Alpine Streams: Aspects of Biocomplexity - Eawag: Personal

Share

Transcript

1 Diss. ETH No. 14601 Alpine Streams: Aspects of Biocomplexity A dissertation submitted to the SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH for the degree of DOCTOR OF NATURAL SCIENCES presented by M ARGIT HIEBER Diplom-Biologist University of the Saarland, Germany born 28 April 1970, Germany accepted on the recommendation of Prof. Dr. J. V. Ward, examiner PD Dr. C. T. Robinson, co-examiner Dr. U. Uehlinger, co-examiner Dr. A. Milner, co-examiner Zurich, Switzerland 2002

2 This thesis was conducted at the Swiss Federal Institute for Environmental Science and Technology (EAWAG/ETH) in the Department of Limnology in Dbendorf, Switzerland. The study was partially funded by a Swiss National Science Foundation Grant (no. 31-50440.97).

3 Chapters 2 and 3 are published Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2002). Are alpine lake outlets less harsh than other alpine streams? Archiv fr Hydrobiologie, 154 (2): 199-223. Hieber, M., Robinson C.T., Rushforth S.R., and U. Uehlinger. (2001). Algal communities associated with different alpine stream types. Arctic, Antarctic, and Alpine Research 33 (4): 447-456. Chapters 5 and 6 have been submitted Hieber, M., Robinson C.T., and Uehlinger U. Seasonal and diel patterns of invertebrate drift in different types of alpine streams. Freshwater Biology, submitted. Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. Colonization dynamics of macroinvertebrates in alpine streams and lake outlets. Journal of the North American Benthological Society, submitted. Chapter 4 is in preparation Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. Macroinvertebrates in alpine streams: community patterns in relation to the habitat templet.

4 Contents Summary ........................................................................................ 1 Zusammenfassung ............................................................................. 5 1. Introduction ............................................................................... 9 2. Are alpine lake outlets less harsh than other alpine streams? .....................17 3. Algal communities associated with different alpine stream types ................47 4. Macroinvertebrates in alpine streams: community patterns in relation to the habitat templet .....................................................................69 5. Seasonal and diel patterns of invertebrate drift in different types of alpine streams.......................................................................... 109 6. Colonization dynamics of macroinvertebrates in alpine streams and lake outlets................................................................................... 137 7. Synopsis ................................................................................. 165 Terms ........................................................................................ 171 Acknowledgments .......................................................................... 173 Curriculum Vitae ............................................................................ 175

5 Summary 1 Summary Habitat heterogeneity and benthic algal and macroinvertebrate communities were investigated in different alpine streams in the Swiss Alps, comprising the following four types: rhithral streams, rhithral lake outlets, kryal streams, and kryal lake outlets. The main objectives of the study were: (1) to characterize the habitat and biota of different alpine stream types; (2) to relate the biotic communities of these diverse habitats with respective environmental conditions; and (3) to investigate the drift and colonization patterns of benthic invertebrates in relation to respective habitat characteristics of the different stream types. Special emphasis was put on the influence of an upstream lake and the differences between lake outlets and alpine streams not affected by lakes. Chapter 1 presents a general introduction on the importance and the distinct characteristics of alpine environments and stream systems and introduces the questions and goals of each chapter. Chapter 2 focuses on the spatial and temporal heterogeneity in environmental conditions that determine the habitat templet of the different stream types. A primary goal was to examine whether alpine lake outlets are relatively benign lotic habitats in the harsh alpine environment, as expected from studies of lowland lake outlets. Indeed, the examined alpine lake outlets differed from respective streams, for example, in having higher maximum water temperatures and lower daily temperature fluctuations. However, the presence of a glacier and the glacial flow pulse were primary determinants of the environmental conditions characterizing both kryal streams and kryal lake outlets: low channel stability, low water temperature, and distinct seasonal changes in water chemistry. Spatial and temporal patterns of benthic algal communities and their relation to env ironmental conditions are described in chapter 3. Algal communities were generally similar among alpine streams and lake outlets and were dominated by diatoms, blue-green algae, and Hydrurus foetidus (chrysophyte) at most sites. However, algal communities tended to be more diverse and less seasonally variable at the rhithral lake outlets, being mainly characterized by the diatom genera Amphora, Denticula, Fragilaria,

6 2 Summary Gomphonema, Nitzschia, and Synedra and the blue-green algae Oscillatoria and Phormidium. Kryal streams, in contrast, generally had the fewest algal taxa and were dominated by Chamaesiphon (blue-green) and Hydrurus foetidus. Thus, algal communities reflected the differences in habitat characteristics of the different alpine stream types, i.e., the seasonality of glacial melt and the stabilizing effect of rhithral lakes. The following chapters focus on different aspects of benthic macro- invertebrates in relation to the habitat characteristics of the different stream types. Chapter 4 describes distribution patterns and the structure of the macro- invertebrate communities. Although assemblage composition was similar among alpine stream types, benthic invertebrates showed distinct patterns in the relative contribution of individual taxa among rhithral streams, rhithral lake outlets and kryal sites, but with no separation between kryal streams and kryal lake outlets. Dominant taxa at all sites were the Chironomidae. Rhithral streams contained the most diverse assemblages, being inhabited by both non-insect taxa as well as many EPT (Ephemeroptera, Plecoptera, Trichoptera) taxa and other dipteran groups. Rhithral lake outlets primarily had higher water temperature and standing crop of algae than rhithral streams, and thus provided a more favorable habitat for non-insect taxa, as exemplified by high densities of Oligochaeta, Nemathelminthes and Crustacea. The presence of a glacier influenced the invertebrate communities of both kryal streams and kryal lake outlets sites; both had generally low taxon richness but surprisingly high abundances of Ephemeroptera and Plecoptera. Drift is a major dispersal mechanism that enables aquatic organisms to escape unf avorable conditions, move among patches and colonize new habitats. In chapter 5, I report on the research that investigated the seasonal and diel patterns of drifting invertebrates in the different alpine stream types. Density and taxon richness of drifting invertebrates were unexpectedly low at all sites. On average, less than 30 % of the identified benthic taxa were found in the drift with Chironomidae dominating both benthic and drift samples. Although drift densities tended to be highest in spring at rhithral sites and in autumn at kryal sites, no distinct and consistent seasonal patterns were evident. In contrast to common patterns known from studies of lowland streams, densities of drifting

7 Summary 3 invertebrates also showed no diel periodicity; this may be a general feature of high altitude streams. Invertebrate drift in alpine streams appeared to be primarily influenced by abiotic habitat conditions. Colonization, as an important ecosystem process for population persistence and community maintenance, was examined in the research reported in chapter 6. Colonization of small-scale patches (i.e., cages) was rapid at all sites, showing no significant increases in density or taxon richness after only eight days. However, colonization dynamics differed among streams as well as among individual cages. Multiple regression analysis indicated that local stream conditions such as flow regime primarily influenced the diversity of colonists, whereas the small-scale distribution of benthic organic matter significantly affected the density of colonists. Despite the high colonization rates, high beta diversities (taxa turnover) among sampling dates and a low proportion of the available taxa being present in the cages suggested that community assembly was still in nonequilibrium even after 30 days (i.e., assemblages were in a continuous state of redistribution). The low density and richness of drifting invertebrates further indicated that active larval movement through swimming and crawling probably is the dominant mode of colonization of local habitat patches in alpine streams. Thus, colonization of local patches in alpine streams reflects the continuous searching by benthic invertebrates for resources. Chapter 7 presents a synthesis of the individual studies and highlights further research needs. The present results emphasize that the complexity of alpine streams are controlled by the interaction of diverse environmental features acting at different hierarchical levels. The distinctive habitat conditions are further reflected by the structure and distribution of the stream communities. However, the distinctiveness of lake outlet communities declines with increasing elevation and glacial influence.

8 4

9 Zusammenfassung 5 Zusammenfassung In der vorliegenden Arbeit wurden Habitateigenschaften und benthische Algen- und Makro invertebratengemeinschaften verschiedener alpiner Fliessgewsser- typen mit den drei folgenden Hauptzielen untersucht: (1) Habitat und Biota der unterschiedlichen Fliessgewssertypen zu charakterisieren; (2) das Verhltnis zwischen den Lebensgemeinschaften der verschiedenen Habitate und ent- sprechenden Umweltbedingungen zu analysieren; und (3) Drift und Besiedlung durch benthische Invertebraten im Zusammenhang mit den Habitateigenschaften der unterschiedlichen Fliessgewssertypen zu erforschen. Die Fliessgewsser- typen umfassten verschiedene rhithrale (schneeschmelz-gespeiste) und kryale (gletscherbeeinflusste) Flsse und Seeausflsse in den Schweizer Alpen. Insbesondere wurde der Einfluss flussaufwrts gelegener Seen und die Unterschiede zwischen Seeausflssen und Flssen betrachtet. Kapitel 1 gibt einen generellen berblick ber die Bedeutung und besonderen Eigenschaften alpiner Gebiete und Fliessgewsser und stellt die Ziele und Fragestellungen der einzelnen Kapitel vor. Studien im Flachland haben gezeigt, dass das Vorhandensein eines Sees einen stabilisierenden Einfluss auf den jeweiligen Seeausfluss ausbt. In Kapitel 2 sollte die Gltigkeit dieser Hypothese fr alpine Gewsser berprft werden. Zu diesem Zweck wurde die rumliche und zeitliche Heterogenitt der Umwelt- bedingungen untersucht, die le tztendlich das 'Habitat Templet', d.h. die fr einen Organismus spezifischen Habitatbedingungen, der verschiedenen Fliess- gewssertypen bestimmen. Die untersuchten Seeausflsse unterschieden sich von den jeweiligen Flssen durch hhere maximale Wassertemperaturen und niedrigere tgliche Temperatur schwankungen. Das Vorhandensein eines Gletschers und die damit verbundenen jahreszeitlichen Schwankungen der Gletscherschmelze war jedoch der Hauptfaktor, der die Umweltbedingungen und dadurch das Habitat der Gletscherflsse und -seeausflsse durch seinen Einfluss auf Abflussregime, Stabilitt des Flussbetts, Wassertemperatur und Wasser- chemie bestimmte. In Kapitel 3 wurden rumlich-zeitliche Muster benthischer Algen- gemeinschaften der verschiedenen alpinen Fliessgewssern beschrieben und der

10 6 Zusammenfassung Einfluss unterschiedlicher Umweltparameter untersucht. Im allgemeinen hnelten sich die Algengemeinschaften der alpinen Flsse und Seeausflsse und wurden von Kieselalgen, Cyanobakterien und der Goldalge Hydrurus foetidus dominiert. Die Algengemeinschaften der rhithralen Seeausflsse waren jedoch im Vergleich zu den Gletscherstellen artenreicher, zeigten weniger starke jahreszeitliche Vernderungen und setzten sich vorwiegend aus den Kieselalgen Amphora, Denticula, Fragilaria, Gomphonema, Nitzschia und Synedra, und den Cyanobakterien Oscillatoria und Phormidium zusammen. In Gletscherflssen hingegen kamen die wenigsten Taxa vor und die Algengemeinschaften wurden von Chamaesiphon (Cyanobakterien) und Hydrurus foetidus dominiert. Die Unterschiede in den Habitateigenschaften der einzelnen alpinen Fliessgewsser, d.h. die jahreszeitlichen Schwankungen der Gletscherschmelze und der stabilisierende Einfluss rhithraler Seen, spiegelten sich daher in den Algen- gemeinschaften wider. In den folgenden Kapitel wurden verschiedene Aspekte benthischer Makro- invertebraten in bezug auf Habitateigenschaften der verschiedenen Fliess- gewssertypen untersucht. Kapitel 4 befasste sich mit Struktur und Verteilungs- mustern der Makroinvertebratengemeinschaften. Obgleich die Zusammensetzung der Invertebratengemeinschaften innerhalb der einzelnen alpinen Fliess- gewssertypen hnlich war, unterschieden sich die Gemeinschaften stark im relativen Anteil einzelner Taxa zwischen rhithralen Flssen, rhithralen See- ausflssen und den Gletscherstellen. Keine Unterschiede traten jedoch zwischen Gletscherflssen und Gletscherseeausflssen auf. Chironomiden dominierten an allen Probenahmestellen. Rhithrale Flsse enthielten die artenreichsten Gemeinschaften, die sowohl von 'Nicht-Insekten' als auch von Ephemeropteren, Plecopteren, Trichopteren und Dipteren bestimmt wurden. Rhithrale Seeausflsse wurden hauptschlich durch hhere Wassertemperaturen und Algenbiomasse charakterisiert und boten dadurch ein gnstigeres Habitat fr 'Nicht-Insekten', was sich durch hohe Dichten an Oligochaeten, Nemathelminthes und Copepoden zeigte. Der Einfluss eines Gletschers kontrollierte die Invertebratengemeinschaften der Gletscherflsse und seeausflsse, die vergleichsweise wenige Arten beherbergten, aber Ephemeropteren und Plecopteren in hheren Abundanzen aufwiesen.

11 Zusammenfassung 7 Drift ist einer der Hauptverbreitungsmechanismen, der es aquatischen Organismen ermglicht, ungnstigen Bedingungen zu entkommen, sich von einem Ort zum nchsten zu bewegen, und neue Habitate zu besiedeln. Kapitel 5 erluterte jahres- und tageszeitliche Muster driftender Invertebraten der unter- schiedlichen Fliessgewssertypen. Im Gegensatz zu Flachlandflssen, in denen meist ausgeprgte zeitliche Driftmuster vorherrschen, waren in den untersuchten alpinen Fliessgewssern keine eindeutigen jahreszeitlichen oder Tagesmuster erkennbar. Insgesamt wurden in der Drift geringe Individuendichten und Artenzahlen ermittelt. Im Durchschnitt fanden sich weniger als 30 % der bestimmten benthischen Taxa in der Drift, wobei Chironomiden sowohl die Drift als auch die benthischen Proben dominierten. Es war eine Tendenz zu erhhten Driftdichten im Frhling an den rhithralen und im Herbst an den kryalen Stellen zu beobachten, jedoch ohne eindeutige jahreszeitliche Muster. Fehlende tages- zeitliche Driftmuster scheinen ein generelles Merkmal alpiner Fliessgewsser zu sein. Die Invertebratendrift in alpinen Fliessgewssern schien hauptschlich durch abiotische Eigenschaften des Habitats beeinflusst zu sein. Kapitel 6 befasste sich mit Besiedlungsmustern, die von grosser Bedeutung fr den Fortbestand von Populationen sind. Die Ergebnisse zeigten, dass klein- rumige Segmente innerhalb einer Probenahmestelle in allen untersuchten Fliessgewssern schnell besiedelt wurden und nach nur 8 Tagen keinen weiteren signifikanten Anstieg in Dichte oder Artenzahl aufwiesen. Unterschiede in den Besiedlungsmustern traten jedoch zwischen den jeweiligen Fliessgewssern sowie den einzelnen in dieser Studie exponierten Substratgehusen auf. Multiple Regressionsanalyse zeigte, dass lokale Fliessgewssereigenschaften, wie zum Beispiel das Abflussregime, den Artenreichtum besiedelnder Invertebraten signifikant beeinflussten, whrend sich kleinrumige Unterschiede in der Verbreitung benthischen organischen Materials signifikant auf die Invertebraten- dichte auswirkten. Hohe beta-Diversitten ('taxa turnover') zwischen den einzelnen Probenahmen und das geringe Vorkommen vorhandener Taxa in den Substratgehusen liessen darauf schliessen, dass sich die Invertebraten- gemeinschaften trotz hoher Besiedlungsraten auch nach 30 Tagen nicht im Gleichgewicht befanden. Das bedeutet, dass sich die Invertebraten in einem kontinuierlichen Zustand der Umverteilung befanden. Die geringe Dichte und

12 8 Zusammenfassung Artenzahl driftender Invertebraten verweist zudem darauf hin, dass Insekten- larven alpiner Fliessgewsser kleinrumige Segmente primr durch aktives Fort- bewegen, z.B. schwimmend und krabbelnd, besiedelten. Daher kann die Besiedlung kleinrumiger Segmente in alpinen Fliessgewssern als kontinuierliche Suche der benthischen Invertebraten nach Ressourcen angesehen werden. Kapitel 7 fasste die Ergebnisse der einzelnen Studien zusammen und unterstreicht die Notwendigkeit fr weitere Forschung in alpinen Gebieten. Die vorliegende Arbeit zeigte, dass alpine Fliessgewsser komplexe Syteme darstellen, die durch das Zusammenspiel verschiedener Umweltfaktoren auf unterschiedlichen hierarchischen Ebenen bestimmt werden. Die besonderen Habitateigenschaften wiederum, beeinflussten massgeblich die Zusammen- setzung und Verteilung der benthischen Organismen. Die Besonderheit von See- ausflssen nimmt jedoch mit ansteigender Hhe und zunehmendem Einfluss eines Gletschers ab.

13 1. Introduction 9 1. Introduction Mountains - sensitive ecosystems The UN General Assembly declared the year 2002 as the 'International Year of Mountains' with the goal "to promote the conservation and sustainable development of mountain regions, thereby ensuring the well-being of mountain and lowland communities" (FAO 2000). The mean altitude of the land area of the Globe is 875 m above sea level, 28 % of which rises above 1000 m and 11 percent above 2000 m (Bandyopadhyay et al. 1997, FAO 2000). High mountains can be found on all continents, such as the Andes in South America, the Rocky Mountains in North America, Mount Kilimanjaro and Mount Kenya in Africa, or the Alpine-Himalayan Systems extending east-west from Asia across the Caucasus and the Armenian Highlands to the Alps in Europe. High mountain areas often are equated with alpine regions. The alpine life zone, however, is defined as the zone between the treeline and the permanent snowline regardless of altitude. Treeline ranges in elevation from near sea level close to the poleward ends to above 4000 m a.s.l. around the equator, and thus the alpine life zone can be found at all latitudes (Krner 1999). The term 'Alpine', in contrast, refers to the European Alps at any elevation (Ward 1994). In the European Alps, the alpine zone is situated approximately between 2000 and 3500 m a.s.l., and thus represents a high mountain area. Mountains are often defined as adverse and harsh environments as they are characterized by steep slopes resulting in a frequent mass transfer downslope, intense weathering, and extreme climatic conditions, such as low temperatures and high solar radiation, and frequent natural disturbances such as avalanches, floods and landslides (Mani 1990, FAO 2000). Yet, coupled with limited access and their unique features, mountains are of substantial global importance. They influence and determine the climatology and circulation patterns for large parts of the world, are one of the main sources of fresh water, contain a rich biological diversity with high degrees of endemism, and act as a hub of cultural integrity and heritage (Stone 1992, FAO 2000). However, mountains are highly vulnerable to human and natural ecological imbalance, and thus are sensitive indicators of climatic change (Stone 1992, McGregor et al. 1995). Yet, there is a

14 10 1. Introduction lack of knowledge on mountain ecosystems, despite the fact that specific information on their ecology, natural resource potential and socio-economic activities is essential (United Nations Commission on Sustainable Development 1992). 'Le chateau d'eau' mountains as water towers Mountains are called the 'water towers of the world' (i.e., 'le chateau d'eau') because of their huge storage capacity that provides between 30 and 95 % of the total downstream fresh water (Liniger et al. 1998). Sources of all major rivers in the World are located in mountains (Bandyopadhyay et al. 1997). In the Swiss Alps alone, four major European rivers originate the Rhine, the Rhone, the Danube (via the River Inn) and the Po (via the River Ticino). These rivers carry 67 % of the annual precipitation in Switzerland (an average 1456 mm) out of the country (Bandyopadhyay et al. 1997). For example, the Swiss mountains although covering only 23 % of the total catchment area, contribute between 30 % (in winter) and 70 % (in summer) to the total discharge of the Rhine into the North Sea (Liniger et al. 1998). Within Switzerland, 136 km2 of water is stored in lakes and reservoirs and another 74 km2 in alpine glaciers, collectively representing five times the total annual river outflow from Switzerland (Liniger et al. 1998). However, the use of water for human needs such as drinking water, irrigation and hydro-electric power, requires water abstraction and the construction of reservoirs. Over the last centuries, anthropogenic impacts, together with climate change, significantly altered most stream ecosystems and will have marked impacts in the future (Brittain and Milner 2001). To date, virtually no major Alpine stream or river has remained untouched with 79 % of Alpine streams affected by hydro- electric power alone (Stone 1992, Tdter 1998). Yet, most undisturbed running water segments occur in mountainous regions at high elevation (Ward 1992). Alpine streams heterogeneous habitats Despite the generally extreme climatic conditions and steep slopes, mountain areas represent a heterogeneous environment characterized by great spatial and

15 1. Introduction 11 temporal variability. In initial limnological studies in mountain environments, Steinmann (1907) had already distinguished glacial and non-glacial fed high mountain streams, separating them from 'middle mountain' streams. Steffan (1971), and later Ward (1994), further classified alpine streams into three main types based on the origin of their water: krenal segments fed by groundwater, rhithral segments dominated by snowmelt and rainfall, and kryal segments fed by glacial meltwater (Fig. 1-1). Because of the differences in the water sources and the associated seasonality in ice and snow melt, these alpine stream types differ substantially in their habitat characteristics. Typically, channel stability, as well as accumulated annual temperature (degree days), decreases not only with increasing elevation, but also from krenal to kryal streams as a result of the seasonal and diel fluctuations in flow and temperature regimes and sediment supply (Ward 1994). Lake outlets, being defined as an ecotone forming the longitudinal transition zone between lentic and lotic habitats, represent another unique habitat that differs from other stream types in its abiotic characteristics and associated biotic communities. The water body of the upstream lake tends to buffer strong fluctuations in discharge and temperature regime, and often provides high concentrations of organic material transported downstream (e.g., Illies 1956, Brnmark and Malmqvist 1984). Thus far, most studies have been conducted in low- and midland lake outlets. Extending Ward's (1994) typology, alpine lake outlets can be viewed as subtypes of either rhithral or kryal streams (Fig. 1-1). Glacier Kryal Figure 1-1 Alpine stream types with their major Rhithral Krenal water sources (modified after Ward 1994 and Freder 1999).

16 12 1. Introduction They are expected to attenuate high flow and temperature variations, and mitigate the seasonality in physical-chemical characteristics, thus leading to a relatively stable channel. They are, therefore, assumed to provide more benign habitats for benthic organisms in an otherwise harsh lotic environment (Milner and Petts 1994, Burgherr and Ward 2000). Surprisingly little information exists on alpine lake outlets. Alpine stream biota diverse communities The majority of organisms in alpine streams are benthic (in the sense of being closely associated with the substratum) and comprise biofilm assemblages, macrophytes, invertebrates and some fishes (Ward 1992). The zoobenthic and algae communities of alpine streams are remarkably similar worldwide and, in general, are dominated by distinctive insect and diatom families. The organisms living in alpine environments are subject to unique habitat characteristics, and both plants and animals show diverse adaptations comprising morphological, physiological and behavioral modifications such as dark pigmentation in response to the strong ultraviolet radiation, reduced body size, morphological adaptations to maintain position in fast flowing water, and prolonged development because of cold temperatures, and generally low nutrient and food resources (Mani 1990, Ward 1992). However, early limnologists (Steinmann 1907, Thienemann 1926) had already noticed that mountain streams consist of different biotopes each with a distinct biocoenoses. Kryal streams, for example, are characterized by organisms adapted to high current velocities, water temperatures near 0 C and high seasonal and diel flow fluctuations. The biota typically comprises sparse algae, often dominated by Hydrurus foetidus, and a few species of the insect order Diptera (dominated by the chironomid genus Diamesa). Most species in kryal streams are highly cosmopolitan but restricted in their longitudinal distribution (i.e., stenozonal) (Ward 1994). Communities in rhithral streams, in contrast, consist of species that are moderately cosmopolitan but longitudinally more widely distributed (i.e., euryzonal). They are relatively species rich and include diverse algae, insect and other invertebrate taxa, as well as macrophytes and fishes (Ward 1994).

17 1. Introduction 13 Relatively recently, the perception of distinct species traits as adaptations to the habitat templet (sensu Southwood 1977) resulted in an increased use of freshwater biota to monitor changes in water quality and for assessing the impact of anthropogenic influences (Hawkins et al. 2000). Thus, although little information of this sort exists on the biotic communities of alpine stream systems (Brittain and Milner 2001), knowledge on their communities, ecological needs and the environmental variables determining distribution patterns is necessary. Aspects of biocomplexity - objectives of this study The lack of a strong empirical foundation describing patterns in stream insect biodiversity is disheartening because streams are among the most threatened ecosystems on Earth. (Vinson and Hawkins 1998) There is a great need to improve the current monitoring of mountain water. (Liniger et al. 1998) Motivated by the strong demand for information on alpine stream ecology, we were interested in investigating the habitat characteristics and associated biota in different alpine stream types. Two major goals of community ecology are to recognize patterns in species assemblages and to understand the processes that determine these patterns (Townsend 1989). In this respect, I focused on: (1) the characterization of the habitat and biota of these threatened but still hardly known ecosystems; (2) the relation between the biotic communities of these diverse habitats and their environmental conditions, with special emphasis on the influence of an upstream lake; and (3) the study of mechanistic aspects of the benthic invertebrate communities in respect to the habitat characteristics of the different stream types. The systems studied included alpine streams and lake outlets of rhithral and kryal origin located in three major drainages in the Swiss Alps, situated above the treeline and yet little impacted by humans. Chapter 2 defines habitat characteristics of the different alpine stream habitats that result from an interplay of diverse environmental conditions, and examines spatial and temporal patterns in various environmental parameters.

18 14 1. Introduction Specific attention is given to the thermal properties of the stream types, as well as to differences in the intra-annual chemical and physical properties, channel morphology and channel stability. Chapters 3 and 4 characterize biotic communities (macroinvertebrates and benthic algae, respectively) and their seasonal patterns in relation to environmental conditions of the different alpine stream types. The following two chapters address mechanistic aspects of benthic invertebrate communities regarding drift and colonization patterns. Chapter 5 describes seasonal and diel patterns of invertebrate drift. Drift describes the downstream transport of stream-dwelling organisms and is one mechanism that enables them to escape unfavorable conditions, move among local patches and colonize new habitats. The colonization dynamics of small habitat patches are discussed in chapter 6. Both chapters investigate patterns in the context of differences between stream types and relate them to respective environmental conditions. Finally, chapter 7 summarizes the different results and provides some suggestions for river management in alpine environments. References BANDYOPADHYAY, J., D. KRAEMER, and Z. W. KUNDZEWICZ. 1997. Mountain water resources - the need for their proper assessment. International Academy of the Environment and World Meteorological Organization, Geneva, Switzerland. BRITTAIN, J. E., and A. M. MILNER . 2001. Ecology of glacier-fed rivers: current status and concepts. Freshwater Biology 46:1571-1578. BRNMARK, C., and B. MALMQVIST . 1984. Spatial and temporal patterns of lake outlet benthos. Verhandlungen der Internationalen Vereinigung der Limnologie 22:1986-1991. BURGHERR , P., and J. V. W ARD. 2000. Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung der Limnologie 27:1587-1590. FAO. 2000. International year of mountains. Food and Agriculture Organization of the United Nations, Rome, Italy. FREDER, L. 1999. High alpine streams: cold habitats for insect larvae. Pages 181- 196 in Margesin, R., and F. Schinner (Editors). Cold-Adapted Organisms -

19 1. Introduction 15 Ecology, Physiology, Enzymology and Molecular Biology. Springer, Berlin, Germany. HAWKINS , C. P., R. H. NORRIS , J. GERRITSEN, R. M. HUGHES , S. K. JACKSON, R. K. JOHNSON, and R. J. STEVENSON. 2000. Evaluation of the use of landscape classifications for the prediction of freshwater biota: synthesis and recommendations. Journal of the North American Benthological Society 19:541-556. ILLIES , J. 1956. Seeausflu-Bioznosen lapplndischer Waldbche. Entomologisk Tidskrift 77:138-513. KRNER , C. 1999. Alpine Plant Life: functional plant ecology of high mountian ecosystems. Springer-Verlag, Berlin Heidelberg, Germany. LINIGER , H., R. W EINGARTNER, and M. GROSJEAN. 1998. Mountains of the world: water towers for the 21st century. Paul Haupt AG, Berne, Switzerland. MANI , M. S. 1990. Fundamentals of high altitude biology. Aspect Publications Ltd., London, U.K.. MCGREGOR , G., G. E. PETTS, A. M. GURNELL , and A. M. MILNER . 1995. Sensitivity of alpine stream ecosystems to climate change and human impacts. Aquatic Conservation: Marine and Freshwater Ecosystems 5:233-247. MILNER, A. M., and G. E. PETTS . 1994. Glacial rivers: physical habitat and ecology. Freshwater Biology 32:295-307. SOUTHWOOD, T. R. E. 1977. Habitat, the templet for ecological strategies? Journal of Animal Ecology 46:337-365. STEFFAN, A. W. 1971. Chironomid (Diptera) biocoenoeses in Scandinavian glacier brooks. Canadian Entomologist 103:477-486. STEINMANN, P. 1907. Die Tierwelt der Gebirgsbche - eine faunistisch-biologische Studie. Annales de Biologie Lacustre 2:30-162. STONE , P. B. 1992. The state of the world's mountains - a global report. Zed Books Ltd, London, U.K.. THIENEMANN, A. 1926. Das Leben im Ssswasser. Ferdinand Hirt, Breslau, Deutsches Reich. TDTER , U. 1998. Flsse - gezhmte Natur. Pages 178-182 in CIPRA (Editor). 1. Alpenreport, Daten, Fakten, Probleme, Lsungsanstze. Paul Haupt, Bern, Schweiz.

20 16 1. Introduction TOWNSEND , C. R. 1989. The patch dynamics concept of stream community ecology. Journal of the North American Benthological Society 8:36-50. UNITED NATIONS COMMISSION ON SUSTAINABLE DEVELOPMENT . 1992. Managing fragile ecosystems: sustainable mountain development. United Nations Conference on Environment and Development (UNCED). Earth Summit - Agenda 21, Rio de Janeiro, Brazil. VINSON, M. R., and C. P. HAWKINS . 1998. Biodiversity of stream insects: Variation at local, basin, and regional scales. Annual Review of Entomology 43:271-293. WARD, J. V. 1992. A mountain river. Pages 793-510 in Calow, P., and G. E. Petts (Editors). The Rivers Handbook. Blackwell, Oxford, U.K.. WARD, J. V. 1994. Ecology of alpine streams. Freshwater Biology 32:277-294

21 2. Alpine stream habitat 17 2. Are alpine lake outlets less harsh than other alpine streams? Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2002). Archiv fr Hydrobiologie, 154 (2): 199-223. We investigated the intra-annual chemical and physical properties of 16 lotic systems situated in the Swiss Alps that comprised alpine lake outlets and streams of kryal and rhithral origin. A primary goal of the study was to examine whether alpine lake outlets are less harsh lotic habitats than alpine non-outlet streams, as expected from studies of lowland lake outlets. The presence of a glacier and associated seasonality in glacial melt were primary determinants of environmental conditions in alpine lotic systems through the influence on flow regime, water temperature, and seasonal changes in chemical variables. Lake outlets exhibited higher maximum water temperatures and lower daily temperature fluctuations compared to non-outlet streams. Channel stability of the lotic systems was analyzed using estimates of shear stress, the Pfankuch score (PSI) and a multivariate habitat index (MHI). Indices using shear stress characterized streams with channel slopes = 15 % as least stable and lake outlets with slopes = 4 % as most stable. PSI and MHI mainly separated rhithral sites as stable and kryal sites as unstable, resulting from differences in discharge patterns. None of the indices resulted in distinct differentiation between the habitat stability of lake outlets and respective streams. Assessment of these indices suggested that it is necessary to incorporate both spatial and temporal changes in habitat parameters to provide a valid measure of the stability of a particular alpine stream type. Our results indicated that the habitat characteristics of alpine lotic systems are mainly controlled by the presence of a glacier and its seasonality, and on a lower hierarchical level by the presence of an upstream lake.

22 18 2. Alpine stream habitat Introduction Streams in alpine regions are generally characterized as harsh environments because of short growing seasons, cold turbulent water, and low availability of nutrients and organic matter (Mani 1990, Ward 1994). The thermal regime, being a composite of patterns of absolute temperatures, diel and seasonal amplitudes, as well as rates of change, plays a major role in the ecology and development of aquatic biota (Ward and Stanford 1982). In the Swiss Alps, lake surface water temperature significantly decreased with increasing altitude by about 7 C km-1 (Livingstone et al. 1999). At high altitudes, temperature was found to determine diversity, distribution and abundance patterns, as well as life-history traits of aquatic insects (Freder 1999). Channel stability is another major factor restricting habitat availability for the associated biota (Hynes 1970). Low stability results from steep gradients, extremely variable flows, and highly mobile bed sediments - typical characteristics of headwater streams (Milner and Petts 1994). Despite these general habitat characteristics, alpine lotic systems have been differentiated into distinct types according to their primary water source: krenal streams are fed by groundwater, rhithral streams are snow-fed, and kryal streams are mainly fed by glacial melt (Ward 1994). Rhithral and kryal streams are influenced strongly by seasonal and diel discharge fluctuations leading to turbid water and unstable channels (Milner and Petts 1994). Another stream type with distinct characteristics, alpine lake outlets, has been largely ignored by lotic ecologists (but see Kownacki et al. 1997, Donath and Robinson 2001). Lake outlets can be defined as an 'ecotone' that represents a longitudinal transition zone between lentic and lotic environments (Pinay et al. 1990, Ward 1998). Lowland lake outlets typically show higher water temperature, a more stable discharge regime and channel bed, and higher concentrations of suspended particulate organic material than respective non-outlet streams (Brnmark and Malmqvist 1984, Richardson 1984, Harding 1992, Wotton 1995). Among alpine lotic systems, lake outlets are expected to be moderate in flow and temperature variation and be characterized by relatively stable channels (Milner and Petts 1994). Alpine lake outlets presumably provide more benign habitats for benthic organisms in an otherwise harsh environment. However,

23 2. Alpine stream habitat 19 little is known about the ecology of these distinct habitats (Milner and Petts 1994), and almost no comprehensive year-round studies exist (Burgherr and Ward 2000). Channel stability of alpine streams, glacial streams in particular, is expected to be a major factor influencing instream habitat conditions (Milner and Petts 1994). Indices estimating channel stability mainly are based on measures of critical shear stress or categorization of descriptive stream bottom and bank characteristics (e.g., Pfankuch 1975, Newbury 1984, Duncan et al. 1999). Recent studies have used multivariate analyses to evaluate the overall stability of a site by combining different physical habitat measures into a single multivariate instability score (Death and Winterbourn 1994, Townsend et al. 1997, Burgherr et al. in press). The habitat template (sensu Southwood 1977) of alpine lotic systems, and ultimately the structure and composition of the biota, can be distinguished by a cross-section of physical and chemical parameters. The present data were collected as part of a larger study examining the general ecology of alpine lake outlets and streams of different origin in the Swiss Alps. The primary goal of the present contribution was to classify alpine lake outlets among alpine lotic systems, based on parameters described by the physical and chemical habitat conditions. We specifically asked whether alpine lake outlets are less harsh lotic habitats for associated organisms, as can be expected based on studies of lowland outlets (Richardson and Mackay 1991). Study sites Study sites were situated in the Swiss Alps at or above the treeline with elevations ranging from 1900 to 2500 m a.s.l. (Table 2-1). Sites consisted of alpine lake outlets and non-outlet streams of rhithral and kryal origin situated in 3 of the 4 major drainages in Switzerland (Po, Danube, and Rhine) (Fig. 2-1). Lake surface area ranged from 0.7 to 27 ha. Catchment areas upstream of the study sites ranged from 0.6 to almost 20 km2 (Table 2-1). The largest catchment areas were associated with kryal sites that had 40 to 90 % of the area glaciated. Tschierva and the Lej Roseg outlet originate from two different glaciers but belong to the same catchment of the Roseg River (Tockner et al. 1997). Steinlimi and Steinsee outlet are both in the catchment of the Gadmerwasser, originating

24 20 2. Alpine stream habitat from two adjacent glaciers (King 1987). The Joerisee outlet is rhithral, but the main tributary of this lake is the outlet of a small proglacial lake originating from the Joeriglacier (Kreis 1921). Data on catchment and glaciated area were based on topographic maps updated between 1991 and 1995. Because the glaciers draining into our study streams retreated between 5 and 30 m/y since 1995, the actual glaciated area at the time of the study was less than listed in Table 2-1 (IAHS/UNESCO 1998). Lago Nero, as well as Lago Bianco, are situated in the Cristallina massif and drain into the Bavona River. The land surfaces of all catchments are covered mainly by bare rock, with sparse vegetation of grass and low shrubs of alder (Alnus sp.) and willow (Salix sp.) also present. All sites are in the crystalline area of the Swiss Alps where bedrock mainly consists of granite and gneiss (Spicher 1980). 8 9 10 N ine Rh re Aa Danube 47 SS, JS, SG J G PM M LB, LN LR, TG one Rh 46 0 50 100 km Po Figure 2-1 Map of Switzerland with locations of the study sites () in the Swiss Alps (shaded area). Site notations are defined in Table 2-1.

25 2. Alpine stream habitat 21 Table 2-1 Location and general characteristics of the study sites. Notations of the study sites as further referred to in the text. Origin of the water: k = kryal, r = rhithral; stream type: L = lake outlet, S = (non-lake outlet) stream, d = downstream sampling site of the respective stream or lake outlet. Total catchment area (km 2) and percentage glaciated area (%); D50 = median substratum size; BS = average baseflow, BF = average bankfull flow. Site Notation Origin Stream Drainage Coordinates Elevation Catchment Slope Substrate Water width (m) type eL nW (m a.s.l.) (km2) % (%) D50 (cm) BS BF BF/BS Lago Nero LN r L Po 832'21" 4626'55" 2387 0.8 0 5 23 2.8 3.8 1.4 Puoz Minor PM r L Inn 1001'52'' 4626'23'' 2336 1.8 0 9 21 3.6 5.0 1.4 Lago Bianco LB1 r L Po 831'19" 4627'14" 2076 2.1 0 4 35 6.7 8.6 1.3 B.downstream LB2 r Ld Po 831'40" 4626'54" 2010 3.7 0 11 20 4.2 9.7 2.3 Joerisee JS r(k) L Rhine 958'13" 4646'52" 2489 3.4 11 5 18 8.3 10.3 1.2 J.downstream J1 r Ld Rhine 958'24" 4647'15" 2310 4.6 8 8 17 4.7 5.3 1.1 J.downstream J2 r Ld Rhine 959'03" 4648'26" 1950 10.1 4 4 18 5.9 7.6 1.3 Moesa M1 r S Po 909'12'' 4629'18'' 2300 0.6 0 16 20 2.6 2.9 0.9 M.downstream M2 r Sd Po 909'38'' 4629'13'' 2100 0.9 0 15 15 1.2 1.6 1.3 Gglia G1 r S Inn 944'28'' 4628'47'' 2310 5.8 0 10 20 4.3 6.0 1.4 G.downstream G2 r Sd Inn 944'33'' 4628'20'' 2205 6.5 0 4 23 4.6 4.9 1.1 Lej Roseg LR k L Inn 951'17" 4624'30" 2159 19.3 44 2 31 12.2 16.7 1.4 Steinsee SS k L Rhine 826'02" 4643'44" 1934 7.3 71 4 18 8.0 13.2 1.7 Tschierva TG k S Inn 952'04" 4626'03" 2100 14.7 42 4 17 10.1 18.2 1.8 Steinlimi SG1 k S Rhine 825'00" 4642'51" 2090 3.4 92 2 8 8.9 9.5 1.1 S.downstream SG2 k Sd Rhine 825'47" 4643'35" 1900 7.8 56 11 15 4.7 9.3 2.0 Methods Sixteen sites were sampled monthly (when accessible) for at least 1 year between July 1998 and November 2000. Sampling of the two rhithral streams (Moesa and Gglia) began in summer 1999. Two rhithral lake outlets, two rhithral streams and one glacial-fed stream also were sampled at a downstream site to compare longitudinal changes in habitat properties. Upper stream sites were at similar distances to the source, elevations, and catchment conditions as respective lake outlet sites. At each site a reach of approximately 20 to 30 m was studied. Lake outlets were sampled at a distance of 10 to 30 m below the

26 22 2. Alpine stream habitat lake. In the following text, the term 'stream' refers only to a lotic system without the influence of a lake, a 'lake outlet', in contrast, is a stream flowing from a lake. Physical and chemical measures Catchment area and site elevation were determined from maps (Swiss Map, Bundesamt fr Landestopographie, 1 : 25,000). Slope typically was measured with a clinometer, but where clinometer data were missing it was determined from a topographic map (1 : 25,000) for a 50 to 100 m reach at a sampling site. Substratum size distribution was assessed in the field by measuring the long ('a') axis of, on average, 50 substrate particles from the stream bed. Particles were selected at the tip of every step along a transect to avoid biased sampling. Bankfull and baseflow water width were measured during low flow (autumn) across 5 transects placed 5 to 10 m equidistant within the study reach (Platts et al. 1983). At 8 selected sites (2 sites of each stream type), water depth and velocity (at 0.6 depth; Mini Air 2, 20 mm diameter, Schiltknecht AG, Gossau, Switzerland) were measured across 3 transects at different flows to calculate discharge following Platts et al. (1983). Velocity was integrated over a period of 6 seconds to adjust for variations due to the small propeller diameter and fluctuations in the water current. Continuous discharge records were available only for the Roseg River (gauging station in Pontresina at 1766 m a.s.l., Swiss National Hydrological and Geological Survey) and the Jrisee (temporary gauging station at the outlet at 2489 m a.s.l., Amt fr Umwelt Graubnden) and were standardized to a catchment size of 1 km2 to compare relative instead of absolute discharge patterns. Water temperature was recorded hourly with temperature loggers (Minilog, Vemco, Nova Scotia, Canada) at each sampling site except for 3 downstream sites (LB2, J2 and SG2). Hourly measurements were used to calculate annual temperature range (min - max), diel temperature amplitude, and mean daily temperature. We used the coefficient of variation (CV, in %) to express the annual variation in water temperature. Total annual degree days (DDa) were estimated as the sum of the average daily temperatures over one year. The rate of thermal increase (DD/d) in summer was calculated by computing the slope of

27 2. Alpine stream habitat 23 the linear regression through the accumulated average daily temperature data for 1 to 4 months depending on linearity. Specific conductance (WTW model LF 325, Weilheim, Germany) and turbidity (Cosmos, Zllig AG, Rheineck, Switzerland; in nephelometric turbidity units, NTUs, which is calc ulated by measuring the dispersion of a light beam passed through a sample of water) were measured in the field. Water samples (1 liter) were taken at monthly intervals (when possible), filtered through pre-ashed glass fibre filters (Whatman GF/F filters; 45 mm diameter) and analyzed for the following chemical parameters in the laboratory: ammonium (NH4-N); nitrite (NO2-N); nitrate (NO3-N); total dissolved (TDN) and particulate nitrogen (PN); soluble reactive (SRP), total dissolved (TDP) and particulate phosphorus (PP); dissolved (DOC) and particulate organic carbon (POC); total inorganic carbon (TIC); total suspended solids (TSS) and ash-free dry mass (AFDM). Determination of each parameter followed methods described in detail in Tockner et al. (1997). Assessment of channel stability Channel stability was evaluated based on calculations of different stability indices. The Instability index (ISI), Riffle stability index (RSI), and the percentage of the streambed in motion at bankfull discharge (PBF) are different functions based on measures of shear stress. The shear stress acting on a column of unit area equals the tractive force () and is calculated using the formula = pdS (kg/m2), where p = the density of water (kg/m3), d = depth of flow (m), and S = slope (m/m) (Newbury 1984). The Instability Index indicates the sensitivity of a substrate particle to the tractive force by dividing by the median substrate size (Cobb and Flannagan 1990). Riffle stability index (Kappesser 1993) and the percentage of the streambed in motion at bankfull discharge (Duncan et al. 1999) estimate the percentage of the streambed in motion at bankfull discharge by comparing the critical grain size (Dcrit ) to the cumulative particle size distribution. The critical grain size indicates the largest size of the surface bed material that will be in movement at bankfull discharge and is calculated using the formula Dcrit = 9.821000RS-1 (mm), where R = hydraulic radius (m) and S = slope (m/m). Duncan's PBF additionally accounts for steep bed slopes, heterogeneous sediment sizes, and low water depth to

28 24 2. Alpine stream habitat particle size ratios. The bottom component of the Pfankuch stability score (PSI) classifies the submerged portion of a mountain stream channel into 4 categories of stream bottom stability (= 15 excellent, 16-30 good, 31-45 fair, 46-60 poor) using observations of substrate shape, color, consolidation, distribution, scouring and vegetation (Pfankuch 1975). Normally, the Instability index, the percentage of the streambed in motion at bankfull discharge and the critical grain size are based on the b-diameter of substrate particles. However, we inadvertently calculated these indices with the a-diameter, thus values should be used only to compare relative channel stability within our sites. In addition, a Multivariate Habitat Index (MHI) was calculated to evaluate the overall stability of the 8 selected sites similar to that described by Death and Winterbourn (1994) and Burgherr (in press). This habitat index equals the score for each site using the first axis of a principal component analysis (PCA) from 7 habitat parameters (similar to those used by Death and Winterbourn 1994, Death 1996). PCA orders the sites by reducing the dimensions of multiple parameters in a graphic space in which the axes are gradients of combinations of the parameters (James and Mcculloch 1990). The individual site scores for axis 1, which represents the dominant parameters, were scaled to positive values so that higher MHI values indicated lower overall stability. The Multivariate Habitat Index combines parameters representing channel stability (tractive force, Pfankuch stability score), as well as spatial and temporal habitat heterogeneity (ratio of bankfull to baseflow width, CV in substratum size, and range in water depth, current velocity and water temperature). Statistical analysis Differences in physical-chemical variables between sites and stream types (rhithral and kryal streams, lake outlets, and downstream sites) were analyzed using one-way analysis of variance with time (season) as covariate (ANCOVA) followed by Tukey's HSD post-hoc test after log10(x+1) transformation (Zar 1984). Physical-chemical data of the sampling sites also were analyzed using a principal component analysis (PCA) after log10(x+1) transformation and standardization of variables (CANOCO 4.02).

29 2. Alpine stream habitat 25 Stability indices were standardized by scaling between 0 and 1 using the equation xij = (zij - minj) / (maxj - minj), where zij is the value of the jth index for the ith site, and maxj and minj are the maximum and minimum values for the jth index. Kendall's coefficient of rank correlation K was used to compare the ranking of all sites among each index of stability (PSI, , ISI, RSI, PBF, MHI); significance of correlation was indicated with the p-level. Kendall's coefficient K equals Pearson's correlation coefficient r used for normal distributed variables but is a nonparametric technique similar to Spearman's R, which represents the probability that two variables are ranked in the same order (Sokal and Rohlf 1995). Results Discharge and channel morphology Study sites were characterized by high slopes, ranging from 2 to 16 % with lake outlets generally having lower gradients (Table 2-1). Median substratum size (a-axis) ranged from 8 to 35 cm, and 80 % of all measured particles were between 0.6 Joerisee Lej Roseg Relative discharge (m3 s -1 km-2) 0.4 0.2 0.0 Jan Apr Jul Okt Jan 1998 Figure 2-2 Relative discharge for a standardized catchment size (1 km 2) of the rhithral Jrisee outlet and the kryal Lej Roseg outlet. Gaps are missing discharge values.

30 26 2. Alpine stream habitat 5 cm (10th percentile) and 51 cm (90th percentile) long. Stream baseflow and bankfull water widths ranged from 1.2 to 12 m and 1.6 to 17 m, respectively, with greatest channel widths at the kryal sites. The ratio of bankfull to baseflow water width as a measure of flow fluctuation was, on average, lowest at rhithral lake outlets (= 1.4) (Table 2-1). Average and maximum water velocities, and discharge measurements were highest at the kryal sites (streams and lake outlets) and lowest at the rhithral streams. Single discharge measurements during July to October ranged from 0.01 m3/s at Moesa upstream (catchment < 1 km2) to 4.7 m3/s at Lej Roseg ( catchment > 19 km2). Relative discharge patterns from a representative rhithral (Jrisee) and kryal (Lej Roseg) lake outlet illustrate the flow differences between snow melt fed versus snow and ice melt fed systems (Fig. 2-2). Discharge patterns for both systems during the period of snow melt (May-June) were quite similar, both in their general pattern and in their maximum flow peak relative to a catchment size of 1 km2. Discharge peaked at the beginning of June and rapidly decreased at the end of June for both types. In summer, however, discharge decreased to baseflow at the rhithral lake outlet, whereas it increased again at the kryal lake outlet, reflecting glacial ice melt in summer with peaks attributed to precipitation from rainfall (Fig. 2-2). Thermal patterns Water temperature reached highest maximum values (> 13 C) at rhithral lake outlets and rhithral downstream sites (Table 2-2). Kryal lake outlets had higher maximum temperatures than respective kryal streams. Total annual degree days (DDa) showed similar patterns with highest values for rhithral lake outlets and rhithral streams (> 1000 DDa), below 800 DDa for glacial-influenced lake outlets, and < 400 DDa for kryal streams. The exception was the glacial-influenced Jrisee outlet that differed from the other rhithral lake outlets but corresponded to the kryal lake outlets (Table 2-2). The rate of thermal increase in summer was highest in rhithral lake outlets, including the glacial influenced Jrisee outlet (> 8 DD/d), and it was lowest in kryal streams (< 1.5 DD/d). Diel temperature amplitudes were highest at downstream sites (maximum > 8 C) and lowest at kryal influenced sites (including the rhithral Jrisee outlet;

31 2. Alpine stream habitat 27 Table 2-2 Thermal characteristics of the study sites. Site notations and stream types defined in Table 2-1. Increase = maximal increase per day, DDa = annual degree days, CVa = annual coefficent of variation, Admax = maximum daily temperature amplitude, NA = not available. Site Type Range Increase DDa CVa Admax (C) (DD/d) (%) (C) LN rL 0 - 14.8 10.6 1228 139 3.4 PM rL 0 - 13.5 8.5 NA 196 4.6 LB1 rL 0 - 13.3 8.6 1257 97 5.6 JS rL 0 - 16.2 8.4 735 110 7.4 J2 rSd 0 - 17.7 8.7 812 104 8.1 M1 rS 0 - 12.0 7.6 1224 77 5.7 M2 rSd 0 - 16.6 8.3 1181 102 8.7 G1 rS 0 - 10.7 5.4 987 75 6.9 G2 rSd 0 - 12.8 6.3 1025 109 8.5 LR kL 0 - 9.1 3.9 750 73 3.5 SS kL 0 - 4.4 2.7 594 62 1.7 TG kS 0 - 4.6 1.4 312 85 3.6 SG1 kS 0 - 3.9 1.1 341 101 3.9 maximum < 4 C), but showed no distinct differences between stream and lake outlet sites (Table 2-2). Seasonality in temperature was most pronounced in rhithral lake outlets and least in kryal streams (Fig. 2-3). Temperature CVs were highest at rhithral lake outlets (97 to 196 %) and lowest at kryal lake outlets and rhithral streams (CV < 80 %), although not being significantly different between types (p = 0.095; Table 2-2). Downstream sites had lower CVs than upstream lake outlets but higher CVs than upstream stream sites. Water temperature also displayed strong daily fluctuations during summer (Fig. 2-3, insets). Diel patterns were more pronounced for stream sites (rhithral and kryal) with maximum water tempera- tures in early afternoon (15:00 - 16:00), whereas lake outlet sites displayed less clear patterns with maximum temperatures later in the evening (18:00 - 20:00).

32 28 2. Alpine stream habitat 15 a) LB1 b) LR 10 10 JS SS 5 5 10 0 0 23.8.99 30.8.99 23.8.99 30.8.99 Average daily temperature (C) 5 0 15 c) G1 d) TG M1 10 10 SG1 5 5 10 0 0 23.8.99 30.8.99 23.8.99 30.8.99 5 0 Jul-98 Jan-99 Jul-99 Jan-00 Jul-00 Jul-98 Jan-99 Jul-99 Jan-00 Jul-00 Figure 2-3 Average daily temperature (large graphs) and diel fluctuations during seven days in late August (small graphs; hourly measurements) of two representative study sites of each lotic type: a) rhithral lake outlet, b) kryal lake outlet, c) rhithral stream, and d) kryal stream. Site notations are defined in Table 2-1. Gaps indicate time periods when loggers were exposed to air. Physical-chemical characteristics NO2-N, SRP, TDP, DOC, POC, and AFDM concentrations generally were low at all sites (Table 2-3). However, kryal sites, in general, had higher concentrations of turbidity, NH4-N, NO2-N, SRP, TDP, PP, TSS and AFDM, and a more pronounced seasonality than rhithral sites that was associated with discharge regime. Turbidity significantly differed between rhithral sites (maximum values < 10 NTU except at the Joerisee outlet and its downstream sites) and kryal sites (maximum values > 100 NTU) (p < 0.001) (Table 2-3). At kryal sites, turbidity displayed strong seasonal patterns with maximum values during summer and autumn (Fig. 2-4). Variation was less pronounced in kryal lake outlets relative to kryal streams where turbidity can increase rapidly during times of ice melt

33 2. Alpine stream habitat 29 Table 2-3 Mean physical-chemical parameters of the different stream types: r = rhithral, k = kryal; L = lake outlet, S = (non-lake outlet) stream, d = downstream site. Cond. = specific conductance, NH 4-N = ammonium, NO2-N = nitrite, NO3-N = nitrate, TDN = total dissolved nitrogen, PN = particulate nitrogen; SRP = soluble reactive phosphorus, TDP = total dissolved phosphorus, PP = particulate phosphorus, DOC = dissolved organic carbon, POC = particulate organic carbon, TIC = total inorganic carbon, TSS = total suspended solids, AFDM = ash-free dry mass.; p = significance level (1-way ANCOVA); significant differences of each parameter between stream types are indicated with small letters. Type Cond. Turbidity NH4-N NO2-N NO3-N TDN PN SRP TDP PP DOC POC TIC TSS AFDM (S/cm) (NTU) (g/l) (g/l) (g/l) (g/l) (g/l) (g/l) (g/l) (g/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) p 0.11

34 30 2. Alpine stream habitat high flow, but reached maximum concentrations at rhithral sites in early summer corresponding to snow melt fed discharge. Mean nitrate (NO3-N) concentrations ranged from 86 to 363 g/l and reached maximum values of almost 600 g/l at Tschierva (Table 2-3). Total dissolved nitrogen showed similar patterns as nitrate, although at slightly higher concentrations. Mean particulate nitrogen ranged from 8 to 59 g/l and was highest at the two Jrisee downstream sites and lowest in rhithral streams. Concentrations of particulate phosphorus (PP) and total suspended so lids (TSS) were low at rhithral sites (except Jrisee outlet and its downstream sites), but reached high mean and maximum concentrations in kryal sites. A glacial influence also was perceptible at Jrisee outlet with maximum concentrations of 16 g PP and 24 mg TSS /l, but decreased downstream of the lake outlet (maximum 2 g PP and 4 mg TSS /l). At kryal sites, PP, AFDM and TSS showed a pronounced seasonality with peak concentrations during summer high flow. Total Rhithral lake outlets Kryal lake outlets Rhithral streams Kryal streams 3000 a) LN b) LR c) J1 d) TG 2000 PM SS J2 SG1 Turbidity (NTU) LB1 M1 SG2 400 JS M2 300 G1 G2 200 LB2 100 Specific conductance (S cm -1 ) 0 200 a) b) c) d) 150 100 50 0 100 a) b) c) d) NH4-N (g l -1 ) 80 60 40 20 0 Jul-98 Jul-99 Jul-00 Jul-98 Jul-99 Jul-00 Jul-98 Jul-99 Jul-00 Jul-98 Jul-99 Jul-00 Figure 2-4 Turbidity, specific conductance, and ammonium (NH 4-N) concentrations of all study sites: a) rhithral lake outlets, b) kryal lake outlets, c) rhithral streams, and d) kryal streams. Downstream sites are included in the stream types (c, d). Site notations are defined in Table 2-1.

35 2. Alpine stream habitat 31 inorganic carbon ranged from 0.7 to 11 mg/l and no differences between stream types were obvious (Table 2-3). PCA based on 6 physical-chemical variables illustrated the differences among stream types and the seasonality within variables. A physical-chemical variable was chosen if it had a loading > 0.7 on the first or second axis of a PCA including all variables (turbidity, NO3-N, PP, AFDM), or showed a pronounced seasonality (conductance, NH4-N). Major parameters separating sites on the first axis were turbidity, PP, AFDM (loading = 0.8) and NH4-N (loading > 0.6), and explained 46 % of the variation (Fig. 2-5). These variables distinguished a gradient from rhithral +1.0 0.6 Eigenvalues III III I 0 F1 F2 F3 F4 II III II II I III Turbidity II I PP F2 IV III II II IV I AFDM I Co. III I NH 4 IV NO 3 -1.0 +1.0 F1 Figure 2-5 Ordination from a principal component analysis (PCA) based on the seasonal average of 6 physical-chemical variables for all study sites. Stream types are represented by the symbols: rhithral lake outlets, kryal lake outlets, s rhithral streams, s kryal streams, open symbols represent the respective downstream sites. Arrows indicate variables. The F1 axis explained 46.2 % and the F2 axis 19.4 % of the variation. Site notations are defined in Table 2-1; seasons are indicated for glacial influenced sites: I spring, II summer, III autumn, IV winter; Co. = specific conductance.

36 32 2. Alpine stream habitat streams < rhithral lake downstream sites < rhithral lake outlets < kryal sites. Seasonality was most obvious for glacial influenced sites (on the first axis), whereas rhithral sites showed less seasonality. Spring and summer samples of the kryal sites were characterized by high turbidity, PP, AFDM and NH4-N values, whereas autumn and winter samples were situated closer to the rhithral sites. Separation of the summer sample of the Jrisee outlet and the Steinlimi downstream site indicated the slight influence of the adjacent Jriglacier at Jrisee and the decreasing influence of the Steinlimiglacier downstream. The second and third axes of the PCA were distinguished by nitrate and specific conductance, respectively (loading > 0.9). These axes primarily showed the influence of geology and environment, such as atmospheric input, by separating single sites. Assessment of channel stability The different stability indices varied in their results (Fig. 2-6). Calculations of tractive force (), Instability index (ISI), Critical grain size (Dcrit ), Riffle stability index (RSI), and the percentage of the streambed in motion at bankfull discharge (PBF) gave similar results. The indices and ISI distinguished lake outlets (both rhithral and kryal) and kryal upstream sites as having the most stable channels (corresponding to small circles in Fig. 2-6), and rhithral streams and most downstream sites with the least stable channels. Dcrit , RSI and PBF, which were calculated for 8 selected sites, classified 2 kryal sites (one lake outlet and one upstream site) as most stable and the 2 rhithral streams as having the least stable channels. In general, these indices indicated lake outlets (except Steinsee) to have more stable channels than stream sites, and rhithral stream sites to have the least stable channels (Fig. 2-6). Last, Pfankuch's stability score (PSI) also classified rhithral lake outlet sites as having stable channels ('good': values < 31), but one rhithral stream was classified as having a stable and another as having an intermediate stable channel ('fair': values 31 to 46). Both kryal lake outlets and one kryal stream, however, were determined as having unstable channels ('poor': values > 46). PCA scores for the first axis accounted for 56.5 % of the variation in the 7 habitat parameters used to estimate the Multivariate Habitat Index (MHI).

37 2. Alpine stream habitat 33 Pfankuch's bottom score, range in current velocity and water temperature, ratio of bankfull to baseflow width, and shear stress had factor loadings = 0.7 for axis 1. MHI arranged sites along an overall stability gradient that mainly separated rhithral sites as being more stable (factor scores: 0.4 - 1.5) and kryal sites as being less stable (factor scores: 2.6 - 3.6) similar to the PSI (Fig. 2-6). Lake outlets were not separated from respective stream sites by the MHI, indicating less variation among these stream types. The ranking of the habitat indices along a stability gradient using Kendalls' coefficient of rank correlation K indicated two distinct groups: 1) ranking was significantly correlated between the Pfankuch stability score (PSI) and the Multivariate Habitat Index (MHI) (K = 0.98, p < 0.005), and 2) ranking was similar among slope and the remaining indices based on estimates of shear stress (, ISI, MHI PSI PBF RSI Dcrit ISI LN PM LB1 LB2 JS J1 J2 M1 M2 G1 G2 LR SS TG SG1 SG2 Rhithral Kryal Figure 2-6 Relative values (0 - 1) of the instability indices: Tractive force ( ), Instability index (ISI), Critical grain size (Dcrit), Riffle stability index (RSI), percentage of the streambed in motion at bankfull discharge (PBF), Pfankuch stability score (PSI), Multivariate Habitat Index (MHI). Smallest circle = most stable, largest circle = least stable; empty cells indicate no indices calculated. Sites notations are defined in Table 2-1; lake outlets are shown with underlined notations.

38 34 2. Alpine stream habitat Table 2-4 Kendall's coefficients of rank correlation between slope and the different stability indices. Significant correlations in ranking of all sites are indicated by: * p < 0.05, ** p < 0.005. = Tractive force, ISI = Instability index, Dcrit = Critical grain size, RSI = Riffle stability index, PBF = Percentage of the streambed in motion at bankfull discharge, PSI = Pfankuch stability index, MHI = Multivariate habitat index. Slope ISI Dcrit RSI PBF PSI 0.717** ISI 0.533** 0.622** Dcrit 0.825** 0.909** 0.473 RSI 0.849** 0.643* 0.786* 0.546 PBF 0.735* 0.567* 0.567* 0.539 0.794* PSI - 0.132 - 0.127 - 0.165 - 0.296 - 0.327 - 0.154 MHI 0.463 0.429 0.429 0.327 0.357 0.189 -0.982** Dcrit, RSI, PBF) (Table 2-4). Within the second group the ranking of Dcrit was significantly correlated only with slope and (K > 0.8, p < 0.005) but tended to be similar with ISI, RSI and PBF (K = 0.47, p between 0.06 and 0.1). Ranking of PSI and MHI, in contrast, was very different from all the indices based on estimates of shear stress (K < 0.47, p between 0.1 and 0.6) (Table 2-4). Discussion In general, differences in the habitat characteristics of the studied alpine stream systems were more distinct between kryal and rhithral sites than between lake outlets and streams. Thus, the origin of water (i.e., glacier vs. snow melt or groundwater) and its effects on different parameters such as discharge, temperature, or channel stability primarily determined the habitat characteristics of our study sites. However, water temperature - as a key factor for the aquatic biota - and associated parameters (e.g., annual degree days, daily fluctuations) differed strongly between the stream types and indicated lake outlets, kryal lake outlets in particular, being more benign habitats regarding

39 2. Alpine stream habitat 35 the temperature regime than respective streams. Kryal lake outlets also mitigated the pronounced seasonality in the physical-chemical characteristics of kryal systems whereas no differences in channel stability relative to kryal streams were obvious. Physical-chemical characteristics Main differences in the physical and chemical characteristics of the study sites occurred between kryal and rhithral sites, as alluded to in other studies (see references in Ward 1994, Freder 1999). However, on a second level, differences were found between kryal lake outlets and kryal streams. Catchment area, turbidity, water temperature and flow patterns (velocity and discharge) were influenced directly by the presence of a glacier and seasonality in glacial melt. Lower turbidity values and less pronounced seasonality in parameters associated with glacial melt were characteristic for kryal lake outlets compared to kryal streams, whereas rhithral sites showed no such distinct separation between respective lake outlets and streams. Discharge regime is a primary factor affected by glacial melt that further influences the physical-chemical characteristics of kryal systems. Differences between kryal and rhithral systems are related to seasonal flow patterns, as well as to total annual discharge. A modeled hydrograph of the Roseg river (one of our study sites) resulted in a two-fold difference between glacial and non-glacial derived annual discharge and seasonal differences in flow patterns similar to our results as shown in figure 2-2 (Zappa et al. 2000). Temporal changes in the concentrations of various chemical measures (e.g., specific conductance, ammonium, particulate phosphorus (PP), ash-free dry mass (AFDM) and total suspended solids (TSS) were further affected by the seasonality of glacial melt. As specific conductance of ice-melt is low, values during high summer discharge were low and increased with decreasing discharge (Gurnell and Fenn 1985, Tockner et al. 1997). In general, specific conductance was low, as typical for streams flowing on crystalline rocks of low solubility (also see Kawecka 1980), whereas streams flowing on limestone sediments can reach values > 300 S/cm (e.g., Kann 1978).

40 36 2. Alpine stream habitat In alpine regions, glacial ice and snow melt are the primary sources for organic and inorganic nitrogen compounds (Barica and Armstrong 1971, Howard- Williams et al. 1989, Malard et al. 1999). High values can result from decomposition of organic matter under snow and ice and from atmospheric precipitation due to anthropogenic activity (Tait and Thaler 2000, Kuhn 2001). Observed seasonal patterns of ammonium, nitrate and total dissolved nitrogen at the kryal sites followed the discharge regime with peak nutrient loads at the beginning of the melt season (Kuhn 2001). Significantly higher concentrations of ammonium at kryal sites indicate glacial ice melt as a dominant ammonium source probably resulting from bacterial denitrification activity, whereas similar nitrate concentrations at all stream types suggest high atmospheric input accumulating in snow and being released during snow melt (Malard et al. 1999). High measured concentrations of nitrogen, together with results of nutrient retention experiments, suggest that alpine streams in the European Alps are not N limited (Malard et al. 1999, Robinson et al. in press). Phosphorus, in contrast, is derived from local geology, being dissolved from the sediments (Malard et al. 1999, Goudsmit et al. 2000). Thus, low concentrations as measured in our study sites can result in seasonal P limitation in alpine streams (Robinson et al. in press). Particulate phosphorus (PP), being an important source for soluble reactive phosphorus, was found to be directly associated with seasonal and diel flow pulses of glacial streams (Bretschko 1966, Tockner et al. 1997). The derivation of phosphorus from glacial scour also was confirmed in our study by a positive correlation of PP with suspended solids (r = 0.81). One of the major limiting factors for algal growth in glacial streams is the availability of light, being directly affected by high turbidity during high flows in summer (Uehlinger et al. 1998). Turbidity (associated with PP, AFDM, and TSS) was the primary factor separating kryal and rhithral sites in the multivariate ordination (Fig. 2-5). Kryal sites had higher values of turbidity and suspended particles (PP, AFDM, and TSS), being derived from glacial scour, compared to rhithral sites. Rhithral lake outlets and respective streams, however, showed no significant differences in these patterns. Kryal lake outlets, however, always had lower maximum values both for turbidity and for suspended particles than kryal

41 2. Alpine stream habitat 37 streams. Alpine lakes, being generally oligotrophic, act more as sinks than as sources of particles. This is in contrast to findings for low- and midland lake outlets with often substantially higher concentrations of particles and organic matter in lake outlets relative to their stream counterparts (e.g., Richardson and Mackay 1991, Wotton 1994). Thermal characteristics Temperature is regarded as a key factor in the distribution of aquatic biota in alpine regions (Hynes 1970, Ward and Stanford 1982, Sweeney 1984, Freder 1999). Although there is evidence that aquatic insects are derived from ancestral lines originating from cool, lotic habitats, a thermal threshold must be exceeded to allow aquatic insects to grow (= developmental threshold) and to mature (= maturation threshold) (Ward and Stanford 1982). Lake outlets generally had higher maximum water temperatures and lower daily fluctuations compared to respective stream sites, as is true for low elevation lake outlets (Malmqvist and Brnmark 1984, Richardson and Mackay 1991, Wotton 1995) and high latitude lake outlets (Ulfstrand 1968). Differences in the annual temperature regime were more pronounced among kryal streams; thus, kryal lake outlets were, as suggested by Ward (1994), more rhithral in character by accumulating > 500 annual degree days. In contrast to findings of Livingstone et al. (1999), differences in elevation seemed to have little influence on the temperature patterns of the studied alpine systems. For example, with increasing elevation the Jrisee outlet decreased by 77 DDa (rise of 539 m) whereas Moesa, in contrast, increased by 43 DDa (rise of 200 m). The low annual degree days at the rhithral Jrisee outlet indicated its glacial influence, whereas all other temperature parameters (range, thermal increase during summer, daily maximum amplitude) corresponded to the rhithral lake outlets. In general, lake outlets can originate either from surface-releases or from deep-releases of the lake and thus, differ in their thermal characteristics depending on lake stratification (Wotton 1995). Studies on water temperature of alpine lakes similar to the preceding lakes of the studied outlet streams indicated that the temperature profiles of high-elevated lakes can be highly variable among years depending on the duration of ice cover, as well as on the

42 38 2. Alpine stream habitat meteorological conditions afterwards. Thermal stratification generally is dimictic (Tilzer and Schwarz 1976, Goudsmit et al. 2000), cold-monomictic (Franz 1979) or meromictic (Lago Cadagno, Don et al. 2001). The studied outlets, however, all drain from natural lakes with surface-releases and thus, water temperature of the outlet reflects the lake surface temperature. Livingstone et al. (1999) showed that lake surface water temperatures of nine (non-kryal) lakes ranging from 613 to 2265 m a.s.l. were substantially higher during June to September than corresponding air temperatures with the temperature difference increasing from July (above 2000 m a.s.l.: + 2.8 C) to September (above 2000 m a.s.l.: + 3.6 C). Therefore, the water body of an upstream lake leads to a higher annual heat accumulation (i.e., greater increase in summer and higher annual degree days) and lower daily temperature fluctuations at the outlets than in streams, and thus may provide more benign temperature conditions for alpine stream biota. The only exception in the study of Livingstone et al. (1999) was the high- altitude Hagelseewli (2339 m a.s.l.) that showed lower lake surface water temperatures than respective air temperatures, resulting from a long period of snow cover and strong shading through a high cliff wall south of the lake. Shading reduced the amount of solar radiation, thus, prolonging the period of snow cover. The influence of shading also can be seen at the kryal influenced sites that are orientated north with high mountains protecting the glaciers from solar radiation. Consequently, water temperatures of kryal lake outlets and streams are not only dependent on glacial melt, but also on shading and water turbidity that reduces light transmission and absorption and therefore solar warming. Assessment of habitat stability The ranking of the habitat indices of the study sites along a stability gradient yielded different results. Ranking was significantly correlated 1) between the Pfankuch stability score and the Multivariate Habitat Index, and 2) among the indices based on estimates of shear stress that are also strongly dependent on slope (Table 2-4). As mentioned, slope ranged widely from 2 to 16 %. The Pfankuch stability score (PSI), in contrast, does not include slope and therefore

43 2. Alpine stream habitat 39 showed little correlation (p = 0.49). As a result, indices influenced by slope defined sites with slopes = 15 % as having the least stable channels, whereas PSI classified the kryal lake outlets and the kryal stream TG as having unstable channels (Fig. 2-6); none of our study sites were categorized as highly stable (PSI = 15, excellent). The Multivariate Habitat Index (MHI), which indirectly included slope via shear stress calculations, also was not correlated with slope (p = 0.11) but significantly correlated with PSI, and classified kryal sites as least stable. Although the indices yielded different stability values for most of the sites, none of the two groups resulted in distinct differentiation between lake outlets and respective streams as known from low-elevation systems, where upstream lakes are expected to buffer temporal fluctuations and therefore lead to more stable habitats (Richardson and Mackay 1991). Alpine lotic systems generally are characterized by high channel gradients (Ward 1994). In these systems, highly variable flow regime and current velocities, depending on spring and summer snow and ice melt as well as on autumnal rain and storm events, are the major parameters causing sediment movement and causing reduced channel stability (Milner and Petts 1994, Ward 1994). In contrast to low-elevation systems, consequently, channel stability of alpine systems is primarily determined by discharge and its temporal fluctuations and is less dependent on slope (Ward 1994, Freder 1999). Contrasting stabilities of the rhithral stream Moesa and the kryal sites elucidated the different classification by the two groups: a high slope of 16 % resulted in low stability according to indices based on estimates of shear stress, whereas low discharge values (average 0.1 m3 s-1) and a low bankfull to baseflow ratio (0.9) indicated a more 'stable' site, corresponding to the results of the Multivariate Habitat Index and the Pfankuch stability score. Similarly, low slopes at the kryal sites resulted in a classification as 'stable', whereas high discharge values (0.9 - 4 m3 s-1) and high bankfull to baseflow ratio (1.4 to 1.8) corresponded to the classification as 'unstable'. Estimates of tractive force (), Instability index (ISI), Critical grain size (Dcrit ), Riffle stability index (RSI), and the percentage of the streambed in motion at bankfull discharge (PBF) are based on a few distinct parameters (e.g., channel gradient, depth) that do not account for temporal change. A single parameter is

44 40 2. Alpine stream habitat not likely to provide an all encompassing measure of stream stability, therefore it is necessary to incorporate the spatial and temporal changes in habitat parameters (Hawkins et al. 1993, Death and Winterbourn 1994); a requirement partially fulfilled by the use of a multivariate habitat index as in the present study. More work must be done to improve a stability measurement of alpine systems including spatial and temporal changes in different habitat parameters. Concluding remarks Habitat characteristics of alpine lotic systems are a complex interplay of different factors varying across spatial and temporal scales (Hawkins et al. 1993, Ward 1994, Freder 1999). A dominant habitat determinant of alpine streams was the presence of a glacier. Flow regime and water temperature, both strongly influenced by a glacier and its seasonality in melting, are major factors affecting these habitats and their associated biota (Ulfstrand 1968, Mani 1990, Ward 1994). Seasonal fluctuations of the environmental parameters ameliorated otherwise extreme conditions in kryal streams leading to improved environmental conditions at certain times in the year. For example, a typical 'kryal' system can shift to a more 'rhithral' or even 'krenal' stream type during periods of low discharge in early spring or late autumn (see also Ward et al. 1998, Malard et al. 1999). The presence of an upstream lake had a strong influence on water temperature with a greater increase in summer and lower daily fluctuations. Similar results were found at high latitude lake outlets in Lapland leading to a characteristic invertebrate community (Ulfstrand 1968). Burgherr et al. (in press) concluded from his studies that habitat stability and spatio-temporal habitat heterogeneity are key factors affecting the community structure of glacial streams. The temporal heterogeneity of alpine lotic systems resulting from seasonal discharge fluctuations requires different approaches to fully assess habitat stability. Common stability indices strongly based on slope proved to be unsuitable. In contrast, a multivariate habitat index including several parameters better represented the overall habitat stability, encompassing both the temporal and spatial habitat heterogeneity of alpine streams. Further research should now focus on the relation of the habitat template of these different alpine lotic systems with biotic, e.g. invertebrate

45 2. Alpine stream habitat 41 and algae, patterns and processes and compare the relations in alpine systems with results found at lake outlets at low-elevation and high latitude. Acknowledgements We thank the numerous individuals who contributed to the completion of this study. Special thanks to Michael T. Monaghan, in particular, and Donna Anderson, Andreas Blum, Peter Burgherr, Chregu Dinkel, Michael Dring, Florian Malard, Marcos de la Puente, Heiko Rinderspacher, Christian Rust, Ulrich Donath, and Monika Winder for assistance in the field, and Richard Illi and Bruno Ribi for completion of the chemical analyses in the laboratory. We are grateful to the Swiss Hydrological and Geological Survey and Jakob Grnenfelder from the Amt fr Umwelt Graubnden for providing discharge data, and the communes of Pontresina, Samedan, and Klosters for providing access to the sampling areas. We thank Karl O. Rothhaupt, R. S. Wotton and an anonymous reviewer for constructive comments that greatly improved the manuscript. The study was partially funded by a Swiss National Science Foundation Grant (no. 31-50440.97) examining the ecology of alpine lake outlets. References BARICA , J., and F. A. J. ARMSTRONG . 1971. Contribution by snow to the nutrient budget of some small northwest Ontario lakes. Limnology and Oceanography 16:891-899. BRETSCHKO , G. 1966. Untersuchungen zur Phosphatfhrung zentralalpiner Gletscherabflsse. Archiv fr Hydrobiologie 62:327-334. BRNMARK, C., and B. MALMQVIST . 1984. Spatial and temporal patterns of lake outlet benthos. Verhandlungen der Internationalen Vereinigung der Limnologie 22:1986-1991. BURGHERR , P., and J. V. W ARD. 2000. Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung der Limnologie 27:1587-1590. BURGHERR , P., J. V. WARD, and C. T. ROBINSON. Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps). Journal of the North American Benthological Society (in press).

46 42 2. Alpine stream habitat COBB, D. G., and J. F. FLANNAGAN. 1990. Trichoptera and substrate stability in the Ochre River, Manitoba. Hydrobiologia 206:29-38. DEATH, R. G. 1996. The effect of habitat stability on benthic invertebrate communities: The utility of species abundance distributions. Hydrobiologia 317:97-107. DEATH, R. G., and M. J. W INTERBOURN. 1994. Environmental stability and community persistence: a multivariate perspective. Journal of the North American Benthological Society 13:125-139. DON, C. D., K. W. HANSELMANN, R. PEDUZZI, and R. BACHOFEN. 2001. The meromictic alpine Lake Cadagno: Orographical and biogeochemical description. Aquatic Science 63:70-90. DONATH, U., and C. T. ROBINSON. 2001. Ecological characteristics of lake outlets in Alpine environments of the Swiss Alps. Archiv fr Hydrobiologie 150:207-225. DUNCAN, M. J., A. M. SUREN, and S. L. R. BROWN. 1999. Assessment of streambed stability in steep, bouldery streams: development of a new analytical technique. Journal of the North American Benthological Society 18:445-456. FRANZ, H. 1979. kologie der Hochgebirge. Eugen Ulmer GmbH & Co., Stuttgart, Deutschland. FREDER, L. 1999. High alpine streams: cold habitats for insect larvae. Pages 181- 196 in Margesin, R., and F. Schinner (Editors). Cold-Adapted Organisms - Ecology, Physiology, Enzymology and Molecular Biology. Springer, Berlin, Deutschland. GOUDSMIT , G.-H., G. LEMCKE , D. M. LIVINGSTONE, A. F. LOTTER, B. MLLER, and M. STURM . 2000. Hagelseewli: a fascinating high mountain lake - suitable for palaeoclimate studies? Verhandlungen der Internationalen Vereinigung der Limnologie 27:1013-1022. GURNELL , A. M., and C. R. FENN. 1985. Spatial and temporal variations in electrical conductivity in a pro-glacial stream system. Journal of Glaciology 31:108- 114. HARDING , J. S. 1992. Physico-chemical parameters and invertebrate faunas of three lake inflows and outlets in Westland, New Zealand. New Zealand Journal of Marine and Freshwater Research 26:95-102. HAWKINS , C. P., J. L. KERSHNER , P. A. BISSON, M. D. BRYANT , L. M. DECKER, S. V. GREGORY, D. A. MCCULLOUGH, C. K. OVERTON, G. H. REEVES , R. J. STEEDMAN, and M. K. YOUNG . 1993. A hierarchical approach to classifying stream habitat features. Fisheries 18:3-12. HOWARD -WILLIAMS , C., J. C. PRISCU, and W. F. VINCENT . 1989. Nitrogen dynamics in two Antarctic streams. Hydrobiologia 172:51-61.

47 2. Alpine stream habitat 43 HYNES , H. B. N. 1970. The ecology of running waters. University Press, Liverpool, U.K.. IAHS/UNESCO. 1998. Fluctuations of glaciers 1990-1995. World Glacier Monitoring Service (WGMS), Zurich, Switzerland. JAMES , F. C., and C. E. MCCULLOCH. 1990. Multivariate analysis in ecology and systematics: Panacea or Pandora's Box? Annual Review of Ecology and Systematics 21:129-166. KANN, E. 1978. Systematik und kologie der Algen sterreichischer Bergbche. Archiv fr Hydrobiologie/Supplementband 53:405-643. KAPPESSER , G. 1993. Riffle stability indices. Idaho Panhandle National Forest, Coeur d'Alene, Idaho, U.S.A.. KAWECKA , B. 1980. Sessile algae in European mountain streams, I. The ecological characteristics of communities. Acta Hydrobiologica 22:361-420. KING , L. 1987. Studien zur postglazialen Gletscher- und Vegetationsgeschichte des Sustenpassgebietes. Selbstverlag Giessener Geographische Schriften, Giessen, Deutschland. KOWNACKI, A., E. DUMNICKA , J. GALAS , B. KAWECKA , and K. W OJTAN. 1997. Ecological characteristics of a high mountain lake-outlet stream (Tatra Mts, Poland). Archiv fr Hydrobiologie 139:113-128. KREIS , A. 1921. Die Jriseen und ihre postglaziale Besiedlungsgeschichte. Internationale Revue der gesamten Hydrobiologie und Hydrographie IX:189- 212, 241-288, 451-490. KUHN, M. 2001. The nutrient cycle through snow and ice, a review. Aquatic Sciences 63:150-167. LIVINGSTONE, D. M., A. F. LOTTER, and I. R. WALKER . 1999. The decrease in summer surface water temperature with altitude in Swiss alpine lakes: a comparison with air temperature lapse rate. Arctic, Antarctic, and Alpine Research 31:341-352. MALARD, F., K. TOCKNER, and J. V. W ARD. 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., and C. BRNMARK. 1984. Functional aspects of a lake outlet versus a springfed stream ecosystem. Verhandlungen der Internationalen Vereinigung der Limnologie 22:1992-1996. MANI , M. S. 1990. Fundamentals of high altitude biology. Aspect Publications Ltd., London, U.K.. MILNER, A. M., and G. E. PETTS . 1994. Glacial rivers: physical habitat and ecology. Freshwater Biology 32:295-307.

48 44 2. Alpine stream habitat NEWBURY, R. W. 1984. Hydrologic determinants of aquatic insect habitats. Pages 232-357 in Resh, V. H., and D. M. Rosenberg (Editors). The Ecology of Aquatic Insects. Praeger Publishers, New York, U.S.A.. PFANKUCH, D. J. 1975. Stream reach inventory and channel stability evaluation. U.S.D.A. Forest Service, Missoula, Montana, U.S.A.. PINAY, G., H. DCAMPS , E. CHAUVET, and E. FUSTEC. 1990. Functions of ecotones in fluvial systems. Pages 141-169 in Naimann, R. J., and H. Dcamps (Editors). The Ecology and Management of Aquatic-Terrestrial Ecotones. UNESCO - The Parthenon Publishing Group, Paris, France. PLATTS , W. S., W. F. MEGAHAN, and G. W. MINSHALL. 1983. Methods for evaluating stream, riparian, and biotic conditions. U.S. Departement of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT, U.S.A.. RICHARDSON, J. S. 1984. Effects of seston quality on the growth of a lake-outlet filter feeder. Oikos 43:386-390. RICHARDSON, J. S., and R. J. MACKAY. 1991. Lake outlets and the distribution of filter feeders: an assessment of hypotheses. Oikos 62:370-380. ROBINSON, C. T., U. UEHLINGER, F. GUIDON, P. SCHENKEL , and R. SKVARC. Limitation and retention of nutrients in alpine streams of Switzerland. Verhandlungen der Internationalen Vereinigung fr Theoretische und Angewandte Limnologie (in press). SOKAL , R. R., and F. J. ROHLF 1995. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Company, New York, U.S.A.. SOUTHWOOD, T. R. E. 1977. Habitat, the templet for ecological strategies? Journal of Animal Ecology 46:337-365. SPICHER , A. 1980. Geologische Karte der Schweiz. Schweizerische Geologische Kommission, Bundesamt fr Landestopographie, Wabern, Schweiz. SWEENEY, B. W. 1984. Factors influencing life-history patterns of aquatic insects. Pages 56-100 in Resh, V. H., and D. M. Rosenberg (Editors). The Ecology of Aquatic Insects. Praeger Publishers, New York, U.S.A.. TAIT, D., and B. THALER. 2000. Atmospheric deposition and lake chemistry trends at a high mountain site in the eastern Alps. Journal of Limnology 59:61-71. TILZER , M. M., and K. SCHWARZ. 1976. Seasonal and vertical patterns of phytoplankton light adaptation in a high mountain lake. Archiv fr Hydrobiologie 77:488-504. TOCKNER, K., F. MALARD, P. BURGHERR , C. T. ROBINSON, U. UEHLINGER , R. ZAH, and J. V. WARD. 1997. Physico-chemical characterization of channel types in a glacial

49 2. Alpine stream habitat 45 floodplain ecosystem (Val Roseg, Switzerland). Archiv fr Hydrobiologie 140:433-463. TOWNSEND , C. R., M. R. SCARSBROOK, and S. DOLDEC . 1997. Quantifying disturbance in streams: alternative measures of disturbance in relation to macroinvertebrate species traits and species richness. Journal of the North American Benthological Society 16:531-544. UEHLINGER , U., R. ZAH, and H. BRGI. 1998. The Val Roseg Project: temporal and spatial patterns of benthic algae in an Alpine stream ecosystem influenced by glacier runoff. Pages 419-424 in Kovar, K., U. Tappeiner, N. E. Peters, and R. G. Craig (Editors). Hydrology, Water Ressources and Ecology in Headwaters. IAHS Press, Wallingford, U.K.. ULFSTRAND, S. 1968. Benthic animal communities in Lapland streams. Oikos Supplementum 10:1-116. WARD, J. V. 1994. Ecology of alpine streams. Freshwater Biology 32:277-294. WARD, J. V. 1998. A running water perspective of ecotones, boundaries, and connectivity. Verhandlungen der Internationalen Vereinigung der Limnologie 26:1165-1168. WARD, J. V., P. BURGHERR , M. O. GESSNER , F. MALARD, C. T. ROBINSON, K. TOCKNER, and U. UEHLINGER . 1998. The Val Roseg Project: habitat heterogeneity and connectivity gradients in a glacial flood-plain system. Pages 425-432 in Kovar, K., U. Tappeiner, N. E. Peters, and R. G. Craig (Editors). Hydrology, Water Resources and Ecology in Headwaters. IAHS Press, Wallingford, U.K.. WARD, J. V., and J. A. STANFORD. 1982. Thermal responses in the evolutionary ecology of aquatic insects. Annual Review of Entomology 27:97-117. WOTTON, R. S. 1994. Particulate and dissolved organic matter as food. Pages 235- 288 in W OTTON, R. S. (Editor). The Biology of Particles in Aquatic Systems. CRC Press, Boca Raton, Florida, U.S.A.. WOTTON, R. S. 1995. Temperature and lake-outlet communities. Journal of thermal biology 20:121-125. ZAPPA , M., A. BADOUX, and J. GURTZ. 2000. The application of a complex distributed hydrological model in an highly glaciated alpine river catchment. Horvatic, J. 33rd Conference of International Association for Danube Research, Osijek, Croatia. ZAR, J. H. 1984. Biostatistical analysis. Prentice Hall, Englewood Cliffs, New Jersey, U.S.A..

50 46

51 3. Algal communities in alpine streams 47 3. Algal communities associated with different alpine stream types Hieber, M., Robinson C.T., Rushforth S.R., and U. Uehlinger. (2001). Arctic, Antarctic, and Alpine Research 33 (4): 447-456. We investigated major physical-chemical characteristics and benthic algae of different alpine lotic systems comprising streams and lake outlets of rhithral and kryal origin over an annual cycle. We also evaluated the structure of the algal communities and its relation to environmental characteristics for the different stream types. Algal communities were generally dominated by diatoms, Cyanophyta, and Hydrurus foetidus (Chrysophyceae). Community structure was similar among alpine streams and lake outlets, but more algal taxa occurred in lake outlets (rhithral and kryal) than in kryal streams. Although algae were identified mainly to genera, distinct patterns in community structure were evident. A major environmental determinant of the algal community among stream types was the presence of a glacier and resulting seasonal differences in flow, temperature, and turbidity. An upstream lake was a secondary determinant in buffering seasonal fluctuations in environmental conditions thus leading to greater stability. Algal communities, consequently, were more diverse and less seasonally variable at lake outlets. The diatom genera Amphora, Denticula, Fragilaria, Gomphonema, Nitzschia, and Synedra and the blue-green algae Oscillatoria and Phormidium were characteristic of lake outlets, whereas Chamaesiphon (blue-green) and Hydrurus foetidus were indicative of kryal sites.

52 48 3. Algal communities in alpine streams Introduction Alpine lotic systems generally are characterized by cold temperature (low number of annual degree days), turbulent well-oxygenated water, and high discharge fluctuations. These systems, however, comprise a diversity of stream types. Ward (1994) distinguished high-altitude headwaters as kryal, rhithral, or krenal depending on their primary water source. Kryal streams are primarily fed by glacial meltwater and have low water temperature (30 NTU), and a seasonal and diel fluctuating flow regime that causes unstable channel beds and variable habitat conditions (Milner and Petts 1994, Ward 1994). Rhithral streams are mainly fed by snowmelt and have lower diel variations in discharge, whereas kre nal streams are groundwater fed and represent a stable habitat throughout the year. Another stream type, almost neglected so far, are lake outlets of kryal or rhithral origin. Lowland lake outlets typically show a more stable temperature and discharge regime (Malmqvist and Brnmark 1984), a relatively stable, single- thread channel (Milner and Petts 1994), and higher concentrations of transported organic material (Brnmark and Malmqvist 1984) than respective nonoutlet streams. In alpine regions, an upstream lake is expected to increase water temperature and attenuate floods thus leading to a more stable channel (Milner and Petts 1994); however, little information is available at present. Environmental conditions can influence the algal diversity of lotic systems by constraining particular species. Several studies have related periphyton communities and their species traits with different habitat parameters such as water velocity (Maier 1994, Biggs et al. 1998), disturbance (Kuhn et al. 1981, Grimm and Fisher 1989), water temperature (Kann 1978), substratum (Maier 1994, Cattaneo et al. 1997), water chemistry (Kann 1978, Passy et al. 1999), and nutrient supply (Kuhn et al. 1981, Passy et al. 1999). These relationships have led to the use of periphyton as an ecological indicator in river assessment (see citations in Round 1993, Hill et al. 2000). The successful use of periphyton to assess water quality has been based on different taxonomic levels and functional groups (Round 1993, Steneck and Dethier 1994, Chessman et al. 1999). Numerous studies on algal communities have been conducted in mid- and lowland streams, whereas little data exist for alpine and glacial streams. Studies

53 3. Algal communities in alpine streams 49 of glacial streams are mainly from arctic and antarctic regions that belong to a different biogeographical zone (Elster and Svoboda 1996, Alger et al. 1997). Further, the few studies of high-mountain lotic freshwaters were completed mostly on streams without an upstream lake (Brun 1965, Kawecka 1980, 1981, Maier 1994, Uehlinger et al. 1998, Passy et al. 1999, Vavilova and Lewis 1999), with little data from high-mountain lake outlets (Kawecka 1965, Kann 1978, Kawecka 1980). We investigated habitat characteristics and associated algal communities of different alpine stream systems that comprised rhithral and kryal streams and respective lake outlets. We expected the glacial influence on kryal systems to have a strong affect on algal community structure, although this influence may be mitigated in kryal lake outlets. In addition, we examined seasonal and longitudinal patterns in algal community structure among the different alpine streams. Study Sites Study sites were situated in the Swiss Alps above treeline at elevations ranging from 1900 to 2500 m a.s.l.; only the kryal stream Grindelwald was below 8 9 10 N ine Rh re Aa Danube 47 SS, JS, SG J G PM M LB, LN LR, TG one Rh 46 0 50 100 km Po Figure 3-1 Map of Switzerland with locations of the study sites in the Swiss Alps (shaded region). Site notations defined in Table 3-1.

54 50 3. Algal communities in alpine streams Table 3-1 Location and general characteristics of the study sites. Ranges are listed as minimum - maximum. k = kryal, r = rhithral, L = lake outlet, S = stream, NTU: nephelometric turbidity units. NA: not available Site Notation Origin Lotic Drainage Elevation Catchment Slope Temperature Turbidity range type (m a.s.l.) (km2) (%) range (C) (NTU) Lago Bianco LB r L Po 2076 2.1 4 0.0 - 15.3 0 .3 - 2 .9 Lago Cadagno LC r L Po 1923 2.2 3 NA 1 .2 - 2 .5 Lago Nero LN r L Po 2387 0.8 5 0.0 - 16.3 0 .2 - 1 .4 Puoz Minor PM r L Danube 2336 1.8 9 0.0 - 16.0 0 .7 - 5 .6 Joerisee JS r(k) L Rhine 2489 3.4 5 0.1 - 16.2 3 .9 - 67 Steinsee SS k L Rhine 1934 7.3 4 1.6 - 4.2 57 - 126 Lej Roseg LR k L Danube 2159 19.3 2 0.0 - 9.1 47 - 361 Moesa MS r S Po 2300 0.6 16 0.8 - 12.0 0 .1 - 3 .2 Tschierva TG k S Danube 2100 14.7 4 0.1 - 4.0 31 - 3284 Morteratsch -upstream MG1 2000 35.0 1 1 - 722 k S Danube 0.0 - 3.2 -downstream MG2 1970 35.5 5 4 - 676 Steinlimi SG k S Rhine 2090 3.4 2 0.1 - 3.9 14 - 109 Grindelwald -upstream GG1 1240 17.3 8 1 - 286 k S Rhine 0.0 - 2.8 -downstream GG2 1210 17.5 6 0 .4 - 373 Lang -upstream LG1 1990 18.6 8 17 - 171 k S Rhone 0.0 - 2.7 -downstream LG2 1910 29.3 6 75 - 171 treeline at 1200 m a.s.l. (Fig. 3-1). Sites consisted of alpine lake outlets and streams of kryal and rhithral origin from the four major drainage areas (Po, Danube, Rhine, and Rhone) in Switzerland (Table 3-1). Catchment size ranged from 0.8 to 35 km2 with kryal sites in the largest catchments and rhithral sites in catchments < 4 km2. Stream slope ranged from 1.3 to 16 %, being higher at stream sites than lake outlets. In the following text, the term "stream type" comprises (1) "streams" - without an upstream lake, and (2) "lake outlets".Three glacial-fed streams (Morteratsch, Grindelwald, and Lang) were sampled at two sites, one close to the glacier mouth and the other ca. 300 to 500 m

55 3. Algal communities in alpine streams 51 downstream. Only one rhithral stream, the Moesa, was sampled, this being from June to October 1999. The outlet of the Joerisee is rhithral, but is influenced by the Joeriglacier that flows into a small lake above the Joerisee. Observations of changes in channel morphology (e.g., lateral channel shift) over the study period indicated that rhithral sites were more physically stable than kryal sites. Among the kryal sites, Morteratsch and Steinlimi represented the more stable streams and the outlet of Lej Roseg appeared more stable than the outlet of the Steinsee where substantial erosion of the ancient end moraine of the Steinglacier occurred. All sites are in the crystalline area of the Swiss Alps with bedrock mainly consisting of granite and gneiss. The catchments are covered mostly by rocks, with sparse vegetation of grass and low shrubs of alder (Alnus sp.) and willow (Salix sp.) also present. Methods Physical and chemical measures On each sampling date, specific conductance (WTW model 325) and turbidity (Cosmos Zllig, in nephelometric turbidity units; NTUs) were measured in the field. A water sample (1 L) was collected, filtered through pre-ashed glass fiber filters (Whatman GF/F filters; 45 mm diameter) and analyzed for the following chemical parameters in the laboratory: ammonium (NH4), nitrate (NO3), nitrite (NO2), dissolved (DN) and particulate nitrogen (PN), soluble reactive phosphorus (SRP), total dissolved (DP) and particulate phosphorus (PP), dissolved (DOC) and particulate organic carbon (POC), total suspended solids (TSS) and ash-free dry mass (AFDM). Determination of each parameter followed the methods described in detail in Tockner et al. (1997). At each site, water temperature was recorded hourly with temperature loggers (Minilog, Vemco). Two-way analysis of covariance (ANCOVA) was used to compare among rhithral vs. kryal sites and streams vs. lake outlets, with date as the covariate (Zar 1984). Algal communities At 16 sites, algal samples were collected seasonally three to six times between July 1998 and October 1999 depending on site accessibility. On each date, we collected 10 stones per site (b axis = 2.8 - 14.0 cm, median = 7.0 cm) from

56 52 3. Algal communities in alpine streams Table 3-2 Classification of algal taxa by relative abundance after transformation of the absolute numbers by 0.5 * log 2. Class Abundance Corresponding visual occurrence 1 0-7 rare 2 8 - 31 rare - common 3 32 - 131 common 4 132 - 527 common - abundant 5 < 528 abundant predominant instream habitats. The stones were stored at minus 25C until processed in the laboratory. For algal identification, periphyton was removed from each stone by brushing with a metal bristle brush. The algal suspension from each group of 10 stones per site and date was composited, an aliquot of 20 to 25 ml was removed and preserved in 2% formalin for later identification and enumeration. Individual subsamples were placed on glass microscope slides and examined directly for the general abundance of algal taxa as classified according to the Index of Relative Species Abundance (Elster and Svoboda 1996) (Table 3-2). In general, if a taxon was represented by only one or a few specimens it was recorded as rare and assigned a value of 1. If a taxon was present in around 5 to 10% of the microscopic examination fields it was recorded as common (value = 3). If a taxon was present in >20% of the examination fields it was recorded as abundant and given a value of 5. Algal taxa from all samples were identified mainly at 400x (maximum 1000x) to genus level. Stream samples were examined using a Zeiss IM microscope, and samples from the lake outlets using a Zeiss RA microscope with Normarski and bright-field optics. In addition, lake outlet samples were examined for identification of diatom species as little data from high-mountain lake outlets exist. Each sample was boiled in concentrated nitric acid, rinsed, mounted in Naphrax mountant, and examined at 1000x oil immersion using a Zeiss RA microscope with Nomarski optics. Counts of 450 to 650 diatom cells were made from each slide to identify diatoms and estimate relative densities. To standardize the analysis of the two data sets (genera for streams, species for lake outlets), the absolute algal numbers of lake outlets were summarized by

57 3. Algal communities in alpine streams 53 genera and then transformed 0.5 * log2 to categorize each taxon as used to describe the abundance patterns for genera as above (Table 3-2). Thus, algal samples were analyzed in two ways: (1) algae of all sites were examined at the genus level based on categories; and (2) diatoms of the lake outlet samples were analyzed based on species level as relative densities from the absolute counts. In the remaining text, the term "genus" refers to the first analysis based on categorical data even though some taxa are listed as species (e.g., Hydrurus foetidus, Hannaea arcus). The term "species", in contrast, refers to the second analysis based on diatom species of lake outlets. For the categorical data, differences among study sites were analyzed using ANCOVA with date as the covariate (Zar 1984). One-way ANCOVA was used to compare among individual sites, whereas two-way ANCOVA was used to compare among lake outlet vs. stream and kryal vs. rhithral types. If differences were detected by ANCOVA, the Tukey HSD test was used to determine which types actually differed (Zar 1984). To compare among types, both upstream and downstream sites were included in the type "kryal stream". We also used indirect PCA to illustrate the differences in community structure among sites in relation to environmental variables. Indirect gradient analysis orders samples from biotic data only, and environmental variables are subsequently used to interpret the biotic ordination; PCA was used as genera displayed a linear response (gradient length < 4 SD) (Ter Braak and Smilauer 1998). PCA was based on the frequency (categorical data) of algal genera comprising >1% of the algal community and environmental variables for each site and date (data log-transformed and centered by taxa). Ordination analysis was performed with Canoco 4 (Ter Braak and Smilauer 1998). Results Physical and chemical characteristics Kryal and rhithral sites, as well as lake outlet and stream sites, differed primarily in maximum water temperature and turbidity (Table 3-1). Temperature at rhithral sites reached almost 17C compared to

58 54 3. Algal communities in alpine streams >3000 NTU (site TG) during glacial melt in summer. Kryal lake outlets were less turbid relative to kryal streams, but still had higher values than rhithral sites (turbidity typically was

59 3. Algal communities in alpine streams 55 lake outlets (P < 0.001), as did richness of kryal and rhithral sites (P = 0.03). However, kryal and rhithral streams showed no differences in genus richness (P = 0.88), but only one rhithral stream was sampled. Kryal streams showed no longitudinal patterns in their genus richness. For example, genus richness increased at Lang and Grindelwald downstream but decreased by one taxon at Morteratsch (Fig. 3-2). MG2 MG1 GG2 GG1 LG2 LG1 PM MS LN TG LR LC LB SG SS JS DIAT HAar CYMB HYfo ACHN LYNG GOMP NAVI FRAG SYNE NITZ DENT OSCI AMPH MEci CALO CHAM PHsp CYCL CLgl genus richness 22 17 18 27 19 16 13 14 11 8 6 3 4 3 2 2 rare rare-common common common-abundant abundan t Figure 3-2 Average abundance of algal genera having a total frequency >1 %, and total genus richness for each site. Sites (notations as defined in Table 3-1) and algal genera (notations as defined in Appendix 3-1) are ordered by decreasing genus richness, and by decreasing number of sites a genus was present.

60 56 3. Algal communities in alpine streams Representatives of the algal genera Achnanthes, Cymbella, Diatoma, Hannaea, Meridion, Navicula (Bacillariophyceae), and Lyngbya (Cyanophyta) were present in all stream types except the rhithral stream (Fig. 3-2). Hydrurus foetidus (Chrysophyceae) occurred only at kryal sites and the Joerisee where it often dominated the algal community. Microspora (Chlorophyta) and Chamaesiphon (Cyanophyta) also were found primarily at kryal sites. Algal genera that occurred only at lake outlets included lentic as well as lotic forms such as Amphora, Cocconeis, Denticula, Epithemia, Eunotia, Nitzschia, Pinnularia (Bacillariophyceae), Cladophora glomerata (Chlorophyta), Chroococcus, Oscillatoria and Phormidium (Cyanophyta), and Audouinella violacea (Rhodophyta) (Fig. 3-2). Mean occurrence for most genera was low; only Cymbella, Hydrurus foetidus, Fragilaria, Calothrix, and Cladophora occurred at one or two sites with high average frequency (category class = 4 or 5). Diatom species of lake outlets The most diatom species were found in the outlet of Lago Cadagno (46 species), 38 species occurred in the kryal Lej Roseg outlet, and 36 diatom species in the rhithral Lago Bianco and Puoz Minor outlets. The lowest number of diatom species occurred in the Steinsee outlet (20) and the rhithral Lago Nero (23) and the Jrisee (23) outlets (Appendix 3-1). Achnanthes minutissima dominated (13 to 99 %) the diatom community at all outlets. Other frequent species were Cymbella minuta, Hannaea arcus, Synedra delicatissima, and S. rumpens. In addition to these, diatom communities of kryal lake outlets were dominated by Gomphonema olivaceum, whereas Amphora perpusilla, Cymbella affinis, and Denticula elegans occurred mainly at rhithral lake outlets. Average species composition was similar between rhithral and kryal lake outlets, but more rare species were identified from rhithral lake outlets. Seasonality in algal community structure Algal genus richness was quite constant during the study period at the rhithral lake outlets. At kryal sites, algal communities had low genus richness during summer and higher richness during late autumn/winter; however, seasonality

61 3. Algal communities in alpine streams 57 20 Rhithral lake outlets 15 10 LB LC 5 JS LN PM 0 20 Kryal lake outlets LR SS Genus richness 15 10 5 0 20 Kryal streams GG2 GG1 15 LG2 Figure 3-3 Seasonal changes LG1 MG2 in total genus richness at all MG1 lake outlet and kryal stream 10 SG TG sites. Kryal streams with two 5 sampling sites are indicated by the same symbol: open = 0 downstream, filled = upstream. Jul-98 Oct-98 Jan-99 Apr-99 Jul-99 Oct-99 Date was most pronounced at kryal lake outlets (Fig. 3-3). The sampling period in the rhithral stream included only 4 mo, thus seasonal patterns are not discussed. Most algal taxa showed seasonal differences in abundance at most sites. An analysis of the average frequency of the top five taxa from each site showed generally high abundances in autumn/winter (Fig. 3-4). Figure 3-4a shows the top five genera of each stream site with four or more sampling dates. Most genera had maximum or high abundance in autumn/winter (e.g., Hannaea arcus, Hydrurus foetidus, Fragilaria spp.) and were present for only a short time (e.g., Achnanthes spp., Diatoma spp.). Only a few genera occurred at high frequencies

62 58 3. Algal communities in alpine streams a) Morteratsch-up (n=4) b) Lago Bianco (n=3) 5 5 4 4 3 3 2 2 1 1 0 0 ACHN CYMB FRAG HAar HYfo ACmn AMpe CYaf CYmi DEel Morteratsch-down (n=4) Lago Nero (n=3) 5 5 4 4 3 3 2 2 1 1 0 0 ACHN CHAM CYMB FRAG HAar ACmn GOpa PHsp SYru SYsp Grindelwald-up (n=4) Puoz Minor (n=3) 5 5 4 4 Categorical abundance 3 3 2 2 1 1 0 0 DIAT HAar HYfo LYNG MIsp ACmn AMpe CYaf CYmi HAar 5 Grindelwald-down (n=4) 5 Jrisee (n=4) 4 4 3 3 2 2 1 1 0 0 CHAM DIAT HAar HYfo LYNG ACmn DEte HAar NIsp PHsp Tschierva (n=4) spring98 Steinsee (n=3) 5 summer98 5 4 autumn98 4 winter98/99 3 spring99 3 2 summer99 2 abundance = 0 1 1 0 0 DIAT HYfo ULOT ACfl ACmn GOod GOol HYfo Lej Roseg (n=5) 5 4 3 2 1 0 ACmn HYfo LYsp OSam PHsp Figure 3-4 Abundance of the top five algal taxa; a) based on genus level for stream sites, and b) based on species level for lake outlet sites. n = sampling dates. Algal notations defined in Appendix 3-1. during spring and summer, whereas most genera were absent or occurred only in low abundances.

63 3. Algal communities in alpine streams 59 An analysis of the average frequency of the top five species for lake outlets ( 3 sampling dates) showed similar seasonal patterns (Fig. 3-4b). Seasonal patterns were most distinct at sites with kryal influence (Joerisee, Lej Roseg, Steinsee) and the rhithral Lago Nero outlet. Most species had maximum abundances in autumn and winter with many absent during spring and summer (e.g., Phormidium, Denticula tenuis, Synedra and Nitzschia species). Species abundance at the rhithral lake outlets seemed more constant over the year maintaining high abundances in summer. Achnanthes minutissima occurred at all lake outlet sites in high abundance throughout the study period with no seasonality. Algal and environmental relationships PCA ordination, combining algal and environmental data, clearly separated streams from lake outlets (Fig. 3-5). Lake outlets were mainly grouped by Achnanthes, Cymbella, Synedra, and Gomphonema (Fig. 3-5a). Rhithral lake outlets were grouped together, whereas kryal lake outlets were more widely distributed resulting from seasonality in algal composition. Seasonality also was evident at the glacially-influenced Joerisee outlet with summer samples near kryal lake outlets, and spring/autumn samples near rhithral sites. Stream sites were widely distributed and mainly distinguished by nondiatom genera Hydrurus foetidus, Lyngbya spp., and Chamaesiphon spp. (Fig. 3-5b). Kryal streams with low genus richness (Lang, Steinlimi, Tschierva) were grouped by Hydrurus foetidus with no differences among dates. Two kryal streams with high richness (Grindelwald, Morteratsch), however, showed high variability related to seasonal changes in community structure and the presence or absence of a single genus (e.g., Achnanthes, Diatoma, Hydrurus foetidus). The rhithral stream had low genus richness and overall low abundances. Distribution of each sampling date per site based on environmental data followed patterns described above, separating rhithral and kryal sites mainly on the first axis that explained 36.1% of the total species-environment relation (Fig. 3-5c). Rhithral sites were distinguished by higher maximum water temperature, DOC, DP, and PN, and kryal sites by higher concentrations of TSS, turbidity, and PP.

64 60 3. Algal communities in alpine streams Eigenvalue % c) F1 29.6 36.1 DN NO 2+NO 3 F2 16.6 20.0 LB LC TSS M LN turbidity TG PM PP MG JS DOC SG SS max. PN GG LR temp. LG a) DP +1 b) LYNG LYNG -1 +1 -1 CHAM CHAM CYMB CYMB DIAT SYNE GOMP DIAT SYNE GOMP ACHN ACHN HYfo HYfo Figure 3-5 Principal component analysis ordination based on the frequency of the algal genera with a total frequency >1 % and the environmental data for each study site and date. a) and b) show the species-site ordination separated for a) all lake outlet sites and b) all stream sites. Symbols represent the different sampling sites (as defined in Table 3-1): black represents rhithral sites, grey represents kryal sites and open grey symbols the respective kryal downstream sites. Arrows indicate algal taxa (notations as defined in Appendix 3-1). c) Ordination of the environmental data with arrows indicating the physical-chemical variables. Eigenvalues and percentage variance of species- environment relation (%) are given for the first two axis (F1, F2). Discussion Algal community structure Algal communities of the studied alpine lotic systems generally had a similar structure among streams and lake outlets, being dominated by diatoms, blue- green algae, and the Chrysophyceae Hydrurus foetidus. These have been found to be dominant taxa in other alpine streams of middle-latitude mountains of the

65 3. Algal communities in alpine streams 61 northern hemisphere (e.g., Kawecka 1980, Vavilova and Lewis 1999). Kryal streams were characterized by a low number of genera and the dominance of Hydrurus foetidus, which is the predominant taxon in alpine kryal streams (Kawecka 1980, Uehlinger et al. 1998) but also common in other alpine streams and lake outlets (Kann 1978, Vavilova and Lewis 1999). Hydrurus foetidus is a widely distributed stenothermic cold-water alga that can resist strong current by forming a gelatinous cover on the substrate (Kawecka 1981, Rott et al. 1999). The most frequent blue-green algae in our study sites were the genera Chamaesiphon and Lyngbya in kryal streams and Oscillatoria and Phormidium in rhithral and kryal lake outlets. Distinct species of Chamaesiphon (e.g., C. fuscus, C. incrustans, C. polonicus) are common in alpine streams and lake outlets, being able to resist winter freezing and summer drought by forming crusts (Kann 1978, Kawecka 1980). Representatives of Lyngbya, Oscillatoria, and Phormidium were dominant taxa in studies of other alpine lake outlets (Kann 1978, Kawecka 1980). Many species of these three genera are typically either free-living or filamentous loosely aggregated forms and are common in standing and slow- flowing waters (Geitler 1971, Rott et al. 1999). Diatom genera common to most of our study sites included Achnanthes spp., Cymbella spp., Diatoma spp., Gomphonema spp., Hannaea arcus, Meridion circulare, and Navicula spp.; taxa widely distributed in all kinds of freshwaters (Brun 1965, Kawecka 1980). Diatoms occurring at kryal sites generally were non- motile with strong attachment abilities that can resist high discharge and abrasion by glacial flour, belonging to the so-called pioneer species with high resilience (e.g., Achnanthes spp., Diatoma spp., Fragilaria spp., Hannaea arcus). Diatoms occurring only in rhithral lake outlets belonged to motile and free-living forms (e.g., Amphora spp., Oscillatoria spp.) and filamentous, loosely aggregated forms (e.g., Navicula spp.). Due to the small sample size, no statement on algal community composition of rhithral streams can be made from our study. Kawecka (1980), however, described the benthic algae of different rhithral streams of European mountains as species rich communities with large numbers of diatom taxa as well as Cyanophyta and Hydrurus foetidus, showing features of lake outlets and lowland streams.

66 62 3. Algal communities in alpine streams Seasonality in algal communities We observed little seasonality in community composition or genera richness, but found seasonal changes in abundances. Many studies of low- and midland streams showed a seasonal succession in algal community structure (see citations in Biggs 1996). Corresponding to our findings, seasonality in algal communities was less pronounced in other studies of high-mountain coldwater streams (Kann 1978, Maier 1994, Kownacki et al. 1997, Passy et al. 1999, Vavilova and Lewis 1999). In general, maximum numbers of green algae (e.g., Ulothrix zonata) were found in summer, blue-green algae in late summer and autumn, and Chrysophyceae (e.g., Hydrurus foetidus) during winter and spring. No distinct seasonal patterns were clear for the diatoms. At our sites, algae showed no seasonal succession in community composition but most genera had high abundances during autumn and winter, being most pronounced at kryal sites. The glacier influenced Joerisee outlet shifted from a rhithral lake outlet assemblage (Amphora perpusilla, Navicula spp., Nitzschia spp., and Synedra rumpens) during late summer and winter to one of a kryal site (lower richness, Hydrurus foetidus) during early summer (July 1998 and 1999) when the lake was affected by high discharge and turbidity from the Joeriglacier (Fig. 3-5). Environmental variables affecting algal communities The predominant determinant of algal community structure in our study was the presence of a glacier. Glacial ice-melt results in strong seasonal fluctuations of flow and temperature that affects channel stability, turbidity (light), and nutrient availability (Milner and Petts 1994). Temperature and light are major factors influencing seasonal changes in algal communities of other alpine systems (Kann 1978, Kawecka 1980). Besides solar intensity, light availability is restricted in kryal systems by high turbidity during summer and, in all systems, by snow cover in winter. Therefore, habitat conditions of kryal systems are most favorable for algae growth during early spring and late autumn (i.e. low flow and high light availability); which is consistent with our findings of algal abundances in kryal sites being highest during autumn and winter. Despite the limiting conditions of kryal systems, algal communities differed among our kryal streams, ranging from taxa-poor sites (e.g., Lang, Tschierva) to taxa-rich communities

67 3. Algal communities in alpine streams 63 similar to those of rhithral sites (e.g., Morteratsch). We suggest that benthic algal characteristics may be dependent on the relative length of the favorable habitat conditions during spring and autumn among different kryal streams. Our data support similar groupings of algal communities as those by Kawecka (1980) who analyzed 70 algal communities in alpine lotic systems of different European mountain ranges. The algal communities differed mainly among polluted and unpolluted streams, but among unpolluted sites algal communities also displayed characteristic structures of glacial streams, lake outlets, and alpine streams of different altitudinal zonation. Accordingly, in our study, another important determinant of benthic algal communities was the presence of an upstream lake. Lake outlets generally buffered temporal fluctuations in environmental parameters, thus leading to a more stable habitat throughout the year. Algal communities of the lake outlets reflected this stability, displaying communities without significant seasonal changes in richness or abundance. Summary and conclusion In summary, algal community composition of alpine streams and lake outlets were generally dominated by diatoms and blue-green algae. Although algal community composition tended to be similar among sites and many taxa occurred in all lotic systems, kryal sites generally were characterized by Hydrurus foetidus (Chrysophyceae) and Chamaesiphon spp. (Cyanophyta), whereas Oscillatoria spp., Phormidium spp. (Cyanophyta), and several diatom genera were indicative of lake outlet communities. The major factor governing the community structure among stream types was the presence of a glacier and the resulting effects on different physical-chemical parameters (e.g., temperature, turbidity, nutrients) and their seasonality. Lakes exerted a stabilizing effect on outlet communities, thus leading to an increased taxonomic richness with less pronounced seasonality. Acknowledgments Special thanks to Michael T. Monaghan in particular, and Andreas Blum, Peter Burgherr, and Marcos de la Puente for assistance in the field; Richard Illi and

68 64 3. Algal communities in alpine streams Bruno Ribi for completion of the chemical analysis in the laboratory; and Esther Keller and Regula Illi for additional identification of the benthic algae. We thank the communes of Pontresina, Samedan, and Klosters for providing access to the sites. We thank Ruben Sommaruga, Eugen Rott and one anonymous reviewer for constructive comments that greatly improved the manuscript. The study was partially funded by a Swiss National Science Foundation Grant (no. 31-50440.97) examining the ecology of alpine lake outlets. References ALGER, A. S., D. M. MCKNIGHT, S. A. SPAULDING, C. M. TATE, G. H. SHUPE, K. A. W ELCH , R. EDWARDS , E. D. ANDREWS, and H. R. HOUSE. 1997. Ecological processes in a cold desert ecosystem: The abundance and species distribution of algal mats in glacial meltwater streams in Taylor Valley, Antarctica. University of Colorado, Boulder, U.S.A.. BIGGS , B. J. F. 1996. Patterns in benthic algae of streams. Pages 31-56 in Stevenson, R. J., M. L. Bothwell, and R. L. Lowe (Editors). Algal Ecology. Academic Press, San Diego, U.S.A.. BIGGS , B. J. F., D. G. GORING , and V. I. NIKORA . 1998. Subsidy and stress responses of stream periphyton to gradients in water velocity as a function of community growth form. Journal of Phycology 34:589-607. BRNMARK, C., and B. MALMQVIST . 1984. Spatial and temporal patterns of lake outlet benthos. Verhandlungen der Internationalen Vereinigung der Limnologie 22:1986-1991. BRUN, J. 1965. Diatomes des Alpes et du Jura. A. Asher & Co., Amsterdam, Netherlands. CATTANEO, A., T. KERIMIAN, M. ROBEGRE, and J. MARTY. 1997. Periphyton distribution and abundance on substrata of different size along a gradient of stream trophy. Hydrobiologia 354:101-110. CHESSMAN, B., I. GROWNS , J. CURREY, and P.-C. N. 1999. Predicting diatom communities at the genus level for the rapid biological assessment of rivers. Freshwater Biology 41:317-331. ELSTER, J., and J. SVOBODA. 1996. Algal diversity, seasonality and abundance in, and along glacial stream in Sverdrup Pass, 79N, Central Ellesmere Island,

69 3. Algal communities in alpine streams 65 Canada. Memoirs of National Institute of Polar Research, Special Issue 51:99- 118. GEITLER , L. 1971. Cyanophyceae. Johnson Reprint Corporation, New York, London. GRIMM , N. B., and S. G. FISHER. 1989. Stability of periphyton and macroinvertebrates to disturbance by flash floods in a desert stream. Journal of the North American Benthological Society 8:293-307. HILL, B. H., A. T. HERLIHY, P. R. KAUFMANN, R. J. STEVENSON, F. H. MCCORMICK, and C. B. JOHNSON. 2000. Use of periphyton assemblage data as an index of biotic integrity. Journal of the North American Benthological Society 19:50-67. KANN, E. 1978. Systematik und kologie der Algen sterreichischer Bergbche. Archiv fr Hydrobiologie/Supplementband 53:405-643. KAWECKA , B. 1965. Communities of benthic algae in the River Bialka and its Tatra tributaries the Rybi Potok and Roztoka. Pages 113-127 in Komitet Zagosp. Ziem Grskich Pan (Editor). Limnol. Invest. in the Tatra Mts and Dunajec River Basin. Krakw, Poland. KAWECKA , B. 1980. Sessile algae in European mountain streams, I. The ecological characteristics of communities. Acta Hydrobiologica 22:361-420. KAWECKA , B. 1981. Sessile algea in European mountain streams, 2. Taxonomy and autecology. Acta Hydrobiologica 23:17-46. KOWNACKI, A., E. DUMNICKA , J. GALAS , B. KAWECKA , and K. W OJTAN. 1997. Ecological characteristics of a high mountain lake-outlet stream (Tatra Mts, Poland). Archiv fr Hydrobiologie 139:113-128. KUHN, D. L., J. L. PLAFKIN, J. JOHN CAIRNS , and R. L. LOWE. 1981. Qualitative characterization of aquatic environments using diatom life-form strategies. Transactions of the American Microscopical Society 100:165-182. MAIER, M. 1994. Die jahreszeitliche Vernderung der Kieselalgengemeinschaft in zwei geologisch unterschiedlichen Fliessgewssern der Alpen und ihre Verteilung auf verschiedenen Substraten. Diatom Research 9:121-131. MALMQVIST , B., and C. BRNMARK. 1984. Functional aspects of a lake outlet versus a springfed stream ecosystem. Verhandlungen der Internationalen Vereinigung der Limnologie 22:1992-1996.

70 66 3. Algal communities in alpine streams MILNER, A. M., and G. E. PETTS . 1994. Glacial rivers: physical habitat and ecology. Freshwater Biology 32:295-307. PASSY, S. I., Y. PAN, and R. L. LOWE. 1999. Ecology of the major periphytic diatom communities from the Mesta River, Bulgaria. International Review of Hydrobiology 84:129-174. ROTT, E., E. PIPP, P. PFISTER , H. V. DAM , K. ORTLER, N. BINDER, and K. PALL. 1999. Indikationslisten fr Aufwuchsalgen in sterreichischen Fliessgewssern. Teil 2: Trophieindikation (sowie geochemische Prferenzen, taxonomische und toxikologische Anmerkungen). WWK, Bundesministerium fr Land- und Forstwirtschaft, Wien, sterreich. ROUND, F. E. 1993. A review and methods for the use of epilithic diatoms for detecting and monitoring changes in river water quality 1993. HMSO, London, U.K.. STENECK, R. S., and M. N. DETHIER . 1994. A functional group approach to the structure of algal-dominated communities. Oikos 69:476-498. TER BRAAK, C. J. F., and P. SMILAUER . 1998. CANOCO Reference Manual and User's Guide to Canoco for Windows: Software for Canonical Community Ordination (version 4). Microcomputer Power, Ithaca NY, Wageningen, Netherlands. TOCKNER, K., F. MALARD, P. BURGHERR , C. T. ROBINSON, U. UEHLINGER , R. ZAH, and J. V. WARD. 1997. Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv fr Hydrobiologie 140:433-463. UEHLINGER , U., R. ZAH, and H. BRGI. 1998. The Val Roseg Project: temporal and spatial patterns of benthic algae in an Alpine stream ecosystem influenced by glacier runoff. Pages 419-424 in Kovar, K., U. Tappeiner, N. E. Peters, and R. G. Craig (Editors). Hydrology, Water Ressources and Ecology in Headwaters. IAHS Press, Wallingford, U.K.. VAVILOVA , V. V., and W. M. J. LEWIS . 1999. Temporal and altitudinal variations in the attached algae of mountain streams in Colorado. Hydrobiologia 390:99- 106. WARD, J. V. 1994. Ecology of alpine streams. Freshwater Biology 32:277-294. ZAR, J. H. 1984. Biostatistical Analysis. Prentice Hall, Englewood Cliffs, New Jersey, U.S.A..

71 Appendix 3-1 Average frequency of algal taxa presented in the figures. Notations of sites are defined in Table 3-1. A full species list is available from EAWAG divison / class Taxon name and Authority notation LB LC LN PM JS LR SS MS TG MG1 MG2 SG GG1 GG2 LG1 LG2 Cyanophyta: Calothrix spp. CALO 5 2 1 1 1 Chamaesiphon spp. CHAM 1 1 3 3 1 Lyngbya spp. LYNG 1 1 1 1 3 1 3 1 1 1 Lyngbya species LYsp 1 1 1 3 Oscillatoria spp. OSCI 1 1 1 2 3 1 Oscillatoria cf. amphibia Agardh OSam Phormidium sp. PHsp 3 1 3 3 Chrysophyceae: Hydrurus foetidus (Villars) Trevisan HYfo 2 5 3 5 3 2 3 2 2 2 2 Bacillariophyceae: Cyclotella spp. CYCL 2 2 1 1 Achnanthes spp. ACHN 2 2 2 3 3 3 2 3 2 1 Achnanthes flexella (Ktzing) Brun ACfl 1 1 1 1 1 2 Achnanthes minutissima Ktzing ACmn 5 5 5 4 5 4 3 Amphora spp. AMPH 2 2 2 1 2 Amphora perpusilla Grunow AMpe 2 2 2 1 2 3. Algal communities in alpine streams Cymbella spp. CYMB 2 2 1 2 2 2 1 4 2 1 1 Cymbella affinis Ktzing CYaf 3 3 1 3 1 1 Cymbella minuta Hilse CYmi 2 3 2 2 2 2 1 Denticula spp. DENT 3 1 1 2 2 1 Denticula elegans Ktzing DEel 3 1 1 2 1 Denticula tenuis Ktzing DEte 2 2 Diatoma spp. DIAT 1 3 1 1 1 1 1 1 2 1 1 1 2 1 1 Fragilaria spp. FRAG 1 1 1 2 1 4 2 Gomphonema spp. GOMP 1 1 1 1 1 1 2 3 1 Gomphonema olivaceoides Hustedt GOod 1 2 Gomphonema olivaceum (Hornemann) Brbisson GOol 1 1 1 2 Gomphonema parvulum Ktzing GOpa 1 1 2 1 1 2 2 Hannaea arcus (Ehrenberg) Patrick HAar 1 1 3 2 2 1 3 2 1 2 1 Meridion circulare (Greville) Agardh MEci 1 1 1 3 1 Navicula spp. NAVI 1 1 1 1 1 1 1 3 1 Nitzschia spp. NITZ 1 1 1 1 1 1 1 Nitzschia species NIsp 2 3 1 Synedra spp. SYNE 1 2 2 1 1 2 1 1 Synedra rumpens Ktzing SYru 1 2 2 1 1 2 1 Synedra species SYsp 3 1 Chlorophyta: Cladophora glomerata (Linn) Ktzing CLgl 1 4 Microspora sp. MIsp 1 2 Ulotrichale ULOT 1 1 1 67

72 68

73 4. Invertebrate habitat relationship 69 4. Macroinvertebrates in alpine streams: community patterns in relation to the habitat templet Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. in preparation. Alpine lotic systems comprise a variety of stream types differing in their environmental conditions and biotic communities. We examined and compared macroinvertebrate communities of four different alpine stream types including streams and lake outlets of rhithral and kryal origin in relation to respective habitat characteristics. Although assemblage composition was similar among stream types, invertebrates showed distinct patterns in the relative contribution of individual taxa among rhithral streams, rhithral lake outlets and kryal sites, with no separation between kryal streams and kryal lake outlets. Dominant taxa at all sites were the Chironomidae. However, rhithral lake outlets also had high densities of non-insect taxa such as Oligochaeta, Nemathelminthes and zooplankton. Kryal sites, although characterized by generally low taxon richness, had surprisingly high abundances of Ephemeroptera and Plecoptera. Rhithral streams contained the most diverse assemblages, being inhabited by both non-insect taxa as well as many Ephemeroptera, Plecoptera and Trichoptera taxa and additional dipteran families. Rhithral lake outlets primarily had higher water temperature and standing crop of algae than rhithral streams, and thus provided a more favorable habitat for non-insect taxa, whereas the glacial influence seemed to be the dominant parameter influencing the invertebrate community of kryal streams and kryal lake outlets. Thus, the different environmental features act as nested "filters", that ultimately dictate the kinds and numbers of species among different alpine streams. Alpine lakes influence environmental conditions of outlets, and therefore, can alter the community assembly, although still being constrained if influenced by a glacier. The distinctiveness of lake outlet communities declines with increasing elevation and glacial influence.

74 70 4. Invertebrate habitat relationship Introduction "Der Wildbach bietet seinen Bewohnern eine Heimat von so ausgeprgtem Character, dass sich dies in der Gestalt und Lebensweise der Bachtiere widerspiegeln muss." Steinmann (1907) Early in the history of limnology, Steinmann (1907) noted that the habitat of streams affects the community composition and respective species traits of organisms. Much later, Southwood (1977) suggested that the habitat provides the templet on which evolution forges characteristic life-history strategies. Following Southwood's (1977, 1988) habitat templet approach, many stream studies have documented a strong relationship between community structure and organism life-history strategies to different habitat characteristics (e.g., Minshall 1988, Death 1995, Statzner et al. 1997, Townsend et al. 1997, Doledec et al. 2000, Habersack 2000). However, because of the hierarchical nature of stream networks, the habitat characteristics of streams act at different scales: at large scales (i.e., the catchment), the geological, lithological and hydrological settings are primary determinants of stream environments, whereas at smaller scales (i.e., mesohabitat), substrate, water depth and velocity, and the input of organic matter play key roles in organism distribution (Frissell et al. 1986). Further, the scaled habitat features can be viewed as nested filters through which species in the regional pool must "pass" to be present at a given locale, thus finally dictating the local distributions of organisms and ultimately community composition (Tonn 1990, Poff 1997). The insularity and distinctiveness of alpine streams (defined as mountain headwaters situated above the treeline) act as major constraints on stream organisms at a coarse level (Mani 1990, Ward 1994). Dominant habitat features of these systems are low temperatures, short growing seasons, an often highly fluctuating discharge regime, turbulent well-oxygenated water, and a low input of organic matter (Mani 1990, Ward 1994, Zah and Uehlinger 2001). Water temperature is a key factor affecting the ecology and evolution of aquatic organisms, and has been ascribed a major role in determining the diversity, distribution and abundance of macro-zoobenthos over altitudinal gradients (Ward and Stanford 1982, Freder 1999). Discharge fluctuations are another

75 4. Invertebrate habitat relationship 71 important habitat characteristics of alpine streams that affects benthic organisms directly (e.g., dislodgement of organisms from the stream bed) as well as indirectly (e.g., by movement of substrate particles) (Ward 1994, Hart and Finelli 1999). At a lower hierarchical level, alpine streams can be differentiated into kryal streams dominated by glacier-melt, rhithral streams being snow-fed, and krenal or spring-fed streams, each with a distinctive set of habitat conditions (Steffan 1971, Ward 1994). A dominant feature defining the ecological condition of these stream types is the origin of the water that generates distinct temporal patterns in flow and temperature. Kryal streams are characterized by temperatures near 0 C, high seasonal and diel discharge fluctuations, high sediment transport and bedload movement, and an usually unstable channel morphology (Milner and Petts 1994, Robinson et al. 2001). Rhithral streams are characterized by a broader temperature range than kryal or krenal systems and a hydrograph dominated by an extended period of snowmelt runoff during early summer, resulting in more stable habitat conditions (Ward 1994). Krenal streams, in contrast, typically show more constant conditions in temperature and flow than the other alpine stream types (Klein and Tockner 2000). Lake outlets, defined as an ecotone forming the longitudinal transition zone between lentic and lotic habitats, represent another distinct stream type (Naiman et al. 1988, Samways and Stewart 1997). Many studies conducted in low- and midland lake outlets found this distinctive habitat to be characterized by relatively stable flow and temperature regimes, and high concentrations of transported organic material (e.g., Illies 1956, Carlsson et al. 1977, Brnmark and Malmqvist 1984, Harding 1994). Alpine lake outlets can be viewed as subtypes of either rhithral or kryal streams. Depending on size, they are expected to attenuate flow and temperature variations and mitigate the pronounced seasonality in physical-chemical characteristics (kryal lake outlets, in particular), thus leading to a relatively stable channel. Lake outlets, therefore, presumably provide more benign habitats for benthic organisms in an otherwise harsh alpine environment (Milner and Petts 1994, Burgherr and Ward 2000, Hieber et al. in press-b).

76 72 4. Invertebrate habitat relationship The first information on alpine streams derived from longitudinal studies that showed a distinct altitudinal zonation in the zoobenthic communities and compositional differences among stream types (e.g., Steinmann 1907, Dodds and Hisaw 1925, Kawecka et al. 1971, Steffan 1971). In the past few years, high- elevation streams, particularly rhithral and kryal systems, and associated invertebrates have attracted much interest. The characteristic macrozoobenthos of rhithral streams includes the four insect orders Ephemeroptera, Plecoptera, Trichoptera and Diptera, plus the non-insects turbellarians, acarines, oligochaetes and nematodes, thus representing a relatively diverse community (Ward 1994). Kryal streams, in contrast, typically are less diverse and show a distinct downstream gradient in the distribution of zoobenthos with chironomid larvae of the genus Diamesa predominating close to the glacier margin and other taxa becoming common further downstream, being associated with increasing channel stability and water temperature (e.g., Steffan 1971, Milner and Petts 1994, Burgherr and Ward 2001, Snook and Milner 2001). Zoobenthic communities of alpine lake outlets, however, have been largely ignored by lotic ecologists (but see Kownacki et al. 1997, Burgherr and Ward 2000, Donath and Robinson 2001). Lowland lake outlet communities generally are dominated by filter feeders such as Simuliidae that are attracted by the high concentrations of transported organic matter from the lake (Illies 1956, Richardson 1984). The few previous studies on alpine lake outlets report low densities of Simuliidae and invertebrate communities not dissimilar from those of other alpine streams. Based on the unique but still little understood nature of alpine streams, we investigated the zoobenthic communities of different alpine stream types including rhithral and kryal streams and lake outlets. Specifically, we determined the dominant environmental characteristics that define the habitat templet for the invertebrate communities and examined whether the diversity and structure of benthic invertebrate assemblages reflected the differences in the habitat templet of the alpine stream types. In particular, we expected lake outlets in general, and kryal lake outlets especially, to represent a less harsh habitat than respective streams, thus having a more temporally constant and diverse community.

77 4. Invertebrate habitat relationship 73 Study sites All study sites were first-order headwater streams in the Swiss Alps above treeline at elevations between 1930 and 2500 m a.s.l. (Table 4-1). Sites consisted of alpine lake outlets and non-outlet streams of rhithral and kryal origin in 3 of the 4 major drainages in Switzerland (Rhine, Danube and Po). Catchment size ranged from 0.6 to almost 20 km2, with the largest catchments associated with kryal sites having 40 to 90 % of their area glaciated. The data on catchment and glaciated area were based on topographic maps updated between 1991 and 1995. Because the glaciers draining into our study streams retreated between 5 and 30 m/y (IAHS/UNESCO 1998), the actual glaciated area at the time of the study was less than listed in Table 4-1. The Joerisee outlet was classified as rhithral, but it was partially influenced by the melt waters of the Joeriglacier that first flows into a small proglacial lake above the Jrisee (Kreis 1921). Lake surface area ranged from 0.7 to 27 ha and was largest for the Table 4-1 Location and general characteristics of the study sites. Notations of the study sites as further referred to in the text. Origin of the water: k = kryal, r = rhithral; stream type: L = lake outlet, S = (non-lake outlet) stream. Catchment includes total area (km2) and percentage glaciated area (%); Pfankuch's index of channel stability: 15 = 'excellent', 16 - 30 = 'good', 31 45 = 'fair', and 46 60 = 'poor'. Site Notation Origin Stream Catchment Elevation Catchment Slope Pfankuch 2 type (m a.s.l.) (km ) % (%) index Lago Nero LN r L Po 2387 0.8 0 5 25 Puoz Minor PM r L Danube 2336 1.8 0 9 29 Lago Bianco LB r L Po 2076 2.1 0 4 29 Joerisee JS r(k) L Rhine 2489 3.4 11 5 24 Moesa M r S Po 2300 0.6 0 16 19 Gglia G r S Danube 2310 5.8 0 10 37 Lej Roseg LR k L Danube 2159 19.3 44 2 49 Steinsee SS k L Rhine 1934 7.3 71 4 50 Tschierva TG k S Danube 2100 14.7 42 4 58 Steinlimi SG k S Rhine 2090 3.4 92 2 37

78 74 4. Invertebrate habitat relationship kryal lakes (12 and 27 ha). The study streams had slopes ranging from 2 to 16 %, median depths of 11 to 28 cm, average baseflow widths of 3 to 12 m, and substrate of a pebble/cobble matrix. Channel stability, estimated as the stream bottom component of the Pfankuch's index (1975), was 'good' ( 30) at the rhithral lake outlets and the rhithral stream M; 'fair' (31 - 45) at the rhithral stream G and the kryal stream SG; and 'poor' (46 - 60) at the kryal lake outlets and the kryal stream TG. The land surface of all catchments are covered mainly by bare rock, with sparse vegetation of grass and low shrubs of alder (Alnus sp.) and willow (Salix sp.) also present. All sites are in the crystalline area of the Swiss Alps where bedrock mainly consists of granite and gneiss (Spicher 1980). More detailed information on the habitat characteristics of the sites is given in Hieber et al. (in press-b). Methods Field collections Ten sites were sampled monthly to bimonthly (when accessible) between June 1998 and September 2000 comprising on average two annual cycles. On each visit, we collected benthic invertebrates, benthic organic matter, periphyton and seston, and recorded physical and chemical water characteristics. On each date, three quantitative samples (Hess-sampler, 0.04 m2, 100-m mesh) of benthic invertebrates and organic matter were collected from each site and water depth and velocity (Mini Air 2, Schiltknecht AG, Gossau, Switzerland) at 0.6 depth was measured to characterize habitat conditions for each sample. Due to strong snow cover, some sites on certain dates were not suited for sampling with a Hess-sampler, in which case benthic invertebrates were sampled qualitatively using a kick-net (100-m mesh). In addition, qualitative samples were collected seasonally from most common habitat types over a period of 5 minutes. This method provided a semi-quantitative assessment of the benthic community to compare both invertebrate richness and density. All benthic samples were preserved with 2 to 4 % formalin for later analysis. Periphyton was sampled by collecting ten stones (b axis = 2.8 - 14.0 cm, median = 7.0 cm) from predominant instream habitats at each site. The stones were stored at -25C until processed in the laboratory. Four samples of

79 4. Invertebrate habitat relationship 75 transported organic (POM) and inorganic particulate matter (PIM) (seston) were collected using a nylon drift net (100-m mesh) for 60 to 180 s at approximately 0.6 depth. The period of sampling depended on clogging of the net by transported sediments, particularly in kryal sites, or organic particles. Velocity was recorded at the net aperture to standardize seston to unit volume of water (m3) and the contents of each sample were stored at - 25C until processed in the laboratory. Physical and chemical measures On each sampling date, specific conductance (WTW model LF 325, Weilheim, Germany) and turbidity (Cosmos, Zllig AG, Rheineck, Switzerland; in nephelometric turbidity units, NTU) were measured in the field at each site. A water sample (1 L) was collected from each site, filtered through pre-ashed glass fiber filters (Whatman GF/F filters; 45 mm diameter) and analyzed for the following chemical parameters in the laboratory: ammonium (NH4), nitrate (NO3), nitrite (NO2), dissolved (DN) and particulate nitrogen (PN), soluble reactive phosphorus (SRP), total dissolved (DP) and particulate phosphorus (PP), dissolved (DOC) and particulate organic carbon (POC), total suspended solids (TSS) and ash-free dry mass (AFDM). Determination of each parameter followed the methods described in detail in Tockner et al. (1997). Water temperature was recorded hourly at each site with temperature loggers (Minilog, Vemco, Nova Scotia, Canada). Total annual degree days (DDa) were calculated as the sum of the average daily temperatures over one year. We calculated and averaged the DDa values for each possible period of 12 months per site (e.g., July '98 to July '99, August '98 to August '99, etc.). Missing data, e.g., due to exposure to air, were estimated using the rate of thermal increase or decrease (DD/d) of the preceding and following months. The overall rate of thermal change (DD/d) was calculated by computing the slope of the linear regression through the accumulated average daily temperature data for 1 to 4 preceding months, depending on linearity.

80 76 4. Invertebrate habitat relationship Organic matter Benthic organic matter (BOM) was estimated from each Hess sample standardized to unit area (m2). After removing the invertebrates, the remaining benthic material from each sample was dried at 60C, weighed to the nearest 0.1 mg, ashed for 3 h at 500C, and reweighed to calculate BOM as ash-free dry mass (AFDM). Periphyton was removed from each stone by scrubbing with a metal bristle brush. Two aliquots of 3 to 5 ml of the algal suspension from each stone were filtered through pre-ashed glass fibre filters (Whatman GF/F) for determination of chlorophyll a content and AFDM. Chlorophyll a was extracted in 90 % ethanol by boiling at 70 C for 10 min and measured with a reversed phase HPLC and a subsequent diode array-detector (Bio-Tek, Basel, Switzerland) (Meyns et al. 1994). Ash-free dry mass was determined as the difference between the weight of the filter with the algal suspension after drying at 60 C and the weight after ashing for 3 h at 500 C. We calculated the surface area of each stone as the area of an ellipsoid based on measurements of the a- and b-axis following Uehlinger (1991) to standardize chlorophyll a and AFDM concentrations to unit area (m2). Each seston sample was filtered through a weighed pre-ashed glass fiber filter (Whatman GF/F), dried at 60C and weighed to the nearest 0.1 mg to determine total transported particulate matter (PM). Each filter was then ashed for 3 h at 500C and reweighed to calculate the particulate organic matter (POM) as AFDM. Particulate inorganic matter (PIM) was estimated as the difference between PM and POM. Invertebrates In the laboratory, we counted all invertebrates from each kick net and Hess sample and identified them to the lowest possible level using a dissecting microscope. Although many groups were identified to species level, all further analyses were based on genus or (sub-) family level to allow inclusion of young instar larvae that were too small to be identified to species. Regardless, many genera were present as only one species, thus resulting in very similar patterns.

81 4. Invertebrate habitat relationship 77 Hydrozoa, Nemathelminthes, Oligochaeta, Hydracarina, Cladocera, Copepoda, Ostracoda and Collembola were not identified further. Functional feeding guilds were assigned according to Moog (1995). Density and richness of total zoobenthos, EPT (Ephemeroptera, Plecoptera, Trichoptera) and non-insect taxa, and Simpson's index of diversity (D) were determined at two levels: for each sample, and for each sampling date per site (density and Simpson's D based on average densities of the three replicate samples). Densities are given as individuals/m2 for Hess-samples and as individuals/5min sample for kick-net samples. Simpson's index of diversity (D) was calculated using the formula D = (ni(ni-1))/(N(N-1)), where ni = the number of individuals in the ith species and N = the total number of individuals. High D values indicate a more equal distribution of taxa, even though the total number of taxa present may be low. In contrast, high taxon richness associated with low D values indicate a strong dominance by one or a few taxa and the presence of many rare taxa. Simpson's index is weighted towards the most abundant species and is less sensitive to species richness than the Shannon-Wiener index (Magurran 1991). In addition, we analyzed beta diversity (taxa turnover among sampling dates or sites), gamma diversity (total taxon richness per site), and fit of four species abundance models (geometric, log series, broken stick, truncated log normal) for each site. Beta diversity is a measure of similarity between samples of different dates or sites. Species abundance distributions use the number of taxa present and their relative abundances, representing a progression from a geometric series where a few species are dominant and the remainder fairly uncommon, through a log-series and log-normal distribution where species of intermediate abundance become more common, to a broken stick model where species are equally abundant (Magurran 1991). Diversity analyses and test of fitness to the various distribution models were performed using Pisces 2.3 (Lymington, UK). Statistical analyses Coefficient of variation (CV, in %) indicates the relative amount of variation between different populations or sampling sets and was calculated as a measure of temporal change among sampling dates (Sokal and Rohlf 1995). As the sampling design was unbalanced between the different sites, we used analysis of

82 78 4. Invertebrate habitat relationship covariance (ANCOVA) to test for differences among sites with date as the covariate, and ANOVA to test for differences among dates within each site (Sokal and Rohlf 1995). If differences were detected (significance level of p = 0.05), the Tukey HSD test was used to determine which values actually differed. All data were ln (x+1) transformed to meet the assumptions of a normal distribution (Sokal and Rohlf 1995). Differences in the macroinvertebrate community composition per site were analyzed using Ward's cluster analysis based on standardized average densities. Total invertebrate densities of both Hess and kick-net samples, were averaged by the number of respective samples taken and afterwards standardized as (raw value mean) / standard deviation per taxon. Standardized scores have a mean = 0 and a standard deviation = 1, thus, the influence of differences in absolute densities per taxon among the different sites is reduced and the data resemble more an absence/presence matrix. Macroinvertebrate densities also were analyzed using principal component analysis (PCA) to determine general spatial and temporal patterns among the study sites. The applicability of a PCA to the data set was tested calculating the length of gradient using DCA. A gradient length 4 SD indicates an unimodal response of the data and requires a correspondence analysis, however, the data clearly showed linear responses (length of gradient 2.6 SD). We analyzed invertebrate data by means of a covariance matrix PCA after ln(x+1) transformation and centering to reduce strong intertaxonomic differences in densities. Because of the two different sets of invertebrate data (i.e. qualitative kick-net and quantitative Hess samples), two separate analyses were performed. Invertebrate densities (no/m2) of the replicate Hess samples for each sampling date and site were averaged per taxon, whereas invertebrate communities of the kick-nets are kept as density/sample. Multivariate analyses were computed using the ADE-4 software (Thioulouse et al. 1997). Macroinvertebrate density, genus richness (total, EPT and non-insect taxa richness) and Simpson's index of diversity were analyzed using regression against different environmental parameters. To assess for general environmental parameters explaining differences among macroinvertebrate communities, we included the various invertebrate indices per sampling date and site in a

83 4. Invertebrate habitat relationship 79 stepwise multiple regression. Some environmental parameters were excluded if they correlated strongly with another (r > 0.8). Regression was calculated first (1) with different instream variables measured at each sampling date including water chemistry, mean daily temperature, turbidity, velocity, water depth, transported particulate material (TOM, and TIM), and benthic organic matter (BOM and periphyton AFDM), to test for variables that explained temporal changes in the various invertebrate indices at a small scale (sample location) within each stream, and second (2) with the same instream variables measured at each sampling date plus three general stream site variables that remained constant for the different sampling dates within a stream site (i.e., slope, DDa, and Pfankuch stability index) to assess the importance of between stream characteristics. In the regression models, the limit for regression was set for an adjusted r2 = 0.3 and the significance level at p = 0.05, thus regression models with adjusted r2 0.3 and/or p 0.05 were interpreted as having no explanatory value. In multiple regression analyses, the adjusted r2-value represents the coefficient of multiple determination r2 divided by the respective degrees of freedom, and the standardized regression coefficient Beta equals the regression coefficient B for the variables standardized to a mean of 0 and a standard deviation of 1 (StatSoft 1995). Thus, the adjusted r2 allows one to compare among regressions with different numbers of variables and the magnitude of Beta allows a comparison of the relative contribution of each independent variable. All statistical analyses were performed using Statistica 5.1 (StatSoft 1995), if not otherwise stated. Results Habitat characteristics Physical-chemical characteristics: Detailed analyses regarding physical, chemical, morphological and stability patterns characterizing the study streams were previously described (Hieber et al. 2001, Hieber et al. in press-b), therefore only a brief summary follows (Table 4-2). In general, water temperature was lowest at the kryal streams (maximum 4.6 C, 300 DDa), slightly higher at the kryal lake outlets (maximum 9.1 C, 540 to 670 DDa), and highest at the rhithral streams and lake outlets (maximum 11 to 17 C, > 900 DDa).

84 80 4. Invertebrate habitat relationship Table 4-2 Physical and chemical characteristics of each study site (notations defined in Table 4-1). Velocity and water depth characterize habitat conditions for the Hess samples. Values are listed as mean standard deviation, except for water temperature being listed as range (minimum maximum) and average standard deviation accumulated annual degree days (DDa). NO2+NO3 = nitrite + nitrate; PP = particulate phosphorus. Site Water temperature Conductance NO2+NO3 Turbidity PP Velocity Water depth (C) DDa (S/cm) (g/L) (NTU) (g/L) (m/s) (cm) LN 0 - 14.8 913 * 13 2 171 42 1.1 0.5 1.9 0.9 0.23 0.13 17 2 PM 0 - 13.5 1225 68 10 188 36 1.9 1.8 2.8 1.6 0.36 0.26 11 4 LB 0 - 15.3 1351 49 56 21 200 51 1.7 2.4 2.7 2.3 0.28 0.26 29 9 JS 0 - 16.8 957 59 20 5 127 46 17 19 5.0 4.3 0.42 0.21 18 7 M 0 - 13.0 1244 34 29 2 354 33 0.7 0.8 0.8 0.5 0.27 0.25 11 5 G 0 - 10.7 904 14 76 26 177 52 0.6 0.3 0.8 0.8 0.41 0.31 20 9 LR 0 - 9.1 670 41 43 18 192 39 143 86 32 16 0.48 0.30 29 12 SS 0 - 6.4 544 18 43 19 141 50 90 25 24 9 0.54 0.34 26 12 TG 0 - 4.6 301 43 49 31 305 121 378 947 69 88 0.53 0.39 25 11 SG 0 - 4.0 226 6 22 12 134 63 94 120 40 43 0.56 0.28 22 11 * Temperature data only for one year; DDa calculated based on data from 8 months with missing data estimated by regressions (see text for details). Study sites had mean values of specific conductance ranging from 13 to 76 S/cm, nitrite+nitrate concentrations from 127 to 354 g/l, and particulate nitrogen from 7 to 29 g/l, with no distinct patterns among stream types. Concentrations of turbidity, ammonium, nitrite, soluble reactive, total dissolved and particulate phosphorus, and ash-free dry mass, generally were higher at kryal sites than rhithral sites. For example, average turbidity was less than 2 NTU at rhithral sites but reached average concentrations of 90 to 380 NTU at the kryal sites (p 0.02). Concentrations of particulate phosphorus were most strongly correlated with turbidity (r = 0.99), being significantly highest at kryal ( 24 g PP /L) and lowest at rhithral sites (< 3 g PP /L) (p < 0.001). The glacial influence also was obvious at the Jrisee outlet (average of 17 NTU and 5 g PP /L), a site partially influenced by glacial discharge during summer (Table 4-2). Seasonality of physical-chemical parameters was more pronounced at kryal than at rhithral

85 4. Invertebrate habitat relationship 81 sites, being associated with the discharge regime. Turbidity, particulate phosphorus, total suspended solids, ash-free dry mass and ammonium showed a pronounced seasonality with peak concentrations during summer high flow at kryal sites and during early summer, corresponding to snow melt fed discharge, at rhithral sites. Seasonal patterns were opposite for specific conductance, showing low values during high summer discharge at kryal sites but higher values during late summer under baseflow conditions at rhithral sites (Hieber et al. in press-b). Average velocity, measured above each Hess sample, ranged from 0.23 m/s to 0.56 m/s and was highest at the kryal sites ( 0.48 m/s) (Table 4-2). Average water depth ranged from 0.11 to 0.29 m and was highest at the lake outlets LB, LR and SS (> 0.25 m) and lowest at the rhithral sites ( 0.2 m, except the rhithral lake outlet LB). In general, velocity and water depth were lowest, although non- significantly, during winter and highest during summer. Organic matter: Benthic organic matter (BOM) ranged on average from less than 2 g/m2 at the kryal sites to 14 g/m2 at the rhithral L. Bianco outlet (Fig. 4-1). Concentrations of BOM were markedly higher at the rhithral sites (average 3 to 14 g/m2) compared to the kryal sites and reached peak concentrations > 30 g/m2 at the rhithral L. Bianco outlet. Periphyton organic matter (AFDM) and chlorophyll a were positively correlated with each other (r = 0.79). Periphyton was lowest at the kryal sites (on average 10 g AFDM and 20 mg chl.a /m2) and highest at the rhithral lake outlets LB, PM and JS (on average > 20 g AFDM and 34 mg chl.a /m2). At the rhithral streams, however, periphyton was as low as at the kryal site s, whereas BOM concentrations were more similar to rhithral lake outlets. Transported particulate organic matter (POM) ranged on average from 0.01 to 0.1 g/m3. Concentrations of POM were highest at the rhithral streams and the kryal stream TG but showed no distinct differences between kryal and rhithral sites (Fig. 4-1). Seasonal patterns were most distinct for periphyton AFDM at kryal sites with peak concentrations in late autumn / early winter, whereas no such distinct patterns were obvious for BOM and POM.

86 82 4. Invertebrate habitat relationship 40 Benthic organic matter (g AFDM/m2) 30 20 10 0 LB LN PM JS G M LR SS TG SG * 50 Periphyton (g AFDM/m2) 40 30 20 10 0 LB LN PM JS G M LR SS TG SG 0.4 Seston (g AFDM/m3) 0.3 0.2 0.1 0.0 LB LN PM JS G M LR SS TG SG Figure 4-1 Concentrations of benthic (from Hess samples), periphytic and transported particulate organic matter (as AFDM) at 10 different sites (notations defined in Table 4-1). Box plots show the median (solid line), mean (dotted line), 25 th and 75th percentile (), 10th and 90th percentile (low and high whiskers), and 5th and 95th percentile () for all measurements during the study period. * concentrations only for 2 sampling dates available.

87 4. Invertebrate habitat relationship 83 Organic matter concentrations at rhithral sites showed high spatial (same sampling date) as well as temporal variation, but without consistent patterns. However, note that the sampling of organic matter was not possible during winter. Invertebrates A total of 67 taxa based on species, and 50 taxa based on genus level, was found during the study period. Of the 67 taxa, 8 species belonged to 2 Ephemeroptera families, 10 species to 6 Plecoptera families, 15 species to 4 Trichoptera families, 21 taxa to 12 Diptera families, and the remaining 13 taxa to Collembola, Coleoptera, and the non-insect classes (Appendix 4-1). The most common taxa were the Turbellaria Crenobia alpina, Nemathelminthes, Oligochaeta, the Ephemeroptera Baetis alpinus and Rhithrogena spp., the Plecoptera Leuctra spp., and the chironomid subfamilies Diamesinae and Orthocladiinae. Community composition: Total community composition based on standardized average densities clustered the study sites into 3 main groups: (1) the two rhithral streams, (2) the rhithral lake outlets LN, PM and LB, and (3) all glacial- influenced sites, i.e., the kryal sites and the Jrisee outlet (Fig. 4-2). The last 16 14 Linkage distance 12 10 8 6 4 M G LN PM LB TG SG SS LR JS rhithral kryal r(k) Figure 4-2 Cluster diagram of the 10 study sites (notations defined in Table 4-1) based on standardized average invertebrate abundances of all taxa. Lake outlets are indicated with underlined notations.

88 84 4. Invertebrate habitat relationship Table 4-3 Number of total and EPT taxa, Whittaker's beta diversity (Beta), Simpson's index of diversity (D), average density from Hess-samples, and fit to the four main species distribution models (NO: p-value < 0.05) for each study site (notations and types defined in Table 4-1). Indices are listed as total per site and (CV as %) of sampling dates at the genus/family level. Taxon EPT Beta Simpson's Density Geometric Log Log Broken Site Type richness richness diversity diversity D (ind./m2) series series normal stick LN rL 20 (36) 6 (65) 1.32 3.7 (25) 11387 (49) NO 0.56 0.61 NO PM rL 24 (21) 7 (52) 0.76 2.8 (27) 5640 (15) NO 0.23 0.14 NO LB rL 25 (22) 8 (40) 0.91 1.8 (54) 29361 (73) NO 0.24 0.60 NO JS rL 23 (24) 9 (40) 1.18 3.9 (65) 3528 (31) NO 0.56 0.67 NO M rS 34 (20) 13 (24) 0.78 3.9 (47) 4846 (132) NO 0.12 0.35 NO G rS 35 (15) 13 (20) 0.75 3.9 (27) 5164 (82) NO 0.17 0.62 NO LR kL 22 (26) 11 (28) 1.49 3.5 (36) 2199 (90) NO 0.62 0.73 NO SS kL 12 (82) 3 (150) 2.89 1.1 (47) 670 (128) NO 0.33 0.94 NO TG kS 17 (65) 6 (90) 2.86 1.3 (49) 3375 (227) NO 0.48 0.93 NO SG kS 12 (29) 6 (47) 1.85 1.2 (22) 4337 (66) NO 0.32 0.57 NO two groups were closer to each other, whereas the rhithral streams were most separate. Jrisee outlet was grouped with the kryal sites indicating its glacial influence, but within this group it was most distinct from the others. Within the kryal sites, lake outlets and streams showed no distinct separation. The two rhithral streams had the highest number of taxa (34 and 35 taxa from the total of 50 taxa, based on genus level). Between 20 and 25 taxa occurred at the rhithral lake outlets and the kryal Roseg outlet and less than 18 taxa at the other kryal sites (Table 4-3). About one third of the taxa present at each site belonged to the EPT (Ephemeroptera, Plecoptera, Trichoptera) with the most EPT taxa occurring at the rhithral streams (13 EPT taxa, = 38 %) and the least at the kryal Steinsee outlet (3 EPT taxa, = 25 %). Temporal changes in both total and EPT taxon richness were lowest in the rhithral streams (CV < 25 %) and highest in the kryal Steinsee outlet and the kryal stream TG (CV 65 %). At the remaining sites, changes in total taxon richness were relatively low (CV 36 %), but higher in number of EPT taxa (CV 40 %). Taxa turnover among sampling

89 4. Invertebrate habitat relationship 85 dates (Beta) showed similar patterns, being lowest at the rhithral sites (Beta 1.0) and highest in the kryal stream TG and the kryal Steinsee outlet (Beta 2.5). However, the number of days and seasons sampled differed among sites, therefore somewhat affecting the calculations of temporal changes. Corresponding to taxon richness, Simpson's index of diversity (D) was highest at the rhithral streams (D = 3.9) and lowest at the kryal sites (D 1.3) (Table 4- 3). Diversity also was high at the two lake outlets LN and JS (D 3.7) although taxon richness was 23. Diversity at the rhithral L. Bianco outlet (D = 1.8), in contrast, was almost as low as at the kryal sites. Temporal changes in diversity seemed to be relatively stable with CV 65 % at all sites and showed no distinct differences among sites. Invertebrate densities in the Hess-samples were on average between 3000 and 6000 ind./m2; however, low densities (< 700 ind./m2) occurred at the kryal Steinsee outlet and high densities of > 10,000 with maximum values of > 60,000 ind./m2 were found at two rhithral lake outlets. Temporal changes in density tended to be most distinct at the rhithral streams and the kryal sites (CV > 80 %, except SG) and most stable at the rhithral lake outlets (CV < 50 %). Taxa distribution, based on average densities, followed both a log series as well as a truncated log normal distribution at all sites (p < 0.1), but none of the sites fit the geometric series or the broken stick model (p < 0.001) (Table 4-3). Seasonality: Neither invertebrate density nor taxon richness showed any clear seasonal pattern (Fig. 4-3). In figure 4-3, data from both Hess- as well as kick- net samples are combined to get a more complete picture, and densities are given as ind./sample to standardize data from both sample methods. In general, highest densities were found during August and/or September in the rhithral lake outlets LB and LN, the rhithral stream M and the kryal stream SG; during February in the rhithral stream G; and during late April in the kryal sites TG and LR. However, differences among dates were not significant at any rhithral lake outlet or the kryal stream SG (p > 0.07). Taxon richness generally was low during June and July and increased between August and October at most sites. At the

90 86 4. Invertebrate habitat relationship Hess 20 LB LN Kick-net 2000 15 10 1000 5 0 0 PM JS 20 2000 Date vs LNh Av abundance Date vs LNq Av abundance 15 Date vs LNh Av richness Date vs LNq Av richness 10 1000 5 0 0 Average abundance/sample G M 20 Taxon richness 2000 15 10 1000 5 0 0 LR SS 20 2000 15 10 1000 5 0 0 TG SG 20 2000 15 10 1000 5 0 0 Jun.98 Dec.98 Jun.99 Dec.99 Jun.00 Jun.98 Dec.98 Jun.99 Dec.99 Jun.00 Figure 4-3 Macroinvertebrate abundance (black) and taxon richness (grey) for Hess () and kick-net ( ) samples during the study period at 10 different sites (notations defined in Table 4-1). Vertical bars indicate 1 SD based on 3 replicate Hess samples.

91 4. Invertebrate habitat relationship 87 kryal sites TG and LR, in contrast, highest taxon richness occurred during late April and May, and at the kryal Steinsee outlet during July (Fig. 4-3). However, when data were available for longer than one year, interannual patterns were not consistent (e.g., peak densities in the rhithral stream M during August 2000 of 677 144 ind. and 16 1 taxa, in contrast to 65 4 ind./sample and 13 1 taxa in August 1999). The relative abundance of the main taxonomic groups during the period of high flow (July) and low flow (October) are shown in figure 4-4. Dominant taxa were the chironomids (11 - 100 %), oligochaetes (0 71 %) and heptageniid mayflies (0 65 %). Temporal changes in community composition were most pronounced at the rhithral lake outlets LB and JS, the rhithral streams and the kryal Steinsee outlet, whereas almost no changes were obvious at the remaining sites. At the rhithral sites, relative abundance of Chironomidae decreased from 100 Others Diptera Chironomidae 80 Trichoptera Plecoptera Relative abundance (%) Ephemeroptera 60 40 20 0 c i LN - Jul LN - Oct LR - Jul LR - Oct SG - Oct* a b e g h JS - Jul JS - Oct d G - Jul* G - Oct* SG - Jul f SS - Jul SS - Oct PM - Jul PM - Oct LB - Jul LB - Oct M - Jul* M - Oct* TG - Jul TG - Oct Figure 4-4 Relative abundances (%) of the main invertebrate groups at the 10 study sites (notations defined in Table 4-1) during a period of high (July) and low flow (October) (all samples taken in 1998, except * taken in 1999). Different colors indicate Ephemeroptera, Plecoptera, Trichoptera, Chironomidae, other Diptera, and 'Others' (combining non-insect taxa and the few Coleoptera and Collembola); relative abundances of single taxa within these groups are marked by a black line.

92 88 4. Invertebrate habitat relationship July to October with an increase of the non-insect taxa (Oligochaeta and Copepoda) at the rhithral lake outlets and the Ephemeroptera (Baetis alpinus and small heptageniids) and Plecoptera (Leuctra spp.) at the rhithral streams. At the kryal Steinsee outlet, the relative abundance of the Chironomidae increased from July (45 %) to October (80 %), whereas all other taxa decreased or even disappeared resulting in a decrease in total taxon richness from 10 during July to 3 in October. At the two kryal streams, total taxon richness increased from 1 and 3 in July to 7 in October, however, no pronounced changes in relative abundance was obvious with the Chironomidae strongly dominating the communities during both months (> 85 %) (Fig. 4-4). Principal component analyses based on the macroinvertebrate communities of both qualitative kick-net and quantitative Hess samples resulted in almost identical groupings of sites (Fig. 4-5). PCA clearly distinguished three groups: rhithral streams, rhithral lake outlets and kryal sites and showed no differences between kryal streams and kryal lake outlets. Rhithral streams were distinguished by high taxon densities on both axes: non-insect taxa, such as Oligochaeta, Nemathelminthes and the crustaceans Copepoda, Ostracoda and Cladocera defining the first axis, and Ephemeroptera-Plecoptera-Diptera (EPD) taxa, such as Baetis alpinus, Rhithrogena spp., small unidentifiable heptageniids, Leuctra spp., Limoniidae and Psychodidae on the second axis. Rhithral lake outlets were separated primarily by the non-insect taxa on the first axis, whereas the kryal sites were situated along the second axis following a gradient from Steinsee outlet and Tschierva stream being characterized by the absence of most taxa, to Steinlimi stream and Roseg outlet being characterized by the presence of EPD taxa. Samples of each site and date varied in their position in the ordination, however, and no clear temporal patterns were obvious. Kryal sites showed a tendency to shift along the second axis and rhithral lake outlets along the first axis, each from only a few to no taxa in June, July and August to an increase in EPD taxa and non-insect taxa during late autumn to early spring. However, samples in the same month varied strongly among years, thus not allowing determination of clear temporal patterns.

93 4. Invertebrate habitat relationship 89 a) Qualitative kick-net samples: Copepoda -4.8 -0.94 LN 4.8 -3.3 Oligochaeta 1.5 -0.21 7.8 1.8 JS Ostracoda LB SS PM Nemathelminthes F2: 16.4 % TG Collembola Simulium Chironomidae Rhyacophila Rhabdiopteryx SG LR Psychodidae Nemoura Heptageniidae Leuctra Crenobia alpina M G Rhithrogena Baetis F1: 32.6 % b) Quantitative Hess samples: Copepoda Ostracoda Cladocera 6.3 0.42 LN -13 6.6 Oligochaeta -2.8 0.53 -12 -2.3 SS Nemathelminthes JS TG SG Hydracarina F2: 21.8 % PM LB Chironomidae Rhabdiopteryx Rhyacophila Dictyogenus Collembola Nemoura Protonemura Limoniidae Ecdyonurus LR M Crenobia alpina Leuctra G Baetis Heptageniidae Rhithrogena F1: 29.6 % Figure 4-5 Ordination from a principal component analysis of invertebrate abundances at 10 alpine study sites based on a) qualitative kick-net samples, and b) quantitative Hess samples. In the factor map, the small circles represent the average score for each study site and are connected by lines to actual scores for each sampling date (dates are not labeled as a matter of clearness). Grouping of sites is highlighted with large circles. In the taxa ordination, length and direction of each arrow indicates the contribution of each respective taxon to axes F1 and F2. The F1 axis explained 32.6 % and the F2 axis 16.4 % of the variation for the qualitative samples, and 29.6 % and 21.8 % for the Hess samples.

94 90 4. Invertebrate habitat relationship Functional feeding guilds were dominated by collector-gatherers (on average 40 98 %) at all sites, reflecting the high densities of Chironomidae (except for Tanypodinae being predators) and Oligochatea (Fig. 4-6). Scrapers, mainly represented by ephemeropteran larvae, and shredders, comprising plecopteran and trichopteran larvae, reached highest relative abundances (up to 20 %) at the rhithral and kryal streams and the kryal Roseg outlet. In the rhithral lake outlets, in contrast, the second dominant guild (up to 40 %) included primarily Nemathelminthes and the Crustacea, combined as an 'undefined' group. Filter feeders were least abundant and reached highest values (5 %) only at the rhithral P. Minor outlet. Invertebrate habitat relationship Multiple linear regression of the various invertebrate indices gave different results depending on the scale of the independent (physical-chemical) variables (Table 4-4). In general, the inclusion of between stream characteristics resulted in better regressions (higher adjusted r2), although not necessarily in an increase in statistical significance (p). Density, as well as the number of non-insect taxa, were best explained when regressed only against instream parameters, whereas total and EPT taxon richness and Simpson's index of diversity (D) were best 100 Undefined Predator Filter feeder Average relative abundance (%) 80 Collector-gatherer Scraper Shredder 60 40 20 0 JS LB LN PM G M LR SS TG SG Figure 4-6 Average relative abundance (%) of the main functional feeding guilds at each study site (notations defined in Table 4-1).

95 4. Invertebrate habitat relationship 91 Table 4-4 Results of stepwise multiple regression analyses between macroinvertebrate density and diversity indices and instream environmental variables and instream environmental variables combined with local between stream variables by sampling date and site. Only variables with p < 0.05 are listed following decreasing beta-values with relationship indicated with '+' and ' '. Abbreviations see text. p = significance level of the regression model; bold = regression models with adjusted r2 0.4 and p < 0.05. Dependent variable Instream Instream + between stream Independent variable Adjusted r2 p Independent variable Adjusted r2 p Total density + mean temp., - POC 0.22 < 0.001 + mean temp., - POC 0.23 < 0.001 EPT density + BOM, - TIM 0.12 < 0.001 - Pfankuch, + slope, + BOM 0.21 0.01 Non-insect density + mean temp., + Per.OM, 0.58 0.001 + DDa, + mean temp. 0.70 0.37 - velocity, - turbidity Taxon richness - turbidity, + mean temp. 0.25 < 0.001 + slope, - Pfankuch 0.61 0.03 # EPT taxa 0.08 0.001 + slope, - Pfankuch, + depth 0.36 0.04 # Non-insect taxa + mean temp., - turbidity 0.45 < 0.001 + DDa, + slope, + mean 0.65 0.55 temp., - NO2+NO3 Simpson's D 0.10 0.001 + slope, - Pfankuch 0.39 < 0.001 explained with between stream characteristics. Density and richness of non- insect taxa were both positively related to mean daily temperature (beta 0.4) and negatively, but less strongly, to turbidity (beta -0.2). Additionally, periphyton organic matter (beta = 0.27) and velocity (beta = -0.19) were included in the regression model explaining non-insect density. Total density also was best explained by mean daily temperature (beta = 0.38); however, the adjusted r2 of 0.22 indicates a low explanatory value (Table 4-4). The most significant variables explaining the regression model for total and EPT taxon richness and Simpson's D were slope and the Pfankuch bottom index, indicating the superimposing influence of between stream site compared to within stream characteristics. Slope was positively related (beta 0.43) to the indices reflecting the high taxon richness and diversity at the rhithral streams, whereas Pfankuch index was negatively related (beta -0.26) indicating the low

96 92 4. Invertebrate habitat relationship taxon richness and diversity of kryal sites throughout the sampling period. In addition, water depth also was positively related to EPT taxon richness (beta = 0.30), indicating the influence of temporal changes in instream characteristics. Discussion General differences in habitat characteristics Environmental conditions of the examined alpine streams and lake outlets were strongly influenced by the presence of a glacier and the associated seasonality in glacial melt (also see Hieber et al. in press-b). Thus, differences in discharge regime were a major factor influencing environmental conditions among the alpine stream types. In kryal streams, temporal changes in flow directly control physical chemical parameters such as flow velocity, temperature and nutrients (Hieber et al. in press-b), but also indirectly influence other environmental processes such as standing crop biomass (Milner and Petts 1994, Uehlinger et al. 1998, Tockner et al. 1999). Rhithral streams, in contrast, were less affected by such strong discharge fluctuations, and therefore, were more stable in their environmental conditions, although exhibiting a wider annual temperature range (Kawecka et al. 1971, Lavandier and Dcamps 1984, Ward 1994). Alpine lakes have been found to influence environmental characteristics of outlet streams (Kann 1978, Kawecka 1980, Hieber et al. in press-b). In this study, the main differences between lake outlets and streams were related to water temperature, reaching higher maximum values, accumulating more annual degree days, and having reduced daily fluctuations in lake outlets compared with streams not influenced by lakes. Moreover, kryal lake outlets showed a reduced seasonality of other physical-chemical parameters, such as turbidity, particulate and soluble reactive phosphorus, and total dissolved solids. Rhithral lake outlets had higher periphyton biomass compared to respective streams (also see Hieber et al. in press-b). In addition to seasonal fluctuations in the habitat characteristics, inter- annual differences in environmental conditions such as climate and geomorphology further influence stream habitat conditions (Minshall 1988, Smith et al. 2001). For example, the seasonal patterns in the environmental variables of our study streams often differed between the two years of study. Hydrological

97 4. Invertebrate habitat relationship 93 data from several streams all over Switzerland showed that in general, 1998 was a relatively dry year, whereas discharge was well above-average in 1999 (Landeshydrologie und -geologie 1998, 1999). Thus, the significance of intra- annual biotic patterns can vary among years depending on the superimposing influence of inter-annual environmental conditions. Invertebrates Community composition: Benthic invertebrates showed distinct patterns in their community composition between rhithral streams, rhithral lake outlets and kryal sites, but no separation between kryal streams and respective lake outlets. Regardless, many of the common taxa were present at most study sites and, as shown in other studies on alpine stream systems (e.g., Kawecka et al. 1971, Snook and Milner 2001, Burgherr et al. in press), differences between the communities were based more on the relative contributions of taxa than on their presence or absence. We also found no major differences among the invertebrate communities of the three different drainages, suggesting that invertebrate communities originated from the same species pool. Species distribution models are one way to compare and interpret community patterns (Magurran 1991). Assemblages at all the study sites followed a truncated log normal distribution indicating the communities had a few dominant taxa and several taxa of intermediate abundance. Dominant taxa at all sites were the Chironomidae, comprising up to 100 % of the assemblages at kryal sites. In addition, rhithral lake outlets were characterized by high densities of non-insect taxa such as Oligochaeta, Nemathelminthes and crustaceans. In fact, rhithral streams represented the most diverse sites and accommodated both, non-insect taxa as well as EPT and other dipteran taxa besides chironomids, thus being a combination of the assemblages of both rhithral lake outlets and kryal sites. Most of the common and abundant taxa found in the present study, especially in rhithral streams, are frequently found in other alpine streams throughout central Europe, and included the mayflies Baetis alpinus and Rhithrogena loyolaea, the stoneflies Nemoura mortoni, Leuctra spp. and Protonemura spp., the caddisflies Rhyacophila spp. and various limnephilid species, diverse dipterans, such as Diamesinae, Simuliidae, and Limoniidae, and

98 94 4. Invertebrate habitat relationship the turbellarian Crenobia (Planaria) alpina (e.g., Steinmann 1907, Kawecka et al. 1971, Lavandier and Dcamps 1984, Burgherr et al. in press). Others, however, occurred only locally, being recorded in only a few studies. For example, the stonefly Rhabdiopteryx alpina was common in the kryal stream SG and also was reported from other kryal streams in the Alps by Robinson et al. (2001), but it was rare or even absent at our other sites. In general, it usually is found along the banks and in connected side-arms of the epi- to metarhithron (Moog 1995, Tachet et al. 2000). Apparently, Rhabdiopteryx alpina is adapted to unfavorable periods, e.g., high discharge, by having an adult flight period between late spring and summer (Lubini, pers. communication). However, it still is unclear why Rhabdiopteryx alpina was dominant at only the one kryal stream. Kryal sites generally were characterized by a low taxon richness and a dominance by the chironomid Diamesinae. Other taxa occasionally occurring in these sites included the ephemeropteran families Baetidae and Heptageniidae, the plecopteran families Nemouridae, Leuctridae and Taeniopterygidae, and other dipterans such as Orthocladiinae and Limoniidae. However, taxon richness varied among the kryal sites. Other taxa typically found common in kryal streams (see Milner et al. 2001), occurred only in low abundances at one or two of our kryal sites (e.g., Perlodidae, Simuliidae) or were absent (e.g., Rhyacophilidae, Chironominae, Empididae, Tipulidae). The invertebrate communities of kryal sites also showed no general differences between the lake outlets and streams. A higher taxon richness, a more diverse community and a lower taxa turnover was found at the kryal Lej Roseg outlet compared to its respective kryal stream, thus confirming patterns found by Burgherr and Ward (2000) at these same sites. In contrast, the same number of taxa occurred at the Steinsee outlet and its respective stream and taxa turnover was higher at the lake outlet. Invertebrate communities of rhithral lake outlets differed substantially in their relative abundances from rhithral streams as well as from the kryal sites. Rhithral lake outlet communities were strongly dominated by non-insect taxa such as Oligochaeta, Nemathelminthes, and the crustaceans Copepoda and Ostracoda. The crustaceans are common in lake outlets of all altitudes and can be of both benthic and planktonic origin (e.g., Kreis 1921, Richardson and

99 4. Invertebrate habitat relationship 95 Mackay 1991, Kownacki et al. 1997, Donath and Robinson 2001). Generally, crustaceans derive from the lakes and are accidentals in benthic samples, thus representing a lentic influence on the outlet communities. However, crustaceans that occurred in the rhithral streams indicate the presence of truly benthic copepods and ostracods. In concordance to early findings of Illies (1956) who studied invertebrates in Lapland lake outlets and streams, the same taxa also occurred in rhithral streams, although in lower abundances. For example, Oligochaeta have been reported as an abundant taxon in various kryal and rhithral streams and lake outlets and many of the most dominant species are cosmopolitan (Lavandier and Dcamps 1984, Jacobson et al. 1997, Donath and Robinson 2001, Malard et al. 2001, Milner et al. 2001). Filter feeding taxa have been shown to dominate lake outlet communities (Richardson and Mackay 1991). However, most of these studies have been conducted on low- and midland lake outlets. In the present study, collector- gatherers significantly dominated all sites and filter feeding taxa (here mainly Simuliidae) generally were rare, being abundant only at one rhithral outlet (average relative abundance of 5 %). It appears that the distribution patterns of filter feeding insects in alpine environments is rather inconsistent: filter feeders were abundant in some alpine lake outlets (Bushnell et al. 1987, Harding 1994, Kownacki et al. 1997, Donath and Robinson 2001), but were low in abundance or even absent in others (Kownacka and Kownacki 1972, Harding 1994, Burgherr and Ward 2000). Obviously, the invertebrate communities of alpine lake outlets, being dominated by collector-gatherers, differ substantially from those of low - and midland lake outlets. Seasonal patterns: Various studies have shown that macroinvertebrates in kryal streams clearly display seasonal changes in abundance and diversity that reflect seasonal changes in environmental conditions (Burgherr and Ward 2001, Freder et al. 2001, Robinson et al. 2001), whereas macroinvertebrate communities in rhithral streams are more constant over the annual cycle (Kawecka et al. 1971, Ravizza and Dematteis 1978, Lavandier and Dcamps 1984). In the present study, kryal sites showed the highest taxa turnover, and in general, most taxa were present in late spring. In the rhithral sites, taxa turnover was lower and taxon

100 96 4. Invertebrate habitat relationship richness tended to be highest during late summer / early autumn, although showing no clear general patterns. We also found no clear differences in the seasonality of invertebrate communities between streams and lake outlets. Temporal changes in rhithral lake outlets, reflected by the CV, were higher for taxon richness but lower for density relative to rhithral streams. Among kryal sites, the outlet SS and the stream TG always showed the highest temporal changes in abundance and taxa turnover, resulting from the generally species poor community, but no seasonal patterns were evident. There are several possible explanations: first, sampling extent differed among sites, thus important periods may have been missed at some sites. However, no consistent seasonal patterns were obvious at sites with more extensive sampling, suggesting that other environmental parameters probably influence invertebrate communities in these alpine streams. For example, inter-annual differences in environmental conditions can lead to a shift in intra-annual habitat characteristics as well as invertebrate patterns (Minshall 1988, Smith et al. 2001). Invertebrate communities and habitat characteristics Community patterns among streams were influenced by different habitat characteristics of the individual stream types. Dominant habitat parameters controlling taxon richness and diversity were slope, Pfankuch bottom index of channel stability and mean daily water temperature. Temperature is a major factor controlling differences in community composition as well as individual life history traits (e.g., Steinmann 1907, Dodds and Hisaw 1925, Steffan 1971, Ward and Stanford 1982, Rossaro 1991). It also is one of the most apparent abiotic variables related to changes along altitudinal gradients (Ward 1985, Jacobson et al. 1997). Our study sites differed significantly in their temperature range as well as in their accumulated annual degree days that ranged from 230 to 1350. Mean daily temperature was a main determinant of temporal differences in invertebrate densities, particularly for non-insect taxa, whereas differences in degree days among streams showed no statistical affect. Related to temperature, the period of snow and ice cover (Dodds and Hisaw 1925, Kownacka and Kownacki 1968) and flow interruption by freezing or drying (Dodds

101 4. Invertebrate habitat relationship 97 and Hisaw 1925, Kawecka et al. 1971, Donath and Robinson 2001) further influence invertebrate communities in alpine stream systems. For example, temporary alpine lake outlets have invertebrate communities dominated by taxa able to complete development within a few months such as Simuliidae and Chironomidae, and thus persist during unfavorable periods as terrestrial adults (Kownacka and Kownacki 1972, Kownacki et al. 1997, Donath and Robinson 2001). We do not have definite information on the length of ice and snow cover or possible freezing of the streams. Yet, L. Nero and its outlet were covered by ice and snow the longest, had a low number of degree days, and had a lower taxon richness compared to the other rhithral lake outlets. None of our sites went dry during summer, but freezing may have been possible at the outlet of P. Minor, based on the low average depth (11 cm) and the small lake area (0.7 ha), and thus, taxa with short life cycles (e.g., Simuliidae) may be favored. Lake outlets also have been shown to have higher concentrations of seston that is expected to support filter feeding taxa such as Simuliidae (Richardson and Mackay 1991). In the present study, concentrations of transported particulate organic material were rather low at all sites whereas the inorganic fraction was as high or up to 10 times higher than the organic fraction (Hieber et al. in press- a), and thus probably was the main reason for the low densities of Simuliidae. Another important structuring force in alpine streams, glacial streams in particular, is channel stability and hydraulic stress resulting from strong seasonal and diel discharge fluctuations (Ward 1994, Friberg et al. 2001, Milner et al. 2001), and many models have included channel stability in explaining community pattern (Townsend and Hildrew 1994, Death 1995, Milner et al. 2001). Recently, Milner et al. (2001) suggested that macroinvertebrate occurrence in glacial streams is controlled by maximum water temperature and channel stability. Milner and Petts (1994) predicted the increase in temperature and the attainment of a relatively stable, single-thread, meandering channel below glacial lakes provided more optimal habitat conditions that allowed for an increased taxon richness. Corresponding to the relation of invertebrate patterns with habitat parameters (as suggested by Milner et al. 2001), within the studied kryal sites, taxon richness (based on family level) was highest in the kryal Lej Roseg outlet which had the highest maximum water temperature and the lowest

102 98 4. Invertebrate habitat relationship Pfankuch's index of channel stability, but lowest in the other kryal lake outlet SS despite a markedly higher maximum temperature relative to the kryal streams. The lower numbers of taxa present in the kryal Steinsee outlet and the respective Steinlimi stream compared to the Roseg outlet and the respective Tschierva stream probably was a result of the shorter sampling periods of the Steinsee and Steinlimi due to inaccessibility during winter. However, taxon richness within the two systems showed almost opposite patterns: being higher at the warmer and more stable Roseg outlet than at the respective stream, but in contrast to the predictions of Milner and Petts (1994), being lower at the warmer but less stable Steinsee outlet compared to its respective stream. Thus, in concordance to the results of the multiple regression analysis, the influence of channel stability seemed to be the overriding factor determining the community structure in our study streams, best reflected in the distinct grouping of species rich, stable rhithral sites from species poor, unstable kryal sites. Richness (total, EPT and non-insect taxa) and diversity indices were significantly explained by differences in channel stability and slope between streams, whereas changes in density were best explained by instream differences in mean daily water temperature, corresponding to results obtained for different arctic streams and lake outlets (Friberg et al. 2001). In the present study, however, regression with total density was of low explanatory value (adjusted r2 = 0.22), whereas density of non-insect taxa was significantly positively correlated with mean daily temperature and periphyton biomass and negatively with velocity and turbidity. Copepoda and Ostracoda were mainly derived from upstream rhithral lakes, and thus, occur primarily in low flowing clear habitats. In the present study, high Oligochaeta densities were found attached to moss (during periods of high benthic algal biomass), thus at times of high mean daily water temperatures. Minshall (1984) emphasized that organic matter not only serves as a food source but also acts as substratum on which to live. In alpine streams, benthic algae, moss and macrophytes in particular, may have provided a distinct habitat for non-insect taxa such as Oligochatea. In summary, we found channel stability to be the major habitat characteristic affecting assemblage structure in alpine streams and lake outlets. However, mean daily water temperature and a suitable substratum substantially

103 4. Invertebrate habitat relationship 99 increased the densities of some taxa. Returning to Southwood's habitat templet concept (1977, 1988), certain life-history strategies of macroinvertebrates appear to be characteristic in these alpine environments (Lavandier and Dcamps 1984): (i) some species show a short life cycle during which they can exploit periods of optimal available resources, and are characterized by rapid growth that assures the sustenance of populations during unfavorable periods (e.g., Simuliidae); (ii) other species have life cycles that extend over two or more years during which several generations of the same species can coexist (e.g., Ephemeroptera). This permits them to occupy more kinds of habitat and perhaps make the overall population less susceptible to disturbance or periods of unfavorable environmental conditions. Lake outlets as "hot spots" in alpine environments? We initially expected alpine lake outlets to provide more bengin lotic habitats in an otherwise harsh alpine environment, and thus be inhabited by a temporally more constant and diverse macroinvertebrate assemblage. However, we found that the invertebrate communities in alpine rhithral lake outlets only differed from respective streams in the relative contribution of individual taxa, and not in their overall composition. Here, the glacial influence on channel stability seemed to be the dominant factor affecting the invertebrate community of kryal lake outlets, overriding any ameliorating effect of the lake. Rhithral lake outlets were characterized primarily by a higher water temperature and algal biomass than rhithral streams, thus providing a more favorable habitat for non-insect taxa. Our results indicate that the distinctiveness of lake outlet communities declines with increasing elevation and glacial influence. In conclusion, we can arrange the different environmental features that distinguish the habitats of our alpine stream systems as nested "filters" that "screen" the species in the regional species pool by their functional attributes, thus finally dictating the structure of the invertebrate community within a stream (Fig. 4-7) (Tonn 1990, Poff 1997). At the largest scale, climate determines the inter- and intra-annual hydrologic and thermal regime that influences intermittency and permanence of flow and thus, "filters" the alpine community out of the regional species pool. At the next level, the primary water source, i.e. the presence of a glacier, influences channel

104 100 4. Invertebrate habitat relationship stability and seasonality, leading to a further constraint on population distribution. The channel type represents another "filter". Here, depending on its size, the presence of an upstream lake affects environmental conditions in the outlet, such as water temperature and seasonality, although still being strongly influenced by the present glaciers. Alpine lakes also affect concentrations of transported organic matter, and as they act more as sinks than as sources, change outlet communities from an expected filter feeder dominated to a collector-gatherer dominated assemblage. Finally, within a stream reach, differences in substratum or food resources can favor certain species such as Oligochaeta in the present study. Although not addressed in this study, biotic interactions may further constraint some species distributions. Regional species pool Climate - Hydrology - Permanence Alpine community Primary water source - Kryal - Rhithral - Krenal Figure 4-7 Conceptual model of landscape filters Kryon / Rhithron / Krenon (sensu Tonn 1990, Poff Channel type 1997) determining the - Lake outlet - Main channel community composition in - Braided channel different alpine stream Stream / Lake outlet types. Each hierarchical community level represents an Patch heterogeneity environmental feature and - Substratum - Food resources the dominant habitat characteristics that Local community further "screen" the invertebrate species.

105 4. Invertebrate habitat relationship 101 Acknowledgements Special thanks to Mike Monaghan, Peter Burgherr, Marcos de la Puente, Chregu Dinkel, Christian Rust, Andreas Blum, Ulrich Donath, Florian Malard, Monika Winder, Heiko Rinderspacher, and Donna Anderson for assistance in the field, and Richard Illi, Bruno Ribi and Gabriella Meier-Brgisser for completion of the chemical analysis in the laboratory. Rolf Glatthaar helped in identification of the Simuliidae. We thank the communes of Pontresina, Samedan, and Klosters for providing access to the sites. The study was partially funded by a Swiss National Science Foundation Grant (no. 31-50440.97) examining the ecology of alpine lake outlets. References BRNMARK, C., and B. MALMQVIST . 1984. Spatial and temporal patterns of lake outlet benthos. Verhandlungen der Internationalen Vereinigung der Limnologie 22:1986-1991. BURGHERR , P., and J. V. W ARD. 2000. Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung der Limnologie 27:1587-1590. BURGHERR , P., and J. V. WARD. 2001. Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps). Freshwater Biology 46:1705-1721. BURGHERR , P., J. V. WARD, and C. T. ROBINSON. Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps). Journal of the North American Benthological Society (in press). BUSHNELL , J. H., S. Q. FOSTER , and B. M. W AHLE. 1987. Annotated inventory of invertebrate populations of an alpine lake and stream chain in Colorado. Great Basi n Naturalist 47:501-511. CARLSSON, M., L. M. NILSSON, B. SVENSSON, and S. ULFSTRAND. 1977. Lacustrine seston and other factors influencing the blackflies (Diptera: Simuliidae) inhabiting lake outlets in Swedish Lapland. Oikos 29:229-238.

106 102 4. Invertebrate habitat relationship DEATH, R. G. 1995. Spatial patterns in benthic invertebrate community structure: products of habitat stability or are they habitat specific? Freshwater Biology 33:455-467. DODDS , G. S., and F. L. HISAW . 1925. Ecological studies on aquatic insects. IV. Altitudinal range and zonation of mayflies, stoneflies, and caddisflies. Ecology 6:380-390. DOLEDEC, S., J. M. OLIVIER , and B. STATZNER. 2000. Accurate description of the abundance of taxa and their biological traits in stream invertebrate communities: Effects of taxonomic and spatial resolution. Archiv fr Hydrobiologie 148:25-43. DONATH, U., and C. T. ROBINSON. 2001. Ecological characteristics of lake outlets in Alpine environments of the Swiss Alps. Archiv fr Hydrobiologie 150:207-225. FRIBERG , N., A. M. MILNER, L. M. SVENDSEN, C. LINDEGAARD, and S. E. LARSEN. 2001. Macroinvertebrate stream communities along regional and physico-chemical gradients in Western Greenland. Freshwater Biology 46:1753-1764. FRISSELL , C. A., W. J. LISS , C. E. WARREN, and M. D. HURLEY. 1986. A hierarchical framework for stream habitat classification: viewing streams in a watershed context. Environmental Management 10:199-214. FREDER, L. 1999. High alpine streams: cold habitats for insect larvae. Pages 181- 196 in Margesin, R., and F. Schinner (Editors). Cold-Adapted Organisms - Ecology, Physiology, Enzymology and Molecular Biology. Springer, Berlin, Germany. FREDER, L., C. SCHTZ, M. WALLINGER , and R. BURGER . 2001. Physico-chemistry and aquatic insects of a glacier-fed and a spring-fed alpine stream. Freshwater Biology 46:1673-1690. HABERSACK, H. M. 2000. The river-scaling concept (RSC): a basis for ecological assessments. Hydrobiologia 422:49-60. HARDING , J. S. 1994. Variations in benthic fauna between differing lake outlet types in New Zealand. New Zealand Journal of Marine and Freshwater Research 28:417-427. HART, D. D., and C. M. FINELLI . 1999. Physical-biological coupling in streams: the pervasive effects of flow on benthic organisms. Annual Review of Ecology and Systematics 30:363-395.

107 4. Invertebrate habitat relationship 103 HIEBER, M., C. T. ROBINSON, S. R. RUSHFORTH , and U. UEHLINGER. 2001. Algal communities associated with different alpine stream types. Arctic Antarctic and Alpine Research 33:447-456. HIEBER, M., C. T. ROBINSON, and U. UEHLINGER . Seasonal and diel patterns of invertebrate drift in different types of alpine streams. Freshwater Biology (in press-a.). HIEBER, M., C. T. ROBINSON, U. UEHLINGER, and J. V. W ARD. Are alpine lake outlets less harsh than other alpine streams? Archiv fr Hydrobiologie (in press-b). IAHS/UNESCO. 1998. Fluctuations of glaciers 1990-1995. World Glacier Monitoring Service (WGMS), Zurich, Switzerland. ILLIES , J. 1956. Seeausflu-Bioznosen lapplndischer Waldbche. Entomologisk Tidskrift 77:138-513. JACOBSON, D., R. SCHULTZ, and A. ENCALADA. 1997. Structure and diversity of stream invertebrate assemblages: the influence of temperature with altitude and latitude. Freshwater Biology 38:247-261. KANN, E. 1978. Systematik und kologie der Algen sterreichischer Bergbche. Archiv fr Hydrobiologie/Supplementband 53:405-643. KAWECKA , B. 1980. Sessile algae in European mountain streams, I. The ecological characteristics of communities. Acta Hydrobiologica 22:361-420. KAWECKA , B., M. KOWNACKA , and A. KOWNACKI . 1971. General characteristics of the biocoenosis in the streams of the Polish High Tatras. Acta Hydrobiologica 13:465-476. KLEIN, B., and K. TOCKNER. 2000. Biodiversity in springbrooks of a glacial flood plain (Val Roseg, Switzerland). Verhandlungen der Internationalen Vereinigung der Limnologie 27:704-710. KOWNACKA , M., and A. KOWNACKI . 1968. The influence of ice cover on bottom fauna in the Tatra streams. Acta Hydrobiologica 10:95-102. KOWNACKA , M., and A. KOWNACKI. 1972. Vertical distribution of zoocenoses in the streams of the Tatra, Caucasus and Balkan Mts. Verhandlungen der Internationalen Vereinigung der Limnologie 18:742-750. KOWNACKI, A., E. DUMNICKA , J. GALAS , B. KAWECKA , and K. W OJTAN. 1997. Ecological characteristics of a high mountain lake-outlet stream (Tatra Mts, Poland). Archiv fr Hydrobiologie 139:113-128.

108 104 4. Invertebrate habitat relationship KREIS , A. 1921. Die Jriseen und ihre postglaziale Besiedlungsgeschichte. Internationale Revue der gesamten Hydrobiologie und Hydrographie IX:189- 212, 241-288, 451-490. LANDESHYDROLOGIE UND -GEOLOGIE. 1998, 1999. Hydrologisches Jahrbuch der Schweiz. Eidgenssisches Departement des Innern, Bern, Schweiz. LAVANDIER , P., and H. DCAMPS . 1984. Estaragne. Pages 237-264 in Whitton, B. A. (Editor). Ecology of European Rivers. Blackwell Scientific Publications, Oxford, U.K.. MAGURRAN, A. E. 1991. Ecological diversity and its measurements. Chapman & Hall, London, U.K.. MALARD, F., M. LAFONT , P. BURGHERR , and J. V. W ARD. 2001. A comparison of longitudinal patterns in hyporheic and benthic oligochaete assemblages in a glacial river. Arctic Antarctic and Alpine Research 33:457-466. MANI , M. S. 1990. Fundamentals of high altitude biology. Aspect Publications Ltd., London, U.K.. MEYNS , S., R. ILLI, and B. RIBI . 1994. Comparison of chlorophyll-a analysis by HPLC and spectrophotometry: where do the differences come from? Archiv fr Hydrobiologie 132:129-139. MILNER, A. M., J. E. BRITTAIN, E. CASTELLA , and G. E. PETTS. 2001. Trends of macroinvertebrate community structure in glacier-fed rivers in relation to environmental conditions: a synthesis. Freshwater Biology 46:1833-1847. MILNER, A. M., and G. E. PETTS . 1994. Glacial rivers: physical habitat and ecology. Freshwater Biology 32:295-307. MINSHALL, G. W. 1984. Aquatic insect-substratum relationships. Pages 358-400 in Resh, V. H., and D. M. Rosenberg (Editors). The Ecology of Aquatic Insects. Praeger Publishers, New York, U.S.A.. MINSHALL, G. W. 1988. Stream ecosystem theory: a global perspective. Journal of the North American Benthological Society 7:263-288. MOOG, O. 1995. Fauna Aquatica Austriaca. Wasserwirtschaftskataster, Bundes- ministerium fr Land- und Forstwirtschaft, Wien, sterreich. NAIMAN, R. J., H. DCAMPS, J. PASTOR , and C. A. JOHNSTON. 1988. The potential importance of boundaries to fluvial ecosystems. Journal of the North American Benthological Society 7:289-306.

109 4. Invertebrate habitat relationship 105 PFANKUCH, D. J. 1975. Stream reach inventory and channel stability evaluation. U.S.D.A. Forest Service, Missoula, Montana, U.S.A.. POFF, N. L. 1997. Landscape filters and species traits: towards mechanistic understanding and prediction in stream ecology. Journal of the North American Benthological Society 16:391-409. RAVIZZA, C., and E. R. DEMATTEIS . 1978. Mountain brooklet stonefly-fauna of northern Apennine. Atti della Societ italiana di Scienze naturali e del Museo civico di Storia naturale di Milano 119:229-242. RICHARDSON, J. S. 1984. Effects of seston quality on the growth of a lake-outlet filter feeder. Oikos 43:386-390. RICHARDSON, J. S., and R. J. MACKAY. 1991. Lake outlets and the distribution of filter feeders: an assessment of hypotheses. Oikos 62:370-380. ROBINSON, C. T., U. UEHLINGER , and M. HIEBER. 2001. Spatio-temporal variation in macroinvertebrate assemblages of glacial streams in the Swiss Alps. Freshwater Biology 46:1663-1672. ROSSARO , B. 1991. Chironomids and water temperature. Aquatic Insects 13:87-98. SAMWAYS , M. J., and D. A. B. STEWART. 1997. An aquatic ecotone and its significance in conservation. Biodiversity and Conservation 6:1429-1444. SMITH , B. P. G., D. M. HANNAH, A. M. GURNELL, and G. E. PETTS . 2001. A hydrogeomorphological context for ecological research on alpine glacial rivers. Freshwater Biology 46:1579-1596. SNOOK, D. L., and A. M. MILNER. 2001. The influence of glacial runoff on stream macroinvertebrate communities in the Taillon catchment, French Pyrenees. Freshwater Biology 46:1609-1623. SOKAL , R. R., and F. J. ROHLF 1995. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Company, New York, U.S.A.. 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. SPICHER , A. 1980. Geologische Karte der Schweiz. Schweizerische Geologische Kommission, Bundesamt fr Landestopographie, Wabern, Schweiz.

110 106 4. Invertebrate habitat relationship STAT SOFT , I. 1995. STATISTICA for Windows (Computer program manual). Tulsa, OK. STATZNER , B., K. HOPPENHAUS , M.-F. ARENS , and P. RICHOUX. 1997. Reproductive traits, habitat use and templet theory: a synthesis of world-wide data on aquatic insects. Freshwater Biology 38:109-135. STEFFAN, A. W. 1971. Chironomid (Diptera) biocoenoeses in Scandinavian glacier brooks. Canadian Entomologist 103:477-486. STEINMANN, P. 1907. Die Tierwelt der Gebirgsbche - eine faunistisch-biologische Studie. Annales de Biologie Lacustre 2:30-162. TACHET, H., P. RICHOUX, M. BOURNAUD, and P. USSEGLIO -POLATERA. 2000. Invertbrs d'eau douce. CNRS ditions, Paris, France. THIOULOUSE, J., D. CHESSEL , S. DOLDEC, and J.-M. OLIVIER . 1997. ADE-4: a multivariate analysis and graphical display software. Statistics and Computing 7:75-83. TOCKNER, K., F. MALARD, P. BURGHERR , C. T. ROBINSON, U. UEHLINGER , R. ZAH, and J. V. WARD. 1997. Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv fr Hydrobiologie 140:433-463. TOCKNER, K., D. PENNETZDORFER , N. REINER, F. SCHIEMER , and J. V. W ARD . 1999. Hydrological connectivity, and the exchange of organic matter and nutrients in a dynamic river-floodplain system (Danube, Austria). Freshwater Biology 41:521-535. TONN, W. M. 1990. Climate change and fish communities: a conceptual approach. Transactions of the American Fisheries Society 119:337-352. TOWNSEND , C. R., S. DOLDEC, and M. R. SCARSBROOK. 1997. Species traits in relation to temporal and spatial heterogeneity in streams: a test of habitat templet theory. Freshwater Biology 37:367-387. TOWNSEND , C. R., and A. G. HILDREW . 1994. Species traits in relation to a habitat template for river systems. Freshwater Biology 31:265-275. UEHLINGER, U. 1991. Spatial and temporal variability of the periphyton biomass in a prealpine river (Necker, Switzerland). Archiv fr Hydrobiologie 123:219- 237.

111 4. Invertebrate habitat relationship 107 UEHLINGER , U., R. ZAH, and H. BRGI. 1998. The Val Roseg Project: temporal and spatial patterns of benthic algae in an Alpine stream ecosystem influenced by glacier runoff. Pages 419-424 in Kovar, K., U. Tappeiner, N. E. Peters, and R. G. Craig (Editors). Hydrology, Water Resources and Ecology in Headwaters. IAHS Press, Wallingford, U.K.. WARD, J. V. 1985. Thermal characteristics of running waters. Hydrobiologia 125:31-46. WARD, J. V. 1994. Ecology of alpine streams. Freshwater Biology 32:277-294. WARD, J. V., and J. A. STANFORD. 1982. Thermal responses in the evolutionary ecology of aquatic insects. Annual Review of Entomology 27:97-117. ZAH, R., and U. UEHLINGER . 2001. Particulate organic matter inputs to a glacial stream ecosystem in the Swiss Alps. Freshwater Biology 46:1597-1608.

112 108 4. Invertebrate habitat relationship Appendix 4-1. List of all taxa found at any site. FFG = functional feeding guild; A = scraper, C = collector, F = filter feeder; P = predator, S = shredder (after Moog 1995). Empty cell: taxon not found; 0: taxon found only once at the respective site; +: 64 individuals/average sample. Taxa FFG JS LB LN PM G M LR SS TG SG Cnidaria Hydrozoa P ++ Plathelminthes Crenobia alpina DANA P + ++ 0 ++ ++ ++ + + 0 Nemathelminthes + ++ ++ ++ + + + 0 + + Oligochaeta C ++ +++ +++ ++ + ++ 0 + + + Arachnida P + + ++ + + + 0 + Bivalvia Sphaerium corneum LINNAEUS F + Crustacea Cladocera ++ ++ Copepoda ++ ++ +++ ++ + ++ 0 Ostracoda ++ ++ ++ ++ + ++ Collembola 0 + + ++ ++ + + + + Coleoptera Coleoptera sp.1 0 Ochthebius sp. - adult A + 0 Helophorus sp. - larvae A 0 Ephemeroptera Baetis alpinus P ICTET AC + + ++ +++ ++ ++ + ++ Ecdyonurus spp. A 0 0 + ++ + Rhithrogena spp. A + ++ ++ ++ + ++ ++ Plecoptera Nemoura mortoni RIS S ++ ++ + Nemurella pictetii KLAPALEK AC + 0 Protonemura spp. S + + ++ + + + ++ + Capnia vidua KLAPALEK S + Leuctra spp. S + 0 + + ++ ++ + ++ ++ Rhabdiopteryx alpina KHTREIBER S + + + + +++ Dictyogenus alpinum PICTET P + + ++ + 0 Isoperla spp. P 0 ++ + Perlodes intricatus P ICTET P 0 + 0 + + Siphonoperla montana P ICTET P 0 Trichoptera Hydropsyche spp. F 0 + Rhyacophila spp. P + + + ++ Acrophylax zerberus BRAUER S + + + 0 + 0 0 + Annitella obscurata / S + Chaetopteryx villosa/fusca S + Drusus spp. A(PF) + + + Plectrocnemia conspersa CURTIS P 0 Diptera Ceratopogonidae P + Chironomidae ++ +++ +++ +++ +++ ++ ++ ++ ++ ++ Diamesinae C ++ ++ + ++ + +++ +++ ++ +++ Orthocladiinae C ++ +++ ++ ++ +++ ++ ++ + Tanypodinae P + + + Chironominae C ++ +++ + + + + 0 Dixidae F + Dolichopodidae P 0 + + Empididae P + + + Limoniidae + + + + + + + + + + Dicranota spp. P 0 + Rhypholophus spp. C + + Psychodidae C + ++ 0 Simuliidae F 0 ++ ++ + + + Prosimulium spp. F 0 ++ + + + + Simulium spp. F 0 0 + + 0 Stratiomyidae SA 0 Thaumaleidae AC 0 + Tipuliidae S + 0 Diptera sp.1 0

113 5. Invertebrate drift in different alpine streams 109 5. Seasonal and diel patterns of invertebrate drift in different alpine stream types Hieber, M., Robinson C.T., and Uehlinger U. Freshwater Biology, submitted. We examined the seasonal and diel patterns of invertebrate drift in relation to seston and various habitat characteristics in two each of four different kinds of alpine streams (rhithral lake outlets, rhithral streams, kryal lake outlets and kryal streams). Samples were collected at four times of the day (dawn, midday, dusk and midnight) during 3 seasons (spring, summer and autumn). Habitat characteristics differed mainly between rhithral and kryal sites with kryal sites having higher discharge and turbi- dity, lower water temperature, and higher concentrations of ammonium, particulate and soluble reactive phosphorus. Seasonality in habitat charac- teristics was most pronounced for kryal streams with autumn samples being more rhithral in character. Concentrations of seston were lowest in the glacial-influenced lake outlets and slightly higher in the stream sites; no seasonal or diel patterns were evident. Densities of drifting invertebrates averaged less than 100 m-3 and were lowest (less than 10 m-3) at three of the four kryal sites. Taxon richness and diversity were lowest at rhithral lake outlets. On average, less than 30 % of the identified benthic taxa were found in the drift, with Chironomidae dominating drift as well as benthic communities. Drifting invertebrates showed no consistent seasonal pattern. However, densities tended to be highest in spring at rhithral sites and in autumn at kryal sites. No diel periodicity in density was found at any site and the lack of diel pattern may be a general feature of high altitude streams. Glacially influenced habitat parameters were a major factor affecting drift in these alpine streams, whereas no clear differences were observed between streams and lake outlets. Our findings indicate that invertebrate drift in alpine streams is primarily influenced by abiotic habitat conditions, and therefore, substantially differs from patterns observed in low- and midland streams.

114 110 5. Invertebrate drift in different alpine streams Introduction Drift is defined as the downstream transport of stream-dwelling organisms (Waters, 1972). Most studies of drift have focused on macroinvertebrates, but downstream transport also is important for other aquatic biota such as periphyton, larval fishes, and amphibians. Since the discovery of invertebrate drift, various studies have examined the relationship of environmental parameters affecting drift and its ecological significance (e.g., Waters, 1972; Statzner, Dejoux and Elouard, 1984; Brittain and Eikeland, 1988; Allan, 1995). As a result, different terms describing the causes and types of drift have been introduced in the literature, such as 'catastrophic drift', 'behavioral drift', 'active drift', 'constant drift', 'background drift' and 'distributional drift'. Drift is a complex phenomenon influenced by both abiotic factors (e.g., velocity, discharge, water chemistry, temperature, and photoperiod) as well as biotic factors (e.g., benthic density, food resources, predation and competition) that still poses many unanswered questions, and thus a deterministic typology is still under discussion (Statzner et al., 1984; Brittain and Eikeland, 1988; Allan, 1995). For instance, Statzner et al. (1984) found partly contrary relationships of single factors affecting drift (even for the same genus), suggesting that drift of an organism is influenced by the interaction of a large number of factors arranged hierarchically. The ecological significance of drift has been related primarily to behavioral parameters such as locomotion and predator avoidance. Drift enables organisms to escape unfavorable conditions, colonize new habitats, and distribute among patchy habitats (Allan 1995, Brittain and Eikeland 1988). In general, the downstream transport of invertebrates is not constant but varies over time. Many studies have demonstrated an annual and diel periodicity in invertebrate drift and have related these temporal patterns to a variety of abiotic and biotic factors. For example, seasonal patterns in invertebrate drift frequently have been explained by benthic density, species-specific life cycles and seasonal changes in water temperature and flow (Hildebrand, 1974; Brittain and Eikeland, 1988), whereas diel drift periodicity usually is related to predation by visually hunting drift-feeding predators and endogenous circadian rhythms such as locomotory rhythms and foraging behavior (Waters, 1972; Flecker, 1992; Huhta, Muotka and Tikkanen, 2000). However, although many studies repeatedly

115 5. Invertebrate drift in different alpine streams 111 confirmed the diel and/or annual periodicity of drift, others have found no or reduced periodicity depending on several factors, such as the species investigated (Anderwald, Konar and Humpesch, 1991; Waringer, 1992), body size of the drifting invertebrates (Allan, 1984), taxon of predator present (Huhta et al., 1999), or altitude (Brewin and Ormerod, 1994). Most drift studies to date have been conducted in low- and midland streams, whereas little is known about drift in high-elevation or alpine streams (but see Lavandier and Dcamps, 1984; Ilg et al., 2001; Robinson, Tockner and Burgherr, 2002). Alpine streams, defined as being situated above the treeline, differ in many of their environmental conditions from low- and midland streams. They are generally characterized by steep slopes, cold turbulent and oxygen-rich water, low channel stability, low resource availability in respect to nutrients and organic matter, and short growing seasons (Mani, 1990; Ward, 1994; Hieber et al., in press). Seasonality is dominated by a long winter period with water temperatures close to 0 C and often snow-covered or frozen channels, a period of high discharge that restricts primary production during spring and summer, and a relatively stable and productive period during autumn (Uehlinger, Zah and Brgi, 1998). Disturbance in alpine streams generally is related to floods. Seasonal discharge peaks lead to substrate movement and unstable channels, and are more pronounced in kryal (glacier-fed) streams during summer ice-melt than in rhithral (snow-fed) streams. However, alpine lake outlets can ameliorate fluctuations in flow and temperature and therefore are expected to have a more stable channel than respective streams, and thus provide a more stable and favorable habitat for alpine stream invertebrates (Milner and Petts, 1994; Hieber et al., in press). Among the few studies of drift in high-elevation streams, we are aware of only two studies that compared drift patterns from different alpine streams. Ilg et al. (2001) and Robinson et al. (2002) each compared drift patterns in streams of a glacial flood plain. Both studies found a high spatial heterogeneity in diversity and abundance of drifting organisms that was dependent on stream type. They found highest diversity and abundance of drifting invertebrates with lowest seasonal changes in the groundwater-fed channels, in contrast to the low diversity and abundance in the glacial streams. Robinson et al. (2002) further

116 112 5. Invertebrate drift in different alpine streams found that seasonal patterns in drifting invertebrates differed among stream types but also among taxa, and thus suggested that seasonality in drift reflects differences both in flow regime and in the life cycles of particular taxa inhabiting a site. The goal of our study was to examine seasonal and diel differences in drifting invertebrates and seston (total organic and inorganic transported particulate matter) among different kinds of alpine streams (lake outlets and non-outlet streams of both rhithral and kryal origin) and to relate the drift patterns to differences in habitat characteristics comprising various local abiotic and biotic factors. We expected the harsh conditions of kryal sites to reduce drift relative to rhithral sites in concordance to findings of Ilg et al. (2001) and Robinson et al. (2002), and the more stable conditions in lake outlets to increase diversity and abundance of drifti ng invertebrates relative to respective streams. We also assumed that temporal patterns of drift depend on fluctuations in flow, and thus would be more pronounced in kryal than rhithral sites, and in streams relative to lake outlets. Study sites Study streams were in the Swiss Alps above treeline at elevations ranging from 1930 to 2500 m a.s.l. (Table 5-1). Eight streams were studied comprising two each of rhithral lake outlets, rhithral streams, kryal lake outlets, and kryal streams, situated in 3 major drainages (Rhine, Inn, Ticino) of the Swiss Alps. Catchment size ranged from 0.6 to 19.3 km2, with the largest catchments associated with kryal sites having 42 to 92 % of the area glaciated. The outlet of the Jrisee was classified as rhithral, but was influenced somewhat by the Jriglacier that drains into a small lake above the Jrisee. Data on catchment and glaciated area were based on topographic maps updated between 1991 and 1995. Because the glaciers draining into our study streams retreated between 5 and 30 m/y since 1995, the actual glaciated area at the time of the study was less than listed in Table 5-1 (IAHS/UNESCO, 1998). The studied streams had slopes ranging from 2 to 16 %, median depths of 14 to 27 cm, average widths of 3 to 12 m, and substrate of a pebble/cobble matrix. Channel stability, estimated as the stream bottom component of the Pfankuch's index (1975), was 'good' (

117 5. Invertebrate drift in different alpine streams 113 30) at the rhithral lake outlets and the rhithral stream M; 'fair' (31 - 45) at the rhithral stream G and the kryal stream SG; and 'poor' (46 - 60) at the kryal lake outlets and the kryal stream TG. All sites were in the crystalline area of the Swiss Alps with bedrock mainly of granite and gneiss (Spicher, 1980). The catchments were covered primarily by rocks with sparse vegetation of grass and low shrubs of green alder (Alnus viridis) and willow (Salix sp.) also present. More detailed information on the physical-chemical characteristics of these sites is given in Hieber et al. (in press). Table 5-1 Classification and general characteristics of the study streams. k = kryal, r = rhithral, L = lake outlet, S = stream, Pfankuch's index of channel stability: 15 = 'excellent', 16 - 30 = 'good', 31 45 = 'fair', and 46 60 = 'poor'. Site Notation Origin Stream Elevation Catchment Slope Pfankuch type (m a.s.l.) area (km 2) % glaciated (%) index Lago Bianco LB r L 2076 2.1 0 4 29 Jrisee JS r(k) L 2489 3.4 11 5 24 Gglia G r S 2310 5.8 0 10 37 Moesa M r S 2300 0.6 0 16 19 Lej Roseg LR k L 2159 19.3 44 2 49 Steinsee SS k L 1934 7.3 71 4 50 Tschierva TG k S 2100 14.7 42 4 58 Steinlimi SG k S 2090 3.4 92 2 37 Methods Samples were collected four times over a 24 h period at each site during 3 seasons: 09 - 19 August 1999 (summer), 14 - 27 October 1999 (autumn), and 21 June - 06 July 2000 (spring). Sampling occurred between full moons to reduce the potential influence of bright moon light (Allan 1995). No winter sampling was possible due to inaccessibility to most sites. During each 24 h period, samples

118 114 5. Invertebrate drift in different alpine streams were taken at dawn (shifting from 5:00 in spring to 8:00 in autumn), midday (12:00), dusk (shifting from 21:30 in spring to 18:30 in autumn), and midnight (0:00; no sample at the kryal stream TG). Lake outlets were sampled at a distance of 10 to 30 m below the lake. In the following text, the term 'stream' refers only to a lotic system without the influence of a lake, a 'lake outlet', in contrast, is a stream flowing from a lake. Physical and chemical measures At each site, water temperature and specific conductance were recorded at 30 min intervals during the 24 h sampling period with a temperature (Minilog, Vemco, Nova Scotia, Canada) and conductivity recorder (WTW model 325, Weilheim, Germany). Discharge was measured on each sampling occasion at each site following Platts et al. (1983). Also on each sampling date, we measured turbidity (Cosmos, Zllig AG, Rheineck, Switzerland; in nephelometric turbidity units, NTU) in the field and collected a water sample (1 L) for chemical analysis in the laboratory. In the laboratory, aliquots of each water sample were filtered through pre-ashed glass fiber filters (Whatman GF/F filters) and subsequently analyzed for the following chemical parameters: ammonium (NH4- N), nitrate+nitrite (NO2-N+NO3-N), particulate nitrogen (PN), soluble reactive (SRP) and particulate phosphorus (PP), dissolved (DOC) and particulate organic (POC) and total inorganic carbon (TIC), ash-free dry mass (AFDM) and total suspended solids (TSS) using methods described in Tockner et al. (1997). Seston and drifting macroinvertebrates Samples were collected with a nylon drift net (mesh: 100 m, length: 1 m, aperture: 0.01 m2). Samples were taken at approximately 0.6 total depth. On each sampling event, 4 consecutive replicates were collected at periods ranging from 30 to 180 s depending on clogging of the net by transported sediments or organic material. Velocity (Mini Air 2, Schiltknecht AG, Gossau, Switzerland) was recorded at the net aperture to standardize seston and drift parameters to unit volume of water (m3). The contents of each sample were preserved in 2 - 4 % formalin.

119 5. Invertebrate drift in different alpine streams 115 In the laboratory, invertebrates were counted and identified to genus or family level using a dissecting microscope, and their body lengths measured to the nearest 0.5 mm using an optical micrometer. Seston was filtered through weighed pre-ashed glass fibre filters (Whatman GF/F filters), dried at 60C and weighed to the nearest 0.1 mg to determine total particulate matter (PM), then further ashed for 3 h at 500C to calculate the particulate organic matter (POM). Particulate inorganic matter (PIM) was estimated as the difference between PM and POM. Benthic macroinvertebrates We sampled benthic invertebrates on each site in summer and autumn to compare with the respective drift samples. Benthic invertebrates were sampled with a kick net (100 m mesh) from the most common habitat types over a total period of 5 minutes. This method provided a semi-quantitative assessment of the benthic community. Macroinvertebrates of benthic samples were processed the same way as drift samples and analyzed to the same identification level (genus or family). Total abundance, taxon richness, Shannon-Wiener diversity index (H'), and the relative abundance of the dominant taxon (Chironomidae) were used to compare between drift and benthic inv ertebrate samples. Statistical analysis For statistical analyses all data were ln (x+1) transformed. Analysis of covariance (ANCOVA) was used to compare physical-chemical data. One-way ANCOVA was used to test for significant differences among and within sites (with season and time of day as covariates), and two-way ANCOVA was used to test for seasonal and diel patterns (with site as the covariate) (Sokal and Rohlf, 1995). Physical and chemical data of each sample also were analyzed using principal component analysis (PCA) based on 10 variables to determine a general physical-chemical classification of the study sites and to detect temporal patterns within this classification. Multivariate analysis was computed using ADE-4 software (Thioulouse et al., 1997). Data on seston (POM, PIM) and invertebrate drift were analyzed with a 2-way ANOVA (season and time of day as factors) for each site separately as the

120 116 5. Invertebrate drift in different alpine streams combined data set was unbalanced, and thus did not meet the assumptions for a 3-way ANOVA (Sokal and Rohlf, 1995). Therefore, we adjusted the significance level using a Bonferroni correction to p' = 0.05/8 = 0.00625 (Sokal and Rohlf, 1995). The non-parametric Kruskal-Wallis test, corresponding to an ANOVA by ranks, was used to test for differences in size distribution of the drifting and benthic invertebrates as the data were not normally distributed (Sokal and Rohlf, 1995). If differences were significant, pairwise comparisons were computed using a non-parametric Tukey-type multiple comparison test for unequal numbers of data in each group (Zar, 1984). Kendall's coefficient of rank correlation (tau) was used to compare the ranking of all samples between drift and seston variables (invertebrate abundance, taxon richness, POM, and PIM) and physical and chemical habitat parameters and benthic samples; significance of correlation was indicated at a p-level < 0.05 (Sokal and Rohlf, 1995). Kendall's tau varies from 0 (no concordance) to 1 (complete concordance). Results Physical and chemical characteristics Differences in physical-chemical parameters were most pronounced between rhithral and kryal sites. Kryal sites had higher discharge (average 0.7 to 4 m3/s), lower water temperature (average < 4.0 C), and higher concentrations of NH4-N, SRP and suspended particles (turbidity, PP, TSS) than rhithral sites (average discharge 0.06 to 0.5 m3/s, average temperature 4.2 C). Water temperature was significantly lower in autumn than in summer and spring (p < 0.005). Diel temperature amplitude was lowest in summer (0.7 to 3.7 C) and highest in spring (1.1 to 5.4 C) except for SS and SG where it was highest in autumn (1.7 and 2.0 C, respectively). In general, water temperature showed no significant diel pattern (p = 0.3), although it typically peaked between 11:00 and 15:00. PCA ordination summarized the differences in habitat characteristics and illustrated the distinction between the different stream types, as well as the temporal differences among stream types (Fig. 5-1). The first two axes explained 52.6 % of the variation among samples. As such, kryal sites were separated from rhithral sites on axis-1 by higher concentrations of ammonium (NH4-N) and total

121 5. Invertebrate drift in different alpine streams 117 a) -1 1 1 c) PN -1 Temp. 3 DOC POC Eigenvalue Cond. 2 TIC NH4-N 1 TSS SRP 0 NO2-N+NO3-N Axis b) III II III I II II III II I II I I III I III I II II I II III III III 2.5 I -3.0 3.2 -3.4 Figure 5-1 Ordination of a principal component analysis based on 10 physical-chemical variables measured each season and at 4 times of the day at each site. a) correlation circle, where Cond. = specific conductance, Temp. = water temperature, DOC = dissolved organic carbon, PN = particulate nitrogen, POC = particulate organic carbon, NH4-N = ammonium, TSS = total suspended solids, SRP = soluble reactive phosphorus, NO2-N+NO3-N = nitrite+nitrate, and TIC = total inorganic carbon; b) factor map of the diel and seasonal patterns for the study sites. In the factor map, circles and squares represent the average seasonal score (where I = spring, II = summer, and III = autumn) for each study site and are connected by lines to scores for each time of day (represented by small squares). The F1 axis explained 32.8 % and the F2 axis 19.8 % of the variation. In general, rhithral sites are represented with black, kryal sites with grey symbols, lake outlets with filled, and streams with open symbols, and each symbol represents a respective site. Thus, rhithral lake outlets: l Jrisee, n L.Bianco; rhithral streams: Gglia, Moesa; kryal lake outlets: l L.Roseg, n Steinsee; and kryal streams: Tschierva, Steinlimi. c) Eigenvalues of the axes 1 to 10. See text for more information.

122 118 5. Invertebrate drift in different alpine streams suspended solids (TSS) and lower values of specific conductance, and lake outlets were separated from streams on axis-2, having higher particulate nitrogen (PN) and particulate organic carbon (POC) and lower nitrite+nitrate (NO2-N+NO3-N) concentrations. JS samples were situated between rhithral and kryal lake outlets indicating the partial kryal influence of the adjacent Jriglacier. Seasonal differences were distinct for kryal sites (except LR) with a shift of autumn samples along axis-1 towards rhithral sites with also low concentrations of NH4-N and TSS, whereas rhithral sites showed no clear seasonal patterns. Two-way ANCOVA indicated that discharge, water temperature, NH4-N and TSS were significantly lower and specific conductance and TIC significantly higher in autumn than in spring and summer. None of the measured physical- chemical parameters showed significant differences among times of day (p > 0.31), although PCA ordination showed a tendency on the second axis for dusk and midday samples to have higher concentrations of PN and POC than dawn and midnight samples. Seston Concentrations of transported particulate matter (PM) in drift samples averaged < 0.1 g POM/m3 and < 0.7 g PIM/m3. Patterns of transported POM and PIM were similar (r = 0.95), therefore only POM concentrations are shown in figure 5-2. PM concentrations were lowest in glacial-influenced lake outlets (LR, SS, JS) and slightly higher in stream sites. In the kryal stream TG, particulates reached peak concentrations of > 0.4 g POM/m3 and > 300 g PIM/m3 at midday in spring. ANOVA results for differences among season and time of day revealed different patterns for the different sites (after Bonferroni correction). Concentrations of POM and PIM were highest in spring at both rhithral lake outlets, the rhithral stream M, and the kryal stream TG, but highest during summer or autumn at the other sites. Daily peaks occurred either at midday (at LB, LR, TG) or at dusk (M, SG) (Fig. 5-2).

123 5. Invertebrate drift in different alpine streams 119 0.50 dawn b dawn midday a L.Bianco (rL) Jrisee (rL) 0.40 midday 0.30 duskb dusk 0.10 midnight b midnight 0.08 0.06 0.04 0.02 0.00 b b a summer ab autumn b spring a summer autumn spring 0.50 dawn ac dawn 0.40 midday b Moesa (rS) Gglia (rS) midday 0.30 dusk a dusk 0.10 midnight bc midnight 0.08 0.06 0.04 0.02 POM g/m3 0.00 summer a autumn b spring a summer a autumn b spring ab 0.50 dawn b dawn midday a L.Roseg (kL) Steinsee (kL) 0.40 midday 0.30 dusk b dusk 0.10 midnight b midnight 0.08 0.06 0.04 0.02 0.00 summer ab autumn a spring b summer b autumn a spring b 0.50 dawn b dawn bc 0.40 midday a Tschierva (kS) Steinlimi (kS) midday ac 0.30 dusk b dusk a 0.10 midnight b 0.08 0.06 0.04 0.02 0.00 summer b autumn b spring a summer a autumn b springb Figure 5-2 Concentrations of particulate organic matter (POM) at 4 times of the day during 3 seasons at the 8 study sites. Vertical bars indicate + 1 SD. Small letters indicate significant differences among season and time of day for each site (Bonferroni's corrected p-value); no letters indicate no significant differences. Stream types (in parentheses) are defined in Table 5-1.

124 120 5. Invertebrate drift in different alpine streams Invertebrate assemblages Abundance: Densities of drifting invertebrates averaged < 100 ind./m3, although at the rhithral lake outlet LB and the kryal stream SG ca. 400 ind./m3 were collected in the autumn dusk samples (Fig. 5-3). Lowest densities (< 10 ind./m3) were found at the kryal lake outlets LR and SS and the kryal stream TG in all seasons and times of day (Table 5-2). ANOVA results for differences among season and time of day showed different patterns for the different sites (after Bonferroni correction). Highest densities occurred in spring at the rhithral lake outlet JS and both rhithral streams, and in autumn at LB and SG. Diel differences were significant only for LB and G, being lowest at midday and dawn, respectively (Fig. 5-3). Benthic macroinvertebrate densities (ind./5 min-sample) ranged from 46 (SS summer) to > 2000 (SG autumn) (Table 5-2). Densities generally were lowest at the kryal stream TG and the rhithral lake outlet JS (< 150 ind./5 min-sample) and highest at the kryal stream SG, the rhithral lake outlet LB, and the rhithral stream G. Autumn samples generally had 2 to 12 times higher densities than summer samples except at the kryal lake outlet LR (summer 6 > autumn) and the kryal stream TG (autumn = summer) (Table 5-2). Taxon richness and diversity: The total number of drifting taxa per season ranged from 0 (SS and TG in spring) to 8 (G in summer) with lowest richness at rhithral lake outlets and the two kryal sites SS and TG (except in autumn) and highest richness at rhithral streams (Table 5-2). The Shannon-Wiener index (H') equaled 0 (i.e. one taxon) at the rhithral lake outlets and two kryal sites in summer and was low (< 1) at the other sites, except at the rhithral stream G where H' = 1 in summer and autumn. Within sites, taxon richness and diversity of drifting invertebrates was higher in autumn than in spring or summer at the kryal streams and the kryal lake outlet SS and highest in summer at the rhithral streams and the kryal lake outlet LR. Benthic taxon richness ranged from 3 to 22 taxa per 5 min-sample with diversity (H') reaching values of > 2 (Table 5-2). Similar to the drift, benthic taxon richness and diversity were highest at the rhithral streams and lowest at SS. Seasonal differences were minor and only a higher diversity in autumn at the

125 5. Invertebrate drift in different alpine streams 121 450 dawn bc dawn 400 L.Bianco (rL) Jrisee (rL) midday c midday 350 150 dusk a dusk midnightab midnight 100 50 0 summer c autumn a spring b summer b autumn c spring a 450 400 Moesa (rS) dawn dawn b midday Gglia (rS) midday a 350 150 dusk dusk ab midnight midnight ab 100 50 Number/m3 0 summer b autumn b spring a summer b autumn c spring a 450 400 L.Roseg (kL) dawn Steinsee (kL) dawn 350 midday midday 150 dusk dusk midnight midnight 100 50 0 summer autumn spring summer ab autumn a spring b 450 400 Tschierva (kS) dawn Steinlimi (kS) dawn midday midday 350 150 dusk dusk midnight midnight 100 50 0 a ab b b a c summer autumn spring summer autumn spring Figure 5-3 Densities of drifting invertebrates at 4 times of the day during 3 seasons at the 8 study sites. Vertical bars indicate + 1 SD. Small letters indicate significant differences among season and time of day for each site (Bonferroni's corrected p-value); no letters indicate no significant differences. Stream types (in parentheses) are defined in Table 5-1.

126 122 5. Invertebrate drift in different alpine streams Table 5-2 Average density, cumulative taxon richness, diversity and % Chironomidae of the drifting and benthic invertebrates during summer and autumn. Site notations are defined in Table 5-1; k = kryal, r = rhithral, L = lake outlet, S = stream. Drift Benthos % Chironomidae % Chironomidae Taxon richness Taxon richness Stream type Diversity H' Diversity H' Density 1 Density 2 Season site summer 2 1 0.00 100 388 13 0.47 91 LB rL autumn 102 1 0.00 100 781 13 0.92 21 summer 22 1 0.00 100 70 8 1.11 71 JS rL autumn 2 1 0.00 100 149 8 0.85 26 summer 32 8 1.03 58 512 21 1.61 55 G rS autumn 7 6 0.96 75 969 21 1.70 49 summer 14 7 0.57 87 257 22 2.37 19 M rS autumn 15 2 0.21 94 455 20 2.44 23 summer 7 2 0.55 76 633 10 0.60 85 LR kL autumn 3 2 0.31 91 112 7 1.25 29 summer 2 1 0.00 100 46 4 0.31 93 SS kL autumn 4 4 0.26 95 561 3 0.11 98 summer 2 1 0.00 100 123 10 0.67 85 TG kS autumn 5 2 0.39 87 107 7 1.53 23 summer 25 3 0.03 100 219 7 1.15 44 SG kS autumn 153 4 0.69 92 2397 8 1.26 67 1 Average density (ind./m 3); 2 Density (ind./5 min-sample). kryal streams was correlated with drift. On average, less than 30 % of the taxa present were observed in the drift. Lowest proportions were found at the rhithral lake outlets (less than 15 % of the benthic taxa) where eight to 13 benthic taxa were recorded and only one of which was identified from drift samples.

127 5. Invertebrate drift in different alpine streams 123 Community composition: The dominant taxon in all drift and benthic samples was Chironomidae, making up between 58 and 100 % of drifting individuals and between 19 and 98 % of the benthic individuals (Table 5-2). Other abundant taxa (frequency > 1 %) in the drift samples were the mayflies Baetis spp. and Rhithrogena spp., unidentifiable first or second instar stonefly larvae, the stonefly Rhabdiopteryx alpina (KHTREIBER), and black flies (Simuliidae); these taxa were also abundant in respective benthic samples. Abundant benthic taxa that were absent or occurred sporadically in the drift were Oligochaetae, the turbellarian Crenobia alpina (DANA ), the mayfly Ecdyonurus spp., and the stonefly Leuctra spp. Size distribution: Drifting and benthic invertebrates were mostly small larvae with 75 % of all drifting and 70 % of all benthic individuals being = 3 mm. Body length was not different between benthic and drifting invertebrates (Fig. 5-4a). Both drifting and benthic individuals were significantly smaller in autumn (median body length of 2 mm, 75 % = 3 mm) than in summer (median of 3 mm, 75 % = 6 mm). Diel patterns in size distribution differed among seasons: size distribution of all drifting invertebrates was not different in autumn (p = 0.63) and spring (p = 0.22), but was in summer (p < 0.001). Non-parametric comparison of the different times of day in summer indicated smallest individuals at dawn, whereas the size of drifting individuals was not different at dusk, midnight and midday. Separate analysis of the size distribution of the dominant taxon for the four times of day in summer showed that small Chironomidae larvae (median body length of 2 mm at dawn compared to 3 mm at midday, dusk and midnight) were responsible for the significant differences at dawn (Fig. 5-4b). Separate analysis of the size distribution of the other two frequent taxa between the four times of day resulted in neither seasonal nor diel differences for Baetis alpinus, whereas Rhabdiopteryx alpina occurred only in autumn and was significantly larger (median body length of 4 mm) at midnight than at dusk and dawn (median body length of 3 mm); only one individual was collected at midday.

128 124 5. Invertebrate drift in different alpine streams Rank correlation of drifting invertebrates and habitat and benthic parameters Kendall's rank correlation indicated that drifting invertebrates generally were significantly correlated with most habitat parameters, but not with benthic invertebrates (Table 5-3). Total density and taxon richness of drifting invertebrates were positively correlated with organic seston (POM) and negatively with discharge, turbidity, ammonium (NH4-N), soluble reactive 15 a) Total invertebrates benthic drift a 10 a b b 5 Body length (mm) 0 summer autumn summer autumn (1415) (5619) (847) (3383) 15 b) Drifting Chironomidae 10 b b b a 5 0 dawn midday dusk midnight (124) (298) (144) (165) Figure 5-4 Box plots showing the median (solid line), 25th and 75th percentile (), 10 th and 90th percentile (low and high whiskers), and 5th and 95th percentile () for inverte- brate body lengths. a) All benthic and drifting invertebrates in two seasons, b) Diel differences in body length of the dominant taxon in the drift (chironomid larvae) in summer, when significant differences between times of day occurred. Small letters indicate significant differences using a non-parametric Tukey-type comparison test; n in parentheses.

129 5. Invertebrate drift in different alpine streams 125 Table 5-3 Kendall's coefficient of rank correlation (tau) between drifting invertebrates, seston and selected habitat and benthic (Be-) invertebrate (total density, taxon richnes, and density of Chironomidae) parameters as independent variables. Benthic invertebrate parameters are correlated to diel averages of the drift parameters for each site (n = 16). Significant correlations in ranking are indicated by: ** p < 0.005, * p < 0.05, ns = non-significant (p 0.05). Dependent drift variables Independent variable n Density Taxon richness Organic seston 91 0.26** 0.27** Inorganic seston 91 ns ns Discharge 58 - 0.28** - 0.26** Velocity 91 - 0.22** ns Temperature 91 ns ns Conductance 86 ns 0.24** Turbidity 82 - 0.30** - 0.29** Ammonium 91 - 0.35** - 0.27** Nitrite+nitrate 91 ns ns Particulate nitrogen 91 - 0.16* - 0.28** Soluble reactive phosphorus 91 - 0.27** - 0.27** Particulate phoshporus 91 - 0.32** - 0.38** Dissolved organic carbon 91 ns ns Particulate organic carbon 91 - 0.28** - 0.30** Total inorganic carbon 91 ns 0.15* Ash-free dry mass 91 - 0.22** - 0.28** Total suspended solids 91 - 0.29** - 0.32** Be-density 16 ns ns Be-taxon richness 16 ns ns Be-Chironomidae 16 ns ns phosphorus (SRP) and all particulates (nitrogen (PN), phosphorus (PP), organic carbon (POC), ash-free dry mass (AFDM), and total suspended solids (TSS)) (p < 0.05). These parameters were indicative of kryal sites, particularly in spring and summer (Fig. 5-1). Water temperature, nitrite+nitrate (NO2+NO3), dissolved

130 126 5. Invertebrate drift in different alpine streams organic carbon (DOC) and inorganic seston (PIM), in contrast, were not correlated with any drift parameter (p > 0.05, tau < 0.15). Correlation analysis with benthic invertebrate community patterns was based on diel averaged data of the drift samples per site as benthic samples were taken only once per day. Average drift density and taxon richness per site and day were correlated, but non-significantly (p 0.1), with benthic invertebrate and Chironomidae (only for drift density) density (tau 0.3), but not with benthic taxon richness (p > 0.2, tau 0.2). Discussion The four examined alpine stream types differed in their drift patterns. Drift densities generally were highest in rhithral and lowest in kryal sites, but no differences were obvious between streams and respective lake outlets. However, relatively high densities with single maximum peaks occurred in one kryal stream (SG). Ilg et al. (2001) and Robinson et al. (2002) both compared drift in high-elevation streams of kryal and krenal (i.e., groundwater fed) origin and found drift densities to be highest in stable groundwater sites and to be lowest in larger channels near the glacier. Thus, the main reason for high drift densities in the kryal stream SG probably was because of its more stable habitat conditions (e.g., Pfankuch index indicating a 'fair' stability, no shifting of the channel bed over time and relatively clear water during ice melt) relative to the other kryal sites. In general, densities were similar or higher than published data from other high-elevation streams (e.g., Allan, 1987; Tilley, 1989; Waringer, 1992) and lake outlets (Kownacki et al., 1997). However, these drift densities are not directly comparable as the mesh size of the drift nets varied among studies and some expressed drift only in terms of abundance (e.g., Ilg et al., 2001). Chironomidae dominated the drift in all sites and were a major component of the drift in other high-elevation streams (e.g., Bergey and Ward, 1989; Tilley, 1989; Tockner and Waringer, 1997; Robinson et al., 2002). Drift in kryal streams generally consisted of only few taxa, predominantly Chironomidae (also see Ilg et al., 2001; Robinson et al., 2002). High densities of the stonefly Rhabdiopteryx alpina, as found at the kryal stream SG, also occurred in the kryal main channel

131 5. Invertebrate drift in different alpine streams 127 studied by Robinson et al. (2002). Additionally, at locations further downstream they found seasonally high densities of Baetis alpinus (PICTET) and Simuliidae, taxa rare or absent in the drift at our kryal sites. In concordance to the present study, Robinson et al. (2002) and Ilg et al. (2001) also found generally higher taxon richness in rhithral and krenal streams than in kryal streams. Still, despite a generally higher taxon richness in various rhithral and krenal streams, drifting invertebrates typically were dominated by Chironomidae (Ferrington, 1984; Lavandier and Dcamps, 1984; Bergey and Ward, 1989; Tilley, 1989; Waringer, 1992; Tockner and Waringer, 1997; Pringle and Ramrez, 1998; Robinson et al., 2002), similar to our results, and the mayfly Baetis spp. (Lavandier and Dcamps, 1984; Allan, 1987; Bergey and Ward, 1989; Brewin and Ormerod, 1994; Ilg et al., 2001; Robinson et al., 2002). Drift in rhithral lake outlets has been found to be characterized by high densities of zooplankton and blackflies (Simuliidae) (Brittain and Eikeland, 1988; Kownacki et al., 1997). Zooplankton typically originates from the upstream lake and is expected to contribute to the high quality and quantity of the seston in lake outlets, and thus favor filter feeding invertebrates such as Simuliidae (Richardson and Mackay, 1991). Although zooplankton were not included in our drift analyses, they were observed in high abundances in drift samples from the two rhithral lake outlets LB and JS but not from the kryal lake outlets. Regardless, Simuliidae were rare in both the drift and the benthos of our streams and lake outlets. Kownacki et al. (1997) and Donath and Robinson (2001), in contrast, found high densities of Simuliidae in the drift and/or benthos of temporary high-mountain lake outlets suggesting that invertebrates with a short life cycle, such as Simuliidae, are specifically adapted to these conditions by completing their development within a short period. Seasonal drift patterns Drifting invertebrates showed no consistent seasonal pattern among the examined study sites. However, three of four rhithral sites had a seasonal pattern of high spring-time drift. Similar results of high densities in late spring/early summer and a decrease in late summer have been found in other high-elevation sites (Allan, 1987; Waringer, 1992; Kownacki et al., 1997). Little

132 128 5. Invertebrate drift in different alpine streams information exists on drift in kryal streams. Two recent studies, Ilg et al. (2001) and Robinson et al. (2002) found low drift densities during early summer with an increase in late summer/autumn, and a particularly pronounced peak in the densities of Baetis alpinus and Simuliidae in November. We found this pattern at one of our four kryal sites, but densities always were very low at the other sites. In low-elevation temperate streams, seasonal patterns usually are more pronounced, with minimum densities occurring in winter (e.g., Brittain and Eikeland, 1988; Anderwald et al., 1991). Thus, seasonality in drift seems to decrease with increasing elevation and with increasing distance from the source, respectively. Diel drift patterns Drifting invertebrates showed no consistent diel periodicity in our alpine streams. This result supports the findings of Brewin and Ormerod (1994) and Pringle and Ramrez (1998) who showed diel drift periodicity to be absent or less pronounced in high altitude regions, in contrast to distinct diel periodicities in low altitude regions of the same streams. In a fishless alpine kryal stream, however, Robinson et al. (2002) found distinct diel patterns in which drift density was highest during day, being associated with the summer afternoon peak in discharge. Discharge during the study period at our alpine sites, however, was relatively constant within a season, showing no significant diel differences, and thus, had no influence on diel drift patterns. In low-elevation streams, diel periodicity in invertebrate drift has been observed mainly for the Ephemeroptera, Simuliidae, Trichoptera, Plecoptera and Gammaridae (Allan, 1984; Brittain and Eikeland, 1988; Waringer, 1992), whereas various studies found Chironomidae to be aperiodic (Waters, 1972; Skinner, 1985; Tilley, 1989; Rader, 1997), predominantly night-drifting (Tilley, 1989) or predominantly day-drifting (Ferrington, 1984; Allan, 1987; Brittain and Eikeland, 1988; Tilley, 1989; Waringer, 1992). A probable reason for these apparently contradictory results is the low taxonomic resolution for chironomids, which were generally identified to family despite typically high species richness. The few studies on drifting chironomids at the species level showed distinct diel patterns, with species-specific day-, night- or aperiodic drifters (Ferrington,

133 5. Invertebrate drift in different alpine streams 129 1984; Tilley, 1989). In alpine streams, however, Diamesinae and Orthocladiinae are the most frequent chironomid larvae (Ward, 1994; Milner et al., 2001) and neither group showed diel periodicity (Tilley, 1989). Thus, our data support the hypothesis that diel drift periodicity is less pronounced in high altitude regions. Diel drift periodicity has been shown to be related to the size composition of drifting invertebrates (Waters, 1972; Allan and Russek, 1985). For example, Allan (1984) found significantly larger individuals drifting at night compared to day in July but not in September in a Rocky Mountain stream, suggesting differences in activity patterns among species explained the diel differences. At our sites, significant differences in the size distribution of drifting invertebrates were found only in summer with smaller Chironomidae larvae drifting at dawn, and with larger Rhabdiopteryx alpina larvae drifting at midnight in autumn. However, these significant differences among median body lengths are small (only 1 mm), and drifting invertebrates in our streams generally were dominated by small larvae with the few large individuals (e.g., Rhithrogena spp. and Limoniidae) being found both during day and night. Drift in relation to biotic and abiotic parameters Predators are expected to have a strong influence on drift patterns (Allan, 1995). For example, Flecker (1992) and Brewin and Ormerod (1994) found no diel periodicity in mountain streams that lacked drift-feeding fishes but high nocturnal drift in streams containing fish. It has been hypothesized that large invertebrates are more vulnerable to fish predation and therefore preferentially drift during night (Allan, 1984; Allan, 1995). Predators at our study sites were restricted mainly to Perlodidae stonefly larvae, although the two rhithral lakes are known to contain brown trout (Salmo trutta) and lake trout (Salvelinus namaycush) (Straskrabov, Callieri and Fott, 1999) and, from time to time, we observed fish in their outlets. Drift at the sites with temporary fish presence, however, showed no diel drift patterns in density or in size distribution, suggesting that fish predation played a minor role in diel patterns of drifting invertebrates. Drift density often has been related to benthic density. In contrast to studies of arctic and high-elevation streams finding drift to represent a constant

134 130 5. Invertebrate drift in different alpine streams proportion of benthic density (Hildebrand, 1974; Allan, 1987; Miller and Stout, 1989; Collier and Wakelin, 1992; Brewin and Ormerod, 1994), rank correlation revealed no significant relationship between these variables at our sites. However, we did find a tendency of higher drift densities at sites with higher benthic densities and the most abundant taxon in the drift (Chironomidae) also was the most abundant in benthic samples (also see Ferrington, 1984), suggesting that benthic invertebrates affected drift but were of less importance than abiotic parameters. We found significantly negative correlations between drift density and taxon richness and various abiotic habitat parameters. Water velocity and discharge are major abiotic measures that often are correlated with drift density. Many studies observed a positive correlation between drift density and discharge (Allan, 1987; Williams and Williams, 1993; Tockner and Waringer, 1997), whereas others have not (Soponis and Russell, 1984; Anderwald et al., 1991). In the present study, we found a significantly negative correlation between drift and discharge or velocity, similar to results for arctic and subarctic streams (LaPerriere, 1983; Miller and Stout, 1989). The negative correlation between drift density and discharge/velocity, however, mainly reflected (1) the low densities at the three kryal sites that also had the highest discharge, and (2) the high density during low flow conditions in autumn at LB and SG. A correlation between drift density and discharge for each site separately also revealed negative, although nonsignificant relationships. Drift also was negatively correlated with other abiotic parameters indicative for the period of glacial melt (Tockner et al., 2002), demonstrating the strong influence of a glacier and the generally low densities and taxon richness of drifting invertebrates in the kryal sites. The glacial influe nce also affected concentrations of PM, and thus resulted in higher concentrations of POM and PIM in spring and/or summer in kryal streams (also see Robinson et al., 2002). In the lake outlets, in contrast, concentrations of PM were lowest. Alpine lakes, being generally oligotrophic, act more as sinks than as sources of particles, and thus transport low concentrations of organic seston to their outlets, probably resulting in low densities of filter feeding invertebrates in their outlets. This is in contrast to findings for low- and

135 5. Invertebrate drift in different alpine streams 131 mid-elevation lake outlets with often substantially higher concentrations of particles and organic matter in lake outlets relative to their stream counterparts, thus favoring populations of filter feeding invertebrates (e.g., Richardson and Mackay, 1991). Summary The primary differences in drift patterns among our alpine streams were between rhithral sites with high and kryal sites with very low densities. Sporadically, very high densities occurred in a kryal stream with stable habitat conditions (as found at SG). Contrary to our initial hypotheses, no clear differences were observed between streams and lake outlets. Glacially influenced habitat parameters were a major factor affecting drift in these alpine streams, and thus superimposed a possible effect from benthic invertebrate densities. Seasonal drift patterns showed a tendency to be highest in spring at rhithral sites and in autumn at kryal sites. No diel periodicity was found (perhaps resulti ng from the virtual absence of fish predation) and may be a commonality among high altitude streams. Our findings suggest that invertebrate drift patterns in alpine streams differ from those in low - and midland streams in having a less pronounced seasonal and diel periodicity and being influenced primarily by abiotic habitat and less by biotic parameters. Acknowledgements Special thanks to Michael T. Monaghan, Marcos de la Puente, Peter Burgherr, Chregu Dinkel, Christian Rust, and Heiko Rinderspacher for assistance in the field, and Richard Illi and Bruno Ribi for completion of the chemical analyses in the laboratory. We are grateful to the communes of Pontresina, Samedan, and Klosters for providing access to the sites. We thank two anonymous reviewers for constructive comments that greatly improved the manuscript. The study was partially funded by a Swiss National Science Foundation Grant (no. 31-50440.97) examining the ecology of alpine lake outlets.

136 132 5. Invertebrate drift in different alpine streams References ALLAN, J. D. 1984. The size composition of invertebrate drift in a Rocky Mountain (Colorado, USA) stream. Oikos 43:68-76. ALLAN, J. D. 1987. Macroinvertebrate drift in a Rocky Mountain stream (Colorado, USA). Hydrobiologia 144:261-268. ALLAN, J. D. 1995. Drift. Pages 221-237 in Allan, J. D. (Editor). Stream Ecology - Structure and Function of Running Waters. Chapman and Hall, London, U.K.. ALLAN, J. D., and E. RUSSEK. 1985. The quantification of stream drift. Canadian Journal of Fisheries and Aquatic Sciences 42:210-215. ANDERWALD, P. H., M. KONAR, and U. H. HUMPESCH. 1991. Continous drift samples of macroinvertebrates in a large river, the Danube in Austria. Freshwater Biology 25:461-476. BERGEY, E. A., and J. V. W ARD. 1989. Upstream-downstream movements of aquatic invertebrates in a Rocky Mountain stream. Hydrobiologia 185:71-82. BREWIN, P. A., and S. J. ORMEROD. 1994. Macroinvertebrate drift in streams of the Nepalese Himalaya. Freshwater Biology 32:573-583. BRITTAIN, J. E., and T. J. EIKELAND. 1988. Invertebrate drift - A review. Hydrobiologia 166:77-93. COLLIER, K. J., and M. D. W AKELIN. 1992. Drift of aquatic macroinvertebrate larvae in Manganuiateao River, Central North Island, New Zealand. New Zealand Natural Sciences 19:15-26. DONATH, U., and C. T. ROBINSON. 2001. Ecological characteristics of lake outlets in Alpine environments of the Swiss Alps. Archiv fr Hydrobiologie 150:207-225. FERRINGTON, L. C. J. 1984. Drift dynamics of Chironomidae larvae: 1. Preliminary results and discussion of importance of mesh size and level of taxonomic identification in resolving Chironomidae diel drift patterns. Hydrobiologia 114:215-228. FLECKER, A. S. 1992. Fish predation and the evolution of invertebrate drift periodicity: Evidence from neotropical streams. Ecology 73:438-448. HIEBER, M., C. T. ROBINSON, U. UEHLINGER, and J. V. W ARD. Are alpine lake outlets less harsh than other alpine streams? Archiv fr Hydrobiologie (in press). HILDEBRAND, S. G. 1974. The relation of drift to benthos density and food level in an artificial stream. Limnology and Oceanography 19:951-957.

137 5. Invertebrate drift in different alpine streams 133 HUHTA, A., T. MUOTKA , A. JUNTUNEN, and M. YRJNEN. 1999. Behavioural interactions in stream food webs: the case of drift-feeding fish, predatory invertebrates and grazing mayflies. Journal of Animal Ecology 68:917-927. HUHTA, A., T. MUOTKA , and P. TIKKANEN. 2000. Nocturnal drift of mayfly nymphs as a post-contact antipredator mechanism. Freshwater Biology 45:33-42. IAHS/UNESCO. 1998. Fluctuations of glaciers 1990-1995. World Glacier Monitoring Service (WGMS), Zurich, Switzerland. ILG, C., E. CASTELLA , B. LODS-CROZET , and P. MARMORIER . 2001. Invertebrate drift and physico-chemical variables in the tributaries of the Mutt, a Swiss glacial stream. Archiv fr Hydrobiologie 151:335-352. KOWNACKI, A., E. DUMNICKA , J. GALAS , B. KAWECKA, and K. W OJTAN. 1997. Ecological characteristics of a high mountain lake-outlet stream (Tatra Mts, Poland). Archiv fr Hydrobiologie 139:113-128. LAPERRIERE, J. D. 1983. Alkalinity, discharge, average velocity, and invertebrate drift concentration in subarctic Alaskan streams. Journal of Freshwater Ecology 2:141-151. LAVANDIER , P., and H. DCAMPS . 1984. Estaragne. Pages 237-264 in Whitton, B. A. (Editor). Ecology of European Rivers. Blackwell Scientific Publications, Oxford, U.K.. MANI , M. S. 1990. Fundamentals of high altitude biology. Aspect Publications Ltd., London, U.K.. MILLER, M. C., and J. R. STOUT . 1989. Variability of macroinvertebrate community composition in an Arctic and Subarctic stream. Hydrobiologia 172:111-127. MILNER, A. M., J. E. BRITTAIN, E. CASTELLA , and G. E. PETTS. 2001. Trends of macroinvertebrate community structure in glacier-fed rivers in relation to environmental conditions: a synthesis. Freshwater Biology 46:1833-1847. MILNER, A. M., and G. E. PETTS . 1994. Glacial rivers: physical habitat and ecology. Freshwater Biology 32:295-307. PFANKUCH D. J. 1975. Stream reach inventory and channel stability evaluation. U.S.D.A. Forest Service, Missoula, Montana. PLATTS , W. S., W. F. MEGAHAN, and G. W. MINSHALL. 1983. Methods for evaluating stream, riparian, and biotic conditions. U.S. Departement of Agriculture,

138 134 5. Invertebrate drift in different alpine streams Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT, U.S.A.. PRINGLE, C. M., and A. RAMREZ. 1998. Use of both benthic and drift sampling techniques to assess tropical stream invertebrate communities along an altitudinal gradient, Costa Rica. Freshwater Biology 39:359-373. RADER, R. B. 1997. A functional classification of the drift: traits that influence invertebrate availability to salmonids. Canadian Journal of Fisheries and Aquatic Sciences 54:1211-1234. RICHARDSON, J. S., and R. J. MACKAY. 1991. Lake outlets and the distribution of filter feeders: an assessment of hypotheses. Oikos 62:370-380. ROBINSON, C. T., K. TOCKNER , and P. BURGHERR . 2002. Seasonal patterns in macroinvertebrate drift and seston transport in streams of an alpine glacial flood plain. Freshwater Biology 47 (in press). SKINNER, W. D. 1985. Night-day drift patterns and the size of larvae of 2 aquatic insects. Hydrobiologia 124:283-286. SOKAL , R. R., and F. J. ROHLF 1995. Biometry: the principles and practice of statistics in biological research. 3. W. H. Freeman and Company, New York, U.S.A.. SOPONIS, A. R., and C. L. RUSSELL. 1984. Larval drift of Chironomidae (Diptera) in a north Florida (USA) stream. Aquatic Insects 6:191-199. SPICHER , A. 1980. Geologische Karte der Schweiz. Schweizerische Geologische Kommission, Bundesamt fr Landestopographie, Wabern, Schweiz. STATZNER , B., C. DEJOUX, and J.-M. ELOUARD. 1984. Field experiments on the relationship between drift and benthic densities of aquatic insects in tropical streams (Ivory Coast). Revue d'Hydrobiologie Tropicale 17:319-334. STRASKRABOV , V., C. CALLIERI , and J. FOTT . 1999. The MOLAR Project: atmospheric deposition and lake water chemistry. Journal of Limnology 58:88-106. THIOULOUSE, J., D. CHESSEL , S. DOLDEC, and J.-M. OLIVIER . 1997. ADE-4: a multivariate analysis and graphical display software. Statistics and Computing 7:75-83. TILLEY, L. J. 1989. Diel drift of Chironomidae larvae in a pristine Idaho (USA) mountain stream. Hydrobiologia 174:133-150.

139 5. Invertebrate drift in different alpine streams 135 TOCKNER, K., F. MALARD, P. BURGHERR , C. T. ROBINSON, U. UEHLINGER , R. ZAH, and J. V. WARD. 1997. Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv fr Hydrobiologie 140:433-463. TOCKNER, K., F. MALARD , U. UEHLINGER, and J. V. WARD. 2002. Nutrients and organic matter in a glacial river-floodplain system (Val Roseg, Switzerland). Limnology and Oceanography 47:266-277. TOCKNER, K., and J. A. WARINGER . 1997. Measuring drift during a receding flood: results from an Austrian mountain brook (Ritrodat-Lunz). Internationale Revue der Gesamten Hydrobiologie 82:1-13. UEHLINGER , U., R. ZAH, and H. BRGI. 1998. The Val Roseg Project: temporal and spatial patterns of benthic algae in an Alpine stream ecosystem influenced by glacier runoff. Pages 419-424 in Kovar, K., U. Tappeiner, N. E. Peters, and R. G. Craig (Editors). Hydrology, Water Ressources and Ecology in Headwaters. IAHS Press, Wallingford, U.K.. WARD, J. V. 1994. Ecology of alpine streams. Freshwater Biology 32:277-294. WARINGER, J. A. 1992. The drifting of invertebrates and particulate organic matter in an Austrian mountain brook. Freshwater Biology 27:367-378. WATERS , T. F. 1972. The drift of stream insects. Annual Review of Entomology 17:253-272. WILLIAMS , D. D., and N. E. WILLIAMS . 1993. The upstream/downstream movement paradox of lotic invertebrates: Quantitative evidence from a Welsh mountain stream. Freshwater Biology 30:199-218. ZAR J. H. 1984. Biostatistical Analysis. Prentice Hall, Englewood Cliffs, New Jersey.

140 136

141 6. Colonization in alpine streams 137 6. Colonization dynamics of macroinverte- brates in alpine streams and lake outlets Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. Journal of the North American Benthological Society, submitted. Colonization is an important ecosystem process for population persistence and community maintenance that also has been widely studied by stream ecologists. However, little information exists on colonization patterns and the parameters affecting this process in high-elevation streams. We examined the colonization of habitat patches by benthic invertebrates in eight rhithral or kryal alpine streams and lake outlets in the Swiss Alps. Cages filled with natural substrate were placed in each study site and collected after 3, 8 and 30 days of colonization during summer. In addition, benthic invertebrates were collected with a Hess sampler to determine each site's species pool and with drift nets to investigate the role of drift as a source of colonists. Colonization was rapid at all sites, showing no significant increases in density or taxon richness after only eight days. However, differences in the density and diversity of colonizing invertebrates were found among streams as well as among individual cages. Multiple regression analysis indicated that local stream conditions such as discharge, being highest in kryal streams, primarily influenced the diversity of colonists, whereas the small-scale distribution of benthic organic matter significantly affected the density of colonists. Despite high colonization rates, high beta diversities (taxa turnover) among sampling dates, and a low proportion of the available taxa being present in the cages, suggested that community assembly was still in nonequilibrium even after 30 days, i.e., assemblages were in a continuous state of redistribution. The low density and richness of drifting invertebrates further indicated that active larval movement through swimming and crawling probably is the dominant mode in the colonization of small patches in alpine streams. Thus, we suggest that colonization of small- scale patches in alpine streams reflects the search of benthic invertebrates for resources through continuous moving.

142 138 6. Colonization in alpine streams Introduction Colonization often is viewed as a major process maintaining long-term community dynamics (sensu MacArthur and Wilson 1967), thus being an important measure of ecosystem resilience to disturbance (Mackay 1992). Colonization can be defined as dispersal between habitat patches regardless of the spatial scale examined (Downes and Keough 1998), ranging from biogeographic to extremely local and immediate phenomena (Mackay 1992). In stream ecosystems, two main scales of disturbance and subsequent colonization can be distinguished: (1) mesoscale colonization following major disturbances such as spates and droughts that affect reaches or stream segments, and (2) colonization among patches of habitat after localized disturbances such as the turning of a stone or the loss of a resource patch (Townsend 1989, Downes and Keough 1998). Minshall and Petersen (1985) suggested that, within a stream, single rocks or patches of substrate can act as 'islands' and, analogous to the MacArthur-Wilson model, become colonized by the immigration and establishment of new species with the community moving towards an equilibrium between colonization and local extinction. Colonization of lotic ecosystems by macroinvertebrates occurs by different pathways and the importance of individual pathways differs depending upon the scale that is examined (Townsend 1989). At coarse scales, colonization is dominated by dispersal from temporary refuges such as the hyporheos and different floodplain water bodies, or by recolonization from terrestrial life stages (i.e., adult) resuming an aquatic existence (e.g., Fisher et al. 1982, Mackay 1992). At finer scales, colonization is dominated by the continuous redistribution of zoobenthos by drift from upstream sources, immigration from adjacent patches, or hatching from eggs laid near, on or in the water (e.g., Sheldon 1984, Townsend 1989, Mackay 1992, Giller 1996). Thus far, drift has been attributed as the primary mechanism of immigration and emigration for larval stream insects (Williams and Hynes 1976, Minshall and Petersen 1985, Matthaei et al. 1997, Gayraud et al. 2000). However, Mackay (1992) argued that although drift may be an important relocating mechanism for invertebrates to colonize downstream reaches, many mobile invertebrates can disperse among patches by active swimming and crawling.

143 6. Colonization in alpine streams 139 Colonization is dependent on both the characteristics of the invertebrate community (e.g., density, composition, species traits) as well as of the habitat (e.g., spatial arrangement of habitat, substrate composition, food supply) (Sheldon 1984, Mackay 1992, Lake 2000). Stream biota generally show low resistance (ability to remain the same) to disturbance, but their resilience (ability to recover) typically is high and recovery following hydrological disturbance usually is rapid (Sheldon 1984, Robinson and Minshall 1986, Giller 1996, Matthaei et al. 1996, Elser 1999, Gayraud et al. 2000, Lake 2000). Moreover, many stream invertebrates are adapted to hydrological disturbance via changes in life-history (e.g., spending periods of high discharge as terrestrial adults or eggs) (Fisher et al. 1982), species traits (e.g., morphology) (Townsend et al. 1997), and behavior (e.g., swimming and burrowing behavior) (Holomuzki and Biggs 2000). Colonizers usually are characterized by good dispersal abilities (e.g., flight, swimming, crawling), short and often asynchronous life cycles, and plasticity in feeding behavior. Indeed, many studies of lotic invertebrates note that early colonizers, globally, mainly belong to the families Baetidae, Leptophlebiidae and Chironomidae (as reviewed by Mackay 1992), families with these characteristic traits. In the past few years, alpine streams have attracted much interest and many studies contributed to a better knowledge of the invertebrate communities and their distinctive nature (Kownacki 1991, Ward 1994, Robinson et al. 2001, Burgherr et al. in press). However, thus far very little is known on the functioning and processes of these systems (Brittain and Milner 2001). Streams in alpine regions are often described as 'harsh' environments. For instance, they are subject to extended periods of high flow (snow melt, glacial melt) often associated with bedload transport, and macroinvertebrate assemblages are dominated by taxa (e.g., Baetidae and Chironomidae) adapted to their distinct environmental conditions (Ward 1994, Freder 1999, Usseglio-Polatera et al. 2001). Yet, habitat conditions differ among alpine stream types depending on the origin of water: kryal (glacier-melt) streams are characterized by large diel and seasonal flow fluctuations and sparse food resources; rhithral (snow-melt) streams by seasonal, generally less strong fluctuations in flow, a broader temperature range and higher amounts of organic matter; and krenal

144 140 6. Colonization in alpine streams (groundwater-fed) streams are characterized by relatively constant flow and temperature regimes (Ward 1994). Lake outlets of each stream type can be viewed as subtypes, the lakes influencing downstream flow and temperature regimes, thereby affecting further parameters such as channel stability and food availability (Burgherr and Ward 2000, Donath and Robinson 2001, Hieber et al. in press-b). As a consequence, we were interested in the relation between the distinct environmental characteristics of the different alpine stream types and the respective colonization patterns of benthic invertebrates. Therefore, we examined short-term colonization dynamics of small patches within kryal and rhithral streams and lake outlets, specifically addressing the following questions: How rapidly are new patches in different alpine stream types colonized by benthic invertebrates?; What environmental factors are responsible for observed differences in the colonization patterns among the different stream types?; and lastly, How important is drift as a colonization mechanism in alpine streams? Study Sites We examined colonization during summer in 8 streams comprising 2 each of the following 4 stream types: rhithral lake outlets, rhithral streams, kryal lake outlets, and kryal streams (Table 6-1). The streams were in the Swiss Alps above treeline at elevations between 1930 and 2500 m a.s.l.. Catchment size ranged from 0.6 to 19.3 km2, with the largest catchments associated with kryal sites having 42 to 92 % of their area glaciated. One outlet, that of the Jrisee, was classified as rhithral-kryal because it was influenced somewhat by the melt waters of the Jriglacier that first flows into a small lake above the Jrisee. The data on catchment and glaciated area were based on topographic maps updated between 1991 and 1995. Because the glaciers draining into our study streams retreated between 5 and 30 m/y (IAHS/UNESCO 1998), the actual glaciated area at the time of the study was less than listed in Table 6-1. The study streams had slopes ranging from 2 to 16 %, median depths of 14 to 27 cm, average widths of 3 to 12 m, and substrate of a pebble/cobble matrix. Water temperature recorded during August showed low cumulative degree days (DD) at the kryal streams (37 and 70), slightly higher DD at the kryal lake outlets

145 6. Colonization in alpine streams 141 Table 6-1 Classification and general characteristics of the study sites. r = rhithral, k = kryal, L = lake outlet, S = stream; DD = accumulated degree days, Q = discharge. Site Notation Origin Stream Elevation Catchment Slope Temperature QAugust type (m a.s.l.) area (km2) % glaciated (%) DDAugust (m3/s) Lago Bianco LB r L 2076 2.1 0 4 328 0.2 Jrisee JS r(k) L 2489 3.4 11 5 257 0.7 Gglia G r S 2310 5.8 0 10 177 0.3 Moesa M r S 2300 0.6 0 16 233 0.1 Lej Roseg LR k L 2159 19.3 44 2 116 4.3 a Steinsee SS k L 1934 7.3 71 4 85 1.0 Tschierva TG k S 2100 14.7 42 4 37 2.6 Steinlimi SG k S 2090 3.4 92 2 70 1.1 a Discharge measurement taken in September 2000. (85 and 116), around 200 DD at the rhithral streams, and the highest DD at the rhithral lake outlets (257 and 328). Discharge measurements taken in August 1999 ranged from 0.1 to 0.7 m3/s at the rhithral sites and from 1.0 to 4.7 m3/s at the kryal sites (Table 6-1). Discharge was measured in September 2000 at the kryal Steinsee outlet when flow was much lower than in summer (> 1.0 m3/s). Previous calculations of the Pfankuch stability score and a multivariate habitat stability index separated rhithral sites as stable and kryal sites as unstable (Hieber et al. in press-b). More detailed information on the physical-chemical characteristics of the sites is given in Hieber et al. (in press-b). Methods Field design The study was conducted between July 30 and September 15, 2000. During the sampling period, water temperature was recorded at each site at 60 min intervals with a temperature logger (Minilog, Vemco, Nova Scotia, Canada). Fifteen metal cages (311713 cm, mesh size 1 cm) filled with stones (b-axis 4 to

146 142 6. Colonization in alpine streams 11 cm) were placed in each stream. The stones were collected from each respective site. To simulate stream bottom conditions, the lower half of each cage was filled with stones lacking periphyton and the upper half with stones with periphyton. These latter stones were taken from each stream and had all macroinvertebrates removed prior to being placed in a cage. On day 0, each cage was buried even with the stream bottom along a 15 to 20 m reach in each stream. Water depth and velocity (Mini Air 2, Schiltknecht AG, Gossau, Switzerland) at 0.6 depth was measured above each cage. Five cages were randomly collected at each site after 3, 8 and 30 days. For collection, a 100-m mesh net was placed immediately downstream of each cage during removal to collect invertebrates leaving the cage or being washed out by the current. All stones in each cage were rinsed at stream-side and the associated invertebrates and organic material retained in a 100-m sieve and preserved with 2 to 4 % formalin for later analysis. In addition, one stone from the upper layer of each cage was collected for determination of periphyton biomass. Even though the cages had been buried in the stream bed, some cages were lost during the study, thus resulting in different n per site (see Table 6-2). Five stones of similar size to those in the cages also were collected from the stream bed at each site on day 30 to characterize instream periphyton levels. All stones, from the cages and from the stream bed, were stored at - 25C until processed in the laboratory, usually within three weeks of collection. Three samples of benthic invertebrates also were collected using a Hess-sampler (0.04 m2, 100-m mesh) on day 30 to determine each site's species pool. Benthic organic matter (BOM) was taken from each Hess sample and standardized to unit area (m2). Lastly, 4 samples of drifting invertebrates and transported organic matter (seston) were collected at the beginning (day 0) and the end (day 30) of the colonization period. Samples were collected during daylight at different times, as earlier studies on diel and seasonal drift patterns have shown that time of day had no effect on drift density or taxonomic composition in these alpine streams (Hieber et al. in press-a). Drift and seston were collected using a nylon drift net (100-m mesh) for 60 to 180 s at approximately 0.6 depth. Period of sampling depended on clogging of the net by transported sediments, particularly in kryal sites, or organic material. Velocity was recorded at the net aperture to

147 6. Colonization in alpine streams 143 standardize seston and drift parameters to unit volume of water (m3). The contents of each sample were stored at - 25C until processed in the laboratory. Laboratory analyses In the laboratory, invertebrates from the cages and Hess samples were counted and identified to the lowest possible level (species for most EPT (Ephemeroptera, Plecoptera, Trichoptera), family for Diptera, no further identification for Hydrozoa, Hydracarina, Mollusca, Nemathelminthes, Oligochatea, Copepoda, Ostracoda and Collembola). Total density, alpha diversity (taxon richness) and Simpson's index of diversity (D) were determined for each sample. Then, alpha diversity per site and sampling date, beta diversity (taxa turnover among sites and days), and gamma diversity (total taxon richness per site) were calculated. Beta diversity () is a measure of similarity between samples or sites and ranges from 0 (complete similarity) to 100 (complete dissimilarity). It was used to compare taxon richness among sites, within each site among different sampling dates (cages only) and between different sampling techniques (cages, Hess, drift). To allow comparison of samples with unequal n, a modified formula was used: = (/ -1)/(NS-1)*100, where = the total number of taxa for all samples combined, = average taxon richness of all samples, and NS = the total number of samples (Harrison et al. 1992). Simpson's index of diversity (D) was calculated using the formula D = (ni(ni-1))/(N(N-1)), where ni = the number of individuals in the ith species, and N = the total number of individuals. High D values indicate a more equal distribution of taxa, even though the total number of taxa may be low. In contrast, high taxon richness but low D values indicate a strong dominance by one or a few taxa and the presence of many rare taxa. Simpson's index is weighted towards the most abundant species and is less sensitive to species richness than the Shannon-Wiener index (Magurran 1991). After removing the invertebrates, the remaining benthic material from each sample was dried at 60C, weighed to the nearest 0.1 mg, ashed for 3 h at 500C, and reweighed to calculate the benthic organic matter as ash-free dry mass (AFDM). Periphyton was removed from each collected stone with a metal bristle brush. Two aliquots (10 to 30 ml each) of the algal suspension from each stone

148 144 6. Colonization in alpine streams were filtered through pre-ashed glass fibre filters (Whatman GF/F filters) to determine chlorophyll a content and AFDM. Chlorophyll a was extracted in 90 % ethanol by boiling at 70 C for 10 min and measured with a reversed phase HPLC and a subsequent diode array-detector (Bio-Tek, Basel, Switzerland) (Meyns et al. 1994). Ash-free dry mass was determined as the difference between the weight of the filter with the algal suspension after drying at 60 C and the weight after ashing for 3 h at 500 C. Chlorophyll a and AFDM concentrations were expressed per unit of stone area. The surface area of each stone was calculated as the area of an ellipsoid based on measurements of the a- and b- axes following Uehlinger (1991). Drifting invertebrates in each sample first were identified and counted to the lowest possible level (mainly genus and family) using a dissecting microscope, and total density, alpha diversity and Simpson's index of diversity (D) were calculated. The body length of each animal was measured to the nearest 0.5 mm using an optical micrometer. Each sample then was filtered through a weighed pre-ashed glass fibre filter (Whatman GF/F filters), dried at 60C and weighed to the nearest 0.1 mg to determine total particulate matter (PM). Following this, each filter was ashed for 3 h at 500C and reweighed to calculate the particulate organic matter (POM) as AFDM. Particulate inorganic matter (PIM) was estimated as the difference between PM and POM. Macroinvertebrate biomass was estimated by means of body-length dry-mass relationships according to Meyer (1989) and Burgherr and Meyer (1997). In general, biomass was low (average 0.002 0.006 g/m3) compared to PM concentrations (average 3.9 30.7 g/m3), and thus had little influence on PM values. Statistical analyses All data were ln (x+1) transformed to meet the assumptions of a normal distribution (Sokal and Rohlf 1995). Analysis of variance was used to test for differences between the eight sampling sites and the different days (date) of colonization (days 3, 8 and 30). Because no cages were left on d 30 at the kryal stream TG, the data had to be treated in two different ways. First, ANCOVA (with site and date as factors) was used to test for general differences either

149 6. Colonization in alpine streams 145 between sites (with date as covariate) or between dates (with site as covariate) (significance level of p = 0.05). Additionally, two-way ANOVA (with site and date as independent factors) was used to test for significant differences between the different days (date) of colonization and the Hess samples from each stream within each site. The unbalanced data-set (data at day 30 for only seven sites) thus required the site TG and the remaining sites to be tested separately, and to adjust the significance level using a Bonferroni correction of p' = 0.05/2 = 0.025 (Sokal and Rohlf 1995). If differences were detected by ANCOVA or ANOVA, the Tukey HSD test was used to determine which values were actually different (Sokal and Rohlf 1995). Forward stepwise multiple linear regression was used to test for possible relationships between invertebrate density and Simpson's index of diversity with different habitat variables and time. First, invertebrate density and Simpson's index in each cage and Hess sample were tested by regression with 6 variables, including time of colonization (day), small-scale habitat characteristics (depth, velocity, BOM) and study stream conditions (degree days in August, discharge in August). Second, to include the influence of periphyton as a food resource, daily average values of density and Simpson's index were tested by regression with six variables, describing available food resources (BOM, periphyton AFDM and chlorophyll a), time of colonization, and study stream. The standardized regression coefficient Beta equals the regression coefficient B for the variables standardized to a mean of 0 and a standard deviation of 1. Thus, the magnitude of Beta allows one to compare the relative contribution of each independent variable (StatSoft 1995). All statistical analyses were performed using Statistica 5.1 (StatSoft, Tulsa, USA). Results Depth and velocity Water depth ranged from 3 to 31 cm over cages in each stream on day 0 and from 4 to 41 cm over Hess samples on day 30. In general, water depth was greatest at the lake outlets (average 16 to 21 cm) and lowest at the rhithral streams and the kryal stream SG (average 9 to 13 cm). Water velocity ranged from 0.00 to 1.15 m/s over the cages and from 0.02 to 1.15 m/s over the Hess

150 146 6. Colonization in alpine streams sampling sites. Average velocities were highest at the kryal streams (0.45 and 0.54 m/s) and tended to be lower at the lake outlets (0.31 to 0.38 m/s). LB 30000 a JS G M LR Density (no/m2) SS TG 20000 SG 10000 0 0 10 20 30 30-H 40 18 b 12 Taxon richness 6 0 0 10 20 30 30-H 40 5 c 4 Simpson's diversity 3 2 1 0 0 10 20 30 30-H 40 Colonization time (d) Figure 6-1 Average a) density, b) taxon richness, and c) Simpson's index of diversity of invertebrates in colonization cages during the exposure period and in the Hess samples on day 30 (30-H) at the 8 study sites (notations defined in Table 6-1). Grey: rhithral sites, black: kryal sites, filled symbols: lake outlets, open symbols: streams. Vertical bars indicate + 1 SD, respectively.

151 6. Colonization in alpine streams 147 Invertebrate densities The initial colonization of cages was rapid. Maximum densities in the cages usually occurred within 3 days except at the rhithral lake outlet LB and the kryal stream SG where densities further increased up to day 30 (Fig. 6-1a). The average invertebrate density (individuals/m2) in cages on day 3 ranged from 90 in the kryal stream TG to > 2500 ind./m2 in the rhithral streams (M, G) and the rhithral lake outlet LB. Glacial-influenced lake outlets were situated on the lower end of this range with densities < 400 on day 3. Maximum densities (> 20,000 ind./m2) occurred by day 30 at LB and SG, however, the increase in density between day 3 and day 30 was significant only at the kryal stream SG (ANOVA p = 0.012). In general, ANCOVA showed significantly higher densities in the kryal stream SG, the rhithral lake outlet LB and both rhithral streams compared to the other sites (p < 0.001). Invertebrate densities in the Hess samples were similar to those in the cages at respective sites. For instance, densities in the Hess samples on day 30 were significantly lowest at TG (< 400 ind./m2) and ranged to almost 10,000 ind./m2 at LB (Fig. 6-1a). Densities in the Hess samples were significantly higher than in respective cages only at JS (day 8; p = 0.018). The density and biomass of drifting invertebrates were significantly correlated (r = 0.6; p > 0.01), therefore only density results are presented (Fig. 6-2a). On average, both drift density and biomass were low (= 5 individuals and = 2 g/m3), and no drifting invertebrates were collected at LR. An exception was the rhithral stream G where significantly greater densities were observed than at the other sites (p < 0.001). ANOVA indicated a significantly higher drift density on day 0 (49 ind./m3) th an on day 30 (10 ind./m3) at G, and no significant differences in drift density between days at the other sites (p = 0.2). Community composition An average of 3 to 16 taxa colonized the cages by day 3 with all sites typically peaking in richness by at least day 8 (Fig. 6-1b). Taxon richness was quite stable and, within each site, was significantly higher on day 30 than on day 3 only at SG (p = 0.003). Average, as well as total taxon richness (alpha diversity), was highest in the two rhithral streams (> 10 and 20) and lowest in the kryal stream

152 148 6. Colonization in alpine streams TG and the glacial-influenced lake outlets (= 4 and 10) (Table 6-2). Beta diversities (turnover between sample dates within a site) were low at most sites ( = 21), although being high ( > 50) at the taxa-poor ( = 7) glacial sites TG and 80 a Ephemeroptera 60 Plecoptera Trichoptera 40 Chironomidae Drifting invertebrates (no/m3) 20 Simuliidae others 15 10 5 0 * 1.0 b PIM POM 0.8 Particulate matter (g/m3) 0.6 0.3 0.2 0.1 0.0 * 0 30 0 30 0 30 0 30 0 30 0 30 0 30 0 30 LB JS G M LR SS TG SG Figure 6-2 Average a) density of drifting invertebrates separated as major groups, and b) concentration of particulate organic (POM) and inorganic matter (PIM) in the drift samples at the beginning (0) and end (30) of the colonization period at the 8 study sites (notations defined in Table 6-1). Vertical bars indicate + 1 SD; * no samples taken.

153 6. Colonization in alpine streams 149 SS (Table 6-2). Simpson's index of diversity (D) was significantly higher at the rhithral stream G and the kryal lake outlet LR (mean D > 2.9) (p < 0.001; Fig. 6- 1c). At the other sites, D ranged from 1.6 to 1.9 at the rhithral sites and was < 1.5 at the kryal sites. Simpson's index was not correlated with alpha diversity (r = 0.35), and generally was lowest on day 30 for all sites. However, changes in D between sample days within any one site were not significantly different (p > 0.05). Assemblages in cages at all sites were dominated by the Chironomidae (average relative density = 76 %). Other dominant taxa occurring at some sites were the Oligochaeta (15 % at LB), the Ephemeroptera Baetis spp. (15 % at G) and Rhithrogena spp. (11 % at G, 23 % at LR). Average taxon richness in the Hess samples was significantly lower at the kryal sites TG and SS (= 3) and higher at the two rhithral streams ( 9) (Fig. 6-1b, Table 6-2). Richness values in Hess samples were significantly higher than in respective cages only at the lake outlets JS and LR on day 8 (p < 0.02). Beta diversities between Hess samples on day 30 and respective cages on day 3, 8 or 30 were lowest in the taxa-rich sites (LB, M, G) and highest in the taxa-poor glacial-influenced sites (TG, SS, LR, JS) (Table 6-2). Simpson's index of Hess samples were not significantly different than respective cages, with D 1.5 in the kryal streams and lake outlet SS and D > 4 in the rhithral stream G and the kryal lake outlet LR (Fig. 6-1c). Community composition in the Hess samples was similar to those in respective cages, being dominated by the Chironomidae at all sites (average relative density = 60 %). Oligochaeta were common in LB and JS (33 and 66 %, respectively), the Ephemeroptera Baetis spp. in G (12 %) and Rhithrogena spp. in G and LR (14 and 13 %, respectively). In total, 8 taxa were found in the drift with the Chironomidae being dominant (average = 79 %). Taxon richness in the drift samples was low and varied from 0 (LR) to a maximum of 5 (G) taxa (Table 6-2). As a consequence, taxa turnover (dissimilarity) between cages and drift samples was very high ( = 61 to 100, except 31 at TG). Gamma diversity (total taxon richness per site) ranged from 7 to 15 taxa at the kryal influenced sites and from 21 to 25 at the rhithral streams and lake outlet LB (Table 6-2). The number of taxa present at any one site always exceeded the number of taxa found in either cages (except at SG), Hess or drift

154 150 6. Colonization in alpine streams Table 6-2 Alpha, beta (taxa turnover) and gamma (total taxa per site) diversity for invertebrates in respective cage, Hess and drift samples. Alpha diversity values are means, with total richness in parentheses; beta diversity is calculated among sample days of cages, and as mean of beta between each sample day of cage and each Hess and drift sample day, respectively. Site notations defined in Table 6-1. Site Alpha Beta Gamma Cages Hess Drift Cages Cage-Hess Cage-Drift (n = 15) (n = 3) (n = 8) (n = 3) (n = 3) (n = 6) LB 8 (20) 6 (10) 0.3 (1) 20 29 87 21 JS 3 (9) 6 (10) 0.5 (1) 14 46 75 13 G 12 (23) a 9 (15) 2.3 (5) 14 35 72 25 M 11 (21) b 14 (17) 1.3 (4) 9 31 75 24 LR 4 (10) 7 (11) 0 (0) 25 54 100 15 SS 2 (6) 1 (2) 0.4 (2) 40 60 61 7 e e e TG 2 (7) c 3 (6) 1.0 (2) d 75 60 31 10 SG 6 (14) 4 (5) 0.6 (1) 20 36 81 14 a n = 12; b n = 14; c n = 9; d n = 4; e n = 2. samples. Gamma diversity per site was most similar to the total richness of invertebrates in cages, differing only by a factor of 1 to 1.5, whereas it exceeded total richness in the drift samples by a factor of 3.5 to 21. Potential food resources Benthic organic matter (BOM) accumulating in the cages on day 3 was 0.1 - 0.6 g/m2 at the kryal influenced sites and 1.3 - 2.0 g/m2 at the rhithral sites (Fig. 6- 3a). After 30 days of colonization, BOM concentrations increased slightly to 0.5 - 1.7 g/m2 at the kryal influenced sites but reached a maximum concentration of 7 g/m2 at the rhithral stream G. Changes in BOM concentrations between day 3 and day 30 were not significant within any one site (p 0.18), except at G (p = 0.007). In general, concentrations of BOM were significantly higher in the rhithral streams and lake outlet LB and lower in the kryal stream TG (p < 0.001).

155 6. Colonization in alpine streams 151 15 a LB Benthic organic matter (g/m ) JS 2 G M 10 LR SS TG SG 5 0 0 10 20 30 40 30-H b 30 Periphyton AFDM (g/m2) 20 10 0 0 10 20 30 30-ref 40 60 c Periphyton chlorophyll a (mg/m2) 40 20 0 0 10 20 30 30-ref 40 Colonization time (d) Figure 6-3 Average concentrations of food resources in the colonization cages during the exposure period and in the benthic Hess samples (30-H) and from stream stone samples (30-ref) at the 8 study sites (notations defined in Table 6-1). a) Benthic organic matter, b) periphyton ash-free dry mass (AFDM), and c) periphyton chlorophyll a. Grey: rhithral sites, black: kryal sites, filled symbols: lake outlets, open symbols: streams. Vertical bars indicate + 1 SD.

156 152 6. Colonization in alpine streams Average BOM concentrations in the Hess samples were < 2 g/m2 at all kryal influenced sites and the rhithral stream G, but was 11 g/m2 at the rhithral stream M (Fig. 6-3a). BOM concentrations in the Hess samples were not significantly different than in the cages (p 0.06), except for a significantly lower concentration at G (p = 0.006) and a higher concentration at M (p 0.008). Periphyton ranged on average from < 0.5 to 25 g AFDM/m2 (Fig. 6-3b) and from < 0.5 to 35 mg chlorophyll a /m2 (Fig. 6-3c). ANCOVA indicated significantly higher concentrations of AFDM and chlorophyll a at the lake outlets LB, JS and LR and lower concentrations at the kryal stream TG (p < 0.001). Changes in concentrations of AFDM and chlorophyll a between sample days within any one site were not significantly different. However, chlorophyll a was significantly higher in the stream bed samples than in respective cages at JS, SS and SG (p = 0.016). Seston ranged from 0.002 to 0.04 g POM and from 0.00 to 0.86 g PIM /m3 (Fig. 6-2b). POM concentrations were significantly higher at G, M and SG (= 0.02 g/m3) than at the glacial-influenced lake outlets JS, LR and SS ( 0.005 g/m3) (p 0.002). PIM concentrations were significantly higher at the kryal streams SG and TG and the rhithral stream G (= 0.4 g/m3) than at the other sites (p < 0.001). Within each site, no generally consistent pattern over time occurred (p > 0.26). ANOVA indicated significantly higher concentrations on day 0 (0.04 g POM and 0.61 g PIM/m3) than day 30 (0.03 g POM and 0.15 g PIM/m3) at G, but higher concentrations on day 30 (0.04 g POM and 0.86 g PIM/m3) than day 0 (0.01 g POM and 0.20 g PIM/m3) at SG. Patterns of POM on day 30 were partly correlated to BOM in the cages: e.g., concentrations of both POM and BOM were high at G, M and LB and low at the glacial influenced SS, LR, TG and JS. However, the correlation of POM with BOM concentrations was poor (r2 = 0.29), because of high POM but low BOM concentrations at SG. Factors influencing colonization Multiple linear regression analysis between invertebrate density and Simpson's diversity of each cage and 6 variables describing time of colonization, study

157 6. Colonization in alpine streams 153 Table 6-3 Evaluation of factors influencing invertebrate colonization using multiple linear regression analysis. Six independent variables, including time of colonization (day), small-scale habitat characteristics (depth, velocity, BOM = benthic organic matter) and study stream conditions (DDAugust = accumulated degree days, QAugust = discharge), were stepwise related to invertebrate abundance (adjusted r2 = 0.54; F(6,120) = 25.0; p < 0.001) and Simpson's index of diversity (adjusted r2 = 0.12; F (3,123) = 6.9; p = < 0.001) in each cage and Hess sample. Beta = standardized regression coefficient standard error; p = significance level of respective variable; multiple r2 = increase in r2 by each variable; = variable removed by stepwise regression. Independent Invertebrate abundance Independent Simpson's index of diversity variable Beta p Multiple r2 variable Beta p Multiple r2 BOM 0.49 0.08 < 0.001 0.49 QAugust 0.43 0.10 < 0.001 0.04 QAugust - 0.16 0.08 0.042 0.52 BOM 0.25 0.10 0.013 0.12 Day 0.18 0.07 0.012 0.54 Day 0.21 0.11 0.047 0.14 Depth - 0.15 0.07 0.037 0.55 Depth DDAugust 0.10 0.08 0.219 0.56 DDAugust Velocity Velocity stream conditions and small-scale habitat characteristics indicated that both changes in invertebrate density, as well as diversity, were best explained by BOM concentration, discharge during August and day of colonization (Table 6-3). The model for invertebrate density included all variables except velocity and was highly significant (p < 0.001; adjusted r2 = 0.54), whereas the regression model of diversity included only three variables and had less predictive power (p < 0.001; adjusted r2 = 0.12). For all samples (cages + He ss), BOM concentration best predicted changes in invertebrate density (Beta = 0.49), whereas discharge in August was the best predicting variable for changes in diversity (Beta = 0.43). Velocity had no influence on predicting changes in invertebrate density or diversity and was removed from both regression models (Table 6-3). The inclusion of periphyton AFDM and chlorophyll a in a multiple linear regression based on average values indicated similar results. Changes in average

158 154 6. Colonization in alpine streams invertebrate density per sample day and site were best explained by BOM concentration (p < 0.001), and changes in Simpson's diversity per sample day and site were best explained by discharge in August (p < 0.001). Periphyton AFDM was removed from both regression models, whereas periphyton chlorophyll a significantly contributed to predicted changes in invertebrate densities (p = 0.0014) but was removed from the regression model for Simpson's diversity. Discussion Colonization patterns The colonization of the cages was generally rapid, occurring within 3 to 8 days at all sites. For instance, invertebrate density and taxon richness showed no significant increases after 8 days of colonization (except at the kryal stream SG), and the similarity between cages and Hess samples was high by day 30. Indeed, the colonization of habitat patches in these high-elevation streams was as fast or even faster than in low- and mid-elevation streams where colonization required 4 to 35 d for density and 9 to 28 days for taxon richness to reach equilibrium (Rosenberg and Resh 1982). However, colonization rates are difficult to compare among studies as they are influenced by different factors, such as the experimental design (Giller and Cambell 1989, Kadono et al. 1999), taxa studied (Ciborowski and Clifford 1984), season (Benson and Pearson 1987, Moser and Minshall 1996), and geographic location (Rosenberg and Resh 1982). It also must be noted that reaching equilibrium does not mean a stable community but simply a ratio of arrival (gain) that is equal to disappearance (loss). There were no significant differences between taxon richness in the cages and Hess samples representing the reference instream fauna, but high beta diversities among sampling dates indicated that the community composition was continually changing even after 30 days. On average, only 35 % of all taxa found during the colonization experiment at any one site (gamma diversity) occurred in the cages or in the Hess samples. Miller and Stout (1989) and Clements (1991) also reported fewer taxa in rock-filled baskets than in reference samples from their streams, whereas Winterbottom et al. (1997a) found all taxa present in a small English stream in colonization boxes after only one week of colonization. Ciborowski and Clifford (1984) observed even more taxa and greater densities in

159 6. Colonization in alpine streams 155 colonization trays than in Hess samples from a large river. The high taxa turnover (beta diversity) suggested that alpine stream invertebrates are highly mobile, and thus are continuously being redistributed within a stream. Factors influencing colonization (density + diversity) Colonization was influenced by the interaction of local stream conditions and patch-scale habitat characteristics. Although colonization was generally rapid at all sites, invertebrate density and the number of colonizing taxa differed strongly among the different stream types, as well as among individual cages. For instance, the variance of density and taxon richness among replicate cages collected on a particular day were higher than differences between consecutive sample days, suggesting that, in addition to local stream characteristics, small scale parameters influenced colonization dynamics of individual cages. Multiple linear regression suggested that the amount of benthic organic matter (BOM), discharge during the study, and time of colonization significantly predicted invertebrate density as well as diversity during colonization. Discharge, being an independent variable of each stream, was a primary determinant of diversity, whereas concentration of BOM, representing small-scale patch differences, was a major factor determining invertebrate density among cages within a stream. Other studies have found benthic density (Ciborowski and Clifford 1984, Winterbottom et al. 1997a), current velocity and discharge (Clements 1991, Winterbottom et al. 1997b), and substrate size and texture (Wise and Molles 1979, Khalaf and Tachet 1980) to be major factors determining patterns of colonization within a stream. However, most studies did not distinguish between local (stream) and small-scale habitat (patch) characteristics. For example, benthic density and stream discharge can be viewed as local stream characteristics. Discharge and habitat stability, as a function of discharge, have often been shown to significantly affect the density and number of colonizing taxa among different stream types (e.g., Clements 1991, Winterbottom et al. 1997a, Burgherr et al. in press). In the present study, glacial-influenced sites (except one kryal stream) having highest discharge values tended to have the lowest density and number of taxa. In addition, colonization at the partly glacial-influenced Jrisee outlet was more similar to the kryal lake outlets than

160 156 6. Colonization in alpine streams to the other rhithral lake outlet, indicating the potential influence of the nearby glacier; and rhithral streams with low discharge values had the highest number of taxa although maximum densities were low ( 10,000 ind./m2). Other studies found similar patterns in the invertebrate communities of different alpine stream types with lower numbers of taxa occurring in kryal streams compared to rhithral streams (see citations in Ward 1994). Thus, in alpine streams, the number of taxa colonizing small patches probably is determined by local stream conditions, as reflected in the zoobenthic community of the respective stream. In contrast, substrate size and composition, as well as concentration of BOM, represent small-scale patch characteristics within a stream reach. Flecker and Allan (1984) found that substrate size and composition had a marked effect on insect colonization. However, in explaining the 'preference' of invertebrates for loose substrate, they rejected the hypothesis that observed patterns of colonization were due to differences in refugia, proposing that these patterns are more likely due to different amounts of trapped detritus (also see Mackay 1992). Although several studies have alluded to the importance of resource availability on colonization (e.g., Benson and Pearson 1987, Robinso n et al. 1990, Elser 1999, Lancaster 2000), few have specifically investigated the effect of organic material on small-scale colonization patterns (e.g., Egglishaw 1964, Peckarsky 1980, Shaw and Minshall 1980, Flecker 1984, Doeg et al. 1989). Most of these investigations have demonstrated a significant correlation between the amount of detritus (as FPOM) and invertebrate colonization, and showed that the Chironomidae were strongly influenced by the amount of detritus (Egglishaw 1964, Shaw and Minshall 1980, Flecker 1984). However, Peckarsky (1980) found a significant effect of BOM only on shredder but not total macroinvertebrate density, and Elser (1999) argued that food did not affect colonization because it was not limiting in their streams. Alpine streams, in contrast, generally are low in allochthonous organic resources and in kryal streams, instream algal biomass also is low during summer (Ward 1994, Uehlinger et al. 1998). BOM concentrations in our streams ( 11 g/m2) were on the very low end in a comparison of a diverse array of 31 different streams in which concentrations ranged from 20 to 35,000 g BOM/m2 (Jones 1997). Thus, in the present study,

161 6. Colonization in alpine streams 157 additional food resources in some individual cages probably attracted invertebrates and resulted in small scale differences in the densities of colonists. Importance of drift as a colonization pathway Drift was presumably of minor importance for colonization at our study sites. Corresponding to an earlier drift study in the same streams (Hieber et al. in press-a), very few individuals were found in the drift regardless of time of day. Correspondingly, drift density was not correlated with any parameters of colonization. Many studies, however, have found drift to be a dominant source of colonists, contributing sometimes to > 80 % of invertebrate movements (e.g., Townsend and Hildrew 1976, Minshall and Petersen 1985, Matthaei et al. 1997, Gayraud et al. 2000), with early colonizers, such as Baetidae, Gammaridae, Simuliidae and Chironomidae, also common in the drift (Waters 1972, Williams and Hynes 1976, Mackay 1992). Besides drift, colonization also occurs by other pathways within and from outside the main stream channel. In the short-term colonization of instream habitat patches, upstream movements, crawling within the substrate complex, and random lateral movements have been attributed as other major sources for colonizing invertebrates (Williams and Hynes 1976, Giller and Cambell 1989), whereas colonization from aerial sources, i.e., by flying adult insects, may be more important in long-term colonization dynamics (Mackay 1992, Giller and Malmqvist 1998). Moser and Minshall (1996) observed that the relative importance in mode of colonization varied with season, with drift being more important in spring but drift and crawling being equally important in summer and autumn. Robinson et al. (2002) also found seasonal differences in drift rates among various streams of an alpine glacial floodplain and associated them with changes in the life cycles of particular taxa. Drift rates generally were lowest in spring and progressively increased from summer to autumn. However, drift rates differed substantially between years and among the various stream types, and low densities and no distinct seasonal pattern was found in the upper glacial channel (Robinson et al. 2002), which corresponds to our site TG. Similarly, Hieber et al. (in press-a) found no seasonality, suggesting that the influence of drift on colonization in these systems was probably low throughout the year. In

162 158 6. Colonization in alpine streams contrast, the high beta diversities among sampling dates, and assemblages dominated by mobile taxa such as Chironomidae and Ephemeropterans, indicate that active movement through swimming and crawling may be an important and dominant pathway of larval redistribution and colonization of small patches in these alpine streams. We suggest that the importance of drift as a colonization pathway probably increases with increasing distance from the water source as a function of the longitudinally increasing densities of both drifting and benthic invertebrates (Burgherr and Ward 2001, Robinson et al. 2002). Summary Our data suggest that macroinvertebrate colonization of small-scale patches in alpine streams is determined at different hierarchical levels: colonist density was influenced by the availability of organic matter, whereas the taxon richness of colonists was dependent somewhat on the habitat characteristics associated with a particular stream type. This idea supports the findings of Robinson and Minshall (1998) who concluded that "certain attributes of macroinvertebrate assemblages (density, biomass and production) are structured by environmental factors regulating growth (e.g., water chemistry thermal loading, periphyton), whereas other biotic attributes (e.g., species richness, assemblage composition, and life histories) are more influenced by thermal and flow regimes". In alpine streams, the glacial influence is a key factor reducing taxon richness relative to rhithral sites, whereas small-scale differences in the distribution of food resources strongly affect invertebrate density regardless of a glacial influence. In contrast to low-elevation streams, colonization in alpine streams appears to primarily result from the active movement of the invertebrates by crawling and less than from drift. We suggest that benthic invertebrates in alpine streams are constantly moving as a means to search for more optimal resources, thus resulting in a continuous redistribution and the rapid colonization of small-scale patches. Acknowledgements Special thanks to Peter Burgherr, Chregu Dinkel, Michael Dring and Tanja Dring for assistance in the field, Robert Berger and his team for manufacturing

163 6. Colonization in alpine streams 159 the cages, and Gabriella Meier Brgisser and Richard Illi for completion of the chemical analyses in the laboratory. The communes of Pontresina and Samedan kindly provided access to the sites. The study was partially funded by a Swiss National Science Foundation Grant (no. 31-50440.97) examining the ecology of alpine lake outlets. References BENSON, L. J., and R. G. PEARSON. 1987. The role of drift and effect of season on macroinvertebrate colonization of implanted substrata in a tropical Australian stream. Freshwater Biology 18:109-116. BRITTAIN, J. E., and A. M. MILNER . 2001. Ecology of glacier-fed rivers: current status and concepts. Freshwater Biology 46:1571-1578. BURGHERR , P., and E. I. MEYER. 1997. Regression analysis of linear body dimensions vs. dry mass in stream macroinvertebrates. Archiv fr Hydrobiologie 139:101- 112. BURGHERR , P., and J. V. W ARD. 2000. Zoobenthos of kryal and lake outlet biotopes in a glacial flood plain. Verhandlungen der Internationalen Vereinigung der Limnologie 27:1587-1590. BURGHERR , P., and J. V. WARD. 2001. Longitudinal and seasonal distribution patterns of the benthic fauna of an alpine glacial stream (Val Roseg, Swiss Alps). Freshwater Biology 46:1705-1721. BURGHERR , P., J. V. WARD, and C. T. ROBINSON. Seasonal variation in zoobenthos across habitat gradients in an alpine glacial flood plain (Val Roseg, Swiss Alps). Journal of the North American Benthological Society (in press). CIBOROWSKI , J. J. H., and H. F. CLIFFORD. 1984. Short-term colonization patterns of lotic macroinvertebrates. Canadian Journal of Fisheries and Aquatic Sciences 41:1626-1633. CLEMENTS , W. H. 1991. Characterization of stream benthic communities using substrate-filled trays: Colonization, variability, and sampling selectivity. Journal of Freshwater Ecology 6:209-222. DOEG, T. J., P. S. LAKE , and R. MARCHANT . 1989. Colonization of experimentally disturbed patches by stream macroinvertebrates in the Acheron River, Victoria (Australia). Australian Journal of Ecology 14:207-220.

164 160 6. Colonization in alpine streams DONATH, U., and C. T. ROBINSON. 2001. Ecological characteristics of lake outlets in Alpine environments of the Swiss Alps. Archiv fr Hydrobiologie 150:207-225. DOWNES , B. J., and M. J. KEOUGH . 1998. Scaling of colonization processes in streams: Parallels and lessons from marine hard substrata. Australian Journal of Ecology 23:8-26. EGGLISHAW , H. J. 1964. The distributional relationship between the bottom fauna and plant detritus in streams. Journal of Animal Ecology 33:463-475. ELSER , P. 1999. Use of colonization baskets for the investigation of disturbance phenomena in streams under model conditions. Limnologica 29:120-127. FISHER , S. G., L. J. GRAY, N. B. GRIMM , and D. E. BUSCH . 1982. Temporal succession in a desert stream ecosystem following flash flooding. Ecological Monographs 52:93-110. FLECKER, A. S. 1984. The effects of predation and detritus on the structure of a stream insect community: a field test. Oecologia 64:300-305. FLECKER, A. S., and J. D. ALLAN. 1984. The importance of predation, substrate and spatial refugia in determining lotic insect distributions. Oecologia 64:306- 313. FREDER, L. 1999. High alpine streams: cold habitats for insect larvae. Pages 181- 196 in Margesin, R., and F. Schinner (Editors). Cold-Adapted Organisms - Ecology, Physiology, Enzymology and Molecular Biology. Springer, Berlin, Germany. GAYRAUD, S., M. PHILIPPE, and L. MARIDET . 2000. The response of benthic macroinvertebrates to artificial disturbance: Drift or vertical movement in the gravel bed of two Sub-Alpine streams? Archiv fr Hydrobiologie 147:431- 446. GILLER , P. S. 1996. Floods and droughts: the effects of variations in water flow on streams and rivers. Pages 1-19 in Giller, P. S., and A. A. Myers (Editors). Disturbance and Recovery of Ecological Systems. Royal Irish Academy, Dublin, Ireland. GILLER , P. S., and R. N. B. CAMBELL . 1989. Colonization patterns of mayfly nymphs (Ephemeroptera) on implanted substrate trays of different size. Hydrobiologia 178:59-72.

165 6. Colonization in alpine streams 161 GILLER , P. S., and B. MALMQVIST 1998. The biology of streams and rivers. Oxford University Press, Oxford, U.K.. HARRISON, S., S. J. ROSS , and J. H. LAWTON. 1992. Beta diversity on geographic gradients in Britain. Journal of Animal Ecology 61:151-158. HIEBER, M., C. T. ROBINSON, and U. UEHLINGER . Seasonal and diel patterns of invertebrate drift in different types of alpine streams. Freshwater Biology (in press-a). HIEBER, M., C. T. ROBINSON, U. UEHLINGER, and J. V. W ARD. Are alpine lake outlets less harsh than other alpine streams? Archiv fr Hydrobiologie (in press-b). HOLOMUZKI , J. R., and B. J. F. BIGGS. 2000. Taxon-specific responses to high-flow disturbance in streams: implications for population persistence. Journal of the North American Benthological Society 19:670-679. IAHS/UNESCO. 1998. Fluctuations of glaciers 1990-1995. World Glacier Monitoring Service (WGMS), Zurich, Switzerland. JONES , J. B. 1997. Benthic organic matter storage in streams: influence of detrital import and export, retention mechanisms, and climate. Journal of the North American Benthological Society 16:109-119. KADONO , A., T. YASUHIRO, and T. MUTSUNORI. 1999. Mobility and habitat colonization in stream invertebrates: An experimental study. Japanese Journal of Limnology 60:215-222. KHALAF, G., and H. TACHET. 1980. Colonization of artificial substrata by macro- invertebrates in a stream and variations according to stone size. Freshwater Biology 10:475-482. KOWNACKI, A. 1991. Zonal distribution and classification of the invertebrate communities in high mountain streams. Verhandlungen der Internationalen Vereinigung der Limnologie 24:2010-2014. LAKE , P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North American Benthological Society 19:573-592. LANCASTER , J. 2000. Geometric scaling of microhabitat patches and their efficacy as refugia during disturbance. Journal of Animal Ecology 69:442-457. MACARTHUR , R. H., and E. O. WILSON. 1967. The theory of island biogeography. Princeton University Press, Princeton, NJ, U.S.A..

166 162 6. Colonization in alpine streams MACKAY, R. J. 1992. Colonization by lotic macroinvertebrates: A review of processes and patterns. Canadian Journal of Fisheries and Aquatic Sciences 49:617-628. MAGURRAN, A. E. 1991. Ecological diversity and its measurements. Chapman & Hall, London, U.K.. MATTHAEI, C. D., U. UEHLINGER, E. I. MEYER, and A. FRUTIGER. 1996. Recolonization by benthic invertebrates after experimental disturbance in a Swiss prealpine river. Freshwater Biology 35:233-248. MATTHAEI, C. D., D. W ERTHMLLER, and A. FRUTIGER . 1997. Invertebrate recovery from a bed-moving spate: the role of drift versus movements inside or over the substratum. Archiv fr Hydrobiologie 140:221-235. MEYER, E. 1989. The relationship between body length parameters and dry mass in running water invertebrates. Archiv fr Hydrobiologie 117:191-203. MEYNS , S., R. ILLI, and B. RIBI . 1994. Comparison of chlorophyll-a analysis by HPLC and spectrophotometry: where do the differences come from? Archiv fr Hydrobiologie 132:129-139. MILLER, M. C., and J. R. STOUT . 1989. Variability of macroinvertebrate community composition in an Arctic and Subarctic stream. Hydrobiologia 172:111-127. MINSHALL, G. W., and R. C. PETERSEN. 1985. Towards a theory of macroinvertebrate community structure in stream ecosystems. Archiv fr Hydrobiologie 104:49- 76. MOSER , D. C., and G. W. MINSHALL. 1996. Effects of localized disturbance on macroinvertebrate community structure in relation to mode of colonization and season. American Midland Naturalist 135:92-101. PECKARSKY, B. L. 1980. Influence of detritus upon colonization of stream invertebrates. Canadian Journal of Fisheries and Aquatic Sciences 37:957- 963. ROBINSON, C. T., and G. W. MINSHALL . 1986. Effects of disturbance frequency on stream benthic community structure in relation to canopy cover and season. Journal of the North American Benthological Society 5:237-248. ROBINSON, C. T., and G. W. MINSHALL . 1998. Macroinvertebrate communities, secondary production, and life history patterns in two adjacent streams in Idaho, USA. Archiv fr Hydrobiologie 142:257-281.

167 6. Colonization in alpine streams 163 ROBINSON, C. T., G. W. MINSHALL , and S. R. RUSHFORTH . 1990. Seasonal colonization dynamics of macroinvertebrates in an Idaho stream. Journal of the North American Benthological Society 9:240-248. ROBINSON, C. T., K. TOCKNER , and P. BURGHERR . 2002. Seasonal patterns in macroinvertebrate drift and seston transport in streams of an alpine glacial flood plain. Freshwater Biology 47 (in press). ROBINSON, C. T., U. UEHLINGER , and M. HIEBER. 2001. Spatio-temporal variation in macroinvertebrate assemblages of glacial streams in the Swiss Alps. Freshwater Biology 46:1663-1672. ROSENBERG, D. M., and V. H. RESH . 1982. The use of artificial substrates in the study of freshwater benthic macroinvertebrates. Pages 175-235 in Cairns, J. (Editor). Artificial Substrates. Ann Arbor Science Publishers, Ann Arbor, MI, U.S.A.. SHAW , D. W., and G. W. MINSHALL . 1980. Colonization of an introduced substrate by stream macroinvertebrates. Oikos 34:259-271. SHELDON, A. L. 1984. Colonization dynamics of aquatic insects. Pages 401-429 in Resh, V. H., and D. M. Rosenberg (Editors). The Ecology of Aquatic Insects. Praeger Publishers, New York, U.S.A.. SOKAL , R. R., and F. J. ROHLF 1995. Biometry: the principles and practice of statistics in biological research. 3. W. H. Freeman and Company, New York, U.S.A.. STATSOFT, I. 1995. STATISTICA for Windows (Computer program manual). Tulsa, OK, U.S.A.. TOWNSEND , C. R. 1989. The patch dynamics concept of stream community ecology. Journal of the North American Benthological Society 8:36-50. TOWNSEND , C. R., S. DOLDEC, and M. R. SCARSBROOK. 1997. Species traits in relation to temporal and spatial heterogeneity in streams: a test of habitat templet theory. Freshwater Biology 37:367-387. TOWNSEND , C. R., and A. G. HILDREW . 1976. Field experiments on the drifting, colonization and continuous redistribution of stream benthos. Journal of Animal Ecology 45:759-772.

168 164 6. Colonization in alpine streams UEHLINGER , U. 1991. Spatial and temporal variability of the periphyton biomass in a prealpine river (Necker, Switzerland). Archiv fr Hydrobiologie 123:219- 237. UEHLINGER , U., R. ZAH, and H. BRGI. 1998. The Val Roseg Project: temporal and spatial patterns of benthic algae in an Alpine stream ecosystem influenced by glacier runoff. Pages 419-424 in Kovar, K., U. Tappeiner, N. E. Peters, and R. G. Craig (Editors). Hydrology, Water Ressources and Ecology in Headwaters. IAHS Press, Wallingford, U.K.. USSEGLIO -POLATERA, P., P. RICHOUX, M. BOURNAUD, and H. TACHET . 2001. A functional classification of benthic macroinvertebrates based on biological and ecological traits: application to river condition assessment and stream management. Archiv fr Hydrobiologie. Supplement-Band 139:53-83. WARD, J. V. 1994. Ecology of alpine streams. Freshwater Biology 32:277-294. WATERS , T. F. 1972. The drift of stream insects. Annual Review of Entomology 17:253-272. WILLIAMS , D. D., and H. B. N. HYNES . 1976. The recolonization mechanisms of stream benthos. Oikos 27:265-272. WINTERBOTTOM , J. H., S. E. ORTON, and A. G. HILDREW . 1997a. Field experiments on the mobility of benthic invertebrates in a southern English stream. Freshwater Biology 38:37-47. WINTERBOTTOM , J. H., S. E. ORTON, A. G. HILDREW , and J. LANCASTER . 1997b. Field experiments on flow refugia in streams. Freshwater Biology 37:569-580. WISE, D. H., and M. C. MOLLES , JR. 1979. Colonization of artificial substrates by stream insects: Influence of substrate size and diversity. Hydrobiologia 65:69-74.

169 7. Synopsis 165 7. Synopsis Research in alpine environments The history of limnological alpine research started more than 100 years ago (Steinmann 1907, Zschokke 1908 and citations therein). Early studies at the beginning of the 20th century laid the foundation for the knowledge of alpine stream biota and habitats (Steinmann 1907, Thienemann 1912, Dodds and Hisaw 1925). Since then, both sampling technology and statistical analyses have substantially improved and the study of alpine streams was significantly expanded by including other scientific fields such as climatology, glaciology, and geology (e.g., Lavandier 1974, Lavandier and Mur 1974, Gurnell et al. 1999, Smith et al. 2001). Yet, many processes are still not fully understood and many questions remain unanswered. One major constraint in alpine research is the long winter period impeding accessibility and sampling, thus most studies have been of limited duration. In addition, most studies were of single streams, thus results minimally transferable to other systems. Hence, comparative long-term studies of different alpine stream types are essential for a better understanding of the ecology of alpine streams. My motivation for this study resulted first, from respect for the beauty and uniqueness of alpine streams, and second, from the desire for a better understanding of the complexity of these systems and their inhabitants. Unfortunately, high mountain environments are highly sensitive ecosystems and climatic change and human impact significantly altered and affected most stream ecosystems (Stone 1992, McGregor et al. 1995). In this study, therefore, I wanted to contribute to a better understanding of these complex sy stems and to stress the importance of protecting and preserving the remaining undisturbed stream systems, and restoring those that have been damaged. Alpine streams aspects of biocomplexity Over the past years, most studies have agreed on the common typology of alpine streams as kryal, rhithral and krenal segments (Ward 1994). The present study showed that within a stream type, subtypes exist that markedly differ in their

170 166 7. Synopsis environmental conditions and community patterns. The various environmental features affected the habitats of alpine stream systems at different hierarchical levels, and therefore, can be arranged as nested "filters" that "screen" species from the regional species pool by their biotic traits (sensu Poff 1997), thereby determining the structure of the invertebrate community within a stream (see Fig. 4-7). The glacial influence on channel stability and seasonality is a primary determinant of stream habitat characteristics. In addition, an upstream lake affects environmental conditions and respective temporal fluctuations in the outlet stream, such as water temperature and food resources, yet was still strongly influenced by the presence of a glacier. Thus, differences in the environmental conditions between lake outlets and streams are more pronounced for rhithral than kryal systems. Ward (1994) noted that alpine stream organisms display relatively broad geographical distributions. In this study, stream communities showed no differences among drainages and assemblage composition was similar among the different alpine stream types. However, the benthic communities exhibited distinct patterns in the relative contribution by individual taxa in the different stream types, and thus reflected differences in habitat characteristics. In general, diversity increased from kryal sites to rhithral streams. Kryal sites generally were dominated by the chrysophyte Hydrurus foetidus, the blue-green algae Chamaesiphon and Lyngbya, and species of the chironomid family Diamesinae. Species richness and abundance of benthic algae and macrozoobenthos in kryal sites tended to be lowest during the maximum period of glacial melt in summer. Diatom communities appeared to be more diverse in kryal lake outlets than in kryal streams, but invertebrate communities showed no distinct differences even though kryal lake outlets exhibited more annual degree days, higher temperatures, and lower temporal fluctuations in flow and temperature. Rhithral streams and rhithral lake outlets significantly differed in environmental conditions relative to kryal sites, generally displaying higher water temperatures, less turbid water, less pronounced temporal fluctuations in flow and higher channel stability. Among rhithral sites, streams and lake outlets clearly differed in their habitat characteristics and biotic communities. Algal

171 7. Synopsis 167 Table 7-1 Idealized abiotic and biotic features of alpine stream types (modified from Ward 1994, to reflect the findings of this study). E: Ephemeroptera, P: Plecoptera, T: Trichoptera, D: Diptera. Rhithral Kryal Variable Stream Lake outlet Stream Lake outlet Annual degree days 900 1300 900 1500 < 300 500 - 700 Annual temperature 0 13 0 17 05 09 range (C) Diel temperature High Intermediate Intermediate Low fluctuations Flow regime High seasonal Intermediate seasonal High seasonal + diel Intermediate seasonal fluctuations fluctuations fluctuations + diel fluctuations Transparency (NTU) Clear (0 3) Clear (0 10) Turbid (2 >1000) Turbid (30 400) Channel stability Variable High Low Variable Algae Diverse diatoms and Diverse diatoms and Hydrurus foetidus Hydrurus foetidus blue-green algae blue-green algae Chamaesiphon Chamaesiphon Hydrurus foetidus Moss Lyngbya Lyngbya Sparse diatoms Few diatoms Macroinvertebrates diverse EPTD non-insects: Diamesinae Diamesinae and non-insects Oligochaeta, EP: Baetidae, EP: Baetidae, Nemathelminthes Heptageniidae, Heptageniidae, Chironomidae Leuctridae Leuctridae few EPT communities in rhithral lake outlets consisted of diverse diatoms and blue-green algae as well as seasonally high biomass of moss, and invertebrate assemblages were dominated by chironomids and non-insect taxa such as Oligochaeta, Nemathelminthes and Copepoda. Rhithral streams, in comparison, were characterized by a relatively diverse biotic community that included diverse diatoms, blue-green algae and Hydrurus foetidus as well as Ephemeroptera, Plecoptera, Trichoptera, Diptera and non-insect taxa, thus representing a combination of taxa characteristic to both kryal sites and rhithral lake outlets. A summary of the idealized features of the studied alpine stream types is presented in Table 7-1 based on criteria suggested by Ward (1994).

172 168 7. Synopsis Furthermore, the distinctiveness of alpine stream types and the hierarchical structure of their environmental features influences mechanistic processes dictating invertebrate assembly. For example, in contrast to patterns observed in low- and midland streams, drift displayed no diel periodicity and this may be a general feature of high altitude streams. Instead, analogous to benthic community patterns, density and richness of drifting, as well as colonizing invertebrates, were significantly affected by glacial influence. Within a stream reach, small-scale distribution patterns of food resources substantially affected colonization dynamics of local patches. Thus, alpine streams display complex dynamics controlled by the interaction of diverse environmental features acting at different hierarchical levels. The distinctive habitat conditions are reflected in the structure and distribution of the benthic stream communities. However, the distinctiveness of lake outlet communities declines with increasing elevation and glacial influence. Perspectives The present thesis emphasized the complexity of alpine stream habitats, their associated biota and the relation between them. However, the presented patterns are still far from being complete and definitive. More studies on alpine stream systems are needed to confirm and extend understanding of pattern and process among different alpine stream types. In particular, the following topics need further research: Only two kryal 'lake outlet stream' systems have been investigated, they exhibited somewhat divergent patterns. It remains unclear whether distinct differences between kryal lake outlets and kryal streams exist. Thus, more kryal systems must be rigorously examined to fully encompass the potential variability among kryal streams and lake outlets. Although the studied lake outlets showed distinctive habitat characteristics and biotic assemblages, differences also occurred among lake outlets. Thus, more focus must be put on additional habitat and landscape characteristics, such as size of the lake, period of snow cover (as a function of altitude and shading from the surrounding mountains) and flow permanence, to better understand which factors affect and determine community patterns.

173 7. Synopsis 169 Differences in habitat characteristics and the biota were found among streams as well as between patches within streams, and similarly, temporal patterns occurred within days, between seasons and among years. Thus, different spatial and temporal scales must be examined to understand the interaction between scales and to distinguish the influence of respective scales. Finally, a more intense exchange between limnologists and other disciplines is essential for a holistic understanding of these unique environments. And lastly, the knowledge of natural stream systems must be applied in the ecological management of rivers to maintain natural habitat conditions, and thus sustain native biodiversity in alpine streams. References DODDS , G. S., and F. L. HISAW . 1925. Ecological studies on aquatic insects. IV. Altitudinal range and zonation of mayflies, stoneflies, and caddisflies. Ecology 6:380-390. GURNELL , A. M., P. J. EDWARDS , G. E. PETTS , and J. V. W ARD. 1999. A conceptual model for alpine proglacial river channel evolution under changing climatic conditions. Catena 38:223-242. LAVANDIER , P. 1974. cologie d'un torrent pyrnen de haute montagne II. - Caractristiques physiques. Annales de Limnologie 10:173-219. LAVANDIER , P., and C. MUR . 1974. cologie d'un torrent pyrnen de haute montagne II. - Caractristiques chimiques. Annales de Limnologie 10:275- 309. MCGREGOR , G., G. E. PETTS, A. M. GURNELL , and A. M. MILNER . 1995. Sensitivity of alpine stream ecosystems to climate change and human impacts. Aquatic Conservation: Marine and Freshwater Ecosystems 5:233-247. POFF, N. L. 1997. Landscape filters and species traits: towards mechanistic understanding and prediction in stream ecology. Journal of the North American Benthological Society 16:391-409. SMITH , B. P. G., D. M. HANNAH, A. M. GURNELL, and G. E. PETTS . 2001. A hydrogeomorphological context for ecological research on alpine glacial rivers. Freshwater Biology 46:1579-1596.

174 170 7. Synopsis STEINMANN, P. 1907. Die Tierwelt der Gebirgsbche - eine faunistisch-biologische Studie. Annales de Biologie Lacustre 2:30-162. STONE , P. B. 1992. The state of the world's mountains - a global report. Zed Books Ltd, London, U.K.. THIENEMANN, A. 1912. Der Bergbach des Sauerland. Internationale Revue der Gesamten Hydrobiologie Supplement 4:1-25. WARD, J. V. 1994. Ecology of alpine streams. Freshwater Biology 32:277-294. ZSCHOKKE , F. 1908. Die Resultate der zoologischen Erforschung hochalpiner Wasserbecken seit dem Jahre 1900. Revue der Hydrobiologie 1:221-235.

175 Terms 171 Terms AFDM: ash-free dry mass, a measure of organic matter content. Alpine: pertaining to the European Alps, regardless of altitude. alpine: areas and organisms situated between the treeline and the permanent snowline. benthos: organisms attached to, living on, in or near the stream bottom. DOC: dissolved organic carbon. drift: downstream transport of stream-dwelling organisms. habitat: the locality, site and particular type of local environment occupied by an organism. krenal: pertaining to a stream segment being spring- or groundwater-fed. kryal: pertaining to a stream originating from the meltwater of glaciers. lentic: pertaining to static, calm or slow-moving aquatic habitats, such as ponds and lakes. lotic: pertaining to running-water habitats, such as streams and rivers NTU: nephelometric turbidity units, a measurement of turbidity. PIM: particulate inorganic matter, part of seston. PN: particulate nitrogen. POC: particulate organic carbon. POM: particulate organic matter, part of seston. PP: particulate phosphorus. rhithral: pertaining to the upper reaches of a river being mainly fed by snowmelt and rainfall. seston: the total (organic and inorganic) particulate matter suspended in water, comprising POM and PIM. SRP: soluble reactive phosphorus, PO4-P, ortho-phosphate. TDN: total dissolved nitrogen. TDP: total dissolved phosphorus. TIC: total inorganic carbon. TSS: total suspended solids.

176 172

177 Acknowledgments 173 DANKE ! all den vielen Leuten, die mich whrend den letzten vier Jahren auf die eine oder andere Art untersttzt und begleitet haben. Ganz besonderen Dank an Chris Robinson, der mich mit seiner Begeisterung immer wieder ansteckte und mir bei allen Problemen und Abenteuern zur Seite stand von meinem ersten Hesssampler-Versuch bis zur Verffentlichung, bei wunderschnen Bergtouren, der mit mir Schneestrmen trotzte und im Schnee zeltete, und mir nicht nur ein sehr guter Betreuer sondern auch ein wertvoller Freund wurde. Einen ebenso herzlichen Dank an Urs Uehlinger fr seine kritischen Fragen, seine konstruktiven Vorschlge, seine Begeisterungsfhigkeit und die sehr angenehme Arbeitsatmosphre in unserem Bro. Meinem Doktor- vater Prof. J. V. Ward mchte ich sehr fr die Ausarbeitung und Betreuung des Projektes danken; und Alexander Milner danke ich fr die bernahme des Korreferats. Ganz speziell danken mchte ich folgenden Kollegen und Kolleginnen, die mich nicht nur bei der Arbeit sehr untersttzten sondern mit der Zeit auch sehr enge und kostbare Freunde wurden: Mike Monaghan fr seine Ruhe und unermdliche Untersttzung - sei es bei einer Probenahme mitten in der Nacht, bei strmendem Regen und Gewitter, oder bei dem Versuch bei 3 m Schnee den Fluss zu finden; Peter Burgherr fr seine stete Bereitschaft, mich auch bei Dauerregen und schlechtem Wetter zur Probenahme zu begleiten, mir geduldig bei statistischen Fragen zu helfen und gemeinsam kniffelige Tiere zu bestimmen; Dave Arscott fr seine ansteckende Begeisterung und die vielen hilfreichen und spannenden Diskussionen; und schliesslich Sandra Lass, die sich trotz des fr sie fremden Gebietes der Fliessgewsserkologie fr meine Arbeit interessierte und fr mich immer als eine sehr enge und vertraute Freundin da war. Meine vielen Probenahmentouren bei Wind und Wetter, Sonne und Regen, Schneesturm und Hagel, zu Fuss, mit Ski oder per Helikopter wren nicht mglich gewesen ohne die Hilfe vieler treuer Freunde: Mike Monaghan, Chris Robinson, Peter Burgherr, Chregu Dinkel, Marcos de la Puente, Michel Dring, Urs Uehlinger, Andreas Blum, Heiko Rinderspacher, Christian Rust, Donna Anderson, Uli Donath, Florian Malard, Tanja Dring, Sandra Lass und Monika Winder Danke!

178 174 Acknowledgments Bedanken mchte ich mich auch bei der gesamte Limnologie-Abteilung fr die angenehme Arbeitsatmosphre und im besonderen bei Christiane Rapin, Ccile Claret, Andreas Frutiger, Achim Ptzold, Klement Tockner und Rainer Zah fr die vielen kleinen und grossen Untersttzungen bei wissenschaftlichen ebenso wie persnlichen Fragen, und bei Michel Dring, der mit mir in der Endphase Bro, Freud und Leid teilte. Ein grosses Dankeschn gebhrt dem Chemielabor mit Gabriella Meier Brgisser, Richard Illi und Bruni Ribi, die Hunderte von Wasser- und Chlorophyll-Proben analysierten, und Regula Illi und Esther Keller, die einen Teil meiner Algenproben bestimmten. Ein Danke auch an die EAWAG fr die Bereitstellung diverser Mglichkeiten, im besonderen einen herzlichen Dank an Monika Zemp und das Bibliotheksteam, an Bouziane Outiti und Daniel Pellanda fr die Informtikuntersttzung, an Robert Berger und sein Werkstattsteam und an Hans-Walter Schwaninger, Werner Frischknecht und das restliche Garagenteam. Rolf Glatthaar und Samuel Rushforth haben freundlicherweise bei der Bestimmung der Simuliidae und der Algen geholfen. Dem Bundesamt fr Landestopographie danke ich fr die Verwendung der Schweizer Karte. Abflussdaten wurden von der Schweizer Landeshydrologie und - geologie und vom Amt fr Umwelt Graubnden zur Verfgung gestellt. Die Gemeinden Pontresina, Samedan, und Klosters und Heinz Jossi vom Alpin Zentrum Steingletscher haben uns freundlicherweise Zugang zu den Probenahmestellen gewhrt. Die Arbeit wurde teilweise durch ein Stipendium des Schweizer Nationalfonds (Nr. 31-50440.97) finanziert. Ganz herzlichen Dank meiner Familie, im besonderen meinen Eltern Rotraut und Karl Hieber, fr ihr Vertrauen, ihre Untersttzung und geteilte Begeisterung, und allen weiteren netten Menschen, die in dieser Zeit Freunde blieben, Freunde wurden und hoffentlich Freunde bleiben werden! Danke! ..und schlussendlich mchte ich mich bei all den Eintagsfliegen Steinfliegen Kcherfliegen Zuckmcken Kriebelmcken Stelzmcken Plattwrmern Rundwrmern Wenigborstern Ruderfssern Muschelkrebsen und Wassermilben bedanken, die fr diese Arbeit ihr Leben lassen mussten! - es war eine sehr spannende und lehrreiche Zeit!

179 Curriculum Vitae 175 Curriculum Vitae Margit Hieber born: 28 April 1970 in Aschaffenburg, Germany home address: Heinrich-Heine-Strasse 6, D-88677 Markdorf, Germany e-mail: [email protected] Education 1976 - 1989 Primary school and high school (Gymnasium) in Markdorf, Germany allgemeine Hochschulreife (A-Levels) 1990 - 1998 Study of biology at the University of the Saarland, Germany 1996 - 1997 Diploma: Decomposition and colonization of leaf litter in an upland stream. Limnological Research Center, EAWAG, Kastanien- baum, Switzerland. Advisor: PD Dr. M.O. Gessner 1998 - 2002 Ph.D.: Alpine streams: aspects of biocomplexity. Swiss Federal Institute of Technology (ETH) in Zurich, Switzerland, and Swiss Federal Institute of Environmental Science and Technology (EAWAG) Advisors: Prof. Dr. J.V. Ward, PD Dr. C.T. Robinson, and Dr. U. Uehlinger Professional activities 1989 - 1990 Study abroad in Paris, France, Language course at the Sorbonne, Certificat de Langue Franaise, work as au-pair-girl. 1993-1994 Internship at the 'V. Engelhardt Institute of Molecular Biology of Russian Academy of Sciences' in Moscow, Russia. 1996 - 1997 Teaching assistantship at the University of the Saarland, Germany. 1999 - 2002 ETH-Assistant in the Department of Limnology, EAWAG.

180 176 Curriculum Vitae Publications Hieber, M. and Gessner, M. (2002). Contribution of bacteria, fungi and invertebrate shredders to leaf litter breakdown in a stream: estimates derived from biomass determinations. Ecology 83(4): 1026-1038. Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2002). Are alpine lake outlets less harsh than other alpine streams? Archiv fr Hydrobiologie 154(2): 199-223. Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (in press). Habitat characteristics of alpine streams in Switzerland. Extended abstract. Verhandlungen der Internationalen Vereinigung der Limnologie 28. Monaghan M.T., Hieber, M., Robinson, C.T., Spaak, P., and Ward, J.V. (in press). Spatial patterns of Ephemeroptera, Plecoptera, and Tricoptera diversity in fragmented alpine streams. Verhandlungen der Internationalen Vereinigung der Limnologie 28. Hieber, M., Robinson C.T., Rushforth S.R., and U. Uehlinger. (2001). Algal communities associated with different alpine stream types. Arctic, Antarctic, and Alpine Research 33 (4): 447-456. Robinson C.T., Uehlinger U., and Hieber, M. (2001). Spatio-temporal variation in macroinvertebrate assemblages of glacial streams in the Swiss Alps. Freshwater Biology 46 (12): 1663-1672. Hieber, M. (2000). Revitalisieren von Fliessgewssern: mehr Raum fr Flsse dient Mensch, Tier und Pflanzen. Die Schweizer Gemeinde 2: 13-15. Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2000). Periphyton und Seston in Seeausflssen der Schweizer Alpen. Tagungsbericht 1999 der Deutschen Gesellschaft fr Limnologie (DGL) (Rostock). 230-234. Hieber, M. and Gessner, M. (1998). Abbau und Besiedlung von Fallaub in einem Mittelgebirgsbach: ein Zehnjahresvergleich. Tagungsbericht 1997 der Deutschen Gesellschaft fr Limnologie (DGL) (Frankfurt/M.). 499-503. Shirokova, E.A., Tarussova N.B., Shipitsin A.V., Semizarov D.G., M. Hieber, Krayevsky A.A. (1995). Novel open-chain nucleotides imitating 2',3'-dideoxy- 2',3'-didehydronucleotides: synthesis and substrate properties toward DNA polymerases. Nucleosides & Nucleotides, 14 (3-5): 749-751.

181 Curriculum Vitae 177 Presentations at scientific meetings Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2001) Alpine streams: aspects of biocomplexity. 6. Nationale Tagung zur Alpenforschung, Luzern, Switzerland (poster presentation). Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2001) Habitat characteristics of alpine streams in Switzerland. Societas Internationalis Limnologiae XXVIII Congress, Melbourne, Australia (oral presentation). Monaghan M.T., Hieber, M., Robinson, C.T., Spaak, P., and Ward, J.V. (2001) Spatial patterns of Ephemeroptera, Plecoptera, and Tricoptera diversity in fragmented alpine streams. Societas Internationalis Limnologiae XXVIII Congress, Melbourne, Australia (oral presentation). Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2000) Suspended particulate matter and drifting invertebrates: seasonal and diel patterns in alpine streams and lake outlets. International Symposium: High Mountain Lakes and Streams, Innsbruck, Austria (oral presentation). Hieber, M., Robinson C.T., Rushforth S.R., and Uehlinger U. (2000) Algal assemblages of alpine lotic systems in Switzerland. International Symposium: High Mountain Lakes and Streams, Innsbruck, Austria (poster presentation). Robinson C.T., Uehlinger U., and Hieber, M. (2000) Alpine glacial streams: interglacial variability and seasonality in macroinvertebrate communities. International Symposium: High Mountain Lakes and Streams, Innsbruck, Austria (oral presentation). Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (2000) Alpine streams and lake outlets: seasonal patterns in macroinvertebrate assemblages. 48th annual meeting of the North American Benthological Society, Keystone, USA (oral presentation). Robinson C.T., Uehlinger U., and Hieber, M. (2000) Alpine glacial streams: interglacial variability and seasonality in macroinvertebrate communities. 48th annual meeting of the North American Benthological Society, Keystone, USA (oral presentation). Uehlinger U., Robinson C.T., and Hieber, M. (2000) Alpine glacial streams: habitat templet and seasonal patterns in periphyton. 48th annual meeting of

182 178 Curriculum Vitae the North American Benthological Society, Keystone, USA (oral presentation). Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (1999) Periphyton und Seston in Seeausflssen der Schweizer Alpen. Jahrestagung der Deutschen Gesellschaft fr Limnologie (DGL), Rostock, Germany, (oral presentation). Hieber, M., Robinson C.T., Uehlinger U., and Ward J.V. (1999) Seasonal patterns of periphytic and sestonic biomass in alpine lake outlets and other alpine streams. Symposium for European Freshwater Sciences, Antwerp, Belgium (poster presentation). Hieber, M. and Gessner, M.O. (1997) Abbau und Besiedlung von Fallaub in einem Mittelgebirgsbach: ein Zehnjahresvergleich. Jahrestagung der Deutschen Gesellschaft fr Limnologie (DGL), Frankfurt/M., Germany (poster presentation).

Load More