Arch. Hydrobiol. Spec. Issues Advanc. Limnol. 53, p. 255-274, December 1998 Lake Constance, Characterization of an ecosystem in transition Crustacean zooplankton in Lake Constance from 1920 to 1995: Response to eutrophication and re-oligotrophication Dietmar Straile and Waiter Geller with 9 figures Abstract: During the first three quarters ofthis century, the trophic state ofLake Constance changed from oligotrophic to meso-/eutrophic conditions. The response ofcrustaceans to the eutrophication process is studied by comparing biomasses ofcrustacean zooplankton from recent years, i.e. from 1979-1995, with data from the early 1920s (AUERBACH et a1. 1924, 1926) and the 1950s (MUCKLE & MUCKLE­ ROTTENGATTER 1976). This comparison revealed a several-fold increase in crustacean biomass. The relative biomass increase was more pronounced from the early 1920s to the 1950s than from the 1950s to the 1980s. Most important changes ofthe species inventory included the invasion of Cyclops vicinus and Daphnia galeata and the extinction of Heterocope borealis and Diaphanosoma brachyurum during the 1950s and early 1960s. All species which did not become extinct increased their biomass during eutrophication. This increase in biomass differed between species and throughout the season which re­ sulted in changes in relative biomass between species. Daphnids were able to enlarge their seasonal window ofrelative dominance from 3 months during the 1920s (June to August) to 7 months during the 1980s (May to November). On an annual average, this resulted in a shift from a copepod dominated lake (biomass ratio cladocerans/copepods = 0.4 during 1920/24) to a cladoceran dominated lake (biomass ratio cladocerans/copepods = 1.5 during 1979/95). The biomass of cyclopoid copepods increased strongly during the first half of the year owing to the invasion of Cyclops vicinus, which caused a strong relative decline ofEudiaptomus. In contrast to the pronounced response to eutrophication, crustaceans have not yet shown an unambiguous response to the beginning re-01igotrophication ofLake Constance. Introduction The eutrophication of a lake ecosystem may be regarded as a large-scale ecological experi­ ment, the study of which will offer important insights into the mechanisms structuring eco­ logical communities. Recent research has emphasized the importance of phenomena, such as indirect effects (STRAUSS 1991, WOOTTON 1994), species invasions and extinctions (LODGE 1993), and dynamics of resting stages (HAIRSTON et al. 1995, ADRIAN & DENEKE 1996), which act on large temporal and spatial scales and, hence, are difficult to study within labora- Addresses of the authors: D. Straile, Limno10gical Institute, University of Konstanz, D-78457 Konstanz, Germany. e-mail: [email protected]. - W. Geller, UFZ-Centre for Environ­ mental Research, Institute for Inland Water Research, Magdeburg, D-39I04 Magdeburg, Germany. 0071-1128/98/0053-255 $ 5.00 © 1998 E. Schweizerbart'sche Verlagsbuchhandlung, 0-70176 Stuttgart 256 D. Straile and W. Geller tory and mesocosm experiments. In this respect, the analysis of the "large-scale experiment eutrophication" will complement studies on smaller temporal and spatial scales. Crustacean zooplankton has been shown to respond to an increase in nutrient loading both by an increase in overall abundance and biomass and by pronounced changes in the commu­ nity structure including the new occurrence and loss of species (RAVERA 1980, EINSLE 1983, 1988, DE BERNARDI et al. 1988, GEORGE et al. 1990, NAUWERCK 1991, POLLI & SIMONA 1992, FITZSIMONS & ANDREW 1993). Based on long-term records and comparisons oflakes of different trophy, several workers suggested that rising lake trophy will favour cyclopoid copepods over ca1anoid copepods (GLIWICZ 1969, PATALAS 1972, ROGNERUD & KJELLBERG 1984, LANGELAND & REINERTSEN 1982) and cladocerans over calanoid copepods (PATALAS 1972, ROGNERUD & KJELLBERG 1984). According to pigment concentrations in the sediment, the "experiment eutrophication" started in Lake Constance at the beginning of this century (LENHARD 1994). Since 1952, eutrophication is documented by measurements of total phosphorus concentrations, which in­ creased approximately 10-fold until 1979. Due to massive efforts in sewage removal, total phos­ phorus concentrations declined down to 33% of the maximum values in recent years (see inlet Fig. 1 and GliDE et al. 1998). The response of phytoplankton to eutrophication was evident already in 1935 (GRIM 1955) and increases in biomass and shifts in species composition continued into the 1970s (WALZ et al. 1987, KUMMERLIN 1998). Theresponse ofphytoplankton to re-oligotrophication is equally well documented and consist of a decline in biomass during summer (GAEDKE & SCHWEIZER 1993, GAEDKE 1998a), pronounced shifts in species composition (SOMMER et al. 1993, KUMMERLIN 1998), and a modest decrease in primary productivity only during the most recent years (HAESE et al. 1998). The first descriptions of zooplankton in Lake Constance date back to the middle of the last century (LEYDIG 1860, WEISMANN 1876). Following the pioneering work of RENSEN (1884) in the ocean, ROFER started in 1892 to study quantitatively horizontal and vertical distributions of various zooplankton species in Lake Constance (ROFER 1896). His work was followed by AUERBACH et al. (1924, 1926) during 1919-1924 and ELSTER during 1932-1935 (ELSTER 1936, 1954, ELSTER & SCHWOERBEL 1970). Recognizing the signs ofeutrophication, a moni­ toring program was initiated by the Institut fUr Seenforschung at Langenargen in 1952 which has continued until now (KIEFER & MUCKLE 1959, MUCKLE & MUCKLE-ROTTENGATTER 1976, EINSLE 1977, 1983, 1987, 1988). These studies were accompanied by investigations of the Limnological Institute at Konstanz since the 1970s (LAMPERT & SCHOBER 1978, GELLER 1985, 1989). A combination of these studies yields a long-term time series of outstanding quantity and quality which covers the period of a dramatic increase ofnutrient loadings during the first three quarters ofthis century, but also includes the years of a beginning re-oligotrophi­ cation. To analyse the zooplankton response to increased nutrient loadings, the present study compares long-term averages of zooplankton biomass from three periods, i.e., 1920-1924 (AUERBACH et al. 1924, 1926), 1952-1962 (KIEFER & MUCKLE 1959, MUCKLE & MUCKLE­ ROTTENGATTER 1976), and 1979-1995, on a seasonal basis (Fig. 1). Subsequently, the response to decreasing nutrient loadings will be analysed within the years 1979-1995. Methods From 1979 to 1995, zoop1ankton was collected weekly (biweekly during the winter months, no data for 1983) with a Clarke-Bumpus sampler (mesh-size 140 1Jll1) by vertical hauls from 140 m Crustacean zooplankton 257 4 100 80 60 ~ 3 3 40 , a. NE f- U 20 ~ o en 2 en 2000 ro E :.c0 JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND 1920-1924 1952-1962 1979-1995 Fig. 1. Average monthly crustacean biomass during the three periods of investigation (1920-1924, 1952-1962, and 1979-1995). Inlet shows the development of total phosphorus concentrations during winter mixis (TPMIX) indicating the trophic state of the lake (no consecutive measurements prior to 1951). Data on TPMIX were provided by the International Commission for the Protection of Lake Constance (IGKB 1997). depth at the sampling station of the Limnological Institute at the central part of the Uberlinger See, a fjord-like branch of Upper Lake Constance. During routine measurements, seven taxa were identified: Daphnia hyalina, Daphnia galeata, Bosmina sp., Eudiaptomus gracilis, cyclopoidcopepods, Leptodora kindtii, and Bythotrephes longimanus. The individual taxa were separated into up to 5 size classes. Biomass was calculated from length-dry weight relationships established for Lake Constance (GELLER & MULLER 1985, WOLFL 1991). Data from previous years were taken from MUCKLE & MUCKLE-ROTTENGATTER (1976) who give monthly mean abundances for 1920-24 (theirTable 1, originally published in AUERBACH et al. 1924, 1926) and for 1952-62 (their Tables 3-15). During 1952/62, up to 6 stations within the Uberlinger See were sampled (MUCKLE & MUCKLE-ROTTENGATTER 1976) with a Nansen closing net with mesh sizes of 130 and 200 !lID. AUERBACH et al. (1924,26) used a Nansen closing net with mesh sizes of50 !lID and sampled at the Uberlinger See, but also in the main basin ofUpperLake Constance (Obersee). The taxonomic resolution used in 1920124 and 1979/95 is identical. The biomass of three cyclopoid copepods species (Cyclops abyssorum, Mesocyclops leuckartii, and Cyclops vicinus) distinguished during 1952/62 (MUCKLE & MUCKLE-ROTTENGATTER 1976) was aggregated for the present study. The biomasses ofDiaphanosoma brachyurum andHeterocope borealis, which disappeared in later years, were calculated by assuming carbon weights of 2.5 /-4SC ind- I for Diaphanosoma and 2.7, 5,13, 19/-4SC ind- I for Heterocope inApril, May, June, and July-December, respectively (ELSTER 1936). The biomasses ofothertaxa in 1920124 and 1952/ 62 were calculated from abundances and species-specific average carbon weights obtained in 1979/95 assuming that average carbon weights have not changed between the study periods. 258 D. Straile and W. Geller This will overestimate biomass during the more oligotrophic periods ifthe average body size of individuals and consequently their biomass has increased during eutrophication. However, the uncertainty involved in the calculation of biomasses is probably small compared to the uncertainties regarding the differences in sampling