Arch. Hydrobiol.

148

4

607-624

Stuttgart, July 2000

Recovery of Lake Geneva from eutrophication: quantitative response of phytoplankton Orlane Anneville 1 and Jean Pierre Pelletier With 7 figures

Abstract: During the seventies Lake Geneva (Switzerland-France) became eutrophie. Measures to reduce phosphorus inputs into the lake have been successful as the concentration of phosphorus has progressively decreased since 1981. This paper describes the long-term quantitative response of phytoplankton (primary production, total phytoplankton biomass, and chlorophyll-a) with regard to the decline in soluble reactive phosphorus (SRP) concentration. The annuaI means of phytoplankton parameters do not show the expected decrease. The long-term trends of annual means appeared to be the resultant of compensating trends which occurred in each season. SRP concentrations indicate that phytoplankton may be phosphorus limited only during summer which appears to be the most appropriate season to study the quantitative response of phytoplankton. From 1981 to 1992, the summer algal biomass and chlorophyll-a have been correlated to the late winter concentration of SRP. However, in the recent years, algal biomass and chlorophyll-a increased while phosphorus remained at low concentrations. This paradoxal increase in total biomass was due to the accumulation of inedible filamentous algae. The observed resilience of phytoplankton is a problem in terms of lake management. These results stress the importance of an appropriate time scale study and bring back into question the use of parameters such as primary production, total phytoplankton biomass, and chlorophyll-a for lake management.

Introduction Lake Geneva is a monomictic, deep lake on the French-Swiss border. It is one of the largest lakes in Western Europe (area = 582km2, Zmax = 309m). Lake Geneva has routinely been monitored by the CIPEL since 1957, (CIPEL: International committee for the protection of Lake Geneva). Its total phosphorus concentrations were about 10 JlgPIl prior to 1970. But the total phosphorus concentration increased during the seventies, and the lake was classified as 1 Authors' address: INRA, Station d'hydrobiologie lacustre, 75 avenue de Corzent, BP 511, 74203 Thonon les Bains Cedex, France; e-mail: [email protected] 0003-9136/00/0148-0607 $ 4.50 @ 2000 E. Schweizerbart'sche Verlagsbuchhandlung, D-70176 Stuttgart

608 Orlane Anneville and Jean Pierre Pelletier

eutrophie based on phosphorus concentration (ANNEVILLE & PELLETIER 1999). The total phosphorus concentration reached a maximal value of about 90/lgP/I in 1979, largely due to the increased human population in the catchment area, greater use of phosphate-based detergents and inadequate sewage treatment (RAPIN et al. 1989). This period of eutrophication produced large summer phytoplankton blooms that had a negative impact on the traditional use of the 1ake (BAL V A Y et al. 1984). Sewage began to be treated with iron compounds to remove phosphorus in the second half of the seventies and to reduce phosphorus inflow into the lake. These measures were successful and the total phosphorus concentration in the lake started to decrease at the beginning of the 1980s; it reached a value of 38 /lg/l in 1997 (BLANC et al. 1998). Long-term measurements of chlorophyll-a and total phytoplankton biomass were used to investigate the quantitative response of phytoplankton to the decrease in the lake phosphorus concentration. Since the physical and chemical properties of the lake, such as its morphology, a1kalinity and silica concentration, varied little compared to the phosphorus concentration, the 10ng-term data presented here provide interesting information about the biological mechanisms of reoligotrophication and allow comparison with other lakes. We have measured the long-term changes in phytoplankton and compared them to the change in soluble reactive phosphorus (SRP). The quantitative phytoplankton responses are the annual means of chlorophyll-a (ChIa), total phytoplankton biomass and net primary production (PPr). Since Lake Geneva undergoes the classical annual cycles of phytoplankton abundance, the longterm changes in chlorophyll-a concentrations and total phytoplankton biomass will be considered seasonally.

Materials and methods Chemical and biological parameters were measured once a month from 1974 to 1980 and twice a month from 1981 to 1997 at a site in the middle of the Upper Basin between Lausanne and Evian (Fig. 1). As this station is not directly influenced by local human activity, it was considered to be the most suitable sampling station for studying the trophic evolution of Lake Geneva (MONOD 1984). Soluble reactive phosphorus (SRP) was measured by acid molybdate (AFNOR 1982). Samples were taken at the subsurface and at 2.5, 5, 7.5 and 10 meters; depthintegrated means of phosphorus were calculated by averaging the results over the first 10 meters, where the phytoplankton activity is maximal. Depth-integrated means of chlorophyll-a were calculated over the first 10 meters by averaging samples taken at the subsurface and at l, 2, 3.5, 5, 7.5 and 10 meters. Cells were collected on a Whatman GF/C filter (47 mm) and sonicated. The pigments were extracted with 90 % (v/v) acetone/water and the solution was filtered through a

Recovery of Lake Geneva 609

Fig.l. Lake Geneva, location of the sampling site (SHL2).

glass fiber filter GF/C (25 mm). The chlorophyll-a concentration was measured by spectophotometry (STRICKLAND & PARSONS 1968).

Biomass was determined in water sampled in the 10 first meters with a Pelletier integrating apparatus (INRA patent, 1978). This device was immersed in the lake, where water entered into a sampling tube as the depth increased; the water sampling was govemed by the compression of the air in a bell-shaped reservoir, according to Mariotte's law. Phytoplankton were counted by the UTERMOHL (1958) technique and LOHMANN'S (1908) method was used to estimate the volumes of each species (for details see REV ACLIER 1979). Biomass was then estimated by adding the biovolumes of each species (cell volume x cell number). Plankton organisms less than 50llm long and 1O,0001lm3 were classified as nanoplankton and the rest as microplankton. Primary production was assessed by the incorporation of inorganic 14C (STEEMANNNIELSEN 1952) at nine depths from the surface to 20 meters. 14C-enriched sampIes were incubated in-situ in the middle of the day for a period equivalent to a third of the day length. The samples were filtered (0.8Ilm Millipore cellulose acetate filters), and the 14C assimilated was determined by liquid scintillation counting (Packard TRICARB 4430). As the available inorganic carbon concentration of the samples was known, the carbon uptake was calculated using the Steemann-Nielsen equation. The carbon uptake was then transformed to daily gross primary production. The net production was considered to be 60 % of the gross production (PELLETIER 1983). The data analysed included both annual and seasonal mean values. Seasonal means were calculated to study the long-term trends in the phytoplankton communities at each season. The beginning dates of the seasons were defined each year according to the phosphorus concentration and phytoplankton development. (i) The winter phytoplankton community was defined as the phytoplankton commu nit y that occurred from December to the start of phytoplankton growth. (ii) The spring phytoplankton community followed the winter community until the clear-water phase.

610 Orlane Anneville and Jean Pierre Pelletier (iii) The summer phytoplankton community was considered to take place after the clear-water phase until the start of replenishment of nutrients within the euphotic zone. (iv) The autumn phytoplankton community occurred from the end of summer to No vember.

Results Soluble reactive phosphorus (Fig. 2) The annual mean of total phosphorus integrated over the entire water column (Ptot) increased in the 1970s, then remained stable and started to decrease at the beginning of the 1980s (Fig. 2). The annual mean total phosphorus integrated over the entire water column appeared to be useful for evaluating the changes in phosphorus loading. However, it is not representative of the phosphorus available for phytoplankon growth because of the great depth of Lake Geneva. The change in the phosphorus concentration in the upper layer is directly linked to the phytoplankton activity. The dissolved phosphorus determined the quantity of phytoplankton and, in the short term, it was the residue left in the water after consumption by organisms. The annual mean concentrations of dissolved phosphorus were, therefore, dynamically linked with the phytoplankton and they are often used to de scribe the trophic evolution of lakes.

Fig.2. Long-term trend of the annual mean total phosphorus concentration averaged over the entire water column (0 to 309 m), the late winter soluble reactive phosphorus concentration averaged over the upper 10 meters and the annual mean soluble reactive phosphorus concentration averaged over the upper 10 meters.

Recovery of Lake Geneva

611 The annual mean soluble reactive phosphorus concentration (SRP) in the first 10 meters of Lake Geneva showed a similar trend to that of total phosphorus integrated over the entire column. The annual mean SRP concentration increased and reached a maximum in 1980 (38 IlgPn); they then decreased and are now about 10 IlgP/l. The late winter SRP in the upper layer (O-lOm) is of great importance because it is the stock available for subsequent phytoplankton growth. Its longterm evolution should, therefore, parallel the changes in chlorophyll-a and bio

Fig.3. Long-term seasonal changes in soluble reactive phosphorus (monthly mean) within the 10 upper meters of the water column.

612 Orlane Anneville and Jean Pierre Pelletier mass. Because there had been no complete mixing for about ten years, the win ter mixings that occurred in 1980 and 1981led to great nutrient enrichment of the upper layer. As a consequence, the late win ter SRP concentrations in the first 10 meters increased and were maximal in 1981 (75.8 ~gPIl), whereas the total phosphorus concentration had already started to decrease. The late win ter SRP concentration declined by about 50% between 1981 and 1991 and remained nearly constant thereafter. The SRP concentrations showed great seasonal variations. It was maximal during winter (60-66 ~gP/l between 1979 and 1981; around 20 ~gP/l between 1993 and 1996) because of low phosphorus uptake by phytoplankton and mixing, and low in summer: about 10 ~gP/l before 1980 and below 3 ~gP/l in 1997 (Fig. 3). The lowering of the thermocline and the increased mixing usually caused the SRP concentrations to increase in autuffill. Longer periods of SRP depletion occurred in later years, as a consequence of the decrease in late win ter SRP. According to SAS (1989), most species in a natural environment may be limited by phosphorus as the SRP concentration becomes less than 10 ~gP/l, and a severe reduction in algal growth may occur if SRP concentrations fall below 3 ~gP/l (SUTTLE et al. 1988, GROVER 1989). According to these definitions, phosphorus limitation occurred in Lake Geneva in the summer untill986. The period of phosphorus limitation has started earlier, during late spring, in recent years and persisted longer, until early autumn. Long-term response of phytoplankton Annual trends (Fig. 4) The phosphorus-chiorophyll-a relationship (Chla-P) is a useful concept in Lake Ecology. It has been calculated from studies on lakes and enclosures and has led to the development of many models that are used for lake management (VOLLENWEIDER 1968, DILLON & RIGLER 1974). These models are based on the concept that phytoplankton growth is limited by phosphorus and that, for a given concentration, the capacity of a lake to produce algae depends on the morphology of the lake (BARROIN 1999). The trophic indices used in these models are useful for determining the trophic state of a lake. The most popular classification is that of the OECD, which classifies lakes on the basis of their annual mean phosphorus and chiorophyll-a contents, and their transparency (OECD 1982). Thus a decrease in phosphorus concentration should reduce chiorophyll-a and result in a lower trophic state. But the long-term trends for the algal parameters in Lake Geneva did not support this. - The annual mean chiorophyll-a concentrations fluctuated without any sig nificant trend. The annual means in the 1980s were not lower than those during the period of maximal eutrophication.

Recovery of Lake Geneva 613

Fig.4. Long-term changes in chlorophyll-a, phytoplankton biomass and primary production in the upper layer in Lake Geneva (annual means). - The

annual mean biomass was exceptionally high in 1975 and 1976. The

algal communities were mainly made up of microplankton. Both the microplankton and nanoplankton increased with the increasing late win ter SRP concentration from 1977 to 1981. The phosphorus concentration started to decrease after 1981, but the annual mean biomass did not decrease until 1984. The decrease in annual mean biomass was mainly due to a drop in microplankton, as the nanoplankton biomass remained high. The increasing microplankton biomass since 1990 has led to a reverse trend, with increasing total biomass. - The annual mean net primary production was significantly correlated with the annual mean nanoplankton biomass (r = 0.55, P = 0.05), but there was

614 Orlane Anneville and Jean Pierre Pelletier no significant correlation between the annual mean net primary production and the annual mean total biomass (r = -0.16, P = 0.05). Net primary pro duction increased irregularly until199l, and has since decreased. Seasonal trends The phytoplankton dynamics appeared to show a seasonal pattern. Maximum biomass and chlorophyll-a concentrations occurred in spring and summer. Spring peaks occurred in May prior to 1987, but a non-parametric Mann-Whitney test (p = 0.05) states that chlorophyll-a and biomass maxima have generally appeared earlier (April) since 1987, and they may remain high until May (Fig. 5). The spring and summer periods of development were separated by a

Fig.5. Changes in chlorophyll-a and biomass during spring. White bars: monthly means over the period of high phosphorus concentration (1976-1986); grey bars: monthly means over the period of lower phosphorus concentration (1987-1997).

Recovery of Lake Geneva

615 short period with low chlorophyll-a concentrations, biomass and primary production: the c1ear water phase, generally in June. (i) The winter means of chlorophyll-a concentrations and biomass were very low in Lake Geneva, despite a high SRP concentration. (ii) The spring algal community was dominated by nanoplankton (Fig. 6). The total biomass varied from one year to the next, but the overall trend was upward. Chlorophyll-a concentrations increased until 1987 and then tended to drop, but the concentrations remained higher than during the previous years with high phosphorus concentrations. (iii) The contribution of nanoplankton to the total biomass in summer was low and varied from year to year (Fig. 6). The polynomial regression of the relative contribution of the nanoplankton showed an irregular upward trend (r2 = 0.5) from 1977 to 1991. The nanoplankton biomass increased

Fig.6. a) Long-term changes in seasonal phytoplankton (mean chlorophyll-a and biomass). b) relative contribution of nanophytoplankton and smooth curves obtained by polynomial regressions.

616 Orlane Anneville and Jean Pierre Pelletier from 1976 to 1991, whi1e the microplankton biomass decreased. A new trend appeared in 1992, the nanop1ankton biomass decreased and the microp1ankton biomass increased, although the late winter SRP concentration remained low. Despite such irregu1arities, the changes in the ch1orophyll-a concentrations and biomass may be divided ioto three periods. The chlorophyll-a concentrations and total biomass increased during the eutrophication period and tended to decrease during the second period from 1981 to 1991. These parameters were well corre1ated with the decrease in late win ter SRP concentration (the ch1orophyll-a concentration: r = 0.76, P = 0.05; total biomass: r = 0.89, p = 0.05). Finally, the renewed increase in microplankton biomass in 1992 caused increases in chlorophyll-a concentration and total biomass. (iv) The autumn a1ga1 population was essentially microp1ankton. The mean chlorophyll-a concentrations and total biomass changed much like the summer values. However, the autumn means of chlorophyll-a concentrations and total biomass had more irregu1ar fluctuations and a more delayed decrease, since they started to decrease in the midd1e of the 1980s and continued until 1991. ln spite of the low SRP concentration, the autumn mean chlorophyll-a concentrations and total biomass started to increase in 1992. The total biomass was surprisingly high in 1996 and 1997, due to the increase in microplankton. These long-term seasonal trends caused a great change in the period of the annual algal maximum in Lake Geneva. The summer mean chlorophyll-a concentrations were significantly greater than the spring values during the 1970s and the first half of the 1980s (Fig. 7). There was also a shift in the annual ch1orophyll maximum from summer to spring (p = 0.05, non-parametric MannWhitney test) that paralleled oligotrophication. Although there was a significant increase in spring biomass (p = 0.05), this trend was 1ess evident with the phytoplankton biomass. The data indicate that the spring a1ga1 biomass was as great as the summer algal biomass, except in the last three years when a high peak appeared in late summer and autumn. This discrepancy between chlorophyll-a and biomass trends was due to a shift in the phytoplankton composition during spring. Because of the appearance and the increase in a1gae with a low biomass but high chlorophyll-a, this trend was more obvious for the ch1orophyll-a concentrations than for the biomass.

Discussion Eutrophication has been extensively studied in order to manage prob1ems caused by increased algal growth in lakes. A quantitative approach has been used because the problem is high phytoplankton biomass. Most of these inves

Recovery of Lake Geneva 617

Fig.7. Spring (white bars) and summer (grey bars) phytoplankton parameters: a) mean chlorophyll-a, b) biomass. Boxplots show the overall shape of the spring and summer values during the period of high phosphorus concentration (1974 to 1986) and the period of lower phosphorus concentration (1987 to 1997). The central box shows the data between the "hinges" (quartiles), with the median represented by a line. "Whiskers" go out to the extremes of the data, and very extreme points are shown alone.

tigations were carried out to determine the factors responsible for algal proliferation. They measured aggregate variables such as primary production, algal biomass, and chlorophyll-a concentration. Many investigations have demonstrated that phosphorus is often the limiting nutrient in lakes and explains a large part of phytoplankton variations (SCHINDLER 1977, HECKY & KILHAM 1988). An increase in phosphorus concentration causes an increase in the quantity of phytoplankton. The relationship between P loading and chi orophyll-a concentration (VOLLENWEIDER 1968, 1975, 1976, DILLON & RIGLER

618 Orlane Anneville and Jean Pierre Pelletier 1974, LEE & JONES 1978) was used to classify lakes as oligotrophic or eutrophie, depending on the phosphorus concentration or phytoplankton parameters (OECD 1982). A decrease in phosphorus is thus expected to result in a decrease in phytoplankton biomass. There are many examples (Lake Washington, Mondsee) where phytoplankton biomass responded immediately to reduced nutrients (EDMONDSON & LEHMAN 1981, DOKULIL 1993). POLLI & SIMONA (1992) showed that annual primary productivity in Lake Lugano decreased within 10 years of re-oligotrophication. But little is known about how phytoplankton respond to re-oligotrophication. There are also many examples in which the phosphorus-phytoplankton relationship is not easily detectable. Several studies have shown that algal biomass, measured as chlorophyll-a or algal volume, vâries nonlinearly with phosphorus. It has been suggested that it reflects nonuniform growth and loss rates of one or more components of the total phytoplankton community (MCCAULLEY et al. 1989, WATSON et al. 1992). Studies on the re-oligotrophication of lakes suggest that the mesotrophic stage may not necessarily be distinguished from the eutrophie stage on the basis of annual production or mean phytoplankton biomass (GAMME TER & ZIMMERMANN 1998). Neither the annual mean values of chlorophyll-a concentration nor annual phytoplankton productivity have shown a consistent decline in Lake Constance, in spite of the decrease in phosphorus loading. There was a significant decline in chlorophyll concentration only in summer. The chlorophyll-a concentration decreased by at least 50 %, whereas the summer depth-integrated photosynthetic rates decreased by only 30 % because of reduced self-shading within the water column. It was suggested that grazing by zooplankton, sedimentation and changes in self-shading contributed considerably to the dampened annual response of phytoplankton (TILZER et al. 1991, HASE et al. 1998). Investigations on the first few years of re-oligotrophication in Lago Maggiore indicated that there was a resilience in the biotic response to decreased phosphorus (DE BERNARDI et al. 1988). The decline in total phosphorus was not paralleled by a comparable trend in chlorophyll-a concentration and the phytoplankton biomass remained quite stable for many years in spite of the reduction in phosphorus. The decline in average chlorophyll-a concentrations started in 19861988, as the maximum annual total phosphorus concentrations were falling to around 15/lg/l. Major biologie al changes have appeared since then. The total biovolume has decreased, but the significant decline occurred only in summer (MANCA et al. 1992, RUGGIU 1993, RUGGIU et al. 1998). The fluctuation in nutrient concentration has traditionally been viewed as the major structuring factor in lake ecosystem. An opposing view stated that biomass and productivity of a certain trophic level may be regulated by grazing from the level above (BENNDORF et al. 1984, SHAPIRO & WRIGHT 1984). Both forces (top-down and

Recovery of Lake Geneva

619 bottom-up) have been considered to be important for the food chain during recent years, so that the phytoplankton response to a decrease in phosphorus concentration depends on the trop hic structure (MCQUEEN et al. 1986). Total phosphorus has been a very poor indicator of maximum phytoplankton biomass during the re-oligotrophication of Lake Geneva. It has been c1aimed that the return from eutrophication was likely to involve hysteresis (GAWLER et al. 1988). It appeared that the damped responses of annual chI orophyll-a and total biomass were the cumulative results of compensating changes occurring in each season. Because the Chla/biomass ratio can vary greatly from one group to another, the chlorophyll-a and biomass trends show little discrepancy. Chlorophyll concentrations are maximum during summer and quite high in spring because of the presence of chlorophyll-rich species. As a consequence, the annual changes in chlorophyll-a concentration were mainly due to the changes in spring and summer values. ln contrast, the annual total biomass is mainly determined by summer and autumn values because of the dominance of species with high biomass. Annual means are useful for outlining the global trend of the system, but a seasonal study appears to be more suitable for understanding phytoplankton evolution. (i) Winter: the phytoplankton activity remains low and does not influence an nual means. The phytoplankton is not limited by phosphorus, as the SRP concentrations are high. Low irradiance and mixing are the main controlling factors in Lake Geneva, as in other lakes (REYNOLDS 1984, SOMMER et al. 1986). (ii) Spring: despite the late-winter decrease in SRP-concentration, the spring mean chlorophyll-a and biomass values increase and become higher than the summer values during re-oligotrophication. The early spring SRPconcentration is high enough to allow a strong development of phytoplankton. The factors controlling the start of phytoplankton growth and the intensity of the spring peak are not nutrients but external factors such as hydrological events and the stability of the water column. The influence of zooplankton on the intensity of the phytoplankton peak has not been studied, but the zooplankton appear to be a driving factor in Lake Geneva, as it is responsible for the dec1ine in phytoplankton and the appearance of the c1ear water phase (GA WLER et al. 1986). As the intensity of the spring algal blooms increased over the studied period, the increasing SRP uptake led to P depletion in the upper meters. The trend in SRP concentration in spring was thus controlled by both the late-winter decrease in SRP-concentration and the uptake by phytoplankton. As a consequence, P-limitation has not been restricted to the summer period but has extended to late spring since 1987 (Fig. 3). (iii) Summer: SRP concentration was always less than 10 f.lg/l. As a result, the maximum responses to the re-oligotrophication process occurred during

620 Orlane Anneville and Jean Pierre Pelletier this season when phosphorus may act as a limiting factor (SUTTLE et al. 1988, GROVER 1989, SAS 1989). The decrease in rnicroplankton biomass in the first decade of re-oligotrophication led to a decrease in phytoplankton. Summer chlorophyll-a and summer biomass showed a consistent downward trend that was significantly correlated with late winter SRP concentration. However, there has been a great increase in chlorophyll-a and total biomass in recent years, due to the increase in rnicroplankton biomass. These trends do not support the theoretical ChIa - P relationship and highlight the problem of understanding re-oligotrophication. The increase in these phytoplankton parameters was partly due to the increase in Dinobryon and the early strong development of a filamentous algae not edible by zooplankton (Mougeotia gracillima). Dinobryon has also appeared during the re-oligotrophication of Lake Constance. It has been argued that its mixotrophic properties are an advantage when nutrients are lirnited (GAEDKE 1998, KÜMMERLIN 1998). The accumulation of Mougeofia gracillima explains the observed high total biomass (PELLETIER et al. 1997). This species started to colonise the lake in the 1960s when phosphorus concentrations were about the same as now and Mougeotia gracillima declined during the maximum phosphorus concentrations. Like Dinobryon, Mougeotia gracillima may be more competitive for the present nutrient concentration. But it is impossible to say if its recent proliferation is linked to a drop in phosphorus or if it was caused by other factors, such as meteorological events, the observed water temperature increase, or the decrease in herbivorous biomass (BAL v A y 1997). Such results indicate that the present phosphorus concentrations are still high enough to allow the development of algae. Their morphological characteristics, which reduce the loss by predation, allow these algae to accumulate and produce a high biomass. However, species characteristic of oligotrophic water have reappeared and increased since 1986 (ANNEVILLE et al., in press). These species contribute little to the total biomass, but their increase is in accordance with the hypothesis of a trophic gradient (REYNOLDS 1984), with a decrease in phosphorus concentration resulting in a quantitative decrease in phytoplankton and a qualitative change in the phytoplankton community. (iv) Autumn; there is a poor relationship between the studied phytoplankton parameters and SRP concentrations when the lowering of the thermocline allows the enrichment of the euphotic zone. Very high biomass values due to the development of the non-edible conjugate Mougeotia gracillima and the xanthophyte Tribonema ambiguum have been observed in recent years (RE V ACLIER et al. 1998). This increase may be due to good meteorological conditions and warmer water temperatures. Thus phosphorus does not appear to be the only limiting factor for phytoplankton biomass. However, the decrease in phosphorus concentrations has a more obvious effect on

Recovery of Lake Geneva

621 the change in the dominant species and species composition than on total biomass (PELLETIER et al. 1997).

Conclusions The reduction in phytoplankton biomass is a key objective in lake restoration. This study has shown that the quantitative response of the phytoplankton is not always obviously correlated with phosphorus reduction. This may be viewed as a hysteresis process. However, summer appears to be the most appropriate period in which to examine the consequences of the decrease in phosphorus concentration on quantitative phytoplankton parameters. An interesting feature is the shift of the maximum chlorophyll-a concentration from summer to spring, which seems to indicate the beginning of re-oligotrophication (GAEDKE & SCHWEIZER 1993, SOMMER et al. 1986). However, the unexpected high summer and autumn phytoplankton development in recent years may raise several questions about the impact of meteorological events and the importance of top-down effects on reoligotrophication. This study illustrates that the non-linear changes in chlorophylla or biomass with phosphorus concentration are caused by changes in the phytoplankton composition, with the appearance of species with different properties, such as mixotrophy, and different growth and loss rates. The present work provides evidence that the response of chlorophyll-a and total biomass to a change in phosphorus concentration depends on the composition of the phytoplankton community and may be strongly influenced by the differences in loss rates among the populations. It seems that the qualitative aspects are more sensitive than the quantitative ones to changes in phosphorus concentrations as the number and the biomass of species which are considered to be characteristic of oligotrophic waters have increased in Lake Geneva (ANNEVILLE et al., in press). But it is difficult to say that there is a switch to more phosphorus efficient species that would allow equal production of phytoplankton biomass, as little is known about the ecological needs of the algae. If it is the case, we may question the meaning of quantitative parameters and their effectiveness for studying phytoplankton dynamics. Morphological and functional properties of algae may be useful for understanding the fluctuations observed in chlorophyll-a and total biomass. It might be more instructive to focus on the structure of phytoplankton communities, referring to their properties, in the future. This may allow a better understanding of the full response of the algal populations. Acknowledgements This work was supported by CIPEL, French Ministry of Environment and GIP HydroSystèmes. We thank NADINE ANGELI, VINCENT GINOT, ANNETTE BERARD and Ba

622 Orlane Anneville and Jean Pierre Pelletier

NYSTROM for reading the manuscript critically and OWEN PARKES for checking the English. References A.EN.O.R. (1982): Essais des eaux: Norme française, dosage des orthophosphates, des

polyphosphates et du phosphore total (méthode spectrométrique). 023, AFNOR. lOpp.

-

NFf 90-

ANNEVILLE, O., ANGELI, N., GINOT, V. & PELLETIER, J. P. (in press): Ambiguïté sur l'état trophique du Léman: vers un indice fondé sur les associations d'espèces. Les variables biologiques comme indicateurs de l'état de santé des Hydrosystèmes. - Ed. cemagref, 243 pp. ANNEVILLE, O.

& PELLETIER, J. P. (1999): Vers une amélioration de la qualité des eaux du Léman? Evolution comparée d'indices trophiques. - Actes du colloque "Découvrir

le Léman 100 ans après E-A. Forel". - Ed. Slatkine, pp. 215-227. BALVAY, G. (1997): Evolution du zooplancton du Léman. - Rapp. Comm. Int. Prot. Eaux Léman contre pollut., Campagne 1996, pp. 79-97. BALVAY, G., DRuART, J. C., PELLETIER, J. P., PONGRATZ, E. & REVACLIER, R. (1984):

Plancton - in le Léman, synthèse 1957-1982. CIPEL (Lausanne) 261-303.

BARROIN, G. (1999): Limnologie appliquée au traitement des lacs et des plans d'eau. Les études des agences de l'eau, 62, 215pp. BENNDORF, J., KNESCHKE, H., KOSSATZ, K. & PENZ, E. (1984): Manipulation of the

pelagic food web by stacking with predacious fishes. - Int. Rev. ges. Hydrobiol. 69: 407-428. BLANC, P., CORVI, c., KHIM-HEANG, S. & RAPIN, E (1998): Evolution physico-chimi que des eaux du Léman - Rapp. Comm. int. prot. eaux Léman contre poli ut., Cam pagne 1997, pp. 33-59. DE BERNARD!, R., GIUSSANI, G., MANCA, M. & RUGGIU, D. (1988): Long-term dy namics of plankton communities in Lago Maggiore (N. Italy). - Verh. Int. Verein. Limnol. 23: 729-733. DILLON, P. J. & RIGLER, E H. (1974): The phosphorus-chlorophyll relationship in

lakes. - Limnol. Oceanogr. 19: 767-773. DOKULIL, M. T. (1993): Long-term response of phytoplankton population dynamics to oligotrophication in Mondsee, Austria. - Verh. Int. Verein. Limnol. 25: 657-661. EDMONDSON, W. T. & LEHMAN, J. T. (1981): The effect of changes in the nu trient in

come on the condition of Lake Washington. - Limnol. Oceanogr. 26: 1-29. GAEDKE, U. (1998): Functional and taxonomical properties of the phytoplankton communit y of large and deep Lake Constance: Interannual variability and response to reoligotrophication (1979-1993). - Arch. Hydrobiol., Spec. Issues Advanc. Limnol. 53: 317-333. GAEDKE, U. & SCHWEIZER, A. (1993): The first decade of oligotrophication in Lake Constance. I. The response of phytoplankton biomass and cell size. - Oecologia 93: 268-275. GAMMETER, S. & ZIMMERMANN, U. (1998): Changes in phytoplankton productivity and composition during reoligotrophication in two Swiss lakes. Abstract, SIL Congress 1998, Dublin.

Recovery of Lake Geneva 623

GAWLER, M., BALVAY, G., BLANC, P., DRU ART, 1. C. & PELLETIER, 1. P. (1988): Plankton ecology of Lake Geneva: a test of the PEG-model. - Arch. Hydrobiol. 114: 161-174. GAWLER, M., BLANC, P., DRUART, J. C. & PELLETIER, J. P. (1986): Dynamique de quelques populations majeures du phytoplancton printanier du lac Léman en relation

avec le broutage et les sels nutritifs. - Coll. Nat. CNRS "Biologie des Populations", 285: 412-415. GROVER, 1. P. (1989): Phosphorus dependent growth kinetics of eleven species of

freshwater algae. - Limnol. Oceanogr. 34: 341-348. HAsE, c., GAEDKE, U., SEIFRIED, A., BEEsE, B. & TILZER, M. M. (1998): Phytoplankton response to re-oligotrophication in large and deep Lake Constance: Photosynthetic rates and chlorophyll concentrations. - Arch. Hydrobiol., Spec. Issues Advanc. Limnol. 53: 159-178. HEC KY , R. E. & KILHAM, P. (1988): Nutrient limitation of phytoplankton in freshwater and marine environments: A review of recent evidence on the effects of enrichment. Limnol. Oceanogr. 33: 796-822. KÜMMERLIN, R. E. (1998): Taxonomical response of the phytoplankton community of Upper Lake Constance (Bodensee-Obersee) to eutrophication and re-oligotrophication. -Arch. Hydrobiol., Spec. Issues Advanc. Limnol. 53: 109-117. LEE, G. F. & JONES, R. A. (1978): Eutrophication of water bodies: New insights for an age-old problem. - Environ. Sci. Technol. 12: 900-908. LOHMANN, H. (1908): Beziehung zwischen Biomasse und Stickstoffgehalt des Phyto planktons. - Schweiz. Z. Hydrol. 31: 110-127. MANCA, M., CALDERONI, A. & MOSELLO, R. (1992): Limnological research in Lago

Maggiore: studies on hydrochemistry and plankton. - Mem. Ist. Ital. Idrobiol. 50: 171-200. MCCAULEY, E., DOWNING, J. A. & WATSON, S. (1989): Sigmoid relationships between

nutrients and chlorophyll among lakes. - Cano J. Fish. Aquat. Sci. 46: 11711175. MCQUEEN, D., POST, J. & MILLS, E. (1986): Trophic relationships in freshwater pel agic ecosystems. - Cano J. Fish. Aquat. Sci. 43: 1571-1581.

MONOD, R. (1984): Evolution de l'état du Léman - Le Léman, synthèse 19571982. CIPEL (Lausanne), pp. 67-73. O.E.C.D. (1982): Eutrophisation des eaux: méthodes de surveillance, d'évaluation et

de lutte. - OCDE, Paris, 165 pp. PELLETIER, J. P. (1983): Mesure de la production primaire en milieu aquatique; problè

mes méthodologiques et éléments d'interprétation. - Rev. franç. sci. de l'eau 2: 339-366. PELLETIER, J. P., DRUART, J. C. & REVACLIER, R. (1997): Evolution du phytoplancton

du Léman. 1996,

-

Rapp. Comm. int. prot. eaux Léman contre pollut., Campagne

pp. 69-77. POLLI, B. & SIMONA, M. (1992): Qualitative and quantitative aspects of the evolution of the planktonic populations in Lake Lugano. - Aquat. Sci. 54: 303 -320. RAPIN, F., BLANC, P. & CORVI, C. (1989): Influence des apports sur le stock de phos

phore dans le lac Léman et sur son eutrophisation. - Rev. sci. de l'eau. 2: 721737. REvAcLIER, R. (1979): Examens biologiques des eaux du Léman. Etude du phyto plancton. - Rapp. Comm. int. Prot. eaux Léman contre pollut., Campagne 1978, pp.99-166.

624 Orlane Anneville and Jean Pierre Pelletier

REvAcLIER, R., DRU ART, J. C. & PELLETIER, J. P. (1998): Evolution du phytoplancton du Léman. - Rapp. Comm. int. prot. eaux Léman contre pollut., Campagne 1997, pp.7584.

REYNOLDS, C. S. (1984): The ecology of freshwater phytoplankton. - Cambridge Uni versity Press, 384 pp. RUGGIu, D. (1993): Osservazioni conclusive. Considerazioni generali sull' evoluzione a

lungo termine dei popolamenti planctonici. Fitoplancton. - ln: C.N.R. Istituto Italiano di Idrobiologia: Ricerche sull' evoluzione deI Lago Maggiore. Aspetti limnologici. Campagna 1992 e Rapporto quinquennale 1988-1992 (ed.), Commissione Intemazionale per la protezione delle acque italo-svizzere: 115-116. RUGGIU, D., MORABITo, G., PANZANI, P. & PUGNETTI, A. (1998): Trends and relations among basic phytoplankton characteristics in the course of the long-term oli-

gotrophication of Lake Maggiore (Italy). - Hydrobiologia 369/370: 243-257. SAS, H. (1989): Lake restoration by reduction of nutrient loading: Expectations, Exper

iences, Extrapolations. - Academia Verlag Richarz, St. Augustin FRG. 497 pp. SCHINDLER, D. W. (1977): The evolution of phosphorus limitation in lakes. - Science 195: 260-262. SHAPIRO, J. & WRIGHT, D. (1984): Lake restoration by manipulation: Round Lake, Minnesota, the first two years. - Freshwat. Biol. 14: 371-383. SOMMER, U., GLIWICZ, Z. M., LAMPERT, W. & DUNCAN, A. (1986): The PEG-model of seasonal succession of planktonic events in fresh waters. - Arch. Hydrobiol. (06: 433-471. STEEMANN NIELSEN, E. (1952): The use of radioactive carbon 14C for measuring or

ganic production in the sea. - J. Cons. Iut. Exploit. Mer. 18: 117-140. STRICKLAND, J. D. H. & PARSONS, T. R. (1968): A practical handbook of sea water analysis. - Bull. Fish. Res. Bd Canada. 167,311 pp. SUTTLE, C. A., STOCKNER, J. G., SHORTREAD, K. F. & HARRISON, P. 1. (1988): Time course of size fractionated phosphate uptake: Are larger cells better competitors for pulses of phosphate than smaller cells? - Oecologia 74: 571-576. TILZER, M. M., GAEDKE, U., SCHWEIZER, A., BEEsE, B. & WIESER, T. (1991): Interannual variability of phytoplankton productivity and related parameters in Lake Constance: no response to decreased phosphorus loading? - 1. Plankton Res. 13: 755777. UTERMOHL, H. (1958): Zur Vervollkommung der quantitativen Phytoplankton-Metho dik. - Mitt. Int. Verein. Limnol. 9: 38 pp. VOLLENWEIDER, R. A. (1968): Water management research. - OECD Paris. DAS/CSI/68.27. - (1975): Input-output models with special reference to the phosphorus loading con cept in limnology. - Schweiz. Z. Hydrol. 37: 53-84. - (1976): Advances in defining criticalloading levels of phosphorus in lake eutroph ication. - Mem. Ist. Ital. Idrobiol. 33: 53-83. WATSON, S., McCAULEY, E. & DOWNING, J. A. (1992): Sigmoid relationships between phosphorus, algal biomass, and algal community structure. - Cano J. Fish. Aquat. Sci. 49: 2605-2610. Submitted: 11 June 1999; accepted: 2 December 1999.