Les peuplements zooplanctoniques et piscicoles de la retenue de Sainte-Croix (lac oligotrophe profond) ont été suivis de la mise en eau (1974) à 1986.La phase initiale de colonisation (1974-1975), caractérisée par le calanide Acanthodiaptomus denticornis et la dominance des espèces piscicoles peuplant antérieurement le Verdon lotique (Barbus fluvatilis, Leuciscus cephalus) a été de courte durée. Par la suite, les cladocères sont restés dominants avec cependant de profondes modifications dans la structure du peuplement. En 1977-78, Daphnia longispina était dominante tout au long de l'année (76 %). Plus récemment, cette espèce ne représentait plus que 34 %. des effectifs (1983-84), puis 24 % (1985-1986). La situation actuelle est caractérisée par la présence de Diaphanosoma brachyurum en été et l'abondance des petites formes (bosmines).L'ichtyofaune s'est elle aussi modifiée avec, notamment, une importante population d'ablette (Alburnus cluburnus), exploitant la zone pélagique. Son régime alimentaire est largement zooplanctonophage mais inclut mollusques et insectes lorsqu'elle est capturée en zone littorale.Cette augmentation de la pression de prédation a entraîné un net déplacement du spectre des tailles du zooplancton vers tes classes inférieures à 0,8 mm en 1986 (82 %) alors que les grandes formes ( > à 1 mm) représentaient 60 % des effectifs en 1977-78. Les densités de D. longispina et D. brachyurum décroissent rapidement entre 1983 et 1986.Les conséquences de cette prédation sélective sont discutées (évolution prévisible du système) ainsi que les causes pouvant expliquer l'absence de régulation efficace (densité-dépendance) dans un tel système prédateur-proie.
The aim of this study was to identify relationships between changes that occurred, from 1974 to 1986, both in the limnetic zooplankton and the fish communities of the large reservoir of Sainte-Croix (Southern France).This reservoir (figure 1) is characterized by a capacity of 767 hm3 and a long average renewal period (280 days). It was filled up in 1974. Changes in levels can attain 16 m, especially in winter. During summer, the lake is clearly stratified : surface mater temperatures are about 22-250°C whereas waters below the thermocline are cooler ( < 10° C) but well oxygenated (more than 50 % saturation). According to the mean annual total phosphorus (19 mg/m3) and chlorophyll a (ranging from 0.5 to 2 mg/m3), the lake is oligotrophic.The zooplankton was collected with a Clarke-Bumpus devise sampler (125 µm mesh-size) during both the 1974-1978 and the 1983-1984 periods. More recently (1985-1986), vertical hauls of zooplankton were made with a 50 µm mesh-size aperture net. Only relative abundance of large species (excluding small rotifers) have been taken into account for comparisons between different periods (figure 2). For the last four years (1983-1986), a set of data on filtration efficiency and comparison between sampling methods provides information concerning the evolution of annual mean densities of zooplankton species (figure 3).Successive inventories of the fish species were made in 1976, 1977, 1984 and 1987 at the same location with 15 mm, 27 mm and 50 mm mesh-size nets respectively. Figure 1 shows that abundant autochtonous species (Chondrostoma chondrostoma, Barbus fluviatilis) during the first period (1976-1977) are now relatively scarce, whereas the roach (Rutilus rutilus), an introduced species, is now largely dominant. Likewise, the relative abundance of the bleak, Alburnus alburnus, also increased, especially during the last period (1984-1987). Only this last species was caught by fine mesh-size nets in the pelagic epilimnion. Adults fed mostly on zooplankton but also took more molluscs in the littoral area (stomach analysis).In 1974, one year after an overflow, the calanoid Acanthodiaptomus denticornis represented the majority of zooplankton collected but it disappeared quickly. Since 1976, cladocerans are largely dominant. General evolutive trends show a decrease in relative abundance of Daphnia longispina from 1978 to 1986 whereas Diaphanosoma brachyurum have been recorded since 1983. Quantitative data (figure 3), available for the fast 4 years (1983-1986) show that the population densities of both these two large cladoceran and cyclopoïd species (mainly Macrocyclops albius) declined radically. On the other hand, small species such as Bosmina longirostris, Bosmina coregoni and Asplancha priodonta exhibited the same population densities throughout the fast period. Comparisons of phytoplankton community structure from 1982 to 1986 do not show signicative trends. Thus, variations in resource abundance cannot explain the long-term evolution of zooplankton communities.By comparing the mean annual length frequency distributions of zooplankton in 1977-78 and 1986 (figure 4), a marked decline of large bodied forms is seen. The individuals larger than 1.0 mm represented 60 % of total organisms in 1977-73 against 14.5 % in 1986. Mean annual lenghts were 0.8 mm in 1977-78 and 0.3 mm in 1986. These results are in good agreement with BROOKS and DODSON hypothesis that planktivorous fish can have a profound effect on zooplankton community taxonomics and size structure. They show that the control of structure and abundance of zooplankton by fish predation in new man-made lakes is a reality.Most fish are general predators. As demonstrated by MURDOCH and BENCE (1987), this can explain the instability of the coupled predator-prey system in the pelagic portion of Sainte-Croix reservoir because of : i) the capacity of the bleak to use fat reserves and to change its diet when large zooplankton prey density is too low, ii) the generation time of the predator is much longer than that of planktonic preys. Thus, the classical density-dependence mechanisms do not stabilize the predator-prey system and the main planktonic prey tend to disappear (D. longispina, in our case).Nevertheless, some potentially stabilizing mechanisms could interfere in the system before the local extinction of the prey : compensatory responses by the prey and significant changes in the fish community. There is reason to consider :1) The existence of refuges in space. Vertical size distribution of D. longispina (unpublished data) show that there is a strong vertical size gradient during daylight hours. Larger-bodied reproductive forms (longer than 1 mm) are only present in the deeper layers (20-30 m) whereas, during the night all the organises migrate to the superficial layer.2) The existence of refuge in time. In the past (1976-1978), D. longispina was dominant all year around. More recently (1983-1986), population densities were high only in spring (figure 5). From April to June, bleaks tend to assemble in the littoral zone for reproduction.3) An increase in the bleak mortality rate due to a recent development of large predator populations, such as Esox esox, could permit a reduction or, at least, a stabilization of the bleak population growth rate.
