temperatures exceed 4°C, Baetidae, Nemouridae and Chloroperlidae typically colonize glacial streams. Additional families of Ephemeroptera, Plecoptera and Diptera, as well as Trichoptera, appear with increasing distance downstream. The same authors also suggested that these distinctive downstream patterns can be disrupted by tributaries, proglacial lakes or changes in valley confinement. In chapter 2 a kryal channel and an outlet stream of a proglacial lake were compared. In contrast to kryal habitats, higher water temperatures, and relatively stable, single-thread outlet channels are found below proglacial lakes (Milner & Petts, 1994). The primary objective of this chapter was to assess how different environ¬ mental conditions in these 2 habitat types affected benthic macroinvertebrate com¬ munity structure. Our present knowledge about longitudinal changes in community structure of alpine benthic macroinvertebrates, as conceptualized in the model of Milner & Petts (1994), is largely based on studies that did not consider seasonal dynamics. As a consequence, year-round studies, such as conducted by Lavandier & Décamps (1984) in the Estaragne River in the French Pyrenees, are sorely needed. The main objective of the study in chapter 3 was to examine longitudinal and seasonal patterns in macroinvertebrate distributions and environmental conditions in a glacier-fed stream. Specifically, 3 questions were addressed: (1) Do environmental conditions exhibit distinct patterns with increasing distance from the glacier terminus and across sea- 12 Introduction
sons? (2) Are there predictable spatio-temporal patterns reflected in the benthic com¬ munity structure? (3) Is there concordance among environmental and biotic gradi¬ ents?
Simuliidae are often the first non-chironomid taxon to appear where water temperatures in a glacial stream exceed 2°C (Milner & Petts, 1994; Ward, 1994). However, most studies of glacial stream fauna generally focused on the Chironomidae, Ephemeroptera and Plecoptera, whereas Simuliidae have received little detailed study. Therefore, the goal of chapter 4 was to investigate diversity, distribution patterns and seasonal changes of the Simuliidae fauna in a glacial stream system with empha¬ sis on their longitudinal distribution. As mentioned above, the ecology and dynamics of glacial flood plains have received little attention, although river-floodplain systems in temperate and tropical biomes have been shown to be a key factor determining ecosystem processes and the community structure of biota (Junk et al, 1989). Therefore, the objective of chapter 5 was to determine how spatio-temporal habitat heterogeneity and connectivity gra¬ dients influence distribution patterns of lotie zoobenthos among various habitat (chan¬ nel) types in a glacial flood plain. Many aquatic macroinvertebrates have not adapted to the extreme environ¬ mental conditions in alpine streams (Füreder, 1999). Therefore, the occurrence of a particular species is a direct expression of its ability to tolerate or adapt to the ther¬ mal regime and other population constraints (Ward, 1992; Füreder, 1999). In con¬ trast to the preceding chapters, the focus in chapter 6 was not at the community level but rather at the level of the individual species by addressing how a particular species can successfully disperse and persist in a fragmented and harsh environment. For this purpose, the genetic structure and feeding behavior of an aquatic insect that inhabits specific stream types in an alpine glacial flood plain were examined.
1.5 References
Allan, J.D. (1975) Faunal replacement and longitudinal zonation in an alpine stream. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 19, 1646-1652. Chapin III, F.S. & Körner, C. (1994) Arctic and alpine biodiversity: patterns, causes, and ecosystem consequences. Trends in Ecology and Evolution, 9, 45-47. Chapin III, F.S. & Körner, C. (1995) Patterns, causes, changes and consequences of biodiversity in arctic and alpine ecosystems. Arctic and alpine biodiversity: patterns, causes and ecosystem consequences (eds. F.S. Chapin III & C. Körner), pp. 313-320, Springer-Verlag, Berlin. Dorier, A. (1937) La faune des eaux courantes alpines. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 8, 33-41. Introduction 13
Elgmork, K. & Saether, O.R. (1970) Distribution of invertebrates in a high moun¬ tain brook in the Colorado Rocky Mountains. University of Colorado Stud¬ ies, Series in Biology, 31, 1-55. Franz, H. (1979) Ökologie der Hochgebirge. Ulmer, Stuttgart. Füreder, L. (1999) High alpine streams: cold habitats for insect larvae. Cold adapted organisms. Ecology, physiology, enzymology and molecular biology (eds. R. Berlin. Margesin & F. Schinner), pp. 181-196, Springer-Verlag, Füreder, L., Schütz, C, Burger, R. & Wallinger, M. (1998) High alpine streams as models for ecological gradients. Hydrology, water resources and ecology in headwaters (eds. K. Kovar, U. Tappeiner, N.E. Peters &c R.G. Craig), pp. 387-394, IAHS Press, Wallingford, UK. Gurnell, A.M., Edwards, P.J., Petts, G.E. & Ward, J.V. (1999) A conceptual model for alpine proglacial river channel evolution under changing climatic condi¬ tions. Catena, 38, 223-242. Gurnell, A.M. & Fenn, CR. (1985) Spatial and temporal variation in electrical con¬ ductivity in a pro-glacial stream system. Journal of Glaciology, 31, 108-114. Irons III, J.G., Miller, K. & Oswood, M.W. (1993) Ecological adaptations of aquatic macroinvertebrates to overwintering in interior Alaska (U.S.A) subarctic streams. Canadian Journal of Zoology, 71, 98-108. Junk, W.J., Bayley, P.B. & Sparks, R.E. (1989) The flood-pulse concept in river floodplain systems. Canadian Special Publication of Fisheries and Aquatic Sciences, 106, 110-127. Kownacka, M. & Kownacki, A. (1972) Vertical distribution of zoocenoses in the streams of the Tatra, Caucasus and Balkan Mts. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 18, 742-750. Kownacki, A. (1991) Zonal distribution and classification of the invertebrate com¬ munities in high mountain streams in South Tirol (Italy). Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 24, 2010-2014. Kownacki, A. & Kownacka, M. (1973) The distribution of the bottom fauna in several streams in the Middle Balkan in the summer period. Acta Hydrobiologia, 15, 295-310. Lavandier, P. & Décamps, H. (1984) Estaragne. Ecology of European Rivers (ed. B.A. Whitton), pp. 237-264, Blackwell Scientific Publications, Oxford, UK. Léger, M.L. (1937) Economie biologique générale des cours d'eau alpines. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 8, 5-14. Maisch, M. (1992) Die Gletscher Graubündens. Rekonstruktion und Auswertung der Gletscher und deren Veränderungen seit dem Hochstand von 1850 im Gebiet der östlichen Schweizer Alpen (Bündnerland und angrenzende Regionen). Teil
A: Grundlagen - Analysen - Ergebnisse, Teil B: Verzeichnisse - Datenkataloge
- Gletscherkarten. Geographisches Institut der Universität Zürich, Zürich. 14 Introduction
McGregor, G., Petts, G.E., Gurnell, A.M. & Milner, A.M. (1995) Sensitivity of al¬ pine stream ecosystems to climate change and human impacts. Aquatic Con¬ servation: Marine and Freshwater Ecosystems, 5, 233-247. Milner, A.M. & Petts, G.E. (1994) Glacial rivers: physical habitat and ecology. Fresh¬ water Biology, 32, 295-307. Nordenskjöld, A.E. (1870) Redergörelse for en Expedition till Grönland ar 1870. Ov. Kgl. Vetensk.-Akad. Förh., 10. Prowse, T.D. (1994) Environmental significance of ice to streamflow in cold re¬ gions. Freshwater Biology, 32, 241-259. Röthlisberger, H. & Lang, H. (1987) Glacial hydrology. Glacio-fluvialsediment trans¬
fer. - An alpine perspective (eds. A.M. Gurnell & M.J. Clark), pp. 207-284, John Wiley & Sons, Chichester, UK. Saether, O.A. (1968) Chironomids of the Finse area, Norway, with special reference to their distribution in a glacier brook. Archiv für Hydrobiologie, 64, 426- 483. Steffan, A.W. (1971) Chironomid (Diptera) biocoenoses in Scandinavian glacier brooks. The Canadian Entomologist, 103, 477-486. Steinbock, O. (1934) Die Tierwelt der Gletschergewässer. Jahresbericht des Deutsch- Österreichischen Alpenvereins, 1934, 263-275. Steinmann, P. (1907) Die Tierwelt der Gletscherbäche. Eine faunistisch-biologische Studie. Annales de Biologie Lacustre, 2, 30-150. Thienemann, A. (1912) Der Bergbach des Sauerland. Internationale Revue der gesamten Hydrobiologie Supplement, 4, 1-25. Ward, J.V. (1985) Thermal characteristics of running waters. Hydrobiologia, 125, 31-46. Ward, J.V. (1992) Aquatic insect ecology: 1. biology and habitat. John Wiley & Sons, New York. Ward, J.V. (1994) Ecology of alpine streams. Freshwater Biology, 32, 277-294. Seite Leer / Blank leaf
Zoobenthos of kryal and lake outlet biotopes 17
Chapter 2
Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain
Abstract
Although alpine streams are considered harsh environments, a diversity of stream types was found in a glacial flood plain in the Swiss Alps. Here we compare macroinvertebrate communities between a kryal stream and a glacial lake outlet. Benthic communities displayed similar abundances over time, but differed substan¬ tially in taxonomic composition. Chironomidae dominated the kryal stream, whereas Ephemeroptera, Plecoptera and Trichoptera were abundant in the lake outlet. As¬ semblage differences were attributable to differences in temperature, substrate sta¬ bility, turbidity, and flow regime.
2.1 Introduction
Although alpine stream systems are considered harsh environments, they con¬ sist of a complex mosaic of kryal, krenal and rhithral segments (Ward, 1994; McGregor et al, 1995). Kryal habitats, restricted to the uppermost reaches of gla¬ cial-fed streams, are characterized by temperatures below 4°C and large diel flow fluctuations in summer (Ward, 1994). In contrast, proglacial lakes lead to relatively stable outlet channels with higher temperatures (Milner & Petts, 1994). Here, we report on benthic macroinvertebrate assemblages in a kryal stream and an adjacent glacial lake outlet in Val Roseg, Swiss Alps. Our primary objective was to compare community structure between these two markedly different habitat types.
2.2 Study area
Val Roseg is situated in the Bernina Massif (granite/diorite) of the Swiss Alps. The catchment area is 66.5 km2, of which 30% is glaciated and 40% is bare rock or glacial till without vegetation cover. Elevations range from 1760 m a.s.l. (confluence with Bernina River) to 4049 m a.s.l. (Piz Bernina). The outlet stream from Roseg Lake (0.27 km2), situated at the base of the Roseg Glacier, and the kryal stream flowing from the Tschierva Glacier converge in the upper part of the main flood 18 Zoobenthos of kryal and lake outlet biotopes
in Tockner et al plain to form the Roseg River. A more detailed description is given (1997). the Macroinvertebrates were sampled at 16 sites encompassing both longitu¬ lake dinal gradient and different habitat types. Here we report on 2 sites: the glacial outlet channel and the kryal site (Table 2.1).
of varia¬ Table 2.1 : Characterisation of sampling sites. Values are means ± 1 SD. Coefficients
= tion (CV = SD / mean) in parentheses. Sample size was n = 12 for the lake outlet, and n 10 for the kryal channel (sampling was not possible in February and March 1997).
Lake outlet Kryal channel
General characteristics
Location 50 m below Roseg Lake 600 m from Tschierva Glacier Elevation (m a.s.l.) 2159 2100 Stream order 1 1 Physico-chemical variables Temperature (°C) 1.5 ± 1.6 (1.09) 0.6 ± 0.4 (0.71) Conductivity (u.S/cm) 50.6 ± 16.9 (0.33) 62.8 ±31.8 (0.51) Turbidity (NTU) 70.5 ± 52.2 (0.74) 68.9 ± 75.9 (1.10) Substrata (cm) 11.3 ± 7.4 (0.66) 10.3 ± 5.3 (0.52) Sedimentary organic matter Coarse BOM (g/m2) 0.97 ± 0.89 (0.92) 0.43 ± 0.41 (0.96) Fine BOM (g/m2) 0.66 ± 0.35 (0.53) 0.66 ± 0.32 (0.48)
2.3 Methods and data analysis
Three replicate Hess samples (area = 0.043 m2, 100 |im mesh size ) were collected at monthly intervals from February 1997 to January 1998 (sampling in the kryal channel was not possible in February and March 1997). In the laboratory, macroinvertebrates were sorted and identified to the lowest practical taxonomic level using a dissecting microscope. The remaining material from each sample was dried at 60°C, ashed at 550°C and weighed for determination of benthic organic matter (BOM) as ash-free dry mass (AFDM). Temperature was measured continuously with instream data loggers. Monthly measurements of conductivity and turbidity were taken early in the morning to minimize the effects of daily fluctuations and to make data comparable. Substratum particle size was assessed in the field by measuring the b-axis (width) of 100 stones selected randomly. Zoobenthos of kryal and lake outlet biotopes 19
Differences in mean total abundances of zoobenthos from each site and sam¬
Data were pling date were analysed using two-way analysis of variance (ANOVA). log10- transformed to meet the assumptions of ANOVA prior to analysis. Differences in community structure were determined using non-metric multidimensional scaling (NMDS) of 4th root transformed abundance data using the Bray-Curtis similarity total com¬ measure. Data were transformed to obtain a better representation of the ordination was munity structure, not just the dominant taxa. A two-dimensional plot produced using centroids of the 3 replicates taken from each site on each date. Two- for differ¬ way crossed analysis of similarity (ANOSIM) was used to test significant in ences in community structure. The influence of particular taxa separating among Statistica sites was quantified using the similarity percentage procedure (SIMPER). (Statsoft, 1995) was used for ANOVA and NMDS, the Primer package (Carr, 1997) for the 2 other analyses. Statistical significance was assessed at a probability level of 0.05.
2.4 Results
Average faunal abundances were between 1023 to 8899 ind./m2 in the lake outlet, whereas abundances in the glacial channel varied between 109 and 25,891 ind./m2 (Fig. 2.1). Ephemeroptera and Plecoptera collectively contributed from 33.5% to 84.6% of total zoobenthos in the lake outlet, and from 0% to 43.3% in the kryal channel. The respective percentages for Chironomidae were 6.8 - 54.5% and 56.7 - 98.2% (Fig. 2.2). Other common taxa included Crenobia alpina, Oligochaeta, Nema- toda, Acrophylax zerberus and Limoniidae. The limnephilid caddisfly Acrophylax zerberus was collected only in the lake outlet. Results of two-way ANOVA indicated significant differences in zoobenthic abundances between sites (p = 0.007), dates (p < 0.0001) and the site x date inter¬ action (p < 0.0001). The stress value of 0.203 for the NMDS ordination indicated a fair representation of the original data in 2 dimensions. The resulting ordination plot clearly delineated the lake outlet and the kryal channel. Community structure in the glacial channel was more variable over time than in the lake outlet, where data points were more tightly clustered (Fig. 2.3). This coincided with mean total abun¬ dances (Fig. 2.1). Two-way crossed ANOSIM confirmed that the differences in com¬ munity structure between the 2 sites were significant (Global R = 0.885, p = 0.01). The 6 taxa that contributed most to the separation between the 2 sites (both aver¬ aged across dates) in the SIMPER analysis are listed in Table 2.2. 20 Zoobenthos of kryal and lake outlet biotopes
40000 T a kryal channel a lake outlet
30000
ï 20000 -o
10000
Fig. 2.1: Average abundances of zoobenthos at the 2 sites during the study period. Error bars represent 1 SD.
100 lake outlet
g '^ 'to o a. E 50 o o s a> o i— CD Q.
FMAMJJASONDJ A M J J A S 0 N D J 1997 1998 1997 1998
D other taxa D Chironomidae Plecoptera S Ephemeroptera
Fig. 2.2: Percent composition of zoobenthos in the lake outlet and the kryal channel during the study period. Zoobenthos of kryal and lake outlet biotopes 21
1.5
channel a kryal p9 lake outlet I.O
* • . . ..-'a8 A2 \
0.5 .. / a7 Al CNJ -" 9a ; 7nn6 / A 6*.' ; 10a 12 Ë o.o 3a m CD a5 E •. a4 a * • D8 -0.5
d12 4 n10 n11 -1.0 n5
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Dimension 1
Fig. 2.3: NMDS ordination plot representing centroids of replicate samples. Numbers represent months: 1 = January, 2 = February, etc.
Table 2.2: Results of SIMPER breakdowns. The 6 taxa contributing most to the average dissimi¬ larity between the 2 sites (both averaged across dates) are listed with their percentage contri¬ bution and cumulative contribution to overall dissimilarity.
Taxon Individual contribution Cumulative contribution (%)
Rhithrogena s pp. 13.9 13.9 Chironomidae 11.1 25.0 Oligochaeta 10.8 35.8 Crenobia alpina 8.8 44.6
Protonemura spp. 8.3 52.9
Leuctra spp. 7.3 60.2 22 Zoobenthos of kryal and lake outlet biotopes
2.5 Conclusions
The benthic community of the lake outlet exhibited less variability in total abundances and community structure than the one in the kryal channel. This is in correspondence with the predictions of the model of Milner &c Petts (1994). Chironomidae clearly dominated the kryal stream zoobenthos, whereas Ephemeroptera, Plecoptera and Trichoptera, in addition to Chironomidae, were abun¬ dant in the lake outlet. The lake outlet did not show a predominance of filter-feeding organisms as described for lower elevation lake outlet streams (Lillehammer & Brittain, 1978), but rather a scraper predominance as found in the High Tatra (Kownacki et al, 1997). Assemblage differences are likely to be attributable to differences in tem¬ perature, substrate stability, turbidity, and flow regime. Although kryal streams have harsh conditions, they are by no means "deserts", and they sustain a very character¬ istic benthic fauna.
2.6 References
Carr, M.R. (1997) Primer User Manual Plymouth Marine Laboratory, Plymouth, UK. Kownacki, A., Dumnicka, E., Galas, J., Kawecka, B. & Wojtan, K. (1997) Ecological characteristics of a high mountain lake-outlet stream (Tatra Mts, Poland,). Archiv für Hydrobiologie, 139, 113-128. Lillehammer, A. &C Brittain, J.E. (1978) The invertebrate fauna of the streams in Ovre Heimdalen. Holarctic Ecology, 1, 271-276. McGregor, G., Petts, G. E., Gurnell, A. M. & Milner, A. M. (1995) Sensitivity of alpine steam ecosystems to climate change and human impacts. Aquatic Con¬ servation: Marine and Freshwater Ecosystems, 5, 233-247. Milner, A. M. & Petts, G. E. (1994) Glacial rivers: physical habitat and ecology. Freshwater Biology, 32, 295-307. Statsoft (1995) Statistica for Windows (vol III). Statsoft, Tulsa, USA. Tockner, K., Malard, F., Burgherr, P., Robinson, CT., Uehlinger, U., Zah, R., &C Ward, J.V. (1997) Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv für Hydrobiologie, 140, 433-463. Ward, J. V. (1994) Ecology of alpine streams. Freshwater Biology, 32, 277-294.
Longitudinal and seasonal distribution patterns 25
Chapter 3
Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps)
Abstract
Seasonal changes in longitudinal patterns of environmental conditions and macroinvertebrate community distributions were examined in an alpine glacial stream (Roseg River, Switzerland). Physico-chemical parameters reflected seasonal changes in glacial influence via shifts in water sources and flowpaths (glacial meltwater vs. ground water), and were best described by turbidity, particulate phosphorous and specific conductance. High nitrogen concentrations indicated snowmelt was the main water source in June. Macroinvertebrate densities and taxon richness were highest during spring (4526 ind./m2 and 16 taxa, all sites combined) and late autumn/early winter (8676 to 13,398 ind./m2 with 16 to 18 taxa), indicating these periods may be more favorable for these animals than summer when glacial melting is maximal. Diamesa spp. (Chironomidae) dominated the fauna at the upper 3 sites (>95% of zoobenthos) and were abundant at all locations. Other common taxa at lower sites (M4 to MIO) included other Chironomidae (Orthocladiinae, Tanytarsini), the may¬ flies Baetis alpinus and Rhithrogena spp., the stoneflies Leuctra spp. and Protonemura spp., blackflies (Simulium spp., Prosimulium spp.), and Oligochaeta. Co-inertia analy¬ sis revealed a strong relationship between environmental conditions and benthic macroinvertebrate assemblages. Furthermore, it elucidated temporal variability in longitudinal response patterns, as well as a similarity in temporal patterns among individual sites. Our results suggest that zoobenthic gradients are not related solely to temperature and channel stability. Seasonal shifts in sources and pathways of wa¬ ter (i.e., extent of glacial influence), and periods of favourable environmental condi¬ tions (in spring and late autumn/early winter) also strongly influenced zoobenthic distributions. 26 Longitudinal and seasonal distribution patterns
3.1 Introduction
Treeline and the permanent snowline comprise 2 important boundaries for life in alpine environments. Above treeline the alpine zone is characterized by sedge- and cushion in the mat vegetation that gradually fades into grass hummocks plants of stream can be subnival zone (Landolt, 1992). Three principal types ecosystems dominated distinguished in areas of high altitude (Ward, 1994): kryal (glacier-melt ), krenal (groundwater-fed) and rhithral (seasonal snowmelt dominated). Environmental feature of harshness (cold temperatures and low food resources) is a general alpine and of streams. Because of heterogeneity in the climate, geomorphology hydrology alpine environments, alpine stream networks often form a complex mosaic of kryal, krenal and rhithral segments (Ward, 1994; McGregor et al, 1995). low Kryal stream segments are characterized by (1) year-round temperatures (Tmax = 4°C), (2) large diel flow fluctuations in summer with peaks in late afternoon from glacial melting, (3) highly turbid water from suspended rock flour through glacial scouring, and (4) usually low channel stability close to the glacier, but with increasing stability downstream (Milner & Petts, 1994; Ward, 1994). Furthermore, instream environmental conditions depend on distance from the glacier terminus, of the season, and the contribution of non-glacial water sources. Downstream kryal can extend for a segment, a glacio-rhithral zone, fed by a mixture of water sources, considerable distance, with temporal changes in discharge and temperature reflect¬ ing the relative proportion of glacial influence (Füreder, 1999). the world Alpine streams play an important role in river systems throughout and (Füreder, 1999), and may be more affected by global climate change anthropo¬ genic impacts than mountain streams at lower elevations (Chapin III & Körner, 1994; McGregor et al, 1995). Despite a well documented interest in high mountain streams at the beginning of the 20th century (Steinmann, 1907; Thienemann, 1912; Steinbock, 1934; Dorier, 1937), most later studies have dealt principally with hydrology, glaciology, geomorphology and physico-chemical attributes (e.g., Gurnell & Fenn, 1985; Röthlisberger & Lang, 1987; Prowse, 1994; Gurnell etal, 1999). In contrast, comprehensive studies examining longitudinal distribution patterns of alpine benthic communities are few (e.g., Saether, 1968; Steffan, 1971; Kownacka & Kownacki, 1972; Kownacki, 1991), and typically of limited duration or intensity. Year-round studies, such as conducted by Lavandier & Décamps (1984) in the Estaragne, an alpine headwater stream in the French Pyrenees, are scarce but of major importance. Synthesizing the available literature on the ecology of glacial-fed streams, Milner &c Petts (1994) proposed a qualitative model that relates gradients of zoobenthic communities to temperature and channel stability; both being a function of distance from the glacier terminus and time since déglaciation. Close to the gla¬ cier, where summer temperatures typically do not exceed 2°C, chironomids of the genus Diamesa generally dominate, or are the sole animals found. Where maximum temperatures are between 2 and 4°C, other Diamesinae, Orthocladiinae and Simuliidae typically may be present. When water temperatures exceed 4°C, Baetidae, Nemouridae Longitudinal and seasonal distribution patterns 27 and Chloroperlidae typically colonize glacial streams. Further downstream other Ephemeroptera, Plecoptera and Diptera, as well as Trichoptera, are predicted to modifications of downstream appear. Milner & Petts (1994) also suggested that pat¬ confinement and lakes. For ex¬ terns may result from tributaries, changes in valley ample, Burgherr & Ward (in press) reported distinct differences in the zoobenthic lake outlet stream. communities of a kryal channel and an adjacent proglacial The study of relationships between faunal assemblages and their environ¬ the inherent ment is a central theme of community ecology. Whenever possible, spatial and temporal variation of both the environment and the faunal assemblage should be considered (e.g., Borcard et al, 1992; Franquet et al, 1995) to gain a better understanding of the ecological phenomenon under study (Resh & Rosenberg, 1989). Ordination techniques, designed to summarize and simplify complex data and relat¬ sets, provide a powerful tool for analyzing patterns of biotic assemblages, ing them to measured environmental variables (ter Braak &C Verdonschot, 1995). Recently, co-inertia analysis has been proposed as an alternative to canonical corre¬ spondence analysis for the study of the co-structure between faunistic data and envi¬ ronmental measurements (Dolédec & Chessel, 1994). This technique is particularly suitable for the simultaneous detection of faunistic and environmental features in studies of ecosystem structure. The principal objective of this study was to examine longitudinal patterns in macroinvertebrate assemblages and environmental conditions in the glacial-fed Roseg River, Switzerland. Specifically, the following questions were addressed: (1) Do en¬ vironmental conditions exhibit distinct patterns with increasing distance from the glacier terminus and across seasons? (2) Are there predictable spatio-temporal pat¬ terns reflected in zoobenthic assemblages? (3) Is there concordance among environ¬ mental and biotic gradients? Because the co-structure between fauna and environ¬ ment may vary in space and time, we also present an approach to separate the spatial co-structure among dates and the temporal co-variation among sites.
3.2 Study area
General description
The study was conducted in the Val Roseg located in the Bernina Massif of the Swiss Alps (lat 46°29'28" N, long 9°53'57" E) (Fig. 3.1a). The 11.3 km long Roseg River (Ova da Roseg) drained a catchment area of 66.5 km2, of which 30% was covered by glaciers, 40% was bare rock or glacial till without vegetative cover, and 9% was used for pasture. Bedrock consists of granite and diorite of the Bernina-, Corvatsch- and Sella-Nappes, which all belong to the austroalpine nappes. Elevation ranged from 1760 m a.s.l. (confluence with Bernina River) to 4049 m a.s.l. (Piz Bernina). Precipitation averaged 1600 mm/yr, of which 50% is snow (1951 - 1980, Spreafico et al, 1992). Glacial meltwater from the Tschierva and Roseg glaciers was 28 Longitudinal and seasonal distribution patterns
the primary water source of the Roseg River. Mean annual discharge of the Roseg River was 2.76 m3/s (1955-1997, Swiss Hydrological and Geological Society). Dis¬ charge peaked in late summer when diel flow fluctuations also were highest. In win¬ ter the system was fed largely by groundwater pathways. Treeline was about 2300 m a.s.l. in the Val Roseg, with subalpine coniferous forests restricted to the lower parts of the valley slopes, whereas no trees occurred on the fluvio-glacial gravel of the valley floor. Dominant tree species were larch (Larix decidua Mill.), stone pine (Pinus cembra L.) and mugo pine (P. mugo Turra). Shrubs consisted mainly of green alder (Alnus viridis (Chaix) DC) and willow (Salix spp.). In the lower alpine zone, mat-grass (Nardus striata L.) vegetation was dominant. Curved sedge (Carex curvula All.) characterized the middle and upper alpine zone. In the subnival zone, sparse hummocks of curved sedge polsters occurred, otherwise only lichens and mosses developed.
Study sites
Five distinct reaches characterize the Roseg River: an unstable braided proglacial reach (length 900 m) below Tschierva glacier, a lake outlet stream below proglacial Lake Roseg (length 950 m), a single thread channel incised in glacial till (length 600 m), the main glacial flood plain (length 2750 m), and a canyon-con¬ strained reach that extends downstream to the river mouth at Pontresina (length 7050 m) (Figs. 3.1b, 3.2a). In the main flood plain, 6 distinct channel types can be distinguished (Tockner et al, 1997). Ten sampling sites (Fig. 3.2a) were located along the main channel (thalweg) of the Roseg River from just below terminus of the Tschierva glacier (site Ml) down to the river mouth (site M10), covering an altitudinal gradient from 2150 to 1760 m a.s.l. For comparison with previous work on the hydro-ecology of this glacial river system, sites numbered Ml to M10 in this paper corresponded to codes M-20, M-15, M-10, M-1, M10, M12, M20, M30, M32, and M40 used in previous papers (Tockner et al, 1997; Malard et al, 1999, Ward et al, 1999). Some general site characteristics are summarized in Table 3.1. Samples were collected in June (increas¬ ing discharge), August (high discharge), October (decreasing discharge), and No¬ vember 1997 (low and constant discharge) (Fig. 3.2b). For sites Ml and M2, sam¬ ples were collected only in August and October. Longitudinal and seasonal distribution patterns 29
(b) Lake 1760 m a.s.l Glacier
Forest
4049 m a.s.l.
Fig. 3.1: (a) Geographical location of the Val Roseg study area in the Bernina Massif of the Swiss Alps, (b) Catchment area of the Val Roseg and indication of glaciers, lakes and forests. Black circles indicate locations of sampling sites. 30 Longitudinal and seasonal distribution patterns
(a)
2200
4 6 8 distance from glacier terminus (km)
(b)
1997
Fig. 3.2: (a) Longitudinal and altitudinal location of sampling sites along the Roseg River. Dashed lines indicate reach borders, (b) Annual discharge of the Roseg River in 1997. Arrows indicate sampling dates for environmental variables and macroinvertebrates. Longitudinal and seasonal distribution patterns 31
Table 3.1: General characteristics of the sampling sites. Means and ranges in parentheses are given for near-bed velocity and annual mean temperature. Annual degree days are expressed as centigrade temperature units (CTU).
03 03
n> ~— CD _> -a I—
^_ C3 CO =3 C/3 L_ >- ** oo Lo r~~ LO !_ m m rt- to m CO < -a •— CM 0O CO 03 CTÎ
cm co •* oo oo o en co in co
0O *3" CO CO
g o r- O O CT> o O CD CD CD oo CD CD
- S CD CD CD O CO CD O O- (£3 CO oo r- CT5 r- i E «d- CO »— CO LO CT) < s O O <- N cm CM
CO CT) co "=3- CM CT) CO CD OO O 03 CO r— CO CO LT) CO CO p- o o CD CD CD CD CD CD CD CD CD 03 > CM CT3 CT) LO OO en «3- LT3 OO CD o CM OO CD CM CM 03 -Q O CD CD CD CD CD CD CD CD CD
CD .C/3 LO p» CO CO O LO CO CM CD CT) CO OO «3" LO *3" 03 B CD CD CD O CD CD CD O CD CD
l_ 03 O CO O)
E . O E * ' H— - 03 co O =3 G E «3- CJ3 *3- CO CO CT) CM CT) CM CD CO CM P- CO LO «3- F co CO i_ 5 r~~ OO LTJ CM OO CO co CD 03 CM OO ^- CO px a
cz ^^^ O —; 4-* CO CD r-- CM co en CD ^3" CM OO > oo CM CT) CO CM CD co O CO r— 03 E CD CD CD CD CT) CO p» CM CM CM CM CM CM
"O T3 o o G 03 03 0) a) 75 "cä "co 'ca 'co _£= c _£= "G 'o o a. a. 'ca 'CD 'co 'co "O i— L- k. CD _ca CD 03 +- +-> 4—' 1—1 O o -a CO CO CO CO O) OJ OJ CO CD O o c: er £= O o O o O o 03 i— o o o o cc Q_ D_ cu u_ LL. o o o o
CD 03 CM oo *3" LT) co p» CO CT) ^ CO i 2 32 Longitudinal and seasonal distribution patterns
3.3 Materials and methods
Physico-chemical parameters were collected in the morning at each site and sampling date to minimize the influence of daily discharge fluctuations and ensure data comparability. Turbidity was determined with a portable turbidity meter (Cos¬ were measured mos; Fa. Züllig, Switzerland). Specific conductance and temperature in the field using a pocket conductivity meter (WTW LF323-B; Wissenschaftliche- Technische Werkstätten, Germany). Suspended solids (SS) were estimated according and to APHA (1989). Nitrate (N03-N) was determined according to Downes (1978) modifications by Stöckli (1985). Soluble reactive phosphorous (SRP) was analyzed according to the molybdenum blue method (Vogler, 1965). Particulate nitrogen (PN) and particulate phosphorous (PP) were quantified as N03-N and SRP after digestion with K2S2Og at 121°C (Ebina et al, 1983). Dissolved organic carbon (DOC) was measured by wet oxidation with subsequent acidification and C02 IR-detection, whereas total inorganic carbon (TIC) was measured as C02 with a Horiba IR-detec- tor after samples had been acidified and heated to 860°C. For further details see Tockner et al (1997). Continuous temperature records (1 hr intervals; StowAway XTI temperature loggers, Onset Corporation, USA) were available for 6 of the 10 sites (U. Uehlinger, unpubl. Data). Near-bed current velocities were measured at the Air same positions where benthic samples were taken using a flow meter (Mini 2; Schiltknecht Messtechnik AG, Switzerland). Substrate size was assessed in the field by measuring 100 randomly selected stones (b-axis, width). Initiation of sediment transport was used as a measure of channel stability because it also was considered a useful disturbance threshold for benthic macroinvertebrates. Estimates of dimensionless critical shear stress (based on channel geometry and grain size distri¬ butions; Gessler, 1965) were used to determine the critical discharge at which bed sediments start to move, using the software FLUSSBAU (Laboratory of Hydraulics, Hydrology and Glaciology, Swiss Federal Institute of Technology, Zurich, Switzer¬ land). Three Hess samples (area 0.043 m2, 100 urn mesh size) from randomly se¬ lected locations in riffle/run habitats were collected at each site and date, and pre¬ served in 4% formalin. In the laboratory, macroinvertebrates were sorted, enumer¬ ated and identified to the lowest practical taxonomic level for most groups using a dissecting microscope. However, Chironomidae were only identified to subfamilies and tribes except for Diamesa spp. for which some estimates for species groups were made. Limnephilid caddiesflies also were pooled, as identification to the species level is only possible using fifth instar larvae (Waringer & Graf, 1997). Nematoda, Hydrachnellae, Crustacea and Oligochaeta were not identified further. After removal of invertebrates, the remaining material from each sample was split into 2 fractions (>1 mm and <1 mm), dried at 60°C, weighed, ashed at 550°C and reweighed for determination of coarse and fine benthic organic matter (BOM) as ash-free dry mass (AFDM). Longitudinal and seasonal distribution patterns 33
Macroinvertebrate community analyses included estimates of mean densi¬ measure of diver¬ ties, taxon richness and Simpson's index of concentration (SI) as a
- - and sity. SI = 2 ((n2 n) / (N2 N)), where n is the number of individuals in a species, to N = Zn. As recommended by Rosenzweig (1995), SI values were converted -ln(SI) of such that values increase as the number of species does, independently sample size. in Co-inertia analysis (CIA) was used to simultaneously study the structure the environmental and faunistic data, and to determine if concordance (i.e., co-struc¬ ture) between these 2 independent structures existed (Dolédec & Chessel, 1994). whereas Environmental data were normalized to ensure equal weights for all variables, reduce inter- the faunistic data were log10 (x+1) transformed and centered to strong taxonomic differences in densities. CIA is a two-table ordination method, as is ca¬ nonical correspondence analysis (CCA). However, CIA enables the joint analysis of tables having similar (even low) as well as different numbers of environmental vari¬ ables, species, and/or samples (Dolédec &c Chessel, 1994). In contrast, in CCA a small number of environmental variables is required to predict the faunistic struc¬ ture, otherwise it would be reduced to a simple correspondence analysis of the faunistic table. Last, the temporal co-variation and spatial co-structure of the 2 data sets were examined using within-class co-inertia analysis (Franquet &c Chessel, 1994; Franquet et al, 1995). All multivariate analyses were computed using the ADE-4 software (Chessel & Dolédec, 1996).
3.4 Results
Environmental conditions
A detailed description of the physico-chemical templet of the Val Roseg was previously summarized (Tockner et al, 1997; Malard et al, 1999); therefore, only a selection of parameters distinctive for kryal streams will be presented here. Substrata at all sites consisted mainly of cobbles and boulders. Average particle size (b-axis, width) ranged between 9 to 12 cm. Estimates of critical shear stress indicated bed sediments in the main channel of the Roseg River should start to move if discharge exceeds 8 m3/s. This threshold was exceeded on 41 days in 1997; 15x in June, 6x in July, and 20x in August. Daily mean water temperature in the main channel ranged from 0.01 to 0.92°C at site M2, 0.39 to 4.83°C at M5, and 0.00 to 6.73°C at M10. Annual degree days increased from 154 CTU (Centigrade Temperature Units) at site M2, to 657 CTU at M5, and 945 CTU at M10 (Table 3.1). Nitrate (N03-N) was the dominant nitrogen species (87%) in surface water of the Roseg River (Malard et al, 1999). The average nitrate concentration was highest in June at 421 ± 17 u,g/L (mean ± SD), and between 180 and 266 ug/L on the other sampling dates (Fig. 3.3a). Values of specific conductance (Fig. 3.3b) were lower during the peak of gla¬ cier melting (36 ± 5 uS/cm in August, 38+5 uS/cm in October). Specific conduct- 34 Longitudinal and seasonal distribution patterns
minimal. ance was highest in November (69 ± 7 uS/cm) when glacial melting was still the Intermediate levels were found in June (55 ± 9 uS/cm) when snowmelt was main source of water, but glacier melting already started. Turbidity and particulate phosphorous (PP) (Fig. 3.3c,d) peaked in August with an of due to glacier melting and associated high loads of glacial flour, average 84 ± 11 NTU (Nephelometric Turbidity Units) for turbidity and 49 ± 16 ug/L for PP. In June and October (when discharge was increasing and decreasing, respec¬ and 58 ± 18 tively) intermediate levels were observed (turbidity: 51 ± 13 NTU NTU; PP: 29 ± 7 ug/L and 29 ± 6 ug/L), whereas during low flow in November for minimum values were recorded (13 ± 12 NTU for turbidity, 11 ± 5 jig/L PP). values in the incised Turbidity and to a lesser extent PP reached substantially higher values were reach compared to the proglacial area, whereas further downstream attenuated again. Overall, longitudinal patterns in physico-chemical parameters ex¬ hibited pronounced seasonal changes.
Macroinvertebrates
Total densities of macroinvertebrates ranged from 116 ind./m2 (M2 in Au¬ gust) to 24,380 ind./m2 (MIO in October), varying along both temporal and spatial dimensions. Lowest densities, all sites combined, were found in August (1472 ± 1355 ind./m2), intermediate in June (4526 ± 2723 ind./m2) and November (8676 ± 4434 ind./m2), and highest in October (13,398 ± 6113 ind./m2). Generally, average densities per site increased with increasing distance from the glacier terminus (Table 3.2). CVs were highest in the proglacial reach (Ml to M3), decreased to a minimum at M6, and then increased again downstream to M10. The number of taxa recorded at a site ranged from 1 to 26. Lowest taxon richness was found within the proglacial reach (Ml to M3), followed by a progres¬ sive increase to M6 at the lower end of the flood plain, then similar richness at the 4 lowermost sites (Table 3.2). Overall, Chironomidae, Baetis alpinus Pictet, Rhithrogena spp., Leuctra spp., Protonemura spp., Simuliidae, and Oligochaeta were the pre¬ dominant taxa (Table 3.3). Other taxa generally made up < 1% of the total commu¬ nity, although some exceptions on certain dates and sites occurred (e.g., Rhypholophus sp. comprised 6 to 13% in the proglacial area in August, Crenobia alpina ca. 3% at M3 and M4 in August, and Limnephilidae 7% at MIO in October). The few large individuals of limnephilid Trichoptera were identified as Acrophylax zerberus Brauer Sind Drusus biguttatus Pictet. Taxa not used in multivariate analyses (<0.05% of the assemblage,or semi-aquatic) were not included in Table 3.3. These were Ecdyonurus picteti Meyer-Dür, Dictyogenus spp., Siphonoperla spp., Rhyacophila intermedia McLachlan, Ceratopogonidae, and Collembola. Diamesa spp. was the only genus of Chironomidae found close to the glacier terminus. Orthocladiinae first appeared at M3, but in low abundance. At M4, rela¬ tive densities of Diamesa spp. declined by ca. 30%, whereas Orthocladiinae contri- Longitudinal and seasonal distribution patterns 35
100 (b) _
distance from glacier terminus (km)
Fig. 3.3: Longitudinal patterns of (a) nitrate (N03-l\l), (b) specific conductance, (c) turbidity, and (d) particulate phosphorous (PP) on each sampling date. Dashed lines indicate reach borders. Order of sampling sites and names shown in Fig. 3.2a. 36 Longitudinal and seasonal distribution patterns
buted >1%. Further downstream, the ratio of Diamesa spp. to Orthocladiinae var¬ Chironomidae ied between 1 to 4. Tanytarsini always constituted a minor part of total D. present. Common Diamesa-groups found, were D. steinboecki-gt., latitarsis-gr., D. bertrami-gr., D. einerella-zernyi-gx., and D. cf. incallida. Baetis alpinus was ca. Navàs and R. nivata 2.5x more abundant than Rhithrogena spp. Rhithrogena loyolaea at all while R. Eaton were the 2 predominant Rhithrogena-species sites, alpestris reach. Eaton and R. degrangei Sowa were most abundant in the constrained Simpson's in incised reach and the index was low in the proglacial reach, but increased the reach flood plain, and reached a maximum at sites in the constrained (Table 3.2).
Table 3.2: Means ± 1 SD, and in parentheses coefficients of variation (CV in %) for densities, taxon richness, and Simpson's index (-In SI) of the macroinvertebrate communities at the 10 sampling sites.
Site Density (ind./m) Taxon richness Simpson's index (-In SI) Ml 2632 ± 3536 (134) 5±1 (16) 0.27 ±0.37 (134)
M2 2469 ± 3327 (135) 2±1 (71) 0.21 ±0.29 (141)
M3 5417 ± 8020 (148) 8±1 (6) 0.38 ± 0.39 (104)
M4 5048 ± 4562 (90) 11±4 (36) 0.71 ±0.39 (54) M5 8438± 5237 (62) 16 ±4 (26) 1.11 ±0.41 (37)
M6 7936 ± 4504 (57) 20 ±5 (25) 1.21 ±0.35 (29) M7 8136± 5991 (74) 20 ±4 (20) 1.42 ±0.10 (7) M8 9138± 7074 (76) 19±3 (16) 1.43 ±0.27 (19) M9 8504 ± 7074 (83) 19 + 2 (12) 1.46 ±0.44 (30) M10 8415±10753 (128) 18±5 (30) 1.34 ±0.50 (37)
Multivariate analyses
The co-structure between the environmental and faunistic data sets revealed by co-inertia analysis (CIA) was highly significant, as confirmed by a Monte-Carlo permutation test (p = 0). The first 2 CIA axes explained 81.1% and 14.5% percent of the total inertia. Turbidity, PP and SS were negatively related to factor 1 of CIA, while TIC and specific conductance showed a positive relationship (Fig. 3.4a). Fac¬ tor 2 was related to temperature, N03-N, PN and BOM in the positive region. Axis Fl of the corresponding faunistic structure (Fig. 3.4b) was best explained by Orthocladiinae, Protonemura spp., Prosimulium spp., Tanytarsini, Baetis alpinus, Rhithrogena spp., and Leuctra spp., whereas axis F2 was described by Oligochaeta, Simulium spp., Rhabdiopteryx alpina, and Isoperla spp. leuipnyßuoi pue /buosbss uopnquisip sujaned Lî
9|qei :£'£ a;n|osqv saouBpunqB (zLU/-pui) joj aqi lUBwiuopajd bxb; paioa||Oo ;b aqi BujjdLUBS eouesqB }B b 9}js bjb saijs ßuunp aqi Apnjs "pouad joj aiu jaqio bxbj A|uo aouasajd (x) jo (-)
= suoiiBiAejqqv payodaj esneoeq uo aßejaAB Aein painquiuoo > "%L as pjepuejs 'uorjBiAap joj ipee uoxe} pasn m sajnßu. £ y g-£ pue g-£ bjb m saseinuajed
en CT1 CD CD -J CD on CD CT) O co CD oo OO on CJ1 en co CO CD oo co CO .p» CD CT) -P» CD on PO PO CD oo PO 00 CD on oo CD CO oo on PO PO PO PO P»
CO CD OO oo oo CD PO CD oo CD oo CD OO co CO CT) on PO CT) on PO PO •P» PO -P» CT) -p> -p» CD P» PO PO PO oo oo -p» >^i CT) -p» oo on
oo oo CD oo on 00 CT) 00 CD CO oo 00 00 oo CT) co oo PO CT) oo oo co CD on -p» CD on CO oo P» OO .p» oo on co PO oo PO oo -p» PO PO PO PO oo on on -^1 PO PO
oo co oo PO PO on CO oo CD PO OO oo CD p» on CT) an PO -J CD P. -p» CD oo PO on OO on CD on CT) OO co »J co PO PO PO PO on CT) PO PO PO
CD on -p. on CO -p» -P* OO -p» •p» oo CO on CT) •P» en PO CO PO C/5 co CT) oo CD CT) PO oo PO -«J 00 oo OO CD PO O OO CD r+ CO co -p» O on on CT) CO PO PO PO PO OO
CCD 5" on CD oo -p» oo PO oo CD co oo CT) oo OO P» CO -p» CT) «J oo oo on on CO CO PO oo OO •P» CD oo CD OO CT) 3 PO on co co OO on PO CD oo on CO PO PO oo CU 00
.p» o oo oo CT) PO OO OO oo on CO oo OO oo co -P» co on PO CT) -P» oo CD CT) oo PO -P» PO oo OD 00 PO oo oo PO
oo •P« on OO co O CD PO .£» on CD on -P» CD CD CD O CJ1 -J oo 00 PO PO -p* 00 on OO co on -J
PO .p» CD CD O CD O CD o CD CD CD CD CD CD CD o O -P» CT) on -p» p. oo PO oo
oo CD CD CD CD CD CD CD -P» CD CD CD CD CD CD CD CD CD OO co oo ^ on on PO 00
Q) CD a 3 00
CD CU Ü0 o —* oo —t CD 30 "^ EU -Eu. 3. 3' o' p O "—""
_ -p -T3 p CD a =3 a -o 1— CD ûT 00 00 - oo —1 e/) TD a EU c: . TD a 0) TD tu su a' 00 #—*- 00 3 3 TD i o. 1' CD 5" 5" CD oo 0) 00 3 0) Q) 0) «a c: 3, Co CD O" o eu 1" o CD Oo' 3 o o 3' a. a- o o 3 _: ^ oo CD 3 CD ca 3- 3' o 3" eu 0) =r CD 00 30 p- Other taxa
Crenobia alpinaDana (Cre)
Nematoda (IMmt)
Hydrachnellae(Hyd)
Crustacea (Cm)
RhabdiopteryxalpinaKühtreiber(Rha)
Nemoura spp. (Nem)
Capniasp. (Cap)
Isoperiaspp. (Iso)
Périodes intricataPictet(Per)
ümnephilidae(Lim)
Rhabdomastix sp. (Rhm)
Wiedemannia sp. (Wie)
Dicranota sp. (Die)
Rhypholophussp. (Rhy) Longitudinal and seasonal distribution patterns 39
The co-structure between the fauna and the environment is best illustrated by plotting standardized environmental and associated faunistic scores together on the FlxF2 factorial plane and to link the 2 positions by an arrow (Fig. 3.4c). The length of the arrow is then a measure of the strength of the co-structure; the shorter this distance the better the agreement between the 2 structures. In general, all arrows indicated good agreement between the 2 structures. On the FlxF2 factorial map, 4 overlapping groups could be distinguished that corresponded to the 4 sampling dates. However, the dispersion of sampling sites varied from one date to another. The environmental separation of the June samples on axis F2 was caused by high concen¬ trations of N03-N, whereas the other 3 dates were separated by a gradient of glacial influence along axis Fl. High discharge from glacial melting resulted in high levels of particulate variables in August, whereas glacial melting no longer occurred in November and ground water contributed a large proportion to stream water, as reflected by changes in specific conductance and TIC. The ordination separated samples from June/August and October/November along the positive diagonal (lower left to upper right) (Fig. 3.4c). This is attributable to the lower density and taxon richness of samples from June and August compared to those from October and November. Additionally, there existed a downstream increase in densities and taxon richness within each sampling date. These results demonstrate an important overlap between a temporal effect mainly caused by a changing glacial influence and a different spatial (i.e., longitudinal) response pattern. A spatial typology was investigated by means of a within-date CIA. The exist¬ ence of a significant co-structure was confirmed (permutation test, p = 0), with factor 1 explaining most of the observed co-structure (90.6%), and factor 2 contrib¬ uting only marginally (4.8%). Within-date environmental gradients were related to downstream changes in glacial influence (Fig. 3.5a). Although longitudinal patterns differed among dates, 3 groups of sites were observed (Fig. 3.5c). Sites Ml to M3 in the proglacial reach, and M6 in the lower flood plain together with sites M7 to M10 in the constrained reach formed 2 distinct groups, whereas the group formed by sites M4 and M5 reflected the transition between them. This pattern was most clear in August and October, but less distinct in June, and even weaker in November, indi¬ cating conditions were most homogenous on this date. Sites in the constrained reach (M7 to M10) that were fringed by forest exhibited higher BOM concentrations than sites located further upstream (Fig. 3.5a,b). By comparing Fig. 3.5b and 3.5c, the seasonal shifts in taxonomic composition, most evident at sites in the flood plain and in the constrained reach, could be observed. Last, the temporal stability of macroinvertebrate assemblages at each site as¬ sessed using a within-site CIA had a significant co-structure (p = 0). The first 2 axes explained 97.3% of the total inertia (Fl = 80.1%, F2 = 17.2%). The FlxF2 facto¬ rial plane showed a distinctive seasonal shift for all sampling sites (counter-clockwise from June to November) that became even more distinct with increasing distance from the glacier terminus (Fig. 3.6c). Positions of June samples on the FlxF2 facto¬ rial map (Fig. 3.6c) were related to high concentrations of N03-N (Fig. 3.6a), while 40 Longitudinal and seasonal distribution patterns sample positions in August could be attributed to high levels of turbidity, SS and PP. November samples reflected the substantial contribution of ground water (specific showed a conductance, TIC, and to a lesser extent DOC), whereas October samples transitional shift between the declining glacial influence and the increasing propor¬ tion of ground water to stream water. Faunistically, Simuliidae and Oligochaeta con¬ tributed highly to the total community at downstream sites in June (Fig. 3.6b,c). In October and November, several taxa of Ephemeroptera, Plecoptera and limoniid Diptera were responsible for site separation (Fig. 3.6b,c).
3.5 Discussion
Environmental conditions
Glacier-fed rivers exhibit distinctive characteristics that distinguish them from other lotie systems (Steffan, 1971). Stability was closely related to the hydrograph of the Roseg River. Moving bed sediments reduced channel stability from June to Au¬ gust. During the remaining time of the year discharge did not exceed the threshold of 8 m3/s, indicating channel stability was higher. Maximum annual temperatures in the proglacial reach (Ml to M3) were below 2°C, and thus are considered metakryal (Ward, 1994). The strong increase in maximum annual temperature at M4 can be attributed to an attenuation of environmental conditions resulting from the influ¬ ence of the proglacial lake outlet stream. The hypokryal starts downstream of the confluence and extends to the head of the flood plain. Sites M5 to MIO all have maximum temperatures >4°C, being characteristic of a glacio-rhithral zone (Füreder, 1999). Annual degree days never exceeded 1000 CTU even at the lowermost site. Our data are partially comparable with those of Lavandier (1974) in the River Estaragne (France). He reported 500 CTU at the uppermost site at 2350 m a.s.l. and 1000 CTU at the lowermost site at 1850 m a.s.l. compared to only 154 CTU and 964 CTU, respectively, in this study. Snowmelt was the main source of water in June as indicated by the high concentration of N03-N. As in other mountain areas (Charles, 1991), seasonal pat¬ terns in N03-N concentrations suggest that atmospheric deposition on snow is the main source of nitrogen in the Val Roseg. The discharge peak in August, due to glacial melting of the Tschierva and Roseg glaciers, was associated with high turbid¬ ity and particulate phosphorous values. High concentrations of particulate phospho¬ rous associated with glacial flour have been reported elsewhere (Bretschko, 1966). The peak in turbidity at site M4 is most likely attributable to the proglacial Lake Roseg, because larger sediment particles settle in the lake. Therefore, suspended solids in the lake outlet stream comprise a much higher fraction of smaller particles compared to the kryal stream originating from the Tschierva Glacier, which led to the increase in turbidity below the confluence of the 2 streams. With decreasing discharge in October, specific conductance was still low, indicating that the contri- Longitudinal and seasonal distribution patterns 41
F2 F2 |1.25 '° •Sim (a) (b) Oli«
PN M N03-Njk t kT .4 BOM
Ort» »Pro
Nmt R|lm • • • Prs
\ Cre Tan • Bae • TIC Nem * -0.5 ~2-° •mr Cru Dia pp *Di0 •"> ss^*-- ^^"Specific «A, conductance Turbidit7"^F -2.5 ^ - 2.0 \ Iso • Leu SRP ••Rha \doc
-0.5 -1.0
F2 T3.0
(c) ...... jun Aug
- M9." Oct Nov
M10i ;yM8 1W7£.JM6 Mio. ' M5
J \ M4 / K/16
' M5-<< M7 1-2.0
Fig. 3.4: Results of the co-inertia analysis (CIA) of 11 environmental variables and 24 taxa from 10 sampling sites and 4 dates, (a) Ordination diagram of the 11 normalized environmental vari¬ ables in the CIA. Codes are given in material and methods, (b) Position of taxa on the CIA F1xF2 plane. Codes are given in Table 3.3. (l; Standardized co-inertia scores of the environmental and faunistic data sets projected onto the F1xF2 factorial map. Arrows link environmental scores (beginning of an arrow) to faunistic ones (arrowhead). In 3.4a the longer an arrow and the closer to an axis the stronger the relationship to this factor, whereas arrows in 3.4c measure the strength of the co-structure between the fauna and the environment; so the shorter the arrow the better is the agreement between the 2 structures. 42 Longitudinal and seasonal distribution patterns
F2 F2 (a) 0.5 (b) 0.3 • Oli
SRP • Prs Hyd • Cru Specific Sim« Rhm** conductance BOM Lim#»Rha Dic« • Were £S^N03-N "•- Ue #Rhy TIC »-PP Nmt* Tan Nem •pejf 'Ts3»~ "1'° 1.0 »Diaj 0.5 P!LsS^ ^DOC • Ort Leu* • Cap Pro» T >k.SS
•Wie
^Turbidity • Rhi
Bae
-0.5 -0.3
Fig. 3.5: Within-date CIA (spatial typology), (a) Co-inertia scores of the environmental param¬ eters, and (b) of the taxa on the F1xF2 factorial plane, (c) Standardized co-inertia scores of sampling sites of the environmental and faunistic data sets onto the F1xF2 factorial maps for each sampling date. In 3.5a the longer an arrow and the closer to an axis the stronger the relationship to this factor, whereas arrows in 3.5c measure the strength of the co-structure between the fauna and the environment; so the shorter the arrow the better the agreement between the 2 structures. Longitudinal and seasonal distribution patterns 43
F2 F2 J 1.0 (b) J1.0
NÛ3-NW j »Pro bom; Sim«!
^Specific \ ! TIC/ ' conductance 1 Prs PN ' Rhy • ! • Tan • Bae Oli» jRhm ^ T '• *^-"DOC •^.»Dia #Leu -°-5Cre»[» Ort »Rh 2.0 "Cru ,Per -1.5 PP. 1.5 ] Dic%Cap Rha Nmt |Hyd« »m .• ss^^ 1 Iso SRP Turbidity 1 »Wie 1-0.5 [-0.7
M1 M2 M3 M4
Nov JunV Jun Oct s- if -^--- Aug Aug^ | iOct t>Aug Oct Aug /
M5 f\ M7 M8 Jun/ i Jun\ Jun 1
Nov Nov/ Nov/
Aug Oct Oct Aug Oct Aug^
' F2 M9 M10 3 -2 —|- 2 Fl
Jun \ -1.5 , Jun ^ Nov Nov
Aug Oct Aug Oct /
Fig. 3.6: Within-site CIA (temporal typology), (a) Co-inertia scores of the environmental param¬ eters, and (b) of the taxa on the F1xF2 factorial plane, (c) Standardized co-inertia scores of sampling dates of the environmental and faunistic data sets onto the F1xF2 factorial maps for each sampling site. In 3.6a the longer an arrow and the closer to an axis the stronger the rela¬ tionship to this factor, whereas arrows in 3.6c measure the strength of the co-structure be¬ tween the fauna and the environment; so the shorter the arrow the better the agreement be¬ tween the 2 structures. 44 Longitudinal and seasonal distribution patterns
bution of englacial water had not significantly changed. At low and constant dis¬ charge in November, the contribution of ground water had increased markedly, as shown by high values of specific conductance. Seasonal shifts in sources and flow paths of water in the Val Roseg catchment determine the degree of glacial influence (Malard et al., 1999), and have important implications for the biota.
Benthic macroinvertebrate assemblages
In general, Chironomidae, especially the genus Diamesa, were a predomi¬ nant constituent of the zoobenthos in the glacial stream of Val Roseg. Larvae of the genus Diamesa were the sole inhabitants close to the glacier snout, as reported for many glacier-fed streams in the European Alps (Bretschko, 1969), Tatra Mountains (Kownacka &C Kownacki, 1972), Scandinavia (Steffan, 1971), and the Rocky Moun¬ tains (Elgmork & Saether, 1970). In contrast, glacial streams influenced by water abstraction were devoid of fauna for 200 to 500 m below the glacier snouts (e.g., Petts & Bickerton, 1994). Although, Diamesa chironomids are generally considered cold-stenotherms (Oliver, 1983), some species (e.g., Diamesa steinböcki) may occur at warmer temperatures (Rossaro, 1991). Ephemeroptera, Plecoptera and other taxa, however, also made substantial contributions to benthic assemblages. Within the genus Baetis, only B. alpinus was found. The absence of B. melanonyx Pictet, another common species of this region, is attributed to the high elevation of the Val Roseg and the crystalline substratum (Sartori & Landolt, 1999). These authors report a univoltine life cycle for B. alpinus >1500 m a.s.l. with animals overwintering as larvae, which is consistent with our findings. Vom Rhithrogena species were collected: R. loyolaea andR. nivata were the 2 most abundant species at most sites, whereas R. alpestris and R. degrangei were only abundant at the more downstream sites. These distributions correspond to their altitudinal optima, which are between 1000 and 1800 m a.s.l. for R. alpestris, and between 700 and 1200 m a.s.l. for R. degrangei (Sartori & Landolt, 1999). Rhithrogena loyolaea usually has a univoltine life cycle in the Swiss Alps (Sartori & Landolt, but Lavandier 1999), (1981) showed that larval development requires 3 years above 2100 m in the French Pyrenees. Unfortunately, life cycles of many alpine stream insects are poorly known, although there is a general agreement that the number of generations per year decreases with increasing altitude, and that low temperature at elevation is high likely responsible for later emergence (Ward, 1992). Total zoobenthic density ranged over 2 orders of magnitude, being lowest in
August (116 ind./m2) and highest in October (24,380 ind./m2). Similar values were > found by Lavandier & Décamps (1984) in the French Pyrenees although other stud¬ ies of glacial streams report much lower densities (e.g., Kownacki & Kownacka, Gislason et 1973; al., 1998). However, several factors may be responsible for these differences. in Sampling other studies was typically restricted to the summer period, when sites are accessible. easily Different sampling techniques (qualitative vs. quan- Longitudinal and seasonal distribution patterns 45
titative), as well as differences in mesh size, additionally confound comparability of results. Taxon richness and Simpson's index of diversity showed similar trends. On all sampling dates, a downstream increase of both measures was observed. In August, taxon richness was markedly lower, most likely due to the extremely harsh environ¬ mental conditions. Our results suggest that changes in density and taxon richness are most likely attributable to (1) spatial phenomena (reduced harshness with increasing distance from the glacier), and (2) temporal phenomena (more favourable periods in spring and late autumn/early winter than in summer).
Relationships between fauna and environment
Co-inertia analysis clearly demonstrated that the environmental conditions and benthic community structure were related. Furthermore, it revealed the exist¬ ence of a strong overlapping spatio-temporal pattern. Within-class analyses elimi¬ nate such spatial or temporal effects and enable the examination of changes in the spatial co-structure among sampling dates, and the temporal stability in observed longitudinal patterns. The results of these analyses can be summarized as follows: (1) Seasonal changes in glacial influence (i.e., shifts in sources and flow paths of water) comprised a key factor structuring benthic macroinvertebrate assemblages. (2) Al¬ though longitudinal response patterns of zoobenthos varied across seasons, indi¬ vidual sites exhibited similar temporal patterns. Milner & Petts (1994) proposed a general qualitative model relating zoobenthic gradients to maximum water temperature and channel stability as a func¬ tion of distance from the glacier margin and time since déglaciation. In general, our results conform quite well with this model; however, some substantial deviations were found. For example, Orthocladiinae were collected at M3, although maximum annual temperatures (Tmax) were still <2°C. Ephemeroptera and Plecoptera also were found below a Tmax of 4°C at sites M3 and M4, especially in October and November, but to a lesser extent at other times. In accordance to our findings, Thomas (1975) reported thatß. alpinus can complete its whole life cycle even if water temperature is always between 0 and 3°C. Burgherr & Ward (in press) reported similar findings in another study comparing a kryal stream with a lake outlet channel. Overall, Milner &c Petts' (1994) conceptual model predicted macroinvertebrate distribution patterns accurately during glacial melt in summer, but did not account for seasonal changes in assemblage patterns. However, it is this temporal pattern that is most interesting in these types of systems, but which was not incorporated in their 1994 conceptual model. Consequently, we postulate that, besides temperature and channel stability, a complex interplay of factors determines the distributions of populations in glacial streams. In particular, sources and flowpaths of water, in combination with discharge, reflect changes in glacial influence and strongly modify the habitat templet. 46 Longitudinal and seasonal distribution patterns
Longitudinal distribution patterns of macroinvertebrates also could be a re¬ sult of stream age (Milner, 1987; Milner, 1994). However, such a site-specific tem¬ poral succession can be constrained by alternating cycles of glacial recession and advance. For example, M3 in the Roseg River has been ice-free since 1955, whereas M2 has been since 1971, and Ml was again covered by the glacier between 1979 and 1991. Peaks in abundance and taxon richness in spring and late autumn/early win¬ ter coincide with accrual of benthic algae, especially the chrysophyte Hydrurus foetidus Kirch, enhancing food availability for zoobenthic organisms. During these periods, stream beds are highly stable as shown in this study and reported elsewhere (Füreder, 1999), and sediment transport, shear stress and turbidity are minimal (Uehlinger et al., 1998).
Conclusions
In summary, our results show that zoobenthic communities in the Roseg River exhibit a high spatio-temporal variation due to seasonal shifts in glacial influence. During periods of favourable environmental conditions (spring and late autumn/ early winter) invertebrate density and taxon richness peak, and cold-adapted taxa other than just Diamesa spp. can occur rather close to the glacier terminus. Future modifications and refinements of the conceptual model of Milner & Petts (1994) should account for temporal patterns reported in this study and other investigations (Robinson et al., in press). Other factors to consider include biogeographic patterns (e.g., Diamesa is absent in some regions (Willassen & Cranston, 1986)), endemic species that may co-occur with Diamesa in metakryal stream segments (Brodsky, 1980), glacial streams that remain clear all year (e.g., in the Colorado Cordillera (Ward, 1994)), and differences in geology, topography and riparian vegetation among river basins within and among geographic regions (e.g., Gislason et al., 1998).
3.6 Acknowledgements
We appreciate the field support by Drs. C.T. Robinson and H. Eisenmann. We are especially grateful to Dr. U. Uehlinger for providing unpublished tempera¬ ture data, Mr. R. Uli and Mr. B. Ribi for analysing chemical samples, and Dr. B. for Janecek verifying Diamesa identifications. We also wish to thank our colleagues froni the Val Roseg Project for the excellent collaboration. We are grateful to the of villages Pontresina and Samedan for providing access to the study area, and to Mr. P. Testa and his crew at the Roseg Hotel for their hospitality. Special thanks to Drs. F. Malard and C.T. Robinson, and 2 anonymous reviewers for their valuable cri¬ tiques that improved the manuscript. Longitudinal and seasonal distribution patterns 47
3.7 References
APHA (1989) Standard methods for the examination ofwater and wastewater. Ameri¬ can Public Health Association, Washington DC. Borcard, D., Legendre, P. & Drapeau, P. (1992) Partialling out the spatial compo¬ nent of ecological variation. Ecology, 73, 1045-1055. Bretschko, G. (1966) Untersuchung der Phosphatführung zentralalpiner Gletscherabflüsse. Archiv für Hydrobiologie, 62, 327-334. Bretschko, G. (1969) Zur Hydrobiologie zentralalpiner Gletscherabflüsse. Verhandlungen der Deutschen Zoologischen Gesellschaft, 1968, 741-750. Brodsky, K.A. (1980) Mountain torrents of the Tien Shan. Junk, The Hague. Burgherr, P. &C Ward, J.V. (in press) Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 27. Chapin III, F.S. &c Körner, C. (1994) Arctic and alpine biodiversity: patterns, causes, and ecosystem consequences. Trends in Ecology and Evolution, 9, 45-47. Charles, D.F. (1991) Acidic deposition and aquatic ecosystems. Springer-Verlag, New York. Chessel, D. &c Dolédec, S. (1996) ADE Version 4: HyperCard Stacks and Quick Basic Microsoft programme library for the analysis of environmental data. User's Manual. ESA CNRS 5023, Ecologie des Eaux Douces et des Grands Fleuves, Université Lyon 1, France. Dolédec, S. & Chessel, D. (1994) Co-inertia analysis: an alternative method for studying species-environment relationships. Freshwater Biology, 31,277-294. Dorier, A. (1937) La faune des eaux courantes alpines. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 8, 33-41. Downes, M.T. (1978) An improved hydrazine reduction method for the automated determination of low nitrate levels in freshwater. Water Research, 12, 673- 675. Ebina, L., Tsutsui, T. & Shirai T. (1983) Simultaneous determination of total nitro¬ gen and total phosphorous in water using peroxidisulfate oxidation. Water Research, 17, 1721-1726. Elgmork, K. & Saether, O.R. (1970) Distribution of invertebrates in a high moun¬ tain brook in the Colorado Rocky Mountains. University of Colorado Stud¬ ies, Series in Biology, 31, 1-55. Franquet, E. & Chessel, D. (1994) Approache statistique des composantes spatiales et temporelles de la relation faune-milieu. Sciences de la vie, 317, 202-206. Franquet, E., Dolédec, S. & Chessel, D. (1995) Using multivariate analyses for sepa¬ rating spatial and temporal effects within species-envrionmental relationships. Hydrobiologia, 300/301, 425-431. 48 Longitudinal and seasonal distribution patterns
Füreder, L. (1999) High alpine streams: cold habitats for insect larvae. Cold adapted organisms. Ecology, physiology, enzymology and molecular biology (eds. R. Berlin. Margesin &c F. Schinner), pp. 181-196, Springer-Verlag, Gessler, J. (1965) Der Geschiebetriebbeginn bei Mischungen untersucht an natürlichen Abpflästerungserscheinungen in Kanälen. Mitteilungen der Versuchsanstalt für Wasser- und Erdbau der ETH Zürich, 69, 1-67. Gislason, G.M., Olafsson, J.S. & Adalsteinsson, H. (1998) Animal communities in icelandic rivers in relation to catchment characteristics and water chemistry. Preliminary results. Nordic Hydrology, 29, 129-148. Gurnell, A.M., Edwards, P.J., Petts, G.E. & Ward, J.V. (1999) A conceptual model for alpine proglacial river channel evolution under changing climatic condi¬ tions. Catena, 38, 223-242. Gurnell, A.M. & Fenn, C.R. (1985) Spatial and temporal variation in electrical con¬ 108-114. ductivity in a pro-glacial stream system. Journal of Glaciology, 31, Kownacka, M. & Kownacki, A. (1972) Vertical distribution of zoocenoses in the streams of the Tatra, Caucasus and Balkan Mts. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 18, 742-750. Kownacki, A. & Kownacka, M. (1973) The distribution of the bottom fauna in several streams in the Middle Balkan in the summer period. Acta Hydrobiologia, 15, 295-310. Kownacki, A. (1991) Zonal distribution and classification of the invertebrate com¬ munities in high mountain streams in South Tirol (Italy). Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 24, 2010-2014. Landolt, E. (1992) Unsere Alpenflora. Verlag Schweizer Alpen-Club, Brugg, CH. Lavandier, P. (1981) Cycle biologique, croissance et production de Rhithrogena loyolaea Navàs (Ephemeroptera) dans un torrent pyrénéen de haute montagne. Annales de Limnologie, 17, 163-179. Lavandier, P. & Décamps, H. (1984) Estaragne. Ecology of European Rivers (ed. UK. B.A. Whitton), pp. 237-264, Blackwell Scientific Publications, Oxford, Lavandier, P. (1974) Ecologie d'un torrent pyrénéen de haute montagne. - I. Caractéristiques physiques. Annales de Limnologie, 10, 173-219. Malard, F., Tockner, K. &C 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. McGregor, G., Petts, G.E., Gurnell, A.M. & Milner, A.M. (1995) Sensitivity of al¬ pine stream ecosystems to climate change and human impacts. Aquatic Con¬ servation: Marine and Freshwater Ecosystems, 5, 233-247. Milner, A.M. (1987) Colonization and ecological development of new streams in Glacier Bay National Park, Alaska. Freshwater Biology, 18, 53-70. Longitudinal and seasonal distribution patterns 49
Milner, A.M. (1994) Colonization and succession of invertebrate communities in a new stream in Glacier Bay National Park, Alaska. Freshwater Biology, 32, 387-400. Milner, A.M. &c Petts, G.E. (1994) Glacial rivers: physical habitat and ecology. Fresh¬ water Biology, 32, 295-307. Oliver, D.R. (1983) The larvae of Diamesinae (Diptera: Chironomidae) of the Holarctic region. Keys and diagnoses. Chironomidae of the Holarctic region. Part 1. Larvae, (ed. T. Wiederholm), pp. 115-131, Entomologica Scandina¬ via Supplement No. 19. Petts, G.E. & Bickerton, M.A. (1994) Influence of water abstraction on the macroinvertebrate community gradient within glacial stream systems: La Borgne d'Arolla, Valais, Switzerland. Freshwater Biology, 32, 375-386. Prowse, T.D. (1994) Environmental significance of ice to streamflow in cold re¬ gions. Freshwater Biology, 32, 241-259. Resh, V.H. &c Rosenberg, D.M. (1989) Spatial-temporal variability and the study of aquatic insects. Canadian Entomologist, 121, 941-963. Robinson, C.T., Uehlinger, U. & Hieber, M. (in press) Spatial and temporal varia¬ tion in macroinvertebrate assemblages of glacial streams in the Swiss Alps. Freshwater Biology. Rosenzweig, M.L. (1995) Species diversity in space and time. Cambridge University Press, Cambridge, UK. Rossaro, B. (1991) Chironomids and water temperature. Aquatic Insects, 13, 87-98. Röthlisberger, H. & Lang, H. (1987) Glacial hydrology. Glacio-fluvialsediment trans¬ fer. - An alpine perspective (eds. A.M. Gurnell & M.J. Clark), pp. 207-284, John Wiley & Sons, Chichester, UK. Saether, O.A. (1968) Chironomids of the Finse area, Norway, with special reference to their distribution in a glacier brook. Archiv für Hydrobiologie, 64, 426- 483. Sartori, M. & Landolt, P. (1999) Atlas de distribution des éphémères de Suisse (In- secta, Ephemeroptera). CSCF/SES, Neuchâtel, CH. Spreafico, M., Leibundgut, C. & Weingartner, R. (1992) Hydrological Atlas of Swit¬ zerland. Hydrological and Geological Swiss Survey, Bern. Steffan, A.W. (1971) Chironomid (Diptera) biocoenoses in Scandinavian glacier brooks. The Canadian Entomologist, 103, 477-486. Steinbock, O. (1934) Die Tierwelt der Gletschergewässer. Jahresbericht des Deutsch- Osterreichischen Alpenvereins, 1934, 263-275. Steinmann, P. (1907) Die Tierwelt der Gletscherbäche. Eine faunistisch-biologische Studie. Annales de Biologie Lacustre, 2, 30-150. Stöckli (1985) Die Rolle der Bakterien bei der Regeneration von Nährstoffen aus Algenexkreten und Autolyseprodukten. Ph. D. Thesis No. 7850, Swiss Fed¬ eral Institute of Technology, Zurich, CH. 50 Longitudinal and seasonal distribution patterns
ter Braak, C.J.F. & Verdonschot, P.F.M. (1995) Canonical correspondence analysis and related multivariate methods in aquatic ecology. Aquatic Sciences, 57, 255-289. Thienemann, A. (1912) Der Bergbach des Sauerland. Internationale Revue der gesamten Hydrobiologie Supplement, 4, 1-25. Thomas, A.G.B. (1975) Ephéméroptères du sud-ouest de la France. I. - Migrations d'imagos a haute altitude. Annales de Limnologie, 11, 47-66. Tockner, K., Malard, F., Burgherr, P., Robinson, CT., Uehlinger, U., Zah, R. &C Ward, J.V. (1997) Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv für Hydrobiologie, 140, 433-463. Uehlinger, U., Zah, R. & Bürgi, H. (1998) The Val Roseg project: temporal and spatial patterns of benthic algae in an alpine stream ecosystem influenced by glacier runoff. Hydrology, water resources and ecology in headwaters (eds. K. Kovar, U. Tappeiner, N.E. Peters & R.G. Craig), pp. 419-424, IAHS Press, Wallingford, UK. Vogler, P. (1965) Beiträge zur Phosphatanalytik in der Limnologie. IL Die Bestimmung des gelösten Orthophosphates. Fortschritte der Wasserchemie und ihrer Grenzgebiete, 2, 109-119. Ward, J.V. (1992) Aquatic insect ecology: 1. biology and habitat. John Wiley &C Sons, New York. Ward, J.V. (1994) Ecology of alpine streams. Freshwater Biology, 32, 277-294. 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. Waringer, J. &c Graf, W. (1997) Atlas der österreichischen Köcherfliegenlarven: unter Einschluss der angrenzenden Gebiete. Facultas Universitätsverlag, Wien. Willassen, E. & Cranston, P. (1986) Afrotropical montane midges (Diptera, Chironomidae, Diamesa). Zoological Journal of the Linnean Society, 87, 91- 123. eer/ Simuliidae fauna in a glacial stream system 53
Chapter 4
Diversity, distribution and seasonality of the Simuliidae fauna in a glacial stream system in the Swiss Alps.
Abstract
Spatial and seasonal patterns in the densities and species richness of Simuliidae (blackflies) were examined in the glacier-fed Roseg River, Switzerland. We also in¬ vestigated how selected environmental factors were related to the observed commu¬ nity structure. Overall, 7 blackfly species were found in streams of this glacial flood plain. Total densities and species richness differed significantly among sampling dates, and densities showed a significant downstream increase. Non-metric multidimen¬ sional scaling supported these findings and suggested that seasonal changes were affected strongly by extent of glacial influence. Furthermore, our results were in¬ dicative that environmental harshness was ameliorated by increased spatial hetero¬ geneity, resulting in enhanced overall ecosystem stability in this glacier-fed stream. Low temperatures and channel instability limited Simuliidae colonization close to the glacier terminus. Suspended particles represented the most important food source, as expected, for simuliids in the Roseg River. However, our results indicated that feeding plasticity may play an important role in the feeding behavior of Simuliidae in this glacial stream because of the general paucity and high temporal variability of food sources.
4.1 Introduction
Simuliidae larvae often constitute an important zoobenthic component in lotie ecosystems. For this ubiquitous and species-rich dipteran family, 1720 morphospecies are known worldwide (Crosskey & Howard, 1997). In Middle Eu¬ rope, between 50 and 60 species have been described based on morphological crite¬ ria. However, several species have been recognized as complexes of isomorphic sib¬ ling species, referred to as cytospecies or cytotypes (Zwick, 1993). Identification of sibling species within previously defined morphological species is possible only by analysis of larval polytene chromosomes (Adler, 1987). The Swiss simuliid fauna has been little studied, except for the comprehensive study by Glatthaar (1978). Cur¬ rently, 34 species are known for Switzerland (Merz et al., 1998). 54 Simuliidae fauna in a glacial stream system
Although occurring in almost every lotie habitat type (Burger, 1987; Crosskey, 1990), larvae of Simuliidae often are found in high densities in lake and reservoir outlets (Wotton, 1987; Malmqvist, 1994). As a consequence, Simuliidae can have important functional implications at the ecosystem level. First, they can provide an important trophic link between suspended particles and predators (Malmqvist, 1994). Second, they can alter the size spectrum of fine particulate organic matter, which can affect the retention of organic material because larger fecal material sediments faster than finer particles (Wotton et al., 1998). Ross & Merritt (1987) reviewed the major environmental factors associated with the distribution and population dynamics of larval blackflies. Simuliid associations have been shown to be influenced by a variety of factors such as distance from lake outlets, stream/river size, food supply, substra¬ tum, current velocity, depth, light, physico-chemical conditions (e.g., temperature and pH) and pollution (e.g., Glatthaar, 1978; Corkum &C Currie, 1987; Malmqvist etal., 1999). Larval Simuliidae are easily distinguished from most other benthic inverte¬ brates by their labral fans that are used to filter food from the water column. For this reason, they are often regarded as classic examples of filter-feeding organisms. How¬ ever, filtering is not the only feeding technique. Scraping and collecting-gathering feeding modes have been observed, and filtering is not possible in some genera (e.g., Twinnia, Gymnopais) that lack cephalic fans (Currie & Craig, 1987). Although pré¬ dation by simuliids is probably an opportunistic behavior, there is evidence that lar¬ vae actively capture their prey (Currie &C Craig, 1987). In arctic and alpine tundra streams a depauperate macroinvertebrate fauna, compared to temperate regions and lower elevations, suggests that extreme environ¬ mental conditions in these areas may exclude many species (Ward, 1994; Füreder, 1999). Kryal (glacier-fed) streams represent the most extreme aquatic habitat type in alpine environments; being characterized by year-round low temperatures, high tur¬ bidity due to large amounts of pulverized material suspended in the water, and large diel flow fluctuations in summer resulting from glacial melting (Milner & Petts, 1994; Ward, 1994). Although the stream fauna is greatly reduced, downstream faunal changes are distinct and predictable in kryal biotopes (Milner & Petts, 1994; Ward, 1994). Based on the available literature, Milner & Petts (1994) proposed a qualita¬ tive model that related zoobenthic gradients in glacial rivers primarily to tempera¬ ture and channel stability; both a function of distance from the glacier margin and time since déglaciation. However, most studies investigating the macroinvertebrate communities of kryal streams focused on the Chironomidae (especially Diamesa spp.), and Ephemeroptera Plecoptera, devoting less attention to other taxonomic groups (e.g., Steffan, 1971; Kownacka & Kownacki, 1972; Kownacki, 1991). Even though the Simuliidae are among the first non-chironomid taxa to appear in glacial streams (Milner & Petts, 1994), few studies have looked in detail at the diversity and distri¬ bution patterns of Simuliidae in alpine and arctic glacier-fed streams (Sommerman et al., 1955; Lavandier, 1979). Simuliidae fauna in a glacial stream system 55
The present study examined the diversity and spatio-temporal distribution patterns of the Simuliidae fauna in a glacier-fed stream in the Swiss Alps. Our objec¬ tive was threefold: (1) identification of the Simuliidae species present in the main glacial channel and to survey distinctively different habitat types for additional spe¬ cies; (2) investigation of longitudinal patterns in blackfly associations during differ¬ ent seasons; and (3) examination of relationships between simuliid distributions and measured environmental factors.
4.2 Materials and methods
Study area
Val Roseg is located in the Bernina Massif of the Swiss Alps (lat 46°29'28" N, long 9°53'57" E) (Fig. 4.1). The 11.3 km long Roseg River drains into the River Inn, a major tributary of the Danube. The primary water source of the Roseg River is meltwater from 2 valley glaciers, namely the Tschierva and Roseg glaciers. Its catch¬ ment area is 66.5 km2, of which 30% is covered by glaciers, 40% is bare rock or glacial till without vegetative cover, and 9% is used for pasture. Bedrock consists of granite and diorite. Elevations range from 1760 (confluence with Bernina River) to 4049 m a.s.l. (Piz Bernina), with treeline at about 2300 m a.s.l. Subalpine coniferous forests in the Val Roseg are restricted to the lower parts of the valley slopes, whereas on the fluvio-glacial gravel of the valley floor no trees occur. Dominant tree species are larch (Larix decidua Mill.), stone pine (Pinus cembra L.) and mugo pine (P. mugo Turra). Shrubs mainly consist of green alder (Alnus viridis (Chaix) DC) and willow (Salix spp.). A more detailed description is given in Tockner et al. (1997).
Study sites
Five distinct reaches characterize the Roseg River: an unstable braided proglacial reach (length 900 m) below Tschierva glacier, a lake outlet stream below proglacial Lake Roseg (length 950 m), a single thread channel incised in glacial till (length 600 m), the main glacial flood plain (length 2750 m) with 6 distinctive chan¬ nel types (Tockner et al., 1997), and a canyon-constrained reach that extends down¬ stream to the river mouth at Pontresina (length 7050 m) (Fig. 4.1). Simuliid assemblages from 16 sampling sites (Fig. 4.1) were investigated at increasing (June), high (August), decreasing (October) and low but constant (No¬ vember) flow conditions in 1997. Ten sites were located along the main glacial chan¬ nel of the Roseg River from the snout terminus of the Tschierva glacier down to the river mouth (sites Ml to M10). To check for other species not colonizing the main channel, additional samples were taken from 3 groundwater channels (Gl, G2, G3), 56 Simuliidae fauna in a glacial stream system
2 intermittently-connected channel habitats (12 and 13), and the outlet stream of Lake Roseg (LI). Groundwater channels are fed by the alluvial aquifer or by a tribu¬ tary aquifer, and have no upstream surface connection to any other channel. Inter¬ mittently-connected channels have a permanent surface connection at their down¬ stream ends, but an intermittent surface connection with the main channel at their upstream ends (Tockner et al, 1997). For comparisons with previous work on the hydro-ecology of this glacial river systems, sites labelled Ml to MIO, Gl to G3,12, 13 and LI in this paper correspond to codes M-20, M-15, M-10, M-1, MIO, M12, M20, M30, M32, M40, G-10, G2, G5, 12, X5 and L-10 used in previous papers (Tockner et ai, 1997; Malard et al., 1999; Ward et al., 1999).
Field techniques and laboratory procedures
Physical factors have been shown to strongly effect blackfly distributions, whereas hydrochemical differences show little influence except under extreme con¬ ditions (Glatthaar, 1978; Ross & Merritt, 1987). Therefore, apart from specific con¬ ductance, only factors describing physical and biotic characteristics were considered. Stream water temperature, turbidity, specific conductance, suspended solids, and the ratio of organic suspended solids (OSS) to inorganic suspended solids (ISS) were measured following the methods outlined in Tockner et al. (1997). Distance from the glacier was calculated using global positioning system data. Water depths and near-bed current velocities (Mini Air 2, Schiltknecht Messtechnik AG, Switzerland) were measured at the same positions where benthic samples were taken. Three replicate Hess samples (area 0.043 m2, 100 urn mesh size) were col¬ lected from haphazardly selected locations at each sampling site and date, preserved in 4% formalin and returned to the laboratory, where macroinvertebrates were sorted and identified. Larvae of Simuliidae were separated and subsequently identified to species using the key by Knoz (1965). After removal of invertebrates, the remaining material from each sample was split into 2 fractions (> 1 mm and < 1 mm), dried at 60°C, weighed, ashed at 550°C and reweighed for determination of coarse and fine benthic organic matter (CBOM and FBOM) as ash-free dry mass.
Statistical analyses
Simuliid associations were analyzed for changes in total densities and species richness using two-way analysis of variance (ANOVA), followed by Tukey's Honest Significant Test (HSD) for multiple pairwise comparisons among sampling dates and sites. Prior to analysis, data were log10 (x+1) transformed to ensure normality and homogeneity of variances (Zar, 1996). 57 Simuliidae fauna in a glacial stream system
1760ma.s.l.^
M1(
constrained reach
flood plain
incised reach
proglacial reach
aà4049ma.s.l.
Fig. 4.1: Location of the Val Roseg catchment in the Bernina Massif of the Swiss Alps. The 5 major reaches and a simplified representation of the channel network of the Roseg River are indicated. Black circles represent sampling sites. 58 Simuliidae fauna in a glacial stream system
Non-metric multidimensional scaling (NMDS, Primer Software Package, Carr, of site 1997) was used to classify sites. The raw matrix average species density per the influence of dominant was initially square-root transformed to downweigh spe¬ cies but still retain the quantitative information. Then, NMDS was run, based on the a two-dimensional ordina¬ Bray-Curtis similarity measure, for all pairs of sites. Last, from NMDS faunal tion plot was produced, where distances represented similarity the stress- between sites. The adequacy of an NMDS representation can be judged by < indicate deviation value which, for a two-dimensional solution, must be 0.25 to from random positions. simuliid asso¬ Partial least-squares regression (PLS) was computed to relate rather ciations to environmental factors using SIMCA-P V.8 software. This method, vari¬ than multiple linear regression, was used because it has no bias if explanatory variables is than ables are strongly correlated, or if the number of explanatory larger the number of objects. PLS reduces the number of variables to one or several latent extracted components (Martens &C Naes, 1989), with the number of components and the statistical significance of the models determined using internal cross-valida¬ tion (Wold, 1978). The performance of a PLS model can be deduced from the statis¬ tics Q2 (predicted Y-variation). The influence of environmental variables on simuliid A VIP associations was judged by VIP (variable importance in the projection) values. < a minor influence. > 1 indicates a strong influence, whereas a VIP 0.8 indicates Two univariate PLS models (PLS 1) were tested, using either total density (all species several combined) or the scores of NMDS axis 1 as the dependent variable versus PLS environmental variables as explanatory variables. Additionally, a multivariate (PLS 2) model was run with individual densities of simuliid species as dependent variables and the same environmental variables. Environmental variables in all mod¬ els were those described above. Prior to analysis, variables were log10 (x+1) trans¬ formed to normalize and stabilize the variance (Zar, 1996).
4.3 Results
Diversity and stream type affinities
A total of 1303 Simuliidae larvae were collected from the main channel. The additional sampling in the other habitat types yielded another 311 specimens. Spe¬ cifically, about 82% of all animals collected came from the main channel, 14.5% from groundwater channels, 3.3% from intermittently-connected channels, and only 0.2% from the proglacial lake outlet stream. Pupae were used only for verification of larval identifications, but not included in data analysis. Overall, 7 species were iden¬ tified: Simulium argenteostriatum Strobl, Simulium argyreatum Meigen, Simulium carthusiense Grenier & Dorier, Simulium cryophilum Rubtsov, Simulium trifasciatum Curtis, Prosimulium latimucro Enderlein, and Prosimulium rufipes Meigen. Most species were found primarily in the main channel, although P. latimucro was more 59 Simuliidae fauna in a glacial stream system
abundant in the intermittently-connected channels than in the main channel and was absent from groundwater channels. The only species not found in the main channel and was S. cryophilum, which mainly occurred in intermittently-connected channels sporadically in groundwater channels. The 2 numerically dominant species were S. argyreatum (32.3% of the assemblages) and P. rufipes (44.9%). Simulium cryophilum and S. trifasciatum, contributing 9.1% and 7.7%, respectively, were intermediate, whereas S. carthusiense (3.3%), P. latimucro (2.4%), and S. argenteostriatum (0.3%)
were least abundant.
Longitudinal distribution patterns in the main channel
No Simuliidae were found on any sampling date at the 3 uppermost sites (Ml to M3) in the proglacial reach; therefore, these sites were excluded from subsequent statistical analyses. Total densities (7 main channel sites combined) were lowest in August (30 ± 31 ind./m2) and considerably higher on the other sampling dates (358 ± 3 05 ind./m2 in June, 451± 775 ind./m2 in October, 605 ± 795 ind./m2 in Novem¬ ber). Two-way ANOVA indicated significant differences in total simuliid densities among sites (F656 = 15.26, p < 0.001), dates (F3J6 = 16.71, p < 0.001), and for the = < interaction term (F1256 3.21, p 0.001) (Table 4.1) . All pairwise comparisons among dates were significant (Tukey's HSD, p < 0.05), except for June versus No¬ vember and October versus November. Pairwise comparisons among sites revealed the following patterns: Site M4 in the incised reach and site M5 in the upper part of the glacial flood plain were not significantly different from each other (Tukey's HSD, p > 0.05). However, both sites were significantly different from all sites further downstream (Tukey's HSD, p < 0.05). No significant differences were found be¬ tween sites in the flood plain and the constrained reach (exceptions: M6 versus M9 and M9 versus M10; Tukey's HSD, p < 0.05 for both). The maximum number of species identified at any site and date never ex¬ ceeded 4 species (Table 4.1). Two-way ANOVA for species richness also revealed significant differences among sites (F656 = 5.48, p < 0.001) and dates (F356 = 8.82, p < 0.001), but not for the interaction term (F1256 = 0.64, p > 0.05). Species rich¬ ness differed significantly among all dates (Tukey's HSD, p < 0.05), whereas no significant differences were found among sites (excluding Ml to M3, which lacked simuliids). Species response patterns differed among sampling dates (Fig. 4.2). In June and August, Simulium spp. contributed 80% of total simuliid densities. In October, this proportion was 72%, but Prosimulium spp. abundance had increased greatly at sites M7 to M9. In November, the reverse was observed, with Prosimulium spp. (78%) being the dominant genus. In contrast to all other species, S. argenteostriatum was found only once, this at site M7 in August. A relatively equal distribution of S. argyreatum at sites M6 to MIO was observed in June, whereas the pattern was skewed more downstream in October and November. Neither S. carthusiense nor P. latimucro 60 Simuliidae fauna in a glacial stream system
exhibited distinct longitudinal patterns. Densities of S. trifasciatum were higher in whereas in October the incised reach and the upper part of the flood plain in June, and November density maxima had shifted downstream. Last, P. rufipes also showed abundant at sites in the constrained a quite equal distribution in June, but was more reach in October and November.
Table 4.1: Summary of Simuliidae densities and species richness at sites M4 to M10. Species richness is expressed as number of species. SD = standard deviation.
Densities Species richness
Site mean ± SD (ind./m2) mean (range)
M4 27 ± 39 1 (0-3)
M5 56 ± 97 1 (0-3)
M6 89 ± 122 2 (1-2)
M7 271 ± 264 2 (2-3)
M8 417 ± 526 2 (1-2) M9 1231 ±1060 3 (3-4)
MIO 434 ± 412 2 (0-4)
A stress value of 0.14 for the NMDS indicated a good representation of the original data in the first 2 dimensions of NMDS space. The resulting ordination plot (Fig. 4.3) mainly separated site M4 in the incised reach and sites M5 and M6 in the floodplain reach from sites M7 to MIO in the constrained reach along NMDS axis 2 (indicated by dashed ellipses in Fig. 4.3). However, 3 sites in each cluster did not follow this pattern (underlined site labels in Fig. 4.3). No clear temporal separation among sampling dates was found, although samples from June and August were primarily positioned towards the positive side of NMDS axis 1, while samples from October and November were on the opposite side. Additionally, sites were most dispersed in two-dimensional space in August and November, whereas sites were more tightly clustered close to the origin in June and October. Sites M4 and M8 in August and site M6 in June were isolated from the other sites, and were the only sites where P. latimucro occurred.
The 2 PLS 1 models were not satisfactory, with only 1 significant component for total densities versus environmental variables (Q2X = 0.408, p < 0.05), and no significant component for scores of NMDS axis 1 versus environmental variables (Q\ = -0.063, p >> 0.05). In contrast, the PLS 2 model with individual species den¬ sities as dependent variables yielded 2 significant components (Q21 = 0.356, p < 0.05; Q22 = 0.305, p < 0.05). Highest VIP values for suspended solids, distance from glacier, and turbidity indicated these environmental variables were most im¬ portant in the model, whereas FBOM, current velocity, specific conductance and water depth contributed little (Table 4.2). 61 Simuliidae fauna in a glacial stream system
June August S. argenteostriatum
S. argyreatum • •#•#99
• » • # S. carthusiense » e @
S. trifasciatum ® ^ 0 .
P. latimucro e
P. rufipes • • £
October November S. argenteostriatum
S. argyreatum
S. carthusiense
S. trifasciatum
P. latimucro
P. rufipes
M4 M5 M6 M7 M8 M9 M10 M4 M5 M6 M7 M8 M9 M10
ind./m2
Fig. 4.2: Distribution patterns of simuliid species along the main glacial channel on the 4 sam¬ pling dates. Only sites M4 to M10 are shown because no Simuliidae were found at Ml to M3. The diameters of circles indicate log10 density classes.
4.4 Discussion
Diversity and stream type affinities
and Glacierrfed streams are believed to provide a poor habitat for Simuliidae, as a consequence, relatively few species colonize them (Burger, 1987; Crosskey, 1990). of all However, 7 blackfly species were found in the Roseg River, representing 20% species currently known for Switzerland. We assume that this number of species in the provides an accurate representation of simuliid diversity glacier-fed Roseg River because sampling was conducted during 4 seasons at 16 sites representing 4 62 Simuliidae fauna in a glacial stream system
# June ,-'' M5A A M6 A A August M6 O October *4 A November a a o â aJ ^5M4 M6 Ml'
M7
x 03 fl u oo "O • A o M10 M8 M8 a A-. M9 s • M4
_ M5 M "~ M8 M10A,»M1° M7 A A M9
o . A M8
1 , , , , -2024
NMDS axis 1
Fig. 4.3: Positions of sampling sites in the first 2 dimensions of non-metric multidimensional scaling (NMDS). Sites with similar blackfly communities are situated close together. Dashed ellipses separate sites from the incised and floodplain reaches from those in the constrained reach. Underlined site labels denote outliers from this pattern. No Simuliidae were found at sites M4 and M5 in October, and site M10 in August.
Table 4.2: Cumulated VIP (Variable importance in the projection, VIP (cum)) values for the first 2 components of partial least squares (PLS) regression reflecting the importance of individual environmental parameters to the model for predicting simuliid associations. A VIP > 1 contrib¬ utes most, whereas a VIP < 0.8 has little influence.
Parameter VIP (cum) Suspended solids 1.20 Distance from glacier 1.19 Turbidity 1.13 CBOM 1.12 Temperature 1.09
OSS : ISS 0.95
FBOM 0.82 Current velocity 0.80 Specific conductance 0.79 Water depth 0.75 Simuliidae fauna in a glacial stream system 63
stream types. However, comparisons are difficult because studies of simuliids in gla¬ cier-fed streams are rare, and simuliid larvae often were not identified to species level in other studies of longitudinal distribution patterns. In the River Estaragne in the French Pyrenees, 8 blackfly species occurred (Lavandier, 1979). Ward (1986) reported 7 species in the middle St. Vrain River (USA) fed by the St. Vrain Glaciers. Similar numbers of species also have been found in arctic tundra streams (not gla¬ cier-fed) in Alaska by Hershey et al. (1995), if species exclusively found in lake out¬ lets or higher order streams are excluded. Nevertheless, in specific habitat types, e.g., springs in the glacial flood plain of the Val Roseg, some additional species such as Twinnia hydroides Novak and Simulium crenobium Knoz are likely to occur. Our results suggest that glacier-fed streams, such as the Roseg River, may provide more adequate habitat for Simuliidae than is generally assumed. Larval simuliids found in the Roseg River belonged to the genera Simulium and Prosimulium, whereas blackflies reported from high latitudes in Alaska, Canada, and Siberia were mostly Gymnopais, a genus whose larvae have lost their labral fans and become adapted to a scraping feeding mode. Some species of the genera Prosimulium and Simulium also have been reported in these systems (Burger, 1987; Crosskey, 1990). Another region where blackfly larvae occur in the ice-melt waters of glaciers is in tropical Africa. In glacial streams originating from the summit gla¬ ciers of Mount Kenya and other high East African mountains of tropical latitudes, blackflies of the S. dentulosum group were found (Crosskey, 1990). Simuliidae were about 4x more abundant in the main glacial channel than in other stream types; this is not surprising considering the habitat preferences of the individual species (compare Glatthaar, 1978). The 3 species P. rufipes, S. carthusiense, and S. argenteostriatum typically inhabit high altitude streams. Similarly, 5. argyreatum and S. trifasciatum prefer cold, fast flowing streams, although they have their distri¬ butional optima at lower elevations. In alpine streams with low gradients, P. latimucro also is commonly found. In contrast, S. cryophilum only occurred in groundwater and intermittently-connected channels, in accordance with its known preference for spring-fed habitats (Glatthaar, 1978). Whereas lowland lake or reservoir outlet streams often exhibit high densities of simuliid larvae (e.g., Wotton, 1987) and other filter feeding organisms, the outlet stream of Lake Roseg did not follow this pattern (see Burgherr & Ward, in press). Although the distribution patterns of individual species generally were in good accordance with known habitat affinities (Glatthaar, 1978), temporal responses of larvae to seasonally changing environmental conditions illustrated the variability in distribution patterns through time. This was reflected in changes in density, species richness and assemblage composition among seasons. Therefore, these results em¬ the phasize importance of examining patterns in more than 1 season before drawing general conclusions. This is especially critical for Simuliidae because they are known to exhibit substantial fluctuations in densities within short periods of time (e.g., Kiel, 1996; Matthäi et al, 1996). 64 Simuliidae fauna in a glacial stream system
Longitudinal distribution patterns in the main channel
Along the longitudinal gradient in the main channel, no Simuliidae were found at sites Ml to M3 in the proglacial reach. This is most likely attributable to low temperature (Tmax = 1.86°C, 248 annual degree days at M3) and unstable stream beds. Temperature is a key factor determining the distribution, diversity, and abun¬ dance patterns of aquatic insects (Vannote &c Sweeney, 1980; Ward &C Stanford, 1982; Ward, 1992), and unstable stream beds prevent attachment of simuliid larvae even when the temperature regime is suitable (Crosskey, 1990). As annual maximum temperature exceeds 2°C and the channel becomes more stable downstream, Simuliidae are often the first non-Chironomidae taxa to appear (Milner & Petts, 1994; Ward, 1994). Furthermore, our results indicate that the glacial floodplain section exerts a stabilizing effect as inferred by the significantly higher abundances at sites M6 to M10 compared to those upstream. Seasonal patterns of total simuliid densities showed significant differences among sampling dates. Lowest densities were recorded during maximum glacial melting and high discharge in August, whereas densities were 12 to 27x higher on the other dates. These findings support the view of Tockner et al. (1997), who suggested that increased spatio-temporal heterogene¬ ity enhances overall ecosystem stability and may play an important role in sustaining populations under the harsh conditions found in glacial streams. NMDS confirmed a similar longitudinal trend in assemblage structure as found for total densities. August samples at the periphery of the ordination were clearly separate, whereas the 3 other dates overlapped. However, the more tight clustering of sites in June and October compared to August and November may reflect seasonal changes in discharge. June and October are transition periods between high and low flows, whereas August (high discharge) and November (low but constant discharge) reflect endpoints when environmental conditions are, respectively, more harsh or favourable. This also is in accordance with patterns of total densities that were inter¬ mediate in June and October, low in August, and high in November. In conclusion, we postulate that environmental harshness is ameliorated by increased spatio-tem¬ poral heterogeneity, resulting in enhanced overall ecosystem stability. PLS regression revealed that environmental variables describing glacial influ¬ ence and potential food supply had important implications for simuliid associations. Temperature was a key factor in separating unsuitable habitats for simuliid larvae in the proglacial reach (Ml to M3) from the remaining sites, but was less influential at sites colonized by simuliids, as indicated by PLS regression. Water depth and current velocity contributed little to the PLS regression model, indicating that both param¬ eters were adequate for simuliid associations (Glatthaar, 1978). Suspended solids, distance from the glacier, and turbidity were the 3 most important terms in the PLS regression model, all representing a measure of glacial influence. In addition, dis¬ tance from the glacier is a function of stream size, which has been shown to be an important variable for simuliid associations (Glatthaar, 1978; Malmqvist etal, 1999), and suspended particles also can play an important role for filter feeding organisms. Simuliidae fauna in a glacial stream system 65
Filtering Simuliidae generally ingest organic and inorganic particles < 200
urn, especially between 0.5 - 50 urn (Ross & Merritt, 1987; Crosskey, 1990). How¬ ever, Wotton (1976) showed experimentally that larvae can filter particles of col¬ loidal size (0.091 urn diameter), and it is likely that even dissolved organic matter (DOM) with diameters < 0.5 urn provide a potential food source (Crosskey, 1990). PLS regression suggested that suspended solids (typically < 0.1 mm in diameter) provide an important food source. Although one would expect the ratio of organic suspended solids (OSS, „seston") to inorganic suspended sediments (ISS, „clay par¬ ticles") to be an important variable for simuliids, it had little influence. This may be partially attributable to the negligible quantities of OSS compared to ISS. The ratio of OSSTSS was low and did not appreciably vary (0.15 ± 0.11). However, the short gut-retention time of simuliid larvae, between 20 to 30 min (Ross & Merritt, 1987), allows them to cope with large amounts of non-digestible particles. Benthic organic matter concentrations in the Roseg River are about 1-2 orders of magni¬ tude less than those reported for other lotie systems (see Tockner et al, 1997), suggesting FBOM and CBOM are a less important food source. High concentrations of suspended solids only occurred during summer high flow, whereas glacial melting is minimal from late autumn to late spring, and as a result, levels of suspended solids are low. Steffan (1971) and Ward (1994) pro¬ posed aeolic input of fine organic particles transported by wind (e.g., pollen) as a potential source of organic material for glacial headwaters. Considering the low aerial inputs of organic material in the Val Roseg (Zah et al, in press), this input of organic material seems rather unlikely as a food source for Simuliidae. Similarly, low concentrations of dissolved organic carbon and particulate organic carbon (Tockner et al, 1997) suggest that these 2 carbon sources also are limited. Late spring and early autumn provide suitable periods for accrual of benthic algae in glacial streams (Uehlinger et al, 1998). Larval simuliids are facultative filter feeders, with scraping or grazing as the second most important feeding strat¬ egy (Currie & Craig, 1987). Therefore, periphyton offers a potential alternative food source. Furthermore blackflies may ingest algal filaments > 1 mm in length (Ross & Merritt, 1987; Crosskey, 1990), which may be important considering the predominance of the filamentous chrysophyte Hydrurus foetidus in this glacier-fed stream (Uehlinger et al, 1998). In addition, simuliid larvae also feed on animal matter, including Chironomidae, Ephemeroptera, Plecoptera and most frequently other blackfly larvae (Ross & Merritt, 1987; Crosskey, 1990). Our results suggest that feeding plasticity may play an important role in the feeding behavior of Simuliidae in this glacial stream, as also has been shown for a limnephilid caddisfly in the same system (Burgherr et al, submitted). Flexibility in functional feeding mode has been reported in other studies of blackfly larvae (e.g., Miller et al, 1998), indicating this may in fact be a common phenomenon. 66 Simuliidae fauna in a glacial stream system
4.5 Acknowledgements
We thank Dr. C.T. Robinson for assistance in the field. Chemical analyses
Ribi. We are indebted to Dr. U. were completed by Mr. R. Uli and Mr. B. especially Uehlinger for providing unpublished temperature data. We also wish to thank our colleagues from the Val Roseg Project for the excellent collaboration. We are grate¬ to the ful to the villages of Pontresina and Samedan for providing access study area, and to Mr. P. Testa and his crew at the Roseg Hotel for their hospitality. We appre¬ ciate the helpful comments of D.B. Arscott and Dr. C.T. Robinson that improved the manuscript.
4.6 References
Adler, P.H. (1987) Ecology of black fly sibling species. Black flies. Ecology, popula¬ tion management, and annotated world list (eds. K.C. Kim & R.W. Merritt), pp. 63-76, The Pennsylvania State University, Pennsylvania. Burger, J.F. (1987) Specialized habitat selection by black flies. Black flies. Ecology, population management, and annotated world list (eds. K.C. Kim & R.W. Merritt), pp. 129-145, The Pennsylvania State University, Pennsylvania. Burgherr, P., Robinson, C.T. & Zah, R. (submitted) Genetic structure and feeding plasticity of Acrophylax zerberus (Trichoptera: Limnephilidae) in a Swiss al¬ pine flood plain. International Review of Hydrobiology. Burgherr, P. & Ward, J.V. (in press) Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 27. Carr, M.R. (1997) PRIMER User Manual (Plymouth Routines in Multivariate Eco¬ logical Research). Plymouth Marine Laboratory, Plymouth, UK. Corkum, L.D. & Currie, D.C. (1987) Distributional patterns of immature Simuliidae (Diptera) in northwestern North America. Freshwater Biology, 17, 201-221. Crosskey, R.W. (1990) The natural history of blackflies. John Wiley & Sons, Chich¬ ester, UK. Crosskey, R.W. & Howard, T.M. (1997) A new taxonomic and geographical inven¬ tory of world blackflies (Diptera: Simuliidae). The Natural History Museum, London. Currie, D.C. & Craig, D.A. (1987) Feeding strategies of larval black flies. Black flies. Ecology, population management, and annotated world list (eds. K.C. Kim &c R.W. Merritt), pp. 155-170, The Pennsylvania State University, Pennsylva¬ nia. Füreder, L. (1999) High alpine streams: cold habitats for insect larvae. Cold adapted organisms. Ecology, physiology, enzymology and molecular biology (eds. R. Margesin & F. Schinner), pp. 181-196, Springer-Verlag, Berlin. 67 Simuliidae fauna in a glacial stream system
Glatthaar, R. (1978) Verbreitung und Ökologie der Kriebelmücken (Diptera, Simuliidae) in der Schweiz. Vierteljahrsschrift der Naturforschenden Gesellschaft in Zürich, 123, 71-124. Hershey, A.E., Merritt, R.W., Miller, M.C. (1995) Insect diversity, life history, and trophic dynamics in arctic streams, with particular emphasis on black flies (Diptera: Simuliidae). Arctic and alpine biodiversity: patterns, causes and eco¬ system consequences (eds. F.S. Chapin III & C. Körner), pp. 283-295, Springer, Berlin.
Kiel, E. (1996) Effects of Aufwuchs on colonization by simuliids (Simuliidae, Diptera). Internationale Revue der gesamten Hydrobiologie, 81, 565-576. Knoz, J. (1965) To identification of Czechoslovakian blackflies (Diptera, Simuliidae). Folia Facultatis Scientiarum Naturalium Universitatis Purkynianae Brunensis, 6, 1-54. Kownacka, M. &c Kownacki, A. (1972) Vertical distribution of zoocenoses in the streams of the Tatra, Caucasus and Balkan Mts. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 18, 742-750. Kownacki, A. (1991) Zonal distribution and classification of the invertebrate com¬ munities in high mountain streams in South Tirol (Italy). Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 24, 2010-2014. Lavandier, P. (1979) Ecologie d'un torrent pyrénéen de haute montagne: VEstaragne. Ph. D. Thesis. Université Paul Sabatier, Toulouse, France. 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. Malmqvist, B. (1994) Preimaginal blackflies (Diptera: Simuliidae) and their preda¬ tors in a central Scandinavian lake outlet stream. Annales Zoologici Fennici, 31, 245-255. Malmqvist, B., Zhang, Y. & Adler, P.H. (1999) Diversity, distribution and larval habitats of North Swedish blackflies (Diptera: Simuliidae). Freshwater Biol¬ ogy, 42, 301-314. Martens, H. & Naes, T. (1989) Multivariate calibration. John Wiley & Sons, Chich¬ ester, UK. Matthäi, CD., Uehlinger, U., Meyer, E. & Frutiger, A. (1996) Recolonization of benthic invertebrates after experimental disturbance in a Swiss préalpine river. Freshwater Biology, 35,233 -24 8.
Merz, B., Bächli, G., Haenni, J.-P. & Gonseth, Y. (eds.) (1998) Diptera - Checklist. Fauna Helvetica 1, CSCF/SES, Neuchâtel, CH. Miller, M.C, Kurzhals, M., Hershey, A.E. & Merritt, R.W. (1998) Feeding behavior of black fly larvae and retention of fine particulate organic matter in a high- gradient blackwater stream. Canadian Journal of Zoology, 76, 228-235. 68 Simuliidae fauna in a glacial stream system
Milner, A.M. & Petts, G.E. (1994) Glacial rivers: physical habitat and ecology. Fresh¬ water Biology, 32, 295-307. Ross, D.H. & Merritt, R.W. (1987) Factors affecting larval black fly distributions and population dynamics. Black flies. Ecology, population management, and annotated world list (eds. K.C. Kim & R.W. Merritt), pp. 90-108, The Penn¬ sylvania State University, Pennsylvania. Sommerman, K.M., Sailer, R.I. & Esselbaugh, CO. (1955) Biology of Alaskan black flies. Ecological Monographs, 25, 345-385. Steffan, A.W. (1971) Chironomid (Diptera) biocoenoses in Scandinavian glacier brooks. The Canadian Entomologist, 103, 477-486. Tockner, K., Malard, F., Burgherr, P., Robinson, CT., Uehlinger, U., Zah, R. & Ward, J.V. (1997) Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv für Hydrobiologie, 140, 433-463. Uehlinger, U., Zah, R. & Bürgi, H. (1998) The Val Roseg project: temporal and spatial patterns of benthic algae in an alpine stream ecosystem influenced by glacier runoff. Hydrology, water resources and ecology in headwaters (eds. K. Kovar, U. Tappeiner, N.E. Peters, R.G. Craig), pp. 419-424, IAHS Press, Wallingford, UK. Vannote, R.L. & Sweeney, B.W. (1980) Geographic analysis of thermal equilibria: a conceptual model for evaluating the effect of natural and modified thermal regimes on aquatic insect communities. The American Naturalist, 115, 667- 695.
Ward, J.V. (1986) Altitudinal zonation in a Rocky Mountain stream. Archiv für Hydrobiologie Supplement, 74, 133-199. Ward, J.V. (1992) Aquatic insect ecology: 1. biology and habitat. John Wiley & Sons, New York. Ward, J.V. (1994) Ecology of alpine streams. Freshwater Biology, 32, 277-294. 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. Ward, J.V. & Stanford, J.A. (1982) Thermal responses in the evolutionary ecology of aquatic insects. Annual Review of Entomology, 27, 97-117. Wold, S. (1978) Cross validatory estimation of the number of components in factor and principal components analysis. Technometrics, 20, 397-406. Wotton, R.S. (1976) Evidence that blackfly larvae can feed on particles of colloidal size. Nature, 261, 697.
Wotton, R.S. (1987) Lake outlet blackflies - the dynamics of filter feeders at very high population densities. Holarctic Ecology, 10, 65-72. Wotton, R.S., Malmqvist, B., Muotka, T. & Larsson, K. (1998) Fecal pellets from dense aggregations of suspension feeders: an example of ecosystem engineer¬ ing in streams. Limnology and Oceanography, 43, 719-725. Simuliidae fauna in a glacial stream system 69
Zah, R., Burgherr, P., Bernasconi, S.M. & Uehlinger, U. (in press) Stable Isotope analysis of macroinvertebrates and their food sources in a glacier stream. Freshwater Biology. Zar, J.H. (1996) Biostatistical Analysis. Prentice-Hall, Upper Saddle River, New Jersey. Zwick, H. (1993) Zum Stand der Taxonomie und Determination einheimischer Kriebelmücken (Diptera: Simuliidae). Beiträge zur Taxonomie, Faunistik und Ökologie der Kriebelmücken in Mitteleuropa (Diptera: Simuliidae) (eds. T. Timm & W. Rühm), pp. 37-53, Westarp Wissenschaften, Magdeburg. ;wa? 1 Ç» * n i,. if*7 ? « fv Seasonal variation in zoobenthos across habitat gradients 71 Chapter 5 Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps) Abstract Flood plains of glacial streams are prominent features of alpine landscapes, yet their ecology has been little studied. The goal of the present study was to assess the effects of habitat stability and heterogeneity of streams in a glacial flood plain in the Swiss Alps on the community structure of benthic macroinvertebrates. Zoobenthic assemblages were investigated during different phases of the annual hydrograph in 3 distinct habitat types (main glacial channel, intermittently-connected channels, and groundwater channels). Communities in the main channel exhibited a high temporal variability in total densities and were dominated by Chironomidae, Ephemeroptera, and Plecoptera (> 95% of total zoobenthos). In contrast, groundwater channels displayed a low temporal variability in community structure, but a high within- channel type variability. Habitat stability, spatio-temporal habitat heterogeneity and connectivity appeared to be key factors in determining the community structure of each channel type as well as individual channels within each type in the glacial flood plain. The complex channel network led to a high spatial heterogeneity, resulting in an enhanced overall ecosystem stability. As a consequence, increased levels of biodiversity at the floodplain scale were possible due to the presence of suitable habitat patches for taxa not able to colonize the main channel, and the provision of numerous réfugia for benthic macroinvertebrates during high discharge in summer when environmental conditions in the main channel were unsuitable for most taxa. 72 Seasonal variation in zoobenthos across habitat gradients 5.1 Introduction Human activities are influencing and altering the functioning of ecosystems at a rate and magnitude much faster than „normal evolution" (Chapin III & Körner, 1994). Although alpine and arctic stream ecosystems have been proposed to be par¬ ticularly sensitive to climate change and human impacts via water resource develop¬ ments (McGregor et al, 1995), relatively little is known of their ecology (Ward, 1994). Whether global warming would lead to glacial recession from increased abla¬ tion or glacial expansion from increased precipitation is still under debate (Knight, 1992). Nevertheless, changes in hydrology, temperature, and channel stability are likely to cause predictable changes in the structure of zoobenthic communities (Milner &c Petts, 1994; McGregor et al, 1995). Headwater streams and rivers are important features of alpine landscapes. Climate, hydrology, geology, and geomorphology define the local character of run¬ ning waters. In alpine regions, 3 major types of streams can be distinguished: kryal stream segments fed by glacial meltwater, krenal segments fed by ground water, and rhithral segments dominated by snowmelt (Ward, 1994). Generally, alpine stream networks form a complex mosaic of kryal, krenal and rhithral stream segments (Milner & Petts, 1994; Ward, 1994; McGregor et al, 1995). Kryal habitats are characterized by low temperatures (Tmax < 4°C), large diel flow fluctuations in summer with peaks in late afternoon from glacial melting, high sediment loads and low channel stability (Milner & Petts, 1994; Ward, 1994). How¬ ever, instream environmental conditions depend on distance from the glacier termi¬ nus, season, and the contribution of non-glacial water sources. For example, down¬ stream of kryal segments, a glacio-rhithral zone fed by a mixture of water sources can extend for a considerable distance, with temporal changes in discharge and tem¬ perature reflecting the relative proportion of glacial influence (Füreder, 1999). Few studies have described the longitudinal distribution patterns of benthic macroinvertebrate communities in glacier-fed streams (e.g., Saether, 1968; Steffan, 1971; Kownacka &C Kownacki, 1972; Kownacki, 1991), and comprehensive year- round studies such as in the River Estaragne in the French Pyrenees (Lavandier & Décamps, 1984) are rare. Milner &c Petts (1994) proposed a qualitative model to predict zoobenthic gradients as a function of water temperature and channel stabil¬ ity. In contrast to this longitudinal perspective, the spatio-temporal dynamics of streams within glacial flood plains have received little attention (Ward et al, 1998). In river floodplain systems at low elevations, the pulsing of discharge (flood pulse) is a key factor structuring biotic assemblages (Junk et al, 1989). However, the ecologi¬ cal effects of these discharge pulses are potentially reduced in glacial floodplain sys¬ tems because of low temperatures, diel flow fluctuations in summer, high amounts of suspended materials, and increased bedload transport during summer high flow. Val Roseg, a glacial flood plain in the Swiss Alps, is characterized by a remarkable degree of aquatic habitat heterogeneity attributable to the shifting dominance of Seasonal variation in zoobenthos across habitat gradients 73 subcatchment water sources and flow paths (Tockner et al, 1997; Malard et al, 1999; Ward et al, 1999). As a consequence, different floodplain channels show different degrees of permanency and connectivity. The primary goal of the present study was to assess the effects of habitat heterogeneity among streams in the Val Roseg glacial flood plain on the community structure and composition of benthic macroinvertebrates. Three channel types, based on their hydrological connectivity and their physico-chemical characteristics, were selected that reflected a gradient of increasing habitat (channel) stability: the main glacial channel, intermittently-connected channels, and groundwater channels. The specific hypotheses addressed were: (1) different channel types are inhabited by dis¬ tinct zoobenthic communities; (2) more stable habitats (i.e., groundwater channels) have the lowest fluctuations in community structure and the highest levels of species diversity; and (3) the mosaic of channel types enhances the biodiversity of glacial flood plains. 5.2 Study area General description Val Roseg is situated in the Bernina Massif of the Swiss Alps (lat 46°29'28" N, long 9°53,57" E). Elevation ranges from 1760 m a.s.l. at the lower end of the catchment (confluence with Bernina River) to 4049 m a.s.l. (Piz Bernina). About 30% of the 66.5 km2 drainage basin is covered by glaciers, 40% by bare rock or glacial till without vegetative cover, and 9% is used for pasture. Treeline is at about 2300 m a.s.l., with subalpine coniferous forests restricted to the lower parts of the valley side-slopes, whereas no trees occur on the fluvio-glacial gravel of the valley floor. Dominant tree species are larch (Larix decidua Mill.), stone pine (Pinus cembra L.) and mugo pine (P. mugo Turra). Bedrock consists of granite and diorite of the austroalpine nappes. The 11.3 km long Roseg River (Ova da Roseg) is primarily fed by 2 valley glaciers: the Tschierva (6.2 km2) and Roseg (8.5 km2) glaciers. It exhibits a seasonal flow regime (summer peak) and distinct daily discharge fluctuations dur¬ ing maximum glacier melting in summer that is typical of glacial-meltwater domi¬ nated rivers (Fig. 5.1). Mean annual discharge of the Roseg River is 2.76 m3/s (1955- 1997, Swiss Hydrological and Geological Society). The system is fed largely by ground water in winter (Malard et al, 1999). Five reaches characterize the Roseg River: an unstable braided reach in the recent proglacial area of the Tschierva glacier (length 900 m), a'lake outlet stream below proglacial Lake Roseg (length 950 m), a single thread channel incised in gla¬ cial till 600 (length m), the main glacial flood plain (length 2750 m), and a canyon- constrained reach that extends downstream to the river mouth at Pontresina (length 7050 m) (Fig. 5.2). The main flood plain contains a variety of aquatic habitats, domi- 74 Seasonal variation in zoobenthos across habitat gradients nated by 6 channel types that can be distinguished according to their hydrology and physico-chemical attributes (Tockner et al, 1997). Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Fig. 5.1: Annual discharge pattern of the Roseg River in 1997. The inset represents diel flow fluctuations in summer. Arrows indicate sampling dates. Study sites The study design combined a habitat gradient with longitudinal position. Of the 6 channel types identified in the flood plain, the 3 that were selected are defined as follows (Tockner et al, 1997): (1) Main channel (M): The primary channel (thalweg) fed mostly by glacial meltwater. (2) Intermittently-connected channels (I) have permanent surface connections at their downstream ends, but an intermittent surface connection with the main channel at their upstream ends. (3) Groundwater channels (G) are fed by the alluvial aquifer or by a tributary (hillslope) aquifer, and have no upstream surface connection. These 3 habitat types cover a gradient of increasing habitat stability from the main channel through intermittently-connected channels to groundwater channels. Habitat (channel) types were selected in the proglacial reach (Ml, Gl; ca. 2090 m a.s.l.), in the upper part of the flood plain (M2,12, G2; ca. 2020 m a.s.l.) and in the lower of part the flood plain (M3,13, G3; ca. 2010 m a.s.l.), thus also incorporating a longitudinal perspective (Fig. 5.2). For comparisons with previous work on the Seasonal variation in zoobenthos across habitat gradients 75 hydro-ecology of this glacial river system, sites labelled Ml, M2, M3,12,13, Gl, G2 and G3 in this paper correspond to codes M-10, MIO, M12,12, X5, G-10, G2 and G5, respectively, used in previous papers (Tockner et al, 1997; Malard et al, 1999; Ward et al, 1999). Because intermittently-connected channels are restricted to the floodplain section, no site of this habitat type was found in the proglacial reach. These 8 sampling sites were visited at bimonthly intervals from April to December 1997 (arrows in Fig. 5.1) to account for seasonal changes in discharge, water sources and flow paths. In April the system is mostly fed by ground water, although some snowmelt occurs. Snowmelt is the main source of water at increasing discharge in June, whereas glacial meltwater dominates at high discharge in August. In October, discharge is decreasing but the contribution of englacial water to surface flow is not significantly lower (Malard et al, 1999). December corresponds to a phase of con¬ stant and low discharge with ground water being the most important source of wa¬ ter. For sites 12 in April and 13 in December, samples could not be taken because surface water was not present. At Site Gl in the proglacial reach, flow in April also was too low to take benthic samples. 5.3 Materials and methods Total channel length of the channel network in the main flood plain was used as an indicator of spatial heterogeneity at the floodplain scale. A detailed description of the mapping procedure and subsequent calculations are given in Malard et al (1999). Only selected parameters, from a variety of measured environmental vari¬ ables, are referred to in this paper to illustrate physico-chemical differences among channel types. Substrate size was assessed in the field by measuring 100 randomly selected stones (b-axis, width). Continuous temperature records (1 hr intervals; StowAway XTI temperature loggers, Onset Corporation, USA) were available for 7 of the 8 sites (U. Uehlinger & F. Malard, unpubl. data). Temperature data from the uppermost site in the constrained reach were used as a substitute for the missing site M3. Specific conductance was determined using a portable conductivity meter (WTW LF323-B, Wissenschaftliche-Technische Werkstätten, Germany). Turbidity (NTUs) was measured with a portable turbidity meter (Cosmos, Fa. Züllig, Switzerland). Specific conductance and turbidity were measured in the morning at each site and sampling date to minimize the effects of daily discharge fluctuations. Near-bed cur¬ rent velocities (Mini Air 2, Schiltknecht Messtechnik AG, Switzerland), water depth and substrate heterogeneity were measured at the same positions where benthic sam¬ ples were taken. Substrate heterogeneity, used as an indicator of microhabitat struc¬ ture, was measured with a device modified from Gore (1978), as described in Tockner & Ward (1999). The device was placed within a standard Hess sampler and the * coefficient of variation (CV % = SD/mean 100) of the length of 20 randomly arranged rods above the surface of the plate, indicated the degree of surface hetero¬ geneity. The environmental data collected for each site were initally summarized as 76 Seasonal variation in zoobenthos across habitat gradients ^Switzerland II ^"^ lake outlet stream N incised reach 200 m ,2 m plain ~~~~~-~—.Jff/ j .1 mT^^i G2 13 M3 AW constrained reach Fig. 5.2: Location of the upper Val Roseg catchment in the Bernina Massif of the Swiss Alps, the 5 showing major reaches. Filled squares represent sampling sites. Inset map shows location of catchment in Switzerland. Seasonal variation in zoobenthos across habitat gradients 77 means, standard deviations (SD) and CVs. Habitat (channel) stability was measured using an approach similar to those described in Death & Winterbourn (1994), and Townsend et al. (1997). A multivariate ordination of 7 variables (substrate size, substrate heterogeneity, range of specific conductance, range of turbidity, range of water depth, maximum near-bed velocity, temperature range) was used to define a single „overall" measure of habitat stability for the 8 sampling sites. Each variable was scaled between 0 and 1 using the equation = - - variable for the ith x (z( min) / (max min), where z is the value of the ;th site, min and max are the minimum and maximum value of the jth variable. The resulting matrix was then analyzed by means of a non-centered principal components analy¬ sis. The first component of such an analysis is always unipolar (Noy-Meir, 1973), and can be used as a multivariate instability score. Three Hess samples (area 0.043 m2, 100 um mesh size) from randomly se¬ lected locations were collected at each site on each date (exceptions mentioned above). Samples were preserved with 4% formalin and returned to the laboratory for process¬ ing. In the laboratory, samples were separated into size fractions greater and less than 1 mm. All macroinvertebrates from both fractions were sorted, identified to the lowest taxonomic unit feasible and enumerated. After removal of invertebrates, the remaining material from the coarse (>1 mm) and fine (<1 mm) fraction of each sample was dried at 60°C, weighed, ashed at 550°C and reweighed for determina¬ tion of coarse and fine benthic organic matter (BOM) as ash-free dry mass (AFDM). Macroinvertebrate data for each site were summarized as means, standard devia¬ tions and CVs for total density, taxon richness and Simpson's index. Simpson's in¬ dex (SI) was expressed as -In SI so that values increase as the number of species increases independently of sample size (Rosenzweig, 1995). Data were log10(x+l) transformed to ensure normality and homogeneity of variances (Zar, 1996), and subjected to two-way analysis of variance (ANOVA) to compare differences among sampling sites and dates. Subsequent multiple pairwise comparisons were made us¬ ing Tukey's Honest Significant Test (HSD). However, the factorial design contained missing treatment combinations, because at sites 12 in April and 13 in December no surface water was present, and flow at site Gl also was too low in April to take benthic samples. In general, there is no one best method to analyze such designs. Furthermore, it is often only possible to estimate main effects (and some interac¬ tions) from complete sub-designs (Milliken & Johnson, 1984; Statsoft, 1995; and references therein). Therefore, as a rather conservative approach, a balanced design was achieved by comparing only samples from June, August and October, instead of assigning „zero values" to missing samples. Finally, the faunistic data set was analyzed by means of centered principal components analysis (PCA) (ADE-4 software, Chessel & Dolédec, 1996). Between- class and within-class PCAs were introduced by Foucart (1978), and Benzécri & Benzécri (1986). Dolédec & Chessel (1989) have demonstrated that these methods are well suited to distinguish between seasonal and spatial effects. First, we focused on the spatial typology of sampling sites using a between-site PCA (Dolédec & Chessel, 78 Seasonal variation in zoobenthos across habitat gradients 1991). The between-site PCA of the initial data table is the PCA of the cumulated (by sites) data table. In order to bring out the variability of each sampling site, the inital data were used in the between-site PCA as supplementary individuals. Second, a within-date PCA (removal of the time effect) was performed to illustrate changes in the spatial distribution of sampling sites among dates. The within-date analysis was an analysis of the residuals. 5.4 Results Habitat characteristics On average, the main channel had the lowest mean temperature and also the lowest temperature range, whereas intermittently-connected channels and groundwater channels had higher mean values and encompassed a wider tempera¬ ture range (Table 5.1). Within each channel type, mean temperature and tempera¬ ture range increased with increasing distance from the glacier terminus. Table 5.1: Water temperatures (CC) at sampling sites based on logger data from 1997. Site Daily mean (range) Ml 0.7 (0.0- 1.9) M2 2.9 (0.4- 4.8) M3 2.6 (0.0- 6.1) 12 4.5 (0.0- 7.1) 13 6.0 (0.2- 11.0) Gl 2.9 (0.3- 6.2) G2 4.0 (0.1- 8.1) G3 5.4 (0.5- 9.4) Environmental variables of sampling sites are summarized in Table 5.2. Aver¬ age substrate size was greatest in the main channel (10.3 - 11.7 cm), intermediate in intermittently-connected channels (8.3 - 9.2 cm), and lowest in groundwater chan¬ nels (5.7 - 7.4 cm). Substrate heterogeneity, water depth, and near-bed velocity gen¬ erally decreased from the main channel towards ground water channels. Benthic organic matter (CBOM and FBOM) showed the reverse trend. Specific conductance and turbidity were inversely related; whereas specific conductance generally increased from the main channel towards groundwater channels, turbidity showed an opposite pattern. In addition to this spatial variation, these 2 parameters showed temporal differences. Turbidity was highest in August due to maximum glacier melting, whereas jeuoseas uoijbuba ui soutuaqooz ssojob leyqeu siuaipejß 5Z JI SBAV |BUITUIUI UT X|JB3 §UIjds pUB •J3JUTA\ UJ 'JSBJJU03 Dljpsds 3DUBJ3npUOD p3t|3B3J UIUUIIXBUI S3n|BA UI "J3JUTM 'J3A3AVOH STqj JBJ0dui3J pU3JJ SBM JSOUI p3>[JBUI UT 3I|1 urera 'puuBijo jnq ss3| jusjBddß m isjBAvpunoiS •spuuBifo 8|qex :g's lejoueg soiisuajoeieqo aq^o ßuj|dujes 'seijs sejnsea|/\| pessejdxe se 'sueeiu -puejs pje suojiejAep (as) pue sjuepjueoo jo uojibuba A3) u\ "(% nv sen|BA peseq uo A|q;uoLu -jene 'seße idooxe sjBJlsqns az!s u) = "(001 kl kl ëo co èo CT) cn nj o oo co o -J o oo en —i CO NO ** ^* °° NJ en oo O O O co co eo -P» -P» eo co eo Nj -p» en CT) —» no co CD o 3 o o a. eo co CO £» O co bo bo -a g oo -p. eo CT) oo t CD -p» NJ eo en O CT) -P» CT) o CT) o CO NJ o en oo o en co en CT) NJ -P* en NJ co en NJ CO CT) —» en NO en CT) NJ CO eo CT) —• en CO CO NJ *» oo en 0O OO O O —I CT) o CO o en Ko ip» *» en cn kl —I cn O CT) CO CO en en ;-> O --. —» CT) —» OO p p CO CO NJ NJ o o CO CO O —» co O NJ -p> CO CD -p» ro -P» o OO eo —» ~-J -P» CT) —i oo CT) CO nj co en en -P» cn CT) C71 NJ '-* N3 CO cn co CO OO NJ O0 S 3 O CO O O O p O no en p p CT) co OO CO CO OO O CT) eo CT) cn CO NJ O OO CT) CT) -P" CO a oo NJ si ^8 lp» O CO p» NJ NO O o —» o Ko o O O co NO O •P* O O o o o O O O Ni o o fief en C0 OO nj -p* CT) CD 3- -P» N CD OO -P» co —» -p» co —» —»00 O o O o o o o <=> o o es o 51 T3 • Ko '—* O O CO popp O LO O CT) O co o o -P» p o nj o o co 3"«i co CD CD 3 co ea oo oo co en ~-J o nj nj CT) OO —» *> en fo o -p» —» en — —' —» =T en -p» *> —> co o *» — CO —» *§: i=> o o o o o o o o o o o o o O O CD o oo o o en 00 O CD o O CT) OO co co co no CO NO NO NO NJ co co CD =r g . Cfl co co o O en cr cd Ko co Ln o bo Ko o —» —» e ni" en CT) ja ct) co -P» o 00 CO co -p» co co un ho ~j CO o en co en en .p» P» =3 sp —_ —. —I —. =! =; -»s c£ ^O ^O ->P ^p —t nP Q) —-. cd a m CD CD CD CD *-> a ^- C3 *- O *~ O "^ O "^ CD O (D < "^- ^ ^ m n S m P s s CD r-h" NJ CO 00 CT ct CT 80 Seasonal variation in zoobenthos across habitat gradients Habitat heterogeneity and stability Total inundated channel length was used as a measure of spatial heterogene¬ ity of channels in the glacial flood plain of Val Roseg. Values increased from April decreased in October (12.4 km) to June (22.5 km), peaked in August (23.9 km), then The observed (20.6 km), and reached a minimum in December (11.1 km). pattern paralleled seasonal variations in discharge and changes in water sources and flow pathways (Malard et al, 1999). from The multivariate instability score revealed that habitat stability increased the main channel sites towards the groundwater sites (Fig. 5.3). Both main channel and intermittently-connected sites exhibited a lower within channel type variability (ranges 0.55, respectively, 0.46), compared to groundwater sites (range 0.81). Over¬ habitat all, the main channel sites were clearly separated, whereas the 2 other types showed some overlap. Site 13 had a stability comparable to an average groundwater channel, whereas site G2 resembled more an intermittently-connected channel with regard to its stability. Multivariate instability score 0 12 3 I • 1 • • 1—• • 1 • •! • 1 1 G3 G1 13 G2 12 M1 M3 M2 Fig. 5.3: Multivariate instability scores of sampling sites, based on axis 1 of a non-centered PCA. Higher values indicate a lower overall stability. Macroinvertebrates Average total densities of macroinvertebrates generally increased from main channel sites towards groundwater sites (Table 5.3). However, similar to the multivariate instability score, total densities at site 13 were closer to the groundwater channel type (i.e., G3), whereas total densities at site 12 were in the range of those found at Ml and M2. Site G3 clearly exhibited the lowest fluctuations (CV = 18%), in accordance with the highest stability reported at this site. Furthermore, total den¬ sities were markedly higher at sites in the lower part of the flood plain (M3,13, and G3), compared to those further upstream. Low coefficients of variation at these 3 sites (18% at G3, 51% at 13, and 58% at M3) were also indicative of more stable communities. Main channel sites had low densities during high discharge in August and high densities in spring and late autumn/winter (Figs. 5.4 and 5.5). In contrast, total densities in groundwater channel sites were not reduced in summer, and inter¬ mittently-connected sites also sustained much higher densities at high flow condi¬ tions than main channel sites (Figs. 5.4 and 5.5). Two-way ANOVA for all sites and Seasonal variation in zoobenthos across habitat gradients 81 3 dates (June, August and October) revealed significant differences among sites = = < and the interaction term (F74g 22.9, p < 0.001), dates (F24g 14.6, p 0.001) = < the trends described above when considering (Fi4 48 14.7, p 0.001), confirming all 5 sampling dates. Subsequent pairwise comparisons (Tukey's HSD, a = 0.05) can be summarized as follows: Site Ml was significantly different from all other sites; sites M2 and M3 were not different from each other, but were different from inter¬ mittently-connected and groundwater channel sites; and sites G3 and 13 showed no significant differences, but site G3 was different from the other groundwater chan¬ nel sites and site 12. Taxon richness and Simpson's index exhibited similar patterns to total den¬ sities. Both measures increased from the main channel towards groundwater channel sites. Similarly, this pattern also was modified by a within channel type increase of the respective values, indicating that site G3 in the lower part of the flood plain was most stable. Two-way ANOVAS also revealed significant effects for sites (F748 = 20.1, p < 0.001 for taxon richness; F74g = 46.0, p < 0.001 for Simpson's index), dates (F24g = 16.7, p < 0.001 for taxon'richness; F24g = 49.8, p < 0.001 for Simpson's index), and for the interaction term (F144g = 2.7, p < 0.01 for taxon richness; F144g = 11.2, p < 0.001 for Simpson's index). Subsequent Tukey's HSD (a = 0.05) snowed similar, but less distinctive patterns than described for total densities. Total densities in the main channel showed an inverse relationship to total channel length (Fig. 5.4); i.e., markedly decreased densities during summer high flow conditions when total channel length peaked. In contrast, intermittently-con¬ nected and groundwater channels exhibited the reversed pattern. Both channel types sustained much higher densities in August than reported for the main channel. The absolute minimum in total density in the intermittently-connected channels in De¬ cember was related to the fact that site 13 lacked surface flow, and 12 was almost dry because it became disconnected from the main channel at its upstream end due to low flow conditions. Temporal faunal changes for individual sites are shown in Fig. 5.5. Ephemeroptera, Plecoptera and Chironomidae dominated the macroinvertebrate assemblages at the main channel sites with combined contributions between 73 to 99% of total zoobenthos. Ephemeroptera and Plecoptera markedly increased in rela¬ tive abundance at main channel sites in the flood plain (M2 and M3) in October and December, whereas Chironomidae declined. Site Ml showed no such pattern with Chironomidae being the dominant taxon at all sampling dates. At site 12, the relative abundance of Chironomidae also decreased in October and December, but primarily at the expense of non-Insecta. In contrast, Oligochaeta, Nematoda and Crustacea were predominant at site 13, except in December when Ephemeroptera dominated. The faunal composition of site G3 was markedly different than the other groundwater sites. Site G3 was dominated by Oligochaeta, Nematoda and Crustacea, whereas G2 resembled 12, and Gl showed an intermediate pattern between the main channel and intermittently-connected channels. 82 Seasonal variation in zoobenthos across habitat gradients Table 5.3: Means, standard deviations (SD), and coefficients of variation (CV in %) for macroinvertebrate densities, taxon richness, and Simpson's index. Site Density Taxon Simpson's (ind./m2) richness index Mean 6966 8 0.33 SD 7177 2 0.23 CV (%) 103 26 72 Mean 6309 14 0.79 SD 5214 3 0.44 CV (%) 83 25 55 Mean 10057 19 0.86 SD 5785 4 0.52 CV (%) 58 23 60 Mean 7227 10 0.55 SD 7289 5 0.60 CV (%) 101 49 110 Mean 16824 21 1.22 SD 8545 2 0.24 CV (%) 51 11 19 Mean 11938 16 0.84 SD 12788 1 0.62 CV (%) 107 8 74 Mean 10214 13 0.56 SD 10027 2 0.40 CV (%) 98 16 72 Mean 21557 20 1.67 SD 3952 2 0.14 CV (%) 18 9 9 83 Seasonal variation in zoobenthos across habitat gradients 25000 20000 15000 1 CO OJ 10000 1 -a 15 - 5000 Fig. 5.4: Relationship between total length of all channels in the flood plain of Val Roseg and total zoobenthic densities in the 3 channel types. Total densities represent averages per channel type for each sampling date. Multivariate relationships among channel types The between-site principal components analysis (PCA) accounted for 58% of the total inertia (Fig. 5.6). Factor Fl was best explained by Ostracoda, Harpacticoida, related Nematoda, Oligochaeta and Nemoura spp., whereas factor F2 was primarily to Rhithrogena spp., Baetis spp. and Protonemura spp. (Fig. 5.6a). Overall, taxa were arranged along a krenal to kryal/glacio-rhithral gradient (arrow in Fig. 5.6a). Sites G3, 13 and G2 were clearly separated, whereas groups formed by Ml and 12, and M2, M3 and Gl, respectively, partially overlapped (Fig. 5.6b). Groundwater sites displayed the least differences in community structure over time as suggested by the low dispersion of sampling dates (short bar lengths) around the centroid of each sampling site. In contrast, the other 2 channel types exhibited greater seasonality as indicated by the longer bar lengths. Furthermore, sites within individual channel types were separated as a func¬ tion of distance from the glacier terminus. This separation was primarily attributed to factor Fl for intermittently-connected channels and groundwater channels, and to factor F2 for the main channel. These findings indicate that longitudinal changes in community structure of the former 2 channel types were attributable to changes in the ratio of Chironomidae to non-insect taxa, and in the main channel to changes in the ratio of Chironomidae to Ephemeroptera and Plecoptera. 84 Seasonal variationin zoobenthos across habitatgradients 1 to S m "S-S „•&. » S2 t £ is S-s gf 1 ZOUO.LUo -S 1Ë ^Ja. IIDDI 3 - O P- Fig.5.5: Temporalchangesin percentcompositionof zoobenthos at individualsamplingsites. Bars are averages of 3 replicatesamplesat a particularsiteand date. Numbers on top of bars representaverage totaldensities(ind./m2).No sampleindicateseitherthat the sitelacked sur¬ face flow at this date (!2and 13),or that flow was too low to take benthic samplesusinga Hess sampler(G1). The within-date PCA accounted for 85% of the total inertia (Fig.5.7).Taxa contributingmost to factor Fl were Ostracoda, Harpacticoida,Nematoda, Oligochaetaand Nemoura spp., whereas Rhithrogenaspp., Baetis spp. Leuctra spp. Seasonal variation in zoobenthos across habitat gradients 85 of taxa and Protonemura spp. were most important for factor F2. The arrangement the between-site PCA. Pro¬ displayed a similar gradient (arrow in Fig. 5.7a) as in nounced differences were found among dates (Fig. 5.7b). Sites G3 and 13 formed a different from G3 in October. group from April to June, whereas 13 was markedly in The remaining sites formed 2 groups in April and June, merged together August, and then returned to a more individualistic pattern. Positions of site G3 were most constant on the factorial plane, whereas the other sites moved more, indicating stronger differences in community structure among dates. For example, 13 was no longer similar to G3 in October, which was attributable to a reduction in crustacean densi¬ ties from about 3800 ind./m2 to about 100 ind./m2 and an increase of Ephemeroptera from about 100 ind./m2 to almost 18,000 ind./m2. Overall, spatial variability de¬ creased with increasing discharge, and increased again with decreasing discharge. Fig. 5.6: Ordination plot of the between-site principal components analysis. Factors Fl and F2 explained 66.5%, and 14.7% of the total variance, (a) F1xF2 correlation plot of the 23 macroinvertebrate taxa used in the analysis. The grey arrow indicates a krenal to kryal/glacio- rhithral gradient, (b) F1xF2 factorial map of sites based on their zoobenthic assemblages. Cir¬ cles represent center of classes for sampling sites. The small black squares indicate the position of sites on each sampling date. Abbreviations of taxa: Cre = Crenobia alpina, Nmt = Nema¬ toda, Oli = Oligochaeta, Hyd = Hydrachnellae, Ost = Ostracoda, Har = Harpacticoida, Ephemeroptera: Bae = Baetis spp., Rhi = Rhithrogena spp., Plecoptera: Rha = Rhabdiopterix alpina, Nem = Nerfioura spp., Pro = Protonemura spp.. Leu = Leuctra spp., Cap = Capnia spp., Iso = Isoperia spp., Per = Périodes intricata, Trichoptera: Lim = Limnephilidae, Diptera: Wie = Wiedemannia spp., Die = Dicranota spp., Rhy = Rhypholophus spp., Sim = Simulium spp., Prs = Prosimulium spp., Chi = Chironomidae, Cer = Ceratopogonidae. 86 Seasonal variation in zoobenthos across habitat gradients F2 'Rhi (a) Bae- [1.-0 i Leu Pro'j f'\so wie! P,n Per / Die \ Rha • Rhy Prs ! f Lim« / 'HVd Chi 1 _ _j Cer 1.0 -Nem A*F® • Sim| • Nmt "Har . „ 'Cre Ost 1-0.5 (b) April June August I -M1 -Ml G2i ! l2' • G3 I ! G2 |2 M2« G3 G.3 | 13 I M3 -M2 13 G2* 13 """giT M3j -M2 | M1- j«G1 ÎM3 October December i i2 ( F2 > G3 I 4 1 -G2 -7lj5F1 |G2 M1 I -5 J | "M1 ]G1 G3 I g"i I -I2 ! -M2 M3-! -M2 1 I3-: -M3 Fig. 5.7: Ordination plot of the within-date principal components analysis. Factors F1 and F2 explained 46.5%, and 18.7% of the total variance, (a) F1xF2 correlation plot of the 23 macroinvertebrate taxa used in the analysis. The grey arrow indicates a krenal to kryal/glacio- rhithral gradient, (b) F1xF2 factorial maps for the 5 sampling dates based on their zoobenthic assemblages. For abbreviations of taxa see legend of Fig. 5.6. Seasonal variation in zoobenthos across habitat gradients 87 5.5 Discussion Downstream faunal changes are expected to be more distinctive in kryal stream segments than in krenal or rhithral biotopes (Ward et al, 199S). However, recent and former glacial activities in glacier-fed stream systems lead to the development of flood plains with different habitat (channel) types with different degrees of perma¬ al. As a nency and connectivity (Tockner et al, 1997; Ward et 1999). consequence, longitudinal distribution patterns of lotie zoobenthos are subject to modifications in floodplain sections. However, few data are available about the ecology of glacial flood plains and virtually nothing is known regarding distribution patterns of benthic macroinvertebrates in such systems. Therefore, investigations in the glacial flood plain of Val Roseg were carried out to enhance our fragmentary knowledge. First, it habitat hetero¬ was a necessary prerequisite to assess temporal changes in the spatial geneity and to determine the stability of channel types and individual sampling sites. Then, specific hypotheses about benthic macroinvertebrate community structure in different channel types with regard to habitat stability and spatial heterogeneity were investigated. Connectivity and water sources Turbidity and specific conductance were used as indicators of hydrological connectivity and water source (Malard et al, 1999). These 2 parameters showed an inverse relationship with respect to both space and time. Considering the temporal aspect, turbidity peaked during high discharge in summer when glacial meltwater was the main source of water and decreased with receding discharge towards winter. In contrast, specific conductance was highest in winter, indicating ground water was the dominant source of water at this period of the year. Additionally, these 2 param¬ eters also displayed a spatial pattern. Turbidity decreased and specific conductance increased from main channel sites to groundwater sites. Intermittently-connected channels exhibited conditions similar to the main channel at elevated flow condi¬ tions, when their upstream ends were connected to the main channel. In contrast, at low flow conditions they were disconnected at their upstream ends and primarily fed by ground water. Groundwater channels never received glacial meltwater, even at high discharge, because they do not have an upstream connection with the main channel. Overall, the glacial flood plain of Val Roseg shifts from a glacial meltwater dominated system in summer to a groundwater controlled system in winter (Tockner et al, 1997; Ward et al, 1999; Malard et al, 1999). As a result, favourable environ¬ mental conditions in the main channel only occur in spring and late autumn/early winter. Burgherr & Ward (in press) demonstrated that macroinvertebrate densities and taxon richness also reach their maxima during these periods. In contrast, groundwater channels are not markedly affected by these changes in glacial influ¬ ence, therefore providing more constant and less harsh environments. 88 Seasonal variation in zoobenthos across habitat gradients Spatial habitat heterogeneity and habitat stability Seasonal changes in discharge control the expansion and contraction cycle of the channel network in the glacial flood plain of Val Roseg (Malard et al, 1999). This affects the seasonal fragmentation and reconnection of habitats, a poorly stud¬ ied attribute of stream ecosystems (Stanley et al, 1997). Therefore, total channel length of the channel network in the glacial flood plain was used as an indicator of spatial habitat heterogeneity. Highest spatial heterogeneity was observed during high discharge in summer, whereas it was reduced at low flows. However, spatial hetero¬ geneity also can decrease at extremely high discharge because channels merge and become more similar in their physico-chemical conditions (Tockner et al 1997). Consequently, spatial habitat heterogeneity is expected to ameliorate the negative effects of diel summer flow fluctuations by providing important réfugia for benthic macroinvertebrates (Tockner et al, 1997; Ward et al, 1998). However, the flood plain shifted from a glacial-meltwater dominated system in summer to a groundwater- dominated system in autumn and winter (Malard et al, 1999; Ward et al 1999). Concomitant with a decrease in total channel length and more homogenous physico- chemical conditions, the main channel progressively disappeared into the alluvium at the upper end of the flood plain and surface flow was almost totally restricted to the downstream end of the flood plain (below site M2) (Ward et al, 1999). The permanency in flow, groundwater-like environmental conditions, and accrual of benthic algae in this section of the main channel during winter provided a potential réfugia for species having an univoltine winter life cycle, such as B. alpinus. The fact B. that alpinus density peaked in late winter (25,790 ind./m2 in February 97) sup¬ ports this notion and is in accordance with the findings in other studies (e.g., Brittain, 1982; Kukula, 1997). Habitat stability of different channel types and individual channels within a particular type was assessed by means of a multivariate instability score. Generally, a reduction in habitat stability from the main channel towards groundwater channels was observed. Additionally, groundwater channel sites examined in this study were individually stable, but collectively encompassed a greater spatial heterogeneity than the less stable main channel and intermittently-connected channel sites. This is in accordance with Tockner et al (1997) and Ward et al (1998). Macroinvertebrate community patterns The hypotheses that distinct channel types have distinct zoobenthic commu¬ nities, and that more stable habitats have the lowest fluctuations in community struc¬ ture and the highest levels of species diversity (i.e., taxon richness) generally applied, some site although specific deviations were found. On average, total densities, taxon richness, and Simpson's index increased from main channel through intermittently- connected channels to groundwater channel sites. For example, Craig & McCart Seasonal variation in zoobenthos across habitat gradients 89 (1975), Kownacki (1991) and Füreder et al (1998) reported higher invertebrate densities and species numbers in streams dominated by ground water. Therefore, Craig & McCart (1975) described spring-fed rivers as the „green oasis in the polar environment". However, among channel patterns in the Val Roseg were modified by temporal aspects and within channel differences. Lotie zoobenthos of main channel sites was strongly influenced by temporal shifts in glacial influence (i.e., changes in water sources and flow paths), as well as by a decrease in glacial influence with increasing distance from the glacier, which is in accordance with the findings of Burgherr & Ward (in press). For example, total densities were markedly reduced during summer high flows and maximal glacial melting. Furthermore, Chironomidae constituted the dominant faunal group, ex¬ cept in late autumn/early winter and to a lesser extent in spring when Ephemeroptera and Plecoptera also made substantial contributions. In contrast to the conceptual model of Milner & Petts (1994), both groups occurred at lower maximum water temperatures than the 4°C threshold assumed in the model. By moving within the substrate, benthic invertebrates can markedly alter the thermal regime to which they are exposed because of the steep temperature gradient within the alluvium of Val Roseg (Malard et al, in press, a). Therefore, the presence of Ephemeroptera and Plecoptera in the metakryal zone close to the glacier may be linked to the particular thermal regime of the hyporheic water. In summary, the generally low habitat stabil¬ ity of the main channel results in a characteristic array of species and also restricts greater densities with more taxa to limited periods of favourable environmental con¬ ditions. Additionally, differences along spatial and temporal gradients were related to habitat stability (i.e., channel stability sensu Milner & Petts (1994)). Stability was closely related to the hydrograph of the Roseg River. Moving bed sediments reduced channel stability from June to August. During the remaining time of the year dis¬ charge did not exceed the threshold of 8 m3/s, indicating channel stability was higher (Burgherr & Ward, in press). The 2 intermittently-connected channels exhibited distinctively different pat¬ terns. Site 12 was more similar to the main channel, whereas 13 was more related to G3, as can be seen from biotic metrics such as total densities, taxon richness, and Simpson's index, as well as from the results of the PCAs. The temporary surface connection (June to October) of site 12 at its upstream end to the main channel clearly affected benthic community structure. Although Chironomidae dominated during high flow conditions, as observed for the main channel, total densities re¬ mained much higher than in the main channel (2 to 3Ox). From October to Decem¬ ber total densities showed a drastic decline to less than 200 ind./m2, which is most likely attributable to the lost connection to the main channel. As a consequence the channel became partly dry and site 12 was isolated from the remaining channel net¬ work. In contrast, site 13 was only marginally influenced by the main channel be¬ cause of upwelling ground water throughout the year, except in September and Oc¬ tober (F. Malard, pers. communication), that was reflected in seasonal changes in macroinvertebrate structure. For example, Crustacea, Oligochaeta and Nematoda 90 Seasonal variation in zoobenthos across habitat gradients in were important components from April to August as normally expected groundwater channels. A totally different composition was found in October with Chironomidae Ephemeroptera, and to a lesser extent, Plecoptera and dominating. the This change is most likely attributable to the change in the water source. Overall, degree of connectivity of intermittently-connected channels to the main channel de¬ termines the stability of individual channels, hence affecting distribution patterns of macroinvertebrates. The high within channel type variability in stability observed for groundwater channels also was reflected in benthic community structure. Both, the between-site and the within-date PCA clearly demonstrated that the most stable habitat type also exhibited the lowest fluctuations in community structure. Additionally, average taxon richness combined for all 3 groundwater sites was highest. Furthermore, PCAs and the relative composition of zoobenthos revealed a high within channel type variabil¬ ity as reported above for habitat stability. Site G3 was dominated by permanent krenal habi¬ aquatic taxa, such as Crustacea, Oligochaeta and Nematoda, that prefer tats. Glazier (1991) discussed potential reasons (e.g., flow and thermal constancy, absence of thermal cues, etc.) why a non-emergent life style in spring-fed habitats constitu¬ may be advantageous. In contrast, insects were by far the most important ents of lotie zoobenthos at sites Gl and G2. This discrepancy may be partly attribut¬ able to reduced stability and lower temperatures in the proglacial reach, but seems less likely in the upper part of the flood plain. A major reason why permanent aquatic taxa were almost absent in surface zoobenthos at site G2 could be due to the reduced ecological connectivity in the upper part of the flood plain that resulted in a lack of surface water - groundwater interactions, strongly affecting the occurrence of taxa related to the hyporheos (Malard et al, in press, b). Additionally, site G2 changed its nature from a true groundwater channel to an intermittently-connected channel due to a newly established connection to the main channel on 15 August 1997 (F. Malard, pers. communication). In conclusion, the high within channel variability in benthic assemblages was not exclusively attributable to habitat stability. Spatial heterogeneity and biodiversity The present study yielded some evidence for the hypothesis that the mosaic of channel types enhances biodiversity (i.e., taxon richness) of this glacial flood plain. On average, Simpson's index was higher in intermittently-connected channels and in groundwater channels, suggesting these habitats substantially contribute to an over¬ all increase in biodiversity at the floodplain scale. However, at least 2 key factors are responsible for sustaining a greater taxon richness at the level of the whole flood plain than in particular channel types or individual channels. First, a variety of taxa had their distributional optima in intermittently-con¬ nected and groundwater channels or were totally absent from the main channel. For example, Baetis rhodani and Simulium cryophilum did not occur in the main chan- Seasonal variation in zoobenthos across habitat gradients 91 also showed nel. Oligochaeta, as a representative of the permanent aquatic taxa, pronounced differences between the main channel and groundwater channels; not only in densities, but also in species composition, since several species were not found in the main channel, but did occur in groundwater-fed habitats (compare Klein & Tockner, in press; Malard et al, in press, b). Second, the high spatial habitat heterogeneity provides an increased number of potential réfugia for macroinvertebrates (Sedell et al, 1990; Scarsbrook & Townsend, 1993; Townsend & Hildrew, 1994), thus is expected to ameliorate the negative effects of high discharge in summer associated with diel flow fluctuations and increased channel instability in the Val Roseg flood plain. The relationship be¬ tween total channel length and total macroinvertebrate densities for individual chan¬ nel types suggested that groundwater channels and intermittently-conncected chan¬ nels may serve as important réfugia during high discharge in summer, thus also pro¬ viding a potential recolonization pathway (besides the hyporheos and ovipositing by winged adults) for the main channel where environmental conditions are too ex¬ treme during high discharge in summer. Our results were consistent with the hy¬ pothesis that stream invertebrates accumulate in réfugia during high flow conditions (e.g., Lancaster, 2000). Conclusions In conclusion, habitat stability in concert with spatio-temporal habitat het¬ erogeneity and connectivity appeared to be the primary structuring agents influenc¬ ing the habitat templet for the benthic communities of different channel types, as well as individual channels, in the glacial flood plain of Val Roseg. Furthermore, alpine stream ecosystems are extreme environments located on the declining limb of a harshness-diversity curve (Tockner et al, 1997). Consequently, habitats with the highest stability (e.g., groundwater channels) exhibited low variability in community structure, and high taxon richness. Finally, our investigations demonstrated that the complex channel network patterns in the Val Roseg flood plain enhance overall ecosystem stability and biodiversity by providing numerous important réfugia for benthic macroinvertebrates, which is not the case for glacial streams lacking floodplain sections. 92 Seasonal variation in zoobenthos across habitat gradients 5.6 Acknowledgements We are especially indebted to Drs. U. Uehlinger and F. Malard for providing unpublished temperature data, and Mr. R. Uli and Mr. B. Ribi for analysing chemical samples. We also wish to thank our colleagues from the Val Roseg Project for the excellent collaboration. We are grateful to the villages of Pontresina and Samedan for providing access to the study area, and to Mr. P. Testa and his crew at the Roseg Hotel for their hospitality. Special thanks to Dr. J. Brittain for his valuable critiques that improved the manuscript. 5.7 References Benzécri, J.P. & Benzécri, F. (1986) Pratique de l'analyse de données. Bordas, Paris. Brittain, J.E. (1982) Biology of mayflies. Annual Review of Entomology, 27, 119- 147. Burgherr, P. & Ward, J.V. (in press) Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps). Fresh¬ water Biology. F.S. & C. Chapin III, Körner, (1994) Arctic and alpine biodiversity: patterns, causes, and ecosystem consequences. Trends in Ecology and Evolution, 9, 45-47. Chessel, D. & Dolédec, S. (1996) ADE Version 4: HyperCard Stacks and Quick Basic Microsoft programme library for the analysis of environmental data. User's Manual. ESA CNRS 5023, Ecologie des Eaux Douces et des Grands Fleuves, Université Lyon 1, France. P.C. Craig, & McCart, P.J. (1975) Classification of stream types in Beaufort Sea drainages between Prudhoe Bay, Alaska, and the Mackenzie Delta, N.W.T., Canada. Arctic and Alpine Research, 7, 183-198. Death, R.G. & Winterbourn, M.J. (1994) Environmental stability and community persistence: a multivariate perspective. Journal of the North American Benthological Society, 13, 125-139. S. & Dolédec, Chessel, D. (1989) Rhythmes saisonniers et composantes stationelles en milieu aquatique IL Prise en compte et élimination d'effects dans un tab¬ leau faunistique. Acta Oecologia, 10, 207-232. Dolédec, S. &c Chessel, D. (1991) Recent developments in linear ordination methods for environmental sciences. Advances in Ecology, India, 1, 133-155. T. Sur les suites de Foucart, (1978) tableaux de contingence indexés par le temps. Statistique et Analyse des données, 2, 67-84. R. &c Füreder, L., Schütz, C, Burger, Wallinger, M. (1998) High alpine streams as models for ecological gradients. Hydrology, water resources and ecology in headwaters K. U. (eds. Kovar, Tappeiner, N.E. Peters & R.G. Craig), pp. 387-394, IAHS Press, Wallingford, UK. Seasonal variation in zoobenthos across habitat gradients 93 Füreder, L. (1999) High alpine streams: cold habitats for insect larvae. Cold adapted organisms. Ecology, physiology, enzymology and molecular biology (eds. R. Margesin & F. Schinner), pp. 181-196, Springer-Verlag, Berlin. Glazier, D. S. (1991) The fauna of North American temperate cold springs: patterns and hypotheses. Freshwater Biology, 26, 527-542. Gore, J.A. (1978) A technique for predicting in-stream flow requirements of benthic macroinvertebrates. Freshwater Biology, 8, 141-151. Junk, W.J., Bayley, P.B. & Sparks, R.E. (1989) The flood-pulse concept in river floodplain systems. Canadian Special Publication of Fisheries and Aquatic Sciences, 106, 110-127. Klein, B. & Tockner, K. (in press) Biodiversity in springbrooks of a glacial flood plain (Val Roseg, Switzerland). Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie. Knight, P.G. (1992) Glaciers. Progress in Physical Geography, 16, 85-89. Kownacka, M. & Kownacki, A. (1972) Vertical distribution of zoocenoses in the streams of the Tatra, Caucasus and Balkan Mts. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 18, 742-750. Kownacki, A. (1991) Zonal distribution and classification of the invertebrate com¬ munities in high mountain streams in South Tirol (Italy). Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 24, 2010-2014. Kukula, K. (1997) The life cycles of three species of Ephemeroptera in two streams in Poland. Hydrobiologia, 353, 193-198. Lancaster, J. (2000) Geometric scaling of microhabitat patches and their efficacy as réfugia during disturbance. Journal ofAnimal Ecology, 69, 442-457. Lavandier, P. & Décamps, H. (1984) Estaragne. Ecology of European Rivers (ed. B.A. Whitton), pp. 237-264, Blackwell Scientific Publications, Oxford, UK. 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. (in press, a) Thermal heteroge¬ neity in the hyporheic zone of a glacial floodplain. Canadian Journal of Fish¬ eries and Aquatic Sciences. Malard, F., Lafont, M., Burgherr, P. & Ward, J.V. (in press, b) Hyporheic versus benthic longitudinal patterns of Oligochaetes in a glacial river. Arctic, Ant¬ arctic and Alpine Research. McGregor, G., Petts, G.E., Gurnell, A.M. &c Milner, A.M. (1995) Sensitivity of al¬ pine stream ecosystems to climate change and human impacts. Aquatic Con¬ servation: Marine and Freshwater Ecosystems, 5, 233-247. Milliken, G.A. & Johnson, D.E. (1984) Analysis of'messy data. Volume I: Designed experiments. Van Nostrand Rheinhold Company, New York. 94 Seasonal variation in zoobenthos across habitat gradients Milner, A.M. & Petts, G.E. (1994) Glacial rivers: physical habitat and ecology. Fresh¬ water Biology, 32, 295-307. Noy-Meir, I. (1973) Data transformations in ecological ordination. Some advan¬ tages of non-centering. Journal of Ecology, 61, 329-341. Rosenzweig, M.L. (1995) Species diversity in space and time. Cambridge University Press, Cambridge, UK. Saether, O.A. (1968) Chironomids of the Finse area, Norway, with special reference to their distribution in a glacier brook. Archiv für Hydrobiologie, 64, 426- 483. Scarsbrook, M.R. & Townsend, C.R. (1993) Stream community structure in rela¬ tion to spatial and temporal variation: a habitat templet study of two con¬ trasting New Zealand streams. Freshwater Biology, 29, 395-410. Sedeil, J.R., Reeves, G.H., Hauer, F.R., Stanford, J.A. & Hawkins, C.P. (1990) Role of réfugia in recovery from disturbances: modern fragmented and discon¬ nected river systems. Environmental Management, 14, 711-724. Stanley, E.H., Fisher, S.G. & Grimm, N.B. (1997) Ecosystem expansion and con¬ traction in streams. BioScience, 47, 427-435. Statsoft (1995) Statistica for Windows. Statsoft Inc., Tulsa. Steffan, A.W. (1971) Chironomid (Diptera) biocoenoses in Scandinavian glacier brooks. The Canadian Entomologist, 103, 477-486. Tockner, K., Malard, F., Burgherr, P., Robinson, CT., Uehlinger, U., Zah, R. & Ward, J.V. (1997) Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv für Hydrobiologie, 140, 433-463. Tockner, K. &c Ward, J.V. (1999) Biodiversity along riparian corridors. Archiv für Hydrobiologie, 115, 293-310. Townsend, C.R. &C Hildrew, A.G. (1994) Species traits in relation to a habitat tem¬ plate for river systems. Freshwater Biology, 31, 265-275. Townsend, C.R., Dolédec, S. & Scarsbrook, M.R. (1997) Species traits in relation to temporal and spatial heterogeneity in streams: a test of habitat templet theory. Freshwater Biology, 37, 367-387. Ward, J.V. (1994) Ecology of alpine streams. Freshwater Biology, 32, 277-294. Ward, J.V., Burgherr, P., Gessner, M.O., Malard, F., Robinson, CT., Tockner, K., Uehlinger, U. &c Zah, R. (1998) The Val Roseg Project: habitat heterogeneity and connectivity gradients in a glacial flood-plain system. Hydrology, water resources and ecology in headwaters (eds. K. Kovar, U. Tappeiner, N.E. Pe¬ ters & R.G. Craig), pp. 425-432, IAHS Press, Wallingford, UK. 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. Zar, J.H. (1996) Biostatistical Analysis. Prentice-Hall, Upper Saddle River, New Jersey. mfi Blank leaf Genetic structure and feeding plasticity of A. zerberus 97 Chapter 6 Genetic structure and feeding plasticity of Acrophylax zerberus (Trichoptera: ümnephilidae) in a Swiss alpine flood plain Abstract Alpine glacial flood plains comprise a variety of stream types that have dis¬ tinct habitat conditions and food resources. These different stream types form a complex mosaic in a harsh landscape that constrains dispersal and influences other life history traits of aquatic macroinvertebrates. We examined the genetic structure, energy flow, and food preferences of a freshwater caddisfly (Limnephilidae, Acrophylax zerberus Brauer) from 3 spatially-separated stream types (lake outlet, proglacial spring, and a side-slope springbrook) in a glacial flood plain in the Swiss Alps. Allozyme electrophoresis indicated high levels of heterozygosity for each A. zerberus population, conforming with population genetic models for species living in spatio-temporally heterogenous environments. Allozyme results suggest that there was little migration of A zerberus populations among sites in the current generation. Dispersal by flying adults appears to be the most likely mechanism in this glacial flood plain. Signatures from multiple stable isotope analysis (13C, 15N) indicated ani¬ mals fed on the predominant food resources at a site. Results of food preference experiments also were indicative of a generalist feeding strategy by A zerberus. Feed¬ ing plasticity plays an important role in the successful dispersal and persistence of this patchily distributed species among different stream types in a glacial floodplain mosaic. 98 Genetic structure and feeding plasticity of A. zerberus 6.1 Introduction Southwood (1977; 1988) posited that habitat provides the templet from which evolution forges characteristic life history traits. Based on Southwood's theoretical several work, templets for running-water habitats have been developed to link or¬ ganism response to habitat heterogeneity, stability, and productivity (Stanford & Ward, 1983; Hildrew & Townsend, 1987; Poff & Ward, 1990; Townsend & Hildrew, 1994). Multi-scale models that integrate habitat filters and species traits have been proposed to enhance our understanding and predictive power of popula¬ tion distribution and community organisation in freshwater environments (e.g., Tonn et al, 1990; Poff, 1997). Habitat heterogeneity, in concert with connectivity, has been shown to effect a of variety biological attributes of a population, such as popu¬ lation size (Lande & Barrowclough, 1987), migration rates (Stamps et al, 1987), and extinction rates (Gilpin &c Hanski, 1991). For example, habitat patchiness via often fragmentation coincides with habitat isolation and loss, and thus could poten¬ tially affect the spatial dynamics and persistence of populations (reviewed by Saunders etal, 1991). Patchiness, at most spatial scales, is a characteristic feature of the distribution of most species (Levin, 1994; Wiens, 1997). Hanski (1999) suggests that spatial patchiness is not synonymous with metapopulation structure because the former refers to rather patterns than processes; e.g., similar spatial patterns can be caused by different processes. Furthermore, the metapopulation approach emphasizes discrete habitat patches that are large enough to support local breeding populations, whereas also can occur spatial patchiness within local populations of a metapopulation (Hanski, 1999). Nevertheless, spatial patchiness and the metapopulation approach comple¬ ment rather than conflict with each other (Kareiva, 1990). Streams of alpine glacial flood plains are considered harsh environments; the habitat of physical templet which includes extreme thermal conditions, severe habi¬ tat expansion/contraction cycles, and sparse food resources (Milner & Petts, 1994; Malard et In Ward, 1994; al, 1999). addition, alpine stream networks are frag¬ mented because landscapes they comprise a complex mosaic of stream types having distinct habitat conditions (Ward, 1994; McGregor et al, 1995; Tockner et al, 1997). The habitat mosaic of alpine floodplain landscapes strongly influences the distribution and dispersal of lotie invertebrates. Stream types with suitable condi¬ tions for a taxon are often particular patchily arranged in space or temporally lim¬ ited in availability (sensu Hanski, 1998). Consequently, macroinvertebrate assem¬ within a blages given stream type usually are quite distinct with populations exhibit¬ ing patchy distribution patterns (P. Burgherr, unpublished data). Even in heterogenous landscapes, however, few populations are truly iso¬ lated. Most species are spatially partitioned as local populations that are connected For by dispersal. many species, populations also are distributed among various habi¬ tats of different quality (Pulliam & Dunning, 1997; Diffendorfer, 1998). Although water invertebrates running generally are considered to show a high degree of popu- Genetic structure and feeding plasticity of A. zerberus 99 lation connectivity (Mackay, 1992), the unidirectional flow of water and the hierar¬ chical structure of stream channel networks limit within-stream movement (Meffe & Vrijenhoek, 1988). Further, the presence of natural (e.g., cascades, waterfalls) and anthropogenic (e.g., weirs, impoundments) barriers effectively reduce dispersal of many aquatic organisms. Dispersal also can be restricted by life history attributes, and behavioral or morphological traits such as through differences in mating strategies, physical habi¬ tat preferences, or specifity in diet. Stream insects disperse by adult flight, down¬ stream drift, and swimming or crawling of larvae (Mackay, 1992). While swimming and crawling typically are restricted to small spatial scales for most taxa, adult flight and larval drift are major mechanisms of dispersal at larger spatial scales (Minshall & Petersen, 1985; Mackay, 1992). In contrast to terrestrial species, however, the adults of stream insects are typically short-lived with limited powers of flight (Ward, 1992). Direct measurements of dispersal distances are often difficult and have been reported in only a few studies (e.g., Hershey et al, 1993). However, assessing the level of genetic differentiation among populations provides a powerful alternative approach for measuring dispersal. For example, population genetic theory predicts that high dispersal reduces the genetic differentiation among populations (Hard 1980). Conversely, natural selection and random genetic drift are the driving forces for genetic differentiation among populations under limited dispersal or gene flow (Slatkin, 1985). In the present study we examined the genetic population structure and feed¬ ing behavior of an aquatic insect to gain an understanding of how this species persists in an otherwise harsh and fragmented alpine glacial flood plain. We used a limnephilid caddisfly (Acrophylax zerberus Brauer) that inhabits specific stream types in alpine glacial flood plains. We focused on populations from 3 spatially distinct lotie habi¬ tats in the flood plain in which dispersal among habitat patches is only through adult flight: a glacial lake outlet, a proglacial spring, and a side-slope tributary. A. zerberus provided an optimal study organism because it is a common, univoltine limnephilid, but is constrained to specific stream types in alpine floodplain systems. Consequently, we posited that successful dispersal, colonization, and persistence of A zerberus among these different lotie habitats would require some degree of life history plasticity (Stearns, 1992). Specifically, 3 inter-related topics pertaining to genetic structure and feeding behavior were investigated. First, we examined whether the populations at the 3 sites differed in their genetic structure using allozyme electrophoresis. Ge¬ netic allowed analysis inference on the extent of dispersal among sites. Second, we used multiple stable isotope analyses as a tracer to determine potential sources of energy sustaining A zerberus at a particular site. Last, we conducted food preference experiments in the laboratory to reveal if A. zerberus exhibited site specific feeding behaviors that reflected the predominant food resources at each site. 700 Genetic structure and feeding plasticity of A. zerberus 6.2 Materials and methods Study organism The freshwater caddisfly A zerberus is restricted geographically to the Alps, Pyrenees and Carpathians (Botosaneanu & Malicky, 1978). It mostly inhabits run¬ ning water habitats, but also has been reported from high alpine lakes (Waringer &C Graf, 1995). A. zerberus primarily feeds by shredding coarse particulate organic mat¬ ter, and to a lesser extent by grazing on periphyton and associated material, or prey¬ ing on other invertebrates (Moog, 1995). Most populations are univoltine. Larval development occurs from about July to October (Szczesny, 1986). Overwintering takes place in the pupal stage (Waringer & Graf, 1995; P. Burgherr, personal obser¬ vation), and adults emerge in March/April (Malicky, 1980). Winged adults typically are active from March to July (Tobias & Tobias, 1981; C.T. Robinson &c P. Burgherr, personal observation). Early emerging adults often have been observed running on snow, and brachyptery (reduced wings) is common in certain populations (Waringer & Graf, 1995), especially at higher altitudes. Study area The 11.3 km long Roseg River (Ova da Roseg) is situated in the Bernina Massif of the Swiss Alps (lat 46°29'28" N, long 9°53,57n E). The catchment area is 66.5 km2, of which 30% is covered by glaciers and 40% is bare rock or glacial till without vegetative cover. Bedrock consists of granite and diorite. Elevations range from 1760 m a.s.l. (confluence with Bernina River) to 4049 m a.s.l. (Piz Bernina), and treeline is at about 2300 m a.s.l. The Roseg River is fed primarily by meltwater from the Tschierva and Roseg glaciers. Discharge peaks in late summer from glacial melting; diel flow fluctuations also are highest in summer. In winter, the system is largely fed by ground water with surface flow ceasing in many channels (Malard et al, 1999). Subalpine coniferous forests are restricted to side-slopes in the lower area of the valley; no trees occur on the fluvio-glacial gravel of the valley floor. Dominant tree species include larch (Larix decidua Mill.), stone pine (Pinus cembra L.), and mugo pine (P. mugo Turra). Shrubs mainly consist of green alder (Alnus viridis (Chaix) DC.) and willow (Salix sp.). A more detailed description of the flood plain is given in Tockner et al. (1997). Genetic structure and feeding plasticity of A. zerberus 101 Study sites This study examined A. zerberus populations from 3 markedly different stream types in the Val Roseg flood plain: a lake outlet stream —50 m below Lake Roseg (called „Outlet site"), a proglacial spring below Tschierva Glacier at the base of a glacial moraine („Willow site"), and a side-slope springbrook in the main flood plain („Alder site") (Fig. 6.1). The Outlet and Willow site were separated by a —100 m high glacial moraine of the Tschierva glacier. Aerial distance between these 2 sites was approximately 0.6 km. The Alder site was about 1.6 km from the Outlet site and 1.0 km from the Willow site. Because of the spatial arrangement of the sites in the floodplain landscape (e.g., no upstream connection among sites), the only means of dispersal among sites was by adult flight. In addition, the Outlet and Alder site had permanent flow, whereas the Willow site lacked surface water by early September in 1998. Although the Outlet site was fed primarily by glacial meltwater, discharge extremes were ameliorated by the lake. The other 2 streams were fed by ground water. These sites were selected not only because of their spatial arrangement and differences in flow regime, but also because food resources substantially differed. Differences in quality and quantity of food can be an important constraint on the distribution and abundance of populations (e.g., Becker, 1994; Jacobsen & Friberg, 1995). The filamentous algae Hydrurus foetidus (Kirch.) (Chrysophyceae) and, to a lesser extent, epilithic diatoms were dominate food sources at the Outlet site in spring and late summer/fall. Primary production was suppressed in summer because of high turbidity from glacial flour that reduced light and increased scour. Terrestrial inputs of grass and some willow (Salix sp.) made up potential supplementary food resources. Detritus, willow, and the aquatic moss Fontinalis sp. were predominant food re¬ sources at the Willow site; periphyton was negligible. Green alder (Alnus viridis), larch (Larix decidua), and stone pine (Pinus cembra) were the main allochthonous inputs at the Alder site, whereas instream food resources consisted of epilithic algae and moss. Allozyme electrophoresis Samples of —50 last instar larvae of A. zerberus were collected from each site and date, quick-frozen in liquid nitrogen (-80°C), and returned to the laboratory. Collections took place on 24 August 1998 at all sites, and at the Outlet and Alder sites on 16 September 1998 (surface flow had ceased at Willow). In the laboratory, allozyme electrophoresis was completed using cellulose acetate gels (Titan III, Morwell Diagnostics GmbH, Switzerland). Seven polymorphic loci were examined: phosphoglucose isomerase (Pgi), phosphoglucomutase (Pgm), glycerol-3-phosphate dehydrogenase (G3pdh), malic enzyme (Me), aldehyde oxidase (Ao), aconitate A. zerberus 702 Genetic structure and feeding plasticity of T\1766ma.s.l. Jf 4049 m a.s.l. Fig. 6.1. Location of the Val Roseg catchment in the Bemina Massif of the Swiss Alps, and indication of sampling sites, glaciers and forests. Genetic structure and feeding plasticity of A. zerberus 103 hydratase (Aeon), and malate dehydrogenase (Mdh). Running conditions and stain¬ ing procedures were adapted from Hebert & Beaton (1989). Differences in sample sizes are attributable to natural densities of animals in the field, and exclusion of gels that were difficult to read. Allelic frequencies, ob¬ served (Hobs) and expected (Hex ) heterozygosities, population heterozygosities, de¬ viations from Hardy-Weinberg equilibrium, and FST values were calculated from the scored loci. Average heterozygosities among sites were compared by one-way analy¬ sis of variance (ANOVA). If the ANOVA rejected a multi-sample null hypothesis of equal means (p < 0.05), then pairwise comparisons using Tukey's Honest Signifi¬ cant Test (HSD) were performed (Zar, 1996). Deviations of observed genotypic frequencies from Hardy-Weinberg equilibrium were examined using exact signifi¬ cance probabilities (Weir, 1996). When a heterozygote excess or deficiency was de¬ tected, the more powerful tests recommended by Rousset & Raymond (1995) were used to confirm or reject the respective finding. FST values, a measure of the degree of genetic differentiation among populations, were calculated using the method of Weir &C Cockerham (1984) and tested for significance following Waples (1987). All statistics were performed with the program packages BIOSYS-2 (Swofford & Selander, 1989), GENEPOP 1.2 (Raymond & Rousset, 1995) andStatistica5.1 (Statsoft, 1995). Multiple stable isotopes For isotope analysis, 10 specimens of last instar A. zerberus larvae per site were collected and transferred alive to the laboratory. These animals were starved for at least 24 h to evacuate gut contents and frozen. Epilithic algae (mostly diatoms) were brushed from single stones and treated with IN HCl to remove inorganic car¬ bon. Filamentous algae (Hydrurus foetidus), benthic organic matter („detritus"), se¬ nescent biofilm, and Fontinalis sp. were sampled separately and rinsed thoroughly to remove animals prior to analysis. Samples of terrestrial food resources included grass, needles of Pinus cembra and Larix decidua, and leaves of Alnus viridis and Salix sp. Samples of allochthonous and autochthonous materials present at each site were frozen in the field for transport to the laboratory. All samples for isotope analysis were dried at 60°C and ground. Subsamples of 0.5 to 1.5 mg (from individual A. zerberus larvae, or 2 to 6 of the other materials) were for analyzed stable isotopes of organic carbon (13C) and nitrogen (15N) using a mass spectrometer coupled in continuous flow with a CE-Instruments NCS2500 standard following methods (Peterson & Fry, 1987). Results are reported in the conventional delta (ô) notation as per mil (%o) deviation from isotope standards (Peedee Belemnite limestone (PDB) for 13C/12C and atmospheric nitrogen for 15N/ 14N) using the equation: or = - Ô13C Ô15N (%o) (R . / R 1) x 1000 N ' x sample standard, , ' where R = 13C/12C or 15N/14N. 104 Genetic structure and feeding plasticity of A. zerberus The carbon isotope 13C was used to identify the source of energy. The 513C of animals directly reflects their diets, with only slight enrichment (ca. l%o) during the feeding process (DeNiro & Epstein, 1978; Rau et al, 1983). The nitrogen isotope 15N provides information about feeding and other trophic relationships among ani¬ mals and plants. ô15N increases 3.4 ± 1. l%o between trophic levels (DeNiro & Epstein, 1981; Minagawa & Wada, 1984). Mean isotopic ratios of A. zerberus were com¬ pared among sites by one-way ANOVA, and multiple pairwise comparisons among group means were made using Tukey's Honest Significant Test (HSD). Prior to analy¬ sis, data were log10(x+l) transformed to normalize and stabilize the variance (Zar, 1996). Food preference experiments Food preference of A. zerberus larvae was examined by conducting labora¬ tory experiments in August 1998 and September 1998. Between 25 and 3 0 last instar larvae were collected at each site and transferred alive to the laboratory. Animals from each sampling site were used in the 1st experiment, whereas no specimens from the Willow site were available for the 2nd experiment because it had gone dry. The experiments were performed in an environmental chamber at a constant tempera¬ ture of 4°C using 5 open transparent PVC cylinders (d = 19 cm) as replicate aquaria for each site. Each aquarium contained autoclaved fine gravel and coarse sand as substrata, and were filled with aerated tap water to a depth of 5 cm. Coarse sand was added to provide material for larval case-building. Four plastic petri dishes (d = 5 cm), each with a different food type, were evenly distributed in each aquarium. In the 1st experiment, food types consisted of filaments of Hydrurus foetidus (amount per dish: 3 ml), abscissed needles of larch (0.26 ± 0.04 g dry weight, mean ± 1 SD), circular leaf fragments (d = 1.5 cm) of alder (0.11 ± 0.01 g), and whole willow leaves (0.09 ± 0.02 g). In the 2nd experiment, willow was replaced by the moss Fontinalis sp. (0.18 ± 0.02 g). At the beginning of each experiment, 4 larvae were in each = placed aquarium (n 20 individuals per site). Larvae that died during the experiment were replaced, but usually fewer than 3 died within any one experi¬ ment. The experiments were terminated after 4 and 7 days, respectively, because larvae began to pupate. Observations of larvae were made hourly during „daytime" 8:00 (light on, -18:00). Larvae were assigned to a certain food type when they were within a respective plastic dish, otherwise they were categorized as „None". Counts for the 5 for each aquaria site were pooled and expressed as percentages. Data were arcsin(sqrt(x)) transformed to ensure normality and homogeneity of variance. Two- way ANOVA (site x food type) followed by Tukey's Honest Significant Test (HSD) for multiple pairwise comparisons (Zar, 1996) were used to test for significant dif¬ ferences food among types. An analysis of food quality was not performed, as the experiments were designed to reveal only whether larvae displayed site specific food preferences that corresponded to the different food resources found at each site. Genetic structure and feeding plasticity of A. zerberus 105 6.3 Results Population genetic structure Allele frequencies, expected (Hex ) and observed (Hobs) heterozygosities, popu¬ lation heterozygosities, sample sizes, and deviations from Hardy-Weinberg expecta¬ tions, are summarized in Table 6.1. Mean Hex per locus ranged between 0.30 ± 0.07 (SE) at the Outlet site, 0.32 ± 0.07 (SE^at the Willow site, and 0.36 ± 0.08 (SE) at the Alder site; no significant differences were observed among sites (F21g = = also were not different sites 0.154, p 0.86). Average Hobs significantly among (F2 lg = 0.155, p = 0.86). Deviations from Hardy-Weinberg equilibrium were found for 2 out of 7 polymorphic loci (29%) at the Outlet site, 3 out of 7 (43%) at the Willow site, and 4 out of 6 (67%) at the Alder site. With the exception of Ao at the Outlet site, all deviations indicated a deficiency of heterozygotes (Table 6.1). Specific tests for heterozygote excess or deficiency confirmed the results of the calculated exact probabilities. A jackknife estimate of FST based on all loci was 0.086 ± 0.038 (SD). Significant FST values were observed for all loci except G3pdh, Ao and Mdh (Table 6.2). Stable isotope analysis Stable isotope signatures for A. zerberus larvae and potential food resources collected at each site are presented in Fig. 6.2. Average 813C values of A. zerberus were -24.6%o at the Outlet site, -29.6%o at the Willow site, and -27.8%o at the Alder site. Respective averages for 815N were -5.2%o, -1.9%o, and -2.9%o. Results of ANOVA indicated that carbon and nitrogen isotopic domains of A. zerberus were significantly different among sites (Ô13C: F227 = 21.03, p < 0.001; S15N: F227 = 27.27, p < 0.001). Pairwise comparisons revealed significant differences for 2 out of the 3 possible combinations for ô13C (Tukey's HSD: Outlet vs. Willow p < 0.001; Outlet vs. Alder p < 0.001; Alder vs. Willow p = 0.09), and all combinations for 815N (Tukey's HSD: Outlet vs. Willow p < 0.001; Outlet vs. Alder p < 0.001; Alder vs. Willow p = 0.02). At the Outlet site, average 813C of A. zerberus was nearest to those of grass and Salix sp., followed by epilithic algae (Fig. 6.2). The low concentrations of allochthonous organic matter at this site (R. Zah, unpublished data) suggest larvae fed mainly on epilithon at the Outlet site. Although 815N of A zerberus were —3%o enriched compared to Hydrurus foetidus, the difference in their S13C values sug¬ gested that animals apparently avoided it in the presence of other food resources at this site. Detritus and, to a lesser extent, willow and grass were identified as major food resources at the Willow site (Fig. 6.2). At the Alder site, the identification of potential food resources was more difficult because of similar 813C signatures for the 106 Genetic structure and feeding plasticity of A. zerberus various foods analyzed. However, larvae seemed most likely to feed on epilithon, Alnus viridis and grass, whereas the use of Pinus cembra and Larix decidua as food resources was less clear (Fig. 6.2). Table 6.1: Allele frequencies, observed and expected heterozygosities, population heterozy¬ gosities, sample sizes (n), and deviations from Hardy-Weinberg equilibrium (HWE) using exact probabilities (p = 0.05) of A zerberus larvae, ex = excess of heterozygotes, def = deficiency of heterozygotes. Locus Site Allele Heterozygosity n Deviations from IECC no. 1 2 3 Observed Expected HWE (p-values) Pgi Outlet 0.000 0.964 0.036 0.073 0.070 55 1.000 5.3.1.9 Willow 0.074 0.926 0.000 0.000 0.140 27 0.001 (def) Alder 0.000 1.000 0.000 0.000 51 na Pgm Outlet 0.337 0.640 0.023 0.395 0.482 43 0.338 5.4.2.2 Willow 0.053 0.921 0.026 0.158 0.152 19 1.000 Alder 0.217 0.620 0.163 0.283 0.548 46 0.000 (def) G3pdh Outlet 0.088 0.912 0.176 0.166 17 1.000 1.1.1.8 Willow 0.125 0.875 0.150 0.224 20 0.236 Alder 0.083 0.917 0.167 0.159 12 1.000 Me Outlet 0.696 0.304 0.464 0.427 56 0.751 1.1.1.40 Willow 0.339 0.661 0.107 0.456 28 0.000 (def) Alder 0.778 0.222 0.148 0.352 27 0.007 (def) Ao Outlet 0.404 0.596 0.808 0.491 26 0.001 (ex) 1.2.3.1 Willow 0.417 0.583 0.833 0.530 6 0.394 Alder 0.591 0.409 0.818 0.506 11 0.066 Aeon Outlet 0.038 0.962 0.077 0.075 39 1.000 4.2.1.3 Willow 0.125 0.875 0.250 0.223 24 1.000 Alder 0.440 0.560 0.080 0.503 25 0.000 (def) Mdh Outlet 0.026 0.733 0.241 0.121 0.408 58 0.000 (def) 1.1.1.37 Willow 0.063 0.625 0.313 0.125 0.524 16 0.004 (def) Alder 0.146 0.729 0.125 0.125 0.441 24 0.002 (def) Site Mean heterozygosity per locus ± 1 SE Observed Expected Outlet 0.302 ±0.103 0.303 ± 0.072 Willow 0.232 ±0.104 0.322 ± 0.066 Alder 0.232 ±0.103 0.359 + 0.078 Genetic structure and feeding plasticity of A. zerberus 107 Table 6.2: FST values for each locus separately, and over all loci. NS = not significant, * ** *** = p < 0.05, = p < 0.01, = p < 0.001. Locus Fst Pgi 0.035 ** Pgm 0.070 *** G3pdh -0.031 NS Me 0.169 *** Ao 0.033 NS Aeon 0.258 *** Mdh 0.001 NS Mean 0.086 Jackknife estimate over loci Mean ± 1 SD 0.086 + 0.038 Food preference experiments In the 1st laboratory experiment, two-way ANOVA showed significant differ¬ ences in the relative abundance of A. zerberus larvae among food types (F4420 = 92.8, p < 0.001) and for the site x food type interaction (F8420 = 3.4, p < 0.001), but not among sites (F2420 = 0.1, p > 0.90) (Fig. 6.3a). Post-hoc comparisons revealed that the proportion of animals not feeding on any of the offered food types (None) was significantly higher at all sites (Tukey's HSD, p < 0.05), whereas no significant dif¬ ferences were found among food types within a site (Tukey's HSD, p > 0.05). Indi¬ vidual food types also did not differ significantly among sites (Tukey's HSD, p > 0.05). In the 2nd experiment, two-way ANOVA revealed a significant effect for food tyPe (F4,49o = 71-9> P < °-01)> but not for site (Fi59o = °-°> P > °-") or tne site x food type interaction (F4J90 = 1.7, p = 0.15). The non-significant interaction term indicates no differences either among food types within sites, or within individual food types among sites (Fig. 6.3b). 705 Genetic structure and feeding plasticity of A. zerberus 2 Outlet site 0 senescent -2 Salix biofilm -4 grass Ç -6 -8 tyrfnifi/s h JpHithon -10 2 ' Willow site r 0 grass . i T i ' -2 1 'L -4 CO -6 h D Salix pjfP detritus -8 . Fontinalis -10 i .i ' Alder site Alnus Fontinalis $ * epilithon grass senescent Larix Pinus J] biofilm -10 -40 -35 -30 -25 -20 -15 Ô13C (%o) Fig. 6.2: Plots of dual isotope ratios for A zerberus larvae and potential food resources at each site. Isotopic signatures of A zerberus are given as mean ± 1 SD. Open boxes are terrestrial food resources, and filled boxes are aquatic food resources (mean ± 1 SD). Genetic structure and feeding plasticity of A. zerberus 109 ( a) 60 - T 40 - T (%) T J T T 20 - i\ T /TEE i n l II i DNone 60 - Hydrurus foetidus IS Larix decidua D Salix sp. Q Fontinalis sp. M Alnus viridis 40 - (%) 20 - no data site had dried up Outlet site Willow site Alder site Fig. 6.3: Proportions of A zerberus larvae feeding on different food resources offered during the first (a) and second (b) food preference experiments. Bars represent mean ± 1 SD. In the sec¬ ond experiment Salix sp. was replaced by the moss Fontinalis sp. 770 Genetic structure and feeding plasticity of A. zerberus 6.4 Discussion Allozyme electrophoresis was used to test for significant differences in the genetic structure of A. zerberus among populations. The high levels of heterozygos¬ ity at all sites could be an indication that A. zerberus larvae show some life history plasticity to maintain viable populations in different stream types (Robinson et al, 1992). Predictions from population genetic models also suggest high levels of het¬ erozygosity for species living in spatio-temporally heterogenous environments (Hartl, 1980; Hedrick, 1986). Empirical evidence for such a fact has been reported for fishes (Zimmermann, 1987), Zooplankton (Weider, 1992), and lotie macroinvertebrates (Hughes et al, 1999). A high spatio-temporal variability in envi¬ ronmental conditions has been demonstrated for the Val Roseg flood plain (Tockner et al,1997; Malard et al.,1999), and also that benthic macroinvertebrate communi¬ ties respond to these changes (Burgherr & Ward, in press, b). The overall FST value indicated the possibility of a moderate genetic differentiation among populations. The observed deviations from Hardy-Weinberg equilibrium were high and differed across loci within each local population. This could be a result of the low sample size for some loci, although it is likely that the specific tests for heterozygote excess and deficiency applied here detected real departures from Hardy-Weinberg equilibrium. Excluding this possibility, several other options remain that have been discussed in Schmidt et al. (1995). Following the reasoning of these authors and empirical evi¬ dence from other studies (Bunn & Hughes, 1997; Hughes et al, 1998), we assume that larval populations in the 3 habitats are likely the product of only a few matings by successfully dispersing adults. In conclusion, our findings suggest that there has been little migration of A. zerberus populations among sites in the current genera¬ tion. Of the little empirical data available on movement distances of caddisflies, the adult stage appears to be most important. Long-distance flights ranging from 1-4 km (Neves, 1979) to 16 km (Coûtant, 1982) have been documented, although limited larval dispersal during periods of low, stable flow have been found (Hart &C Resh, 1980; Jackson et al, 1999). For example, Neves (1979) showed that Pycnopsyche guttifer (Walker) (Limnephilidae) moved out of pools during low flow conditions. However, the harsh environmental conditions in the main glacial channel (high tur¬ bidity and bedload movement), and the lack of upstream connections among sites effectively limit dispersal by larvae in our glacial flood plain. Brachyptery (flightlessness) of adults is known for some populations of A. zerberus, and was observed for individuals at the Outlet site (P. Burgherr, personal observation). Brachyptery in insects is known to increase with altitude and latitude (Roff, 1990), and also can be a function of time since déglaciation (Donald, 1985). Our observations support the notion that brachyptery in insects occurs more com¬ monly in stable environments in which long-term persistence is unlikely to require long-distance dispersal (Roff, 1990). For example, the higher temperatures and more stable habitat conditions of the proglacial lake outlet attenuate the otherwise harsh environmental conditions of other glacial streams (Burgherr & Ward, in press, a). Genetic structure and feeding plasticity of A. zerberus 111 On the other hand, macropterous individuals can disperse among isolated habitats, and therefore, have an advantage when environments are spatially or temporally heterogenous (Roff, 1994), as found at the temporary Willow site. Because brachypterous females are more fecund than macropterous females (Roff & Fairbairn, 1991), the brachyptery allele can spread quickly throughout a population within a particular habitat patch. If habitats are spatially isolated, as in the present study, the brachyptery allele can be passed among local populations only if a dispersing macropterous female mates with a brachypterous male. As a consequence, a popula¬ tion homozygous for the brachyptery allele is more prone to extinction when envi¬ ronmental conditions become unsuitable (Roff, 1994). In streams of alpine glacial flood plains, the ratio of brachypterous to macropterous individuals in local populations may be a function of habitat stability, and thus influence long-term dis¬ tribution patterns in population structure. The analysis of multiple stable isotopes provided information of whether food resources limited the distribution of populations within the study sites, and gave insight towards the plasticity in A. zerberus feeding behavior. Traditionally, studies regarding the diet of aquatic insects was based on gut content analyses (Coffman, 1971), field or laboratory observations of feeding, or radio-isotopes as tracers (Rounick & Winterbourn, 1986). These approaches, however, only yield information on recent ingestion, and give no insight on an organism's feeding his¬ tory, trophic status (Gearing, 1991) or food web structure that is provided by stable isotope analysis (Rounick & Winterbourn, 1986; Peterson & Fry, 1987). We also used a site specific perspective in this study (sensu Doucett et al, 1996) to compen¬ sate for some of the limitations of stable isotope analysis suggested in literature (France, 1995). Isotope signatures indicated a site-specific, omnivorous feeding mode for lar¬ vae of A. zerberus, and furthermore, that larvae were opportunistic feeders. This is an important finding, as it indicates the role of behavioural trade offs in sustaining local populations in potentially sub-optimal habitats or in the (re)colonization of isolated habitat patches (sensu Hanski, 1999). Plasticity in diet/feeding minimizes the potential constraints of food resources on the colonization and maintenance of local populations in otherwise suitable habitats. This idea was tested further by assessing the feeding behaviour of larvae in food preference experiments. The laboratory experiments confirmed the existence of a generalist feeding strategy by A. zerberus that was independent of the food re¬ sources available at a specific site. That is, last instar larvae from the different sites similarly consumed available resources in the experimental situation even though these same larvae (as earlier instars) fed on different resources in the field, as inferred from the isotope analyses. These results provide additional evidence that plasticity in feeding behavior is a key factor for allowing dispersal and colonization into lotie habitats with different food resources (quality or quantity), e.g., forested streams with a predominant allochthonous food supply vs. alpine streams with mostly auto¬ chthonous food sources. For example, A. zerberus has been shown to capitalize on high quality food such as introduced leaf litter in an alpine stream (Robinson et al, 112 Genetic structure and feeding plasticity of A. zerberus 1998), or as demonstrated for other stream macroinvertebrates in studies on re¬ source limitation (Dobson & Hildrew, 1992). Relationships between the genetic structure and life history traits of A. zerberus were not assessed within the framework of the present study. However, the rela¬ tively high genetic diversity within A. zerberus populations in the present study can be associated with the observed plasticity in its life history (as implied from A. zerberus feeding behavior). A number of studies have shown positive correlations between genetic variation (e.g., number of alleles per locus, heterozygosity, and polymor¬ phism) and life history traits (Nevo &c Yang, 1979 in the tree frog Hyla arborea savignyi; Mitton &c Lewis Jr., 1989 for a variety of organisms), whereas others found no such associations (Whithehurst 8c Pierce, 1991 in the spotted chorus hogPseudacris clarkii Baird; Hutchings &c Ferguson, 1992 in Salvelinus fontinalis Mitchill). These contrasting findings may reflect interspecific variation in the strength of genetic cor¬ relations among life history traits (Hutchings & Ferguson, 1992). Similarly, aquatic invertebrates lacking a terrestrial stage often exhibit rather high levels of genetic differentiation among populations (Hughes et al, 1995 for the prawn Paratya australiensis Kemp), whereas stream insects having an adult winged stage tend to show little genetic differentiation at larger spatial scales (Schmidt et al, 1995; Bunn &C Hughes, 1997; but see Hughes et al, 1999). Clearly, more research is needed to address the important interplay between genetic structure, life history, and the habi¬ tat templet on the maintenance and structure of macroinvertebrate populations in lotie ecosystems. 6.5 Acknowledgements We are especially grateful to Dr. S.M. Bernasconi of the Geology Institute of ETH Zürich for stable isotope analyses, and T. Dambone-Bösch for electrophoretic analyses. Special thanks to Mr. P. Testa and his crew at the Roseg Hotel for their hospitality, and to the villages of Pontresina and Samedan for providing access to the study area. We appreciate the helpful comments of M.T. Monaghan and J.V. Ward that improved the manuscript. This study was partially funded by a grant from the Swiss National Science Foundation (no. 3100-050444.97/1). Genetic structure and feeding plasticity of A. zerberus 113 6.6 References Becker, G. (1994) Food preference by five trichopteran scrapers. Hydrobiologia, 273, 171-178. Botosaneanu, L. &C Malicky, H. (1978) Trichoptera. Limnofauna Europaea (ed. J. lilies), pp. 333-359, Gustav Fischer Verlag, Stuttgart. Bunn, S.E. & Hughes, J.M. (1997) Dispersal and recruitment in streams: evidence from genetic studies. Journal of the North American Benthological Society, 16, 338-346. Burgherr, P. & Ward, J.V. (in press, a) Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 27. Burgherr, P. & Ward, J.V. (in press, b) Longitudinal and seasonal distribution pat¬ terns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps). Freshwater Biology. Coffman, W.P. (1971) Energy flow in a woodland stream ecosystem: I. Tissue sup¬ port trophic structure of the autumnal community. Archiv für Hydrobiologie, 68, 232-276. Coûtant, C.C. (1982) Evidence for upstream dispersion of adult caddisflies (Trichoptera: Hydropsychidae) in the Columbia River. Aquatic Insects, 4, 61-66. DeNiro, M.J. & Epstein, S. (1978) Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta, 42, 495-506. DeNiro, M.J. & Epstein, S. (1981) Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta, 45, 341-351. Diffendorfer, J.E. (1998) Testing models of source-sink dynamics and balanced dis¬ persal. Oikos, 81, 417-444. M. & Dobson, Hildrew, A.G. (1992) A test of resource limitation among shredding detritivores in low order streams of southern England. Journal of Animal Ecology, 61, 69-77. Donald, D.B. (1985) The wing length of Sweltsa revelstoka (Plecoptera: Chloroperlidae). Canadian Entomologist, 117, 233-239. Doucett, R.R., Barton, D.R., Guiger, K.R.A., Power, G. & Drimmie, R.D. (1996) Comment: Critical examination of stable isotope analysis as a means for trac¬ ing carbon pathways in stream ecosystems. Canadian Journal ofFisheries and Aquatic Sciences, 53, 1913-1915. R. Critical France, (1995) examination of stable isotope analysis as a means for trac¬ ing carbon pathways in stream ecosystems. Canadian Journal ofFisheries and Aquatic Sciences, 52, 651-656. Gearing, J.N. (1991) The study of diet and trophic relationships through natural abundance of 13C. Carbon Isotope Techniques (eds. D.C. Coleman & B. Fry), pp. 201-218, Academic Press, San Diego. 7 74 Genetic structure and feeding plasticity of A. zerberus Gilpin, M.E. & Hanski, I. (1991) Metapopulation dynamics: empirical and theoreti¬ cal investigations. Biological Journal of the Linnean Society, 42, 1-323. Hanski, I. (1998) Metapopulation dynamics. Nature, 396, 41-49. Hanski, I. (1999) Metapopulation ecology. Oxford University Press, Oxford, UK. Hart, D.D. & Resh, V.H. (1980) Movement patterns and foraging ecology of a stream caddisfly larva. Canadian Journal of Zoology, 58, 1174-1185. Hard, D.L. (1980) Principles ofpopulation genetics. Sinauer Associates, Sunderland, Massachusetts. Hebert, P.D.N. & Beaton, M.J. (1989) Methodologies for allozyme analysis using cellulose acetate electrophoresis. A practical handbook. Department of Bio¬ logical Sciences, Great Lakes Institute, University of Windsor, Windsor, On¬ tario. Hedrick, P.W. (1986) Genetic polymorphism in heterogenous environments: a dec¬ ade later. Annual Review of Ecology and Systematics, 17, 535-566. Hershey, A.E., Pastor, J., Peterson, B.J. &c Kling, G.W. (1993) Stable isotopes re¬ solve the drift paradox for Baetis mayflies in an arctic river. Ecology, 74, 2315-2325. Hildrew, A.G. & Townsend, C.R. (1987) Organization in freshwater benthic com¬ munities. Organization of Communities: Past and Present (eds. J.H.R. Gee & P.S. Giller), pp. 347-371, Blackwell Scientific Publications, Oxford, UK. Hughes, J.M., Bunn, S.E., Hurwood, D.A. & Cleary, C. (1998) Dispersal and re¬ cruitment of Tasiagma ciliata (Trichoptera: Tasimiidae) in rainforest streams, south-eastern Australia. Freshwater Biology, 39, 117-127. Hughes, J.M., Bunn, S.E., Kingston, M. & Hurwood, D.A. (1995) Genetic differen¬ tiation and dispersal among populations of Paratya australiensis (Atyidae) in rainforest streams in southeast Queensland, Australia. Journal of the North American Benthological Society, 14, 158-173. Hughes, J.M., Mather, P.B., Sheldon, A.L. & Allendorf, F.W. (1999) Genetic struc¬ ture of the stonefly, Yoraperla brevis, populations: the extent of gene flow among adjacent montane streams. Freshwater Biology, 41, 63-72. Hutchings, J.A. & Ferguson, M.M. (1992) The independence of enzyme heterozy¬ gosity and life-history traits in natural populations of Salvelinus fontinalis (brook trout). Heredity, 69, 496-502. Jackson, J.K., McElravy, E.P. & Resh, V.H. (1999) Long-term movements of self marked caddisfly larvae (Trichoptera: Sericostomatidae) in a California coastal mountain stream. Freshwater Biology, 42, 525-536. Jacobsen, D. & Friberg, N. (1995) Food preference of the trichopteran larva Anabolia nervosa from two streams with different food availability. Hydrobiologia, 308, 139-144. Kareiva, P. (1990) Population dynamics in spatially complex environments: theory and data. Philosophical Transactions of the Royal Society Series B, 330, 175- 190. Genetic structure and feeding plasticity of A. zerberus 115 Lande, R. & Barrowclough, G.F. (1987) Effective population size, genetic variation, and their use in population management. Viable populations for conservation (ed. M.E. Soulé), pp. 87-123, Cambridge University Press, Cambridge, UK. Levin, S.A. (1994) Patchiness in marine and terrestrial systems: from individuals to populations. Philosophical Transactions of the Royal Society Series B, 343, 99-103. Mackay, R.J. (1992) Colonization by lotie macroinvertebrates: a review of processes and patterns. Canadian Journal of Fisheries and Aquatic Sciences, 49, 617- 628. 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. Malicky, H. (1980) Emergenzuntersuchungen am Teichbach und am Schreierbach. Jahresberichte der Biologischen Station Lunz, 3, 113-122. McGregor, G., Petts, G.E., Gurnell, A.M. & Milner, A.M. (1995) Sensitivity of al¬ pine stream ecosystems to climate change and human impacts. Aquatic Con¬ servation: Marine and Freshwater Ecosystems, 5, 233-247. Meffe, G.K. & Vrijenhoek, R.C. (1988) Conservation genetics in the management of desert fishes. Conservation Biology, 2, 157-167. Milner, A.M. & Petts, G.E. (1994) Glacial rivers: physical habitat and ecology. Fresh¬ water Biology, 32, 295-307. Minagawa, M. & Wada, E. (1984) Stepwise enrichment of 15N along food chains: further evidence and the relation between 15N and animal age. Geochimica et Cosmochimica Acta, 48, 1135-1140. Minshall, G.W. & Petersen, R.C. (1985) Towards a theory of macroinvertebrate community structure in stream ecosystems. Archiv für Hydrobiologie, 104, 49-76. Mitton, J.B. & Lewis Jr., W.M. (1989) Relationships between genetic variability and life-history features of bony fish. Evolution, 43, 1712-1723. Moog, O. (ed) (1995) Fauna Aquatica Austriaca: Katalog zur autökologischen Einstufung aquatischer Organismen Österreichs. Bundesministerium für Land- und Forstwirtschaft - Wasserwirtschaftskataster, Wien. Neves, R.J. (1979) Movements of larval and adult Pycnopsyche guttifer (Walker) (Trichoptera: Limnephilidae) along Factory Brook, Massachusetts. The Ameri¬ can Midland Naturalist, 102, 51-58. Nevo, E. & Yang, S.Y. (1979) Genetic diversity and climatic determinants of tree frogs in Israel. Oecologia (Berlin), 41, 47-63. Peterson, BJ. & Fry, B. (1987) Stable isotopes in ecosystem studies. Annual Review of Ecology and Systematics, 18, 293-320. Poff, N.L. (1997) Landscape filters and species traits: towards mechanistic under¬ standing and prediction in stream ecology. Journal of the North American Benthological Society, 16, 391-409. 776 Genetic structure and feeding plasticity of A. zerberus Poff, N.L. & Ward, J.V. (1990) Physical habitat template of lotie systems: recovery in the context of historical pattern of spatiotemporal heterogeneity. Enviromental Management, 14, 629-645. Pulliam, H.R. &C Dunning, J.B. (1997) Demographic processes: population dynam¬ ics on heterogenous landscapes. Principles of conservation biology (eds. G.K. Meffe & C.R. Carroll), pp. 203-233, Sinauer, Sunderland, Massachussetts. Rau, G.H., Mearns, A.J., Young, D.R., Olson, R.J., Schäfer, H.A. & Kaplan, LR. (1983) Animal 13C/12C correlates with trophic levels in pelagic food webs. Ecology, 64, 1314-1318. Raymond, M. & Rousset, F. (1995) GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity, 86, 248-249. Robinson, CT., Gessner, M.O. & Ward, J.V. (1998) Leaf breakdown and associ¬ ated macroinvertebrates in alpine glacial streams. Freshwater Biology, 40, 215-228. Robinson, CT., Reed, L.M. & Minshall, G.W. (1992) Influence of flow regime on life history, production, and genetic structure of Baetis tricaudatus (Ephemeroptera) and Hesperoperla pacifica (Plecoptera). Journal of the North American Benthological Society, 11, 278-289. Roff, D.A. (1990) The evolution of flightlessness in insects. Ecological Monographs, 60, 389-421. Roff, D.A. (1994) Why is there so much genetic variation for wing dimorphism? Researches on Population Ecology (Kyoto), 36, 145-150. Roff, D.A. & Fairbairn, D.J. (1991) Wing dimorphisms and the evolution of migra¬ tory polymorphisms among Insecta. American Zoologist, 31, 243-251. Rounick, J.S. & Winterbourn, M.J. (1986) Stable carbon isotopes and carbon flow in ecosystems. BioScience, 36, 171-177. Rousset, F. & Raymond, M. (1995) Testing for heterozygote excess and deficiency. Genetics, 140, 1413-1419. Saunders, D.A., Hobbs, R.J. & Margules, C.R. (1991) Biological consequences of ecosystem fragmentation: a review. Conservation Biology, 5, 18-32. Schmidt, S.K., Hughes, J.M. & Bunn, S.E. (1995) Gene flow among conspecific populations of Baetis sp. (Ephemeroptera): adult flight and larval drift. Jour¬ nal of the North American Benthological Society, 14, 147-157. Slatkin, M. (1985) Gene flow in natural populations. Annual Review ofEcology and Systematics, 16, 393-430. Southwood, T.R.E. (1977) Habitat, the templet for ecological strategies? Journal of Animal Ecology, 46, 337-365. Southwood, T.R.E. (1988) Tactics, strategies and templets. Oikos, 52, 3-18. Stamps, J.A., Buechner, M. & Krishnan, V.V. (1987) The effect of edge permeability and habitat geometry on emigration from patches of habitat. The American Naturalist, 129, 533-552. Genetic structure and feeding plasticity of A. zerberus 117 Stanford, J.A. &c Ward, J.V. (1983) Insect species diversity as a function of environ¬ mental variability and disturbance in stream systems. Stream Ecology (eds. J.R. Barnes & G.W. Minshall), pp. 265-278, Plenum Press, New York. Statsoft (1995) Statistica for Windows. Statsoft Inc., Tulsa. Stearns, S.C (1992) The evolution of life histories. Oxford University Press, Oxford, UK. Swofford, D.L. & Selander, R.B. (1989) BIOSYS-1. A computer program for the analy¬ sis of allelic variation in population genetics and biochemical systematics. Release 2. University of Illinois, Urbana, Illinois. Szczesny, B. (1986) Caddisflies (Trichoptera) of running waters in the Polish North Carpathians. Acta Zoologica Cracoviana, 29, 501-586. Tobias, W. & Tobias, D. (1981) Trichoptera Germanica: Bestimmungstafeln für die deutschen Köcherfliegen, Teil I: Imagines. Courier Forschungs-Institut Senckenberg, Frankfurt am Main. Tockner, K., Malard, F., Burgherr, P., Robinson, CT., Uehlinger, U., Zah, R. & Ward, J.V. (1997) Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv für Hydrobiologie, 140, 433-463. Tonn, W.M., Magnuson, J.J., Rask, M. & Toivonen, J. (1990) Intercontinental com¬ parison of small-lake fish assemblages: the balance between local and regional processes. The American Naturalist, 136, 345-375. Townsend, C.R. &c Hildrew, A.G. (1994) Species traits in relation to a habitat tem¬ plet for river systems. Freshwater Biology, 31, 265-275. R.S. Waples, (1987) A multispecies approach to the analysis of gene flow in marine shore fishes. Evolution, 41, 385-400. Ward, J.V. (1992) Aquatic insect ecology: 1. biology and habitat. John Wiley &c Sons, New York. Ward, J.V. (1994) Ecology of alpine streams. Freshwater Biology, 32, 277-294. Waringer, J. &c Graf, W. (1995) The larvae of Acrophylax zerberus Brauer, 1867 and Leptotaulius gracilis Schmid, 1955 (Trichoptera: Limnephilidae) from two Austrian mountain brooks. Aquatic Insects, 17, 41-49. Weider, L.J. (1992) Disturbance, competition and maintenance of clonal diversity in Daphnia pulex. Journal of Evolutionary Biology, 5, 505-521. Weir, B.S. (1996) Genetic data analysis II: methods for discrete population genetic data. Sinauer Associates, Sunderland, Massachusetts. B.S. & C.C. Weir, Cockerham, (1984) Estimating F-statistics for the analysis of popu¬ lation structure. Evolution, 38, 1358-1370. Whithehurst, P.H. &C Pierce, B.A. (1991) The relationship between allozyme varia¬ tion and life-history traits of the spotted chorus frog, Pseudacris clarkii. Copeia, 4, 1032-1039. Wiens, J.A. (1997) Metapopulation dynamics and landscape ecology. Metapopulation biology (eds. I.A. Hanski &c M.E. Gilpin), pp. 43-68, Academic Press, San Diego. 7 75 Genetic structure and feeding plasticity of A. zerberus Zar, J.H. (1996) Biostatistical analysis. Prentice-Hall, Upper Saddle River, New Jersey. Zimmermann, E.G. (1987) Relationships between genetic parameters and life-his¬ tory characteristics of stream fishes. Community and evolutionary ecology of North American stream fishes (eds. W.J. Matthews & D.C. Heins), pp. 239- 244, Norman, Oklahoma. fr Ssite \xzf %K^ 1 5 Blan k i >«• w% § pn, 3er/ >*A, I É ï Y eaf Synopsis 121 Chapter 7 Synopsis 7.1 Motivation streams? Why this thesis, or in a broader context, why research on alpine glacial My motivation for doing this research was twofold. First, alpine streams constitute integral landscape components in high mountainous areas, but the diversity and uniqueness of these ecosystems is threatened. Human impacts (i.e., water resource developments, tourism) will be the greatest sources of environmental change in al¬ pine regions in the coming decades, and effects of global warming are predicted to be more pronounced in these areas than in other parts of the world (Chapin III &C Körner, 1994). However, there is little known about patterns of diversity and eco¬ system functioning of alpine and arctic streams (e.g., Milner & Petts, 1994; Scrimgeour et al, 1994; Ward, 1994), which leads to my second motive: inadequate scientific understanding of alpine streams. In a review, Ward (1994) concluded that almost all investigations of alpine streams are of (1) limited intensity and duration, and (2) that most data were collected in kryal segments, whereas krenal biotopes have been widely neglected. Based on these facts, the ultimate goal of the present thesis was to contrib¬ ute to the understanding how spatio-temporal habitat heterogeneity affects distribu¬ tion patterns of macroinvertebrate communities in an alpine glacial stream ecosys¬ tem. The following sections summarize major findings of my research, and view them in the context of the current state of knowledge. Last, I discuss some crucial points of the thesis. 7.2 Longitudinal patterns Downstream zonation patterns in glacier-fed streams are most distinct and predictable, reflecting adaptations of the aquatic fauna (Ward, 1994; Füreder, 1999). The predominance of the chironomid Diamesa in the metakryal zone close to the glacier has been documented by several researchers (e.g., Saether, 1968; Steffan,- 1971; Kownacka & Kownacki, 1972; Kownacki & Kownacka, 1973), and provided the foundation for the conceptual qualitative model of Milner & Petts (1994). They related zoobenthic gradients to temperature and channel stability, as a function of distance from the glacier terminus and time since déglaciation (Fig. 7.1). The find¬ ings in this thesis (chapters 3 and 4) yielded only partial support for Milner & Petts' 122 Synopsis model (1994). Changes in community structure in the main glacial channel during high flow conditions in summer were in accordance with model predictions. In con¬ trast, the model did not fully account for seasonal changes in assemblage structure. In addition to these temporal variabilities in longitudinal response patterns, a simi¬ larity in temporal patterns among individual sites along the longitudinal gradient was observed. Overall, results suggest that zoobenthic gradients are not solely a func¬ tion of temperature and channel stability. Seasonal shifts in water sources (i.e., bal¬ ance between glacial meltwater and ground water), and periods of favourable envi¬ ronmental conditions (in spring and late autumn/early winter) also strongly affect zoobenthic distributions. In conclusion, this thesis demonstrated that only compre¬ hensive year-round studies can lead to a better understanding of the spatial and seasonal dynamics in glacier-fed rivers. However, before developing new or modify¬ ing existing models, more research is clearly needed to encounter the variability among different types of glacial streams; e.g., streams fed by the extant glaciers in the Colorado Cordillera remain clear all year (Ward, 1994), whereas differences in geology, topography and riparian vegetation among river basins can cause consider¬ able differences within and among geographic regions (e.g., Gislason et al, 1998). Milner & Petts (1994) also discussed how tributaries, changes in valley con¬ finement, and lakes disrupt downstream patterns. In chapter 2 the influence of the proglacial Lake Roseg was investigated, by comparing macroinvertebrate assemblages between the lake outlet stream and a kryal stream. As predicted by Milner & Petts (1994) the outlet channel differed in numerous abiotic characteristics from the kryal channel; i.e., temperature was higher, a relatively stable single-threaded channel was attained, and the flow regime was attenuated. This reduction in environmental harsh¬ ness resulted in a more diverse and less fluctuating community structure than in the kryal stream. In other words, a proglacial lake can compress the typical faunal gradi¬ ent observed in a kryal stream (Fig. 7.1). 7.3 Heterogeneity of zoobenthos across habitat gradients Differences in lotie zoobenthos among different alpine stream types have received little attention (e.g., Kownacki, 1991; Füreder et al, 1998), and hardly is anything known regarding spatial dynamics among streams in glacial flood plains (Tockner et al, 1997). Following the rationale of the intermediate disturbance hy¬ pothesis (Connell, 1978) and its adaptation to lotie freshwaters (Ward & Stanford, 1983), alpine flood plains are positioned on the descending limb of the harshness - diversity curve (Fig. 7.2a). Considering seasonal changes in spatial habitat heteroge¬ neity (Tockner et al, 1997; Ward et al, 1998; Malard et al, 1999), a two-dimen¬ sional habitat templet can be obtained for the 3 different habitat types compared in chapter 5 (Fig. 7.2b). From this templet, scale-specific predictions about distribution patterns of their benthic communities can be made. Synopsis 123 • Whole flood plain: High habitat heterogeneity increases overall ecosystem stability by providing numerous réfugia (after Tockner et al, 1997). • Among channel types: More stable habitats (i.e., groundwater channels) exhibit the lowest fluctuations in densities and community structure, and highest levels of species diversity (Burgherr &C Ward, submitted). • Within channel types: Channel types encompassing a greater spatial het¬ erogeneity (e.g., groundwater channels) (Ward et al, 1999) also show distinct differences in their benthic assemblage structure. The study on how habitat heterogeneity affected lotie zoobenthos in the Val Roseg flood plain (see chapter 4) supported the above predictions. However, these findings can only be considered as preliminary. Future studies are needed to test these predictions in glacial floodplain systems with different geology and in different geographical regions. Proglacial lake *£ Glacier T_„ = 2 °C Diamesa spp. Diamesinae, Orthocladiinae, Chironominae, Simuliidae, T =4°C other Diamesinae, Ephemeroptera, Plecoptera, Orthocladiinae, Trichoptera, other taxa Simuliidae Baetidae, Nemouridae Chloroperlidae Chironominae, other Ephemeroptera and Plecoptera, Trichoptera, other taxa Fig. 7.1: Left side: characteristic zonation of macroinvertebrates predicted in a glacial river by the model of Milner & Petts (1994). Right side: modified situation as it was found in the outlet stream of the proglacial Lake Roseg. 124 Synopsis (a) high >- 'c/5 CD > alpine streams low low environmental harshness high (b) high mam o channel CD C intermittently- CD OJ connected O i— o channels CD CO o a. CD groundwater channels O low low spatial heterogeneity high Fig. 7.2.: (a) Alpine streams are positioned on the descending limb of the harshness - diversity curve, (b): Within the glacial flood plain, groundwater channels are characterized by a high spatial but low temporal heterogeneity. In contrast, the main channel and intermittently-con¬ nected channels exhibit a high temporal but low spatial heterogeneity. The diameter of the ellipses represents within channel type variability. Synopsis 125 7.4 Dispersal and life history plasticity The value of studies analyzing distribution patterns and responses of biotic communities across environmental gradients is essential to determine what underly¬ ing processes may control the structure of riverine communities and to generate hypotheses that can be tested experimentally in the field or in the laboratory. Chap¬ ters 2 to 5 of this thesis analyzed patterns in macroinvertebrate community structure in respect to longitudinal and habitat gradients, whereas chapter 6 tested specific hypotheses derived from these investigations. The extreme environmental conditions in alpine glacial streams exert a se¬ vere constraint on successful colonization and persistence for many aquatic macroinvertebrate species; indicating the occurrence of a particular species is a di¬ rect expression of this species' ability to tolerate or adapt to the thermal regime and other life history traits (Ward, 1992; Füreder, 1999). Specifically, the complex channel network formed by different stream types in the glacial flood plain is expected to constrain dispersal and to influence other life history traits of aquatic macroinvertebrates. This array of questions was examined in chapter 6. Results of this study suggested that dispersal of the limnephilid caddisfly Acrophylax zerberus most likely occurs through flying adults in the glacial flood plain of Val Roseg. Fur¬ thermore, results were indicative of a generalist feeding strategy by A. zerberus, sug¬ gesting feeding plasticity may be a key factor for the successful dispersal and persist¬ ence of this patchily distributed species. Additional evidence that feeding plasticity may be a relatively common feature in macroinvertebrate species inhabiting the gla¬ cial flood plain of Val Roseg, was found for Simuliidae (chapter 4), and in a stable isotope study (Zah et al, in press). Clearly, more research is needed to address the important interplay between genetic structure, life history, and the habitat templet on the maintenance and structure of macroinvertebrate populations in harsh envi¬ ronments such as glacial streams. The combination of allozyme electrophoresis or related techniques to examine genetic structure, and of stable isotope analysis and laboratory food preference experiments used here, may provide a promising ap¬ proach for future studies. 7.5 Critical remarks Generally, field studies are hampered by the fact that one can never take as many replicate samples as one would like, and the need to avoid pseudoreplication (e.g., Hurlbert, 1984; Hairston, 1989; Scheiner & Gurevitch, 1993). Some prelimi¬ nary analyses revealed no significant reduction in sample variability if the number of replicate Hess samples was increased from 3 to 5. Samples were generally taken in riffle/run habitats to make samples more comparable. The use of 3 replicates also allowed a higher number of sampling sites to be more intensively investigated. A 126 Synopsis further limiting factor was the use of 100 urn mesh size, instead of the usual 200 u,m or more, which substantially increased sample processing time. In total, 145,196 macroinvertebrates were sorted, enumerated and identified. However, it was worth the effort. First, many studies in alpine glacial streams underestimated total densities because coarser mesh sizes were used. Second, most Nematoda, Harpacticoida and Ostracoda that constitute major components of the benthic communities in groundwater channel habitats would have been missed. Multivariate statistical techniques comprise a thread through the data analy¬ ses in this thesis. These methods provide an invaluable tool for the joint analysis of faunal-environmental relationships. The literature on this topic is tremendous, but for example Dolédec & Chessel (1991), ter Braak & Verdonschot (1995), and Blanc (2000) provide excellent overviews of common methods in aquatic sciences. Clearly, a full understanding of ecology is no longer possible without some knowledge of multivariate analyses. However, multivariate methods are no panacea, they also have opened a Pandora's box (James & McCulloch, 1990). Consequently, their adequacy should be checked in every single case to minimize misapplications and misinterpre¬ tations of results, and they also should be accompanied by more „simple measures" and/or graphical representations of the data, yielding a more complete and compre¬ hensive picture. 7.6 References Blanc, L. (2000) Données spatio-temporelles en écologie et analyses multitableaux: examen d'une relation. Ph. D. Thesis. Université Claude Bernard-Lyon, France. Burgherr, P. & Ward., J.V. (submitted) Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps). Journal of the North American Benthological Society. Chapin III, F.S. & Körner, C (1994) Arctic and alpine biodiversity: patterns, causes, and ecosystem consequences. Trends in Ecology and Evolution, 9, 45-47. Connell, J.H. (1978) Diversity in tropical rain forests and coral reefs. Science, 199, 1302-1310. Dolédec, S. &c Chessel., D. (1991) Recent developments in linear ordination meth¬ ods for environmental sciences. Advances in Ecology, India, 1, 133-155. Füreder, L. (1999) High alpine streams: cold habitats for insect larvae. Cold adapted organisms. Ecology, physiology, enzymology and molecular biology (eds. R. Margesin & F. Schinner), pp. 181-196, Springer-Verlag, Berlin. Füreder, L., Schütz, C, Burger, R. & Wallinger, M. (1998) Higbalpine streams as models for ecological gradients. Hydrology, water resources and ecology in headwaters (eds. K. Kovar, U. Tappeiner, N.E. Peters & R.G. Craig), pp. 387-394, IAHS Press, Wallingford, UK. Synopsis 127 Gislason, G.M., Olafsson, J.S. & Adalsteinsson, H. (1998) Animal communities in Icelandic rivers in relation to catchment characteristics and water chemistry. Preliminary results. Nordic Hydrology, 29, 129-148. Hairston, N.G. (1989) Ecological experiments: purpose, design, and execution. Cam¬ bridge University Press, Cambridge, UK. Huribert, S.H. (1984) Pseudoreplication and the design of ecological field experi¬ ments. Ecological Monographs, 54, 187-211. James, F.C & McCulloch, CE. (1990) Multivariate analysis in ecology and system- atics: panacea or pandora's box. Annual Review of Ecology and Systematics, 21, 129-166. Kownacka, M. & Kownacki, A. (1972) Vertical distribution of zoocenoses in the streams of the Tatra, Caucasus and Balkan Mts. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 18, 742-750. Kownacki, A. (1991) Zonal distribution and classification of the invertebrate com¬ munities in high mountain streams in South Tirol (Italy). Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 24, 2010-2014. Kownacki, A. & Kownacka, M. (1973) The distribution of the bottom fauna in several streams in the Middle Balkan in the summer period. Acta Hydrobiologia, 15, 295-310. 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. Milner, A.M. & Petts, G.E. (1994) Glacial rivers: physical habitat and ecology. Fresh¬ water Biology, 32, 295-307. Saether, O.A. (1968) Chironomids of the Finse area, Norway, with special reference to their distribution in a glacier brook. Archiv für Hydrobiologie, 64, 426- 483. Scheiner, S.M. & Gurevitch, J. (1993) Design and analysis of ecological experiments. Chapman & Hall, New York. Scrimgeour, G.J., Prowse, T.D., Culp, J.M. & Chambers, P.A. (1994) Ecological effects of river break-up: a review and perspective. Freshwater Biology, 32, 261-275. Steffan, A.W. (1971) Chironomid (Diptera) biocoenoses in Scandinavian glacier brooks. The Canadian Entomologist, 103, 477-486. ter Braak, C.J.F. & Verdonschot, P.F.M. (1995) Canonical correspondence analysis and related multivariate methods in aquatic ecology. Aquatic Sciences, 57, 255-289. Tockner, K., Malard, F., Burgherr, P., Robinson, CT., Uehlinger, U., Zah, R. & Ward, J.V. (1997) Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv für Hydrobiologie, 140, 433-463. 128 Synopsis Ward, J.V. (1992) Aquatic insect ecology: 1. biology and habitat. John Wiley Sc Sons, New York. Ward, J.V. (1994) Ecology of alpine streams. Freshwater Biology, 32, 277-294. Ward, J.V., Burgherr, P., Gessner, M.O., Malard, F., Robinson, CT., Tockner, K., Uehlinger, U. & Zah, R. (1998) The Val Roseg Project: habitat heterogeneity and connectivity gradients in a glacial flood-plain system. Hydrology, water resources and ecology in headwaters (eds. K. Kovar, U. Tappeiner, N.E. Pe¬ ters Sc R.G. Craig), pp. 425-432, IAHS Press, Wallingford, UK. Ward, J.V., Malard, F., Tockner, K. Sc 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. Ward, J.V. Sc Stanford, J.A. (1983) The intermediate-disturbance hypothesis: an explanation for biotic diversity patterns in lotie ecosystems. Dynamics ofLotie Ecosystems (eds. T.D Fontaine III Sc S.M Bartell), pp. 347-356, Ann Arbor Science, Ann Arbor, Michigan. Zah, R., Burgherr, P., Bernasconi, S.M. Sc Uehlinger, U. (in press) Stable isotope analysis of macroinvertebrates and their food sources in a glacier stream. Freshwater Biology. w eer/ arr !ea "% {"$• 1 Ç t- Va' w / ? L** ËOSSB^-" f Verdankung 131 Verdankung Ein herzlicher Dank gebührt Prof. Dr. J.V. Ward, der die Arbeit als Doktorvater stets kritisch begleitet und zu deren guten Gelingen viel beigetragen hat. Prof. Dr. J.E. Brittain und Dr. C.T. Robinson haben freundlicherweise das Korrefe¬ rat übernommen und mit ihrer konstruktiven Kritik wesentlich zur Verbesserung dieser Arbeit beigetragen. Florian Malard, Chris Robinson, Klement Tockner, Urs Uehlinger und Rainer Zah von der „Roseg Gruppe" danke ich herzlich für die ausgezeichnete Zusammenarbeit. Ein besonderes Dankeschön gebührt den Gemeinden Pontresina und Samedan, die stets den Zugang zum Untersuchungsgebiet ermöglicht haben, sowie Herrn Plinio Testa und seinen Mitarbeitern vom Hotel Roseggletscher für ihre Gastfreundschaft. Ganz besonders möchte ich mich bei allen Doktorandinnen, Diplomandinnen und Assistentinnen während meiner Dissertationszeit bedanken, deren Unterstützung weit über das Wissenschaftliche herausging, insbesondere Diana Soldo, Mäggi Hieber, Sandra Lass, Rainer Zah, Dave Arscott, Urs Holliger, Frank Sunder, Mike Monaghan, Luana Bottinelli, Monika Winder, Dimitry van der Nat, Nanna Büsing, Birgit Klein und Gabi Meier. Allen Mitgliedern der Abteilung Limnologie und aus den anderen EAWAG-Abtei¬ lungen, die mich in irgendeiner Form unterstützt haben, danke ich ebenfalls. Speziell erwähnt seien: Chris Robinson für die gute Zusammenarbeit im Feld, seine guten Tipps und die kritische Durchsicht meiner Manuskripte. Mäggi Hieber für das Korrekturlesen und die Kontrolle des Layouts. Richard Uli und Bruno Ribi für die Analyse zahlreicher Wasserproben. Tina Dambone-Bösch hatte die elektrophoretischen Analysen durchgeführt. Die Mitarbeiterinnen der EAWAG-Bi- bliothek, der technischen Dienste und der EMPA-Garage, die mich stets hilfsbereit unterstützt haben. Ein besonderer Dank gebührt auch Dr. Stefano Bernasconi vom geologischen Insti¬ tut der ETH Zürich, der die stabilen Isotopen-Analysen durchführte. 132 Verdankung Ganz herzlich danken für ihre Unterstützung und ihr Vertrauen möchte ich auch meinen Eltern, meinem Zwillingsbruder Stefan und allen meinen Freunden und Kol¬ legen, die mich in diesem Lebensabschnitt begleitet haben, insbesondere (in alphabe¬ tischer Reihenfolge): Dave und Yeda Arscott, Stephan Bernath, Heini Eisenmann, Mäggi Hieber, Urs Holliger, Michael Hütte, Barbara Känel, Sandra Lass und Holger Bertram, Florian Malard, Christoph Matthäi, Elisabeth Meier, Mike Monaghan, Fran¬ ziska Pfister, Wolfgang Riss, Chris Robinson, Gaby Schweizer, Frank Sunder, Diana und Seba Soldo, Matthias Voisard, Bettina Wagner, Rainer Zah... und alle, die hier nicht mit Namen aufgeführt sind. Curriculum Vitae 135 Curriculum Vitae Peter Burgherr born 19 August 1969 in Schaffhausen (Switzerland) home address: Breitenaustrasse 163, CH-8200 Schaffhausen, Switzerland e-mail: [email protected], [email protected] Education 1977 - 1983 6 years primary school in Schaffhausen, Switzerland 1983 - 1985 2 years junior high school in Schaffhausen, Switzerland 1985 - 1990 5 years high school (Gymnasium) in Schaffhausen, Switzerland Matura Typus C (A-Levels) 1990 - 1995 Study of Biology at the Swiss Federal Institute of Technology in Zurich (ETH), Switzerland Dipl. Natw. ETH Biologische Richtung (Master of Science) Diploma thesis: Die Struktur der Makroinvertebraten-Gesellschaft und Habitatscharakteristika im Längsverlauf des Neckers (SG) (Structure of the macroinvertebrate community and habitat char¬ acteristics along the river Necker (SG)) Supervision: Prof. Dr. E.I. Meyer 1996 - 2000 Ph. D. Thesis at the Swiss Federal Institute of Technology (ETH) in Zurich, Switzerland conducted at the Swiss Federal Institute of Environmental Science and Technology (EAWAG) Title: Spatio-temporal community patterns of lotie zoobenthos across habitat gradients in an alpine glacial stream ecosystem Supervision: Prof. Dr. J.V. Ward Teaching assistantships 7/1994 -9/1996 ETH-Assistant at the Department of Limnology, EAWAG 2/1998 -3/2000 ETH-Assistant at the Department of Limnology, EAWAG Professional affiliations Schweizerische Gesellschaft für Hydrologie und Limnologie (SGHL) North American Benthological Society (NABS) 136 Curriculum Vitae Publications in peer reviewed journals and books Burgherr, P. Sc Meyer, E.I. (1997) Regression analysis of linear body dimensions vs. 101- dry mass in stream macroinvertebrates. Archiv für Hydrobiologie, 139, 112. Tockner, K., Malard, F., Burgherr, P., Robinson, CT., Uehlinger, U., Zah, R. Sc Ward, J.V. (1997) Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg Switzerland). Archiv für Hydrobiologie, 140, 433-463. Ward, J.V., Burgherr, P., Gessner, M.O., Malard, F., Robinson, CT., Tockner, K., Uehlinger, U. Sc Zah, R. (1998) The Val Roseg project: habitat heterogeneity and connectivity gradients in a glacial flood-plain system. Hydrology, water resources and ecology in headwaters (eds. K. Kovar, U. Tappeiner, N.E. Pe¬ ters Sc R.G. Craig), pp. 425-432, IAHS Press, Wallingford, UK. Eisenmann, H., Burgherr, P. Sc Meyer, E.I. (1999) Spatial and temporal heterogene¬ ity of an epilithic streambed community in relation to the habitat templet. Canadian Journal of Fisheries and Aquatic Sciences, 56, 1452-1460. Robinson, CT. Sc Burgherr, P. (1999) Seasonal disturbance of a lake outlet benthic community. Archiv für Hydrobiologie, 145, 297-315. Robinson, CT., Rushforth, S.R. Sc Burgherr, P. (2000) Seasonal effects of distur¬ bance on a lake outlet algal assemblage. Archiv für Hydrobiologie, 148, 283- 300. Manuscripts in press Burgherr, P. &C Ward, J.V. (in press) Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 27. Burgherr, P. &C Ward, J.V. (in press) Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps). Fresh¬ water Biology. Burgherr, P. Ward, J.V. &c Glatthaar, R. (in press) Diversity, distribution and seasonality of the Simuliidae fauna in a glacial stream system in the Swiss Alps. Archiv für Hydrobiologie. Malard, F., Lafont, M., Burgherr, P. Sc Ward, J.V. (in press) Longitudinal distribu¬ tion of benthic and hyporheic oligochaetes in a glacial river. Arctic, Antarctic and Alpine Research. Zah, R., Burgherr, P., Bernasconi, S.M. Sc Uehlinger U. (in press) Contribution of organic resources for a glacial stream (Val Roseg, Swiss Alps) - a stable iso¬ tope study. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie, 27. Curriculum Vitae 137 Zah, R. Burgherr, P., Bernasconi, S.M. Sc Uehlinger, U. (in press) Isotope analysis of macroinvertebrates and their food sources in a glacier stream. Freshwater Biology. Submitted manuscripts Burgherr, P., Ward, J.V. Sc Robinson, C.T. Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps). Sub¬ mitted to Journal of the North American Benthological Society. Burgherr, P., Robinson, CT. Sc Zah, R. Genetic structure and feeding plasticity of Acrophylax zerberus (Trichoptera: Limnephilidae) in a Swiss alpine flood plain. Submitted to International Review of Hydrobiology. Other publications Burgherr, P. (2000) Struktur der Wirbellosenfauna an der Flusssohle eines voralpinen Fliessgewässers (Necker SG). Berichte der St. Gallischen Naturwissen¬ schaftlichen Gesellschaft, 137, 137-154. Papers presented at scientific meetings Burgherr, P. (1997) Heterogenität der benthischen Makroinvertebraten-Gesellschaft in einem komplexen, alpinen Fliessgewässersystem. Annual meeting of the Swiss Hydrological and Limnological Society (SGHL), La Chaux de Fonds, Switzerland Burgherr, P. Sc Ward, J.V. (1998) Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. 27th SIL congress, Dublin, Ireland. R. S.M. Sc U. Allochthonous Zah, , Burgherr, P., Bernasconi, Uehlinger, (1998) detritus as an organic resource for a glacial flood plain (Val Roseg) - a stable isotope study. 27th SIL congress, Dublin, Ireland. Burgherr, P., Robinson, CT. Sc Zah, R. (1999) Feeding behavior and genetic struc¬ ture of Acrophylax zerberus (Trichoptera: Limnephilidae) populations in an alpine flood plain. Annual meeting of the North American Benthological Society (NABS), Duluth, Minnesota, USA. Burgherr, P. Sc Ward, J.V. (2000) Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine stream (Val Roseg, Swiss Alps). Annual meet¬ ing of the North American Benthological Society (NABS), Keystone, Colo¬ rado, USA. 735 Curriculum Vitae Posters presented at scientific meetings Burgherr, P., Ward, J.V. Sc Tockner, K. (1998) The Val Roseg Project: Heterogene¬ 1998 - Inter¬ ity of surface zoobenthos across habitat gradients. Headwater in Head¬ national conference on: Hydrology, Water Resources and Ecology waters. Meran, Italy. Burgherr, P., Robinson, CT. Sc Rushforth, S.R. (1998) Seasonal effects of distur¬ of the North Ameri¬ bance on lake outlet benthic assemblages. Annual Meeting Prince Edward can Benthological Society (NABS,). Charlotteville, Island, Canada. o@it@ L©@r / Blank leaf