Cet article vise à évaluer les effets à moyen terme des feux et des coupes de forêt sur la biomasse, la richesse spécifique et la composition du zooplancton des lacs de la forêt boréale au Québec. L'étude a été réalisée durant trois ans (1996-1998) après les perturbations sur 3 sous-ensembles de lacs similaires au niveau de leur morphométrie, incluant 7 lacs naturels non perturbés sur leur bassin versant, 7 lacs affectés par des coupes sur plus de 43% de la surface du bassin versant et 7 lacs affectés par des feux sur plus de 50% de la surface du bassin versant. Notre étude indique que le zooplancton a une forte résilience aux perturbations des lacs de la forêt boréale par les feux et les coupes. Les assemblages d'espèces sont très stables et varient peu entre les lacs naturels et les lacs perturbés. Les Rotifères sont le seul groupe affecté par les perturbations; leur richesse spécifique et leur biomasse sont plus fortes dans les lacs perturbés que dans les lacs naturels. Les feux ont un effet d'eutrophisation plus marqué que les coupes. Dans les lacs affectés par les feux, les apports accrus de nutriments favorisent la croissance du phytoplancton puis des rotifères. Toutefois, dans les lacs affectés par les coupes, la forte couleur de l'eau due aux apports accrus en carbone organique dissous est un facteur limitant, et la hausse du phytoplancton et des rotifères est moins accentuée. La richesse spécifique et la biomasse des Crustacés (Cladocères, Copépodes Calanoïdes et Cyclopoïdes) ne varient pas avec les perturbations. Sur l'ensemble du suivi, le niveau de perturbation n'est pas fortement relié aux changements dans la biomasse des groupes taxinomiques qui s'expliquent plutôt par les variations annuelles de la température de l'eau et des concentrations en chlorophylle a.
This paper aims to evaluate mid-term impacts of forest wildfires and harvesting on the biomass, species richness and assemblages of zooplankton communities in boreal lakes in Québec. The research was carried out during a post-perturbation 3-year survey (1996-1998) on three subsets of lakes (natural, cut, burnt lakes). It complements previous short-term studies conducted 1 year after perturbations in the same region by PATOINE et al. (2000, 2002a) on different subsets of lakes, which failed to detect significant effects. This comparative study attempts to minimize the spatial variability among natural and perturbed lakes by selecting lakes with similar morphometric conditions. For this selection, we applied a multivariate approach developed by DUHAIME and PINEL-ALLOUL (2005). Making sure that natural and perturbed lakes have similar morphometric features allows minimization of the confounding effects of lake and watershed conditions on organic carbon and nutrient exports from perturbed watersheds, and subsequently on lake water quality and biota.The initial 3-year survey was conducted on 31 lakes including 15 natural lakes without any perturbation in their watershed over the past 70 years, and 9 burnt lakes with more than 50% of their watershed intensively burnt, and 7 logged lakes with more than 43% of their watershed clear-cut. Lakes were visited three times per summer (May-June, July, August-September) between 6:00 to 15:00h, during the 3 years (1996-98) following the fires and harvesting of 1995. Watershed and lake morphometry, and water quality, and algal communities were monitored and studied by CARIGNAN et al. (2000) and PLANAS et al. (2000). Fish communities were sampled and studied by SAINT-ONGE and MAGNAN (2000). Zooplankton was sampled using a cantilever net (53 µm mesh size) on vertical hauls from 1 m above the sediments to the lake surface. Sampling methods and analysis of zooplankton were previously described by PATOINE et al. (2000, 2002a). We estimated species richness and assemblages, as well as the biomass of each zooplankton group (Rotifera, Cladocera, Copepoda Cyclopoida and Calanoida) and of the total community in each year and each lake. Simple ANOVA were used to evaluate global changes over the mid-term survey in zooplankton attributes among the three subsets of lakes (natural, cut, burnt). RMANOVA (analysis of variance with repeated measures) were used to test the effects of watershed conditions related to perturbations (natural, cut, burnt lakes), the effects of annual variations during the survey and their interaction. Redundancy analysis (RDA) served to evaluate the relationships between changes in the biomass of zooplankton groups during the survey in the three subsets of lakes and the changes in environmental factors and perturbation intensity. Environmental factors retained for the analysis were the drainage ratio (ratio of watershed area to lake area), lake water transparency, concentrations of total phosphorus, total nitrogen and dissolved organic carbon, water temperature in the euphotic zone, chlorophyll a concentrations, and biomass classes of dominant fish (yellow perch, white sucker). The perturbation intensity, expressed as the percentage of watershed area clear-cut or burnt, was also included in the model.Based on their similarity in morphometric variables (drainage ratio, lake volume, area and maximum depth, percentage of wetlands in the watershed), we retained for the study 7 cut lakes (C2, C9, C12, C23, C24, C29, C48), 7 burnt lakes (FBP9, FBP10, FP2, FP24, FP30, FP31, FP32), and 7 natural lakes (N5, N16, N56, N70, N82, N88, N109). Variations in morphometric characteristics were found within each subset of lakes (reference, burnt, logged), and we did not detect significant variation among groups. However, we did find significant differences in water quality variables among subsets of lakes. Thus, in our study, we have minimized the effects of variable morphometry among watersheds and lakes that could confound the responses of lakes to perturbations, while maintaining the same effects of fires and harvesting on water quality, as previously observed (CARIGNAN et al. 2000). This allowed us to test scaling-up effects of these changes on zooplankton communities.Our study shows that zooplankton communities in boreal lakes have strong resilience to environmental perturbations by wildfires and clear-cut logging. Species assemblages were very stable and varied little between natural and perturbed lakes. Some rotifer species such as Keratella taurocephala, known as a species adapted to acidic and humic lakes, had higher abundance in the cut lakes most enriched in dissolved organic carbon. Another rotifer species, Polyarthra vulgaris, adapted to eutrophic lake and reservoir ecosystems, was more abundant in the burnt lakes that were the most enriched in nutrients. Only short life-span and r-strategic species such as rotifers showed increases in species richness in the perturbed lakes compared to natural lakes. Compared to previous studies, our study showed that this increase in the number of rotifer species lasted for the 3-year survey. Rotifers also showed higher biomass in perturbed lakes (especially burnt lakes), but only during the first two years of the survey. In general, fires had greater eutrophication effects than did clear-cut harvest. In lakes impacted by fires, increases in nutrient inputs (total phosphorus, total nitrogen) favoured phytoplankton and rotifer growth, while in lakes impacted by logging, increases in dissolved organic carbon and water color limited the increase of phytoplankton and rotifers. For the crustaceans (Cladocera, Calanoida, Cyclopoida), neither species richness nor biomass was related to watershed perturbations. In general, our study indicates that natural sources of temporal variation between years had greater effects on the biomass of some zooplankton groups than did the perturbations. Overall, increases in Rotifera biomass were primarily related to higher mean summer water temperature whereas increases in Cyclopoida biomass were related to higher chlorophyll a concentrations. Perturbation intensity, nutrient concentrations, and drainage ratio also tended to increase the responses of these zooplankton groups, but their influence over the mid-term survey was not significant. A higher biomass of Calanoida was observed in lakes (mainly natural ones) with high water transparency and low nutrients. Variation in Cladocera biomass was not related to lake enrichment, higher temperature or perturbation intensity. Our mid-term study generally confirms the effects of wildfires and logging, as previously suggested, but not significantly detected by other studies (PATOINE et al. 2000, 2002a) conducted only during the first year after the perturbations. It shows the usefulness of mid-term monitoring to better assess the ecological impacts of natural and anthropogenic perturbations on aquatic ecosystems.
