With the increasing development and extraction of natural resources starting in the 1970s, Canadian freshwater ecosystems were not exempt from degradation recorded at the regional scales (
Christensen et al. 2006;
Kreutzweiser et al. 2013). Among the most common perturbations, hydrological disturbance due to reservoir impoundment, watershed land-use and deforestation, eutrophication and acidic deposition, as well as species invasion have been identified as major stressors of boreal ecosystems (
Keller 2009). However, understanding the individual and combined impacts of such stressors on the ecological integrity and function of boreal lakes is a complex issue due to region specific responses and complex interactions between natural factors and climate (
Schindler 2001). This section provides key case studies using zooplankton as an effective bioindicator to track the effects of major anthropogenic perturbations on boreal lakes and reservoirs.
Reservoir impoundment
Dams and impoundments of rivers created for generation of hydroelectric power had profound impacts on freshwater environments, ecological processes and biota (
Marty et al. 2004). The creation of large reservoirs in northern Québec in the 1970s and their long-term (7-years) ecological monitoring have provided a unique opportunity to assess the impacts of impoundments on zooplankton communities in boreal lakes and rivers (
Pinel-Alloul et al. 1982,
1989). Impoundment of reservoirs has been associated with an increase in zooplankton biomass and changes in community composition in both inundated rivers and lakes. The flooding phase induced a trophic upsurge of the whole planktonic food-web due to the release of nutrients and labile organic matter from inundated soils and vegetation, as observed in other northern reservoirs in Canada (
Duthie and Ostrofsky 1975). Variation in the intensity and length of the trophic upsurge depends on the initial trophic state of the flooded ecosystem but also on landscape characteristics such as soil thickness, type of vegetation and its degradability. In most boreal ecosystems, a return to baseline conditions was observed after 15 to 20 years (
Marty et al. 2004).
Overall, zooplankton species assemblages in reservoirs were similar to those observed in natural lakes and rivers of the James Bay area (
Pinel-Alloul et al. 1979). However, the trophic upsurge induced by flooding constitutes a bottom-up forcing that increases nutrients, and in turn phytoplankton and zooplankton biomass (Table S1
1). Zooplankton biomass was 2–6-fold higher in the new reservoir LG-2 compared to reference lakes (
Pinel-Allout and Méthot 1984). After flooding, cladocerans and rotifers responded rapidly to the trophic upsurge, since they are generally more adapted to variable hydrological characteristics of reservoirs (
Marty et al. 2004). Thus, zooplankton communities within reservoirs are usually dominated by fast-growing organisms (r-strategist species) such as cladocerans and rotifers compared to copepods that have a longer development time (K-strategist species). Increase in zooplankton biomass was related to higher temperature, nutrients and algal biomass in flooded lake sites, and to longer water residence time in flooded river sites (
Méthot and Pinel-Alloul 1987;
Pinel-Alloul et al. 1989). Overall, water residence time was the best predicting variable of zooplankton biomass, followed by water temperature, total phosphorus, chlorophyll
a, and turbidity (
Marty et al. 2004). The gradual increase in zooplankton biomass lasted for the first 4–5 years after impoundment. At short-term, there was no decline in productivity as nutrient concentrations remained stable.
A comparative analysis among reservoirs in northern Québec provided a more comprehensive perspective of long-term changes in zooplankton community in new and older reservoirs, compared to reference lakes (
Marty et al. 2004). This analysis based on the estimation of limnoplankton biomass (zooplankton and large algae > 53 µm) indicated a dampening in bottom-up forcing with years as older reservoirs supported lower limnoplankton biomass than new reservoirs. In general, zooplankton communities benefited from the large inputs of detritus and algal resources in new impoundments, compared to river and lake environments. However, the bottom-up forcing lasted only for less than a decade (5–10 years) as older reservoirs supported similar zooplankton communities and biomass as the reference lakes.
Acidic deposition
The acidification of lake ecosystems is a complex process as it may be caused by natural characteristics of watersheds (granitic rock geology, drainage of organic acids from wetlands) or by anthropogenic sources (dry and wet acidic atmospheric depositions due to mining emissions). During the 1980s, acidification of lakes emerged as a major concern in the Boreal Canadian Shield, and large-scale lake surveys were conducted in northeastern Ontario (
Locke and Sprules 1994;
Yan et al. 1996,
2008;
Keller 2009) and southern Québec (
Pinel-Alloul et al. 1990,
1995). Zooplankton has been shown to be an excellent indicator of water acidification which results in the decline of species richness and changes in species assemblages (
Brett 1989). Acidic lakes (pH < 5.5) were poor in species and characterized by acid-tolerant rotifers (
Keratella taurocephala), and small cladocerans (
Bosmina sp.,
Diaphanosoma sp.) and copepod calanoids (
Leptodiaptomus minutus); they lacked acid-sensitive species such as large cladocerans (
Daphnia) and some other copepods (
Epischura lacustris,
Skistodiaptomus oregonensis,
Diacyclops bicuspidatus thomasi) (
Pinel-Alloul et al. 1990;
Valois et al. 2011). Spatial patterns in zooplankton community structure revealed strong gradients related to natural and anthropogenic acidification. However, the response of zooplankton to recent declines in the rate of acidic deposition from atmospheric sources was confounded by interactions with other factors such as climate, watershed geology, lake morphology, and predator and invasive species (
Pinel-Alloul et al. 1995;
Christensen et al. 2006;
Yan et al. 2008;
Anas et al. 2017;
Couture et al. 2021) (Table S1
1).
Lake recovery from acidification and other multiple stressors remains an important issue for the ecological integrity of boreal lakes (
Valois et al. 2011). After the reduction in Canadian sulfur emissions and the Canada–US air quality agreement, liming experiments were conducted to accelerate lake recovery (
Yan et al. 1996). Although results were encouraging, recovery success varied among lakes and crustacean groups. Crustacean species richness increased by 2-folds after neutralization, but copepods recovered more rapidly than cladocerans (
Yan et al. 2004). The severity of past acidification, water conditions, and biological resistance to colonisation affected lake recovery (
Arnott et al. 2001). For instance, zooplankton community in more acidic lakes (pH 4.5) did not completely recover 15 years after neutralization whereas recovery was completed within 10 years of liming in less acidic lakes (pH 5.7) (
Yan et al. 1996). Paleolimnological studies showed that cladoceran assemblages were more strongly affected by acidification in clear-water lakes with low dissolved organic carbon content than in naturally acidic and dystrophic lakes with high dissolved organic carbon content (
Korosi and Smol 2012). Another important threat for crustacean zooplankton in softwater boreal lakes of Canada is the decline of calcium concentrations consequent to reduced acidic deposition and lower inputs of calcium from catchment (
Jeziorski et al. 2008). With a decline of calcium concentrations near or below 1.5 mg·L
–1, keystone herbivores such as
Daphnia with high calcium demand for moulting might be replaced by the large gelatinous cladoceran
Holopedium gibberum, a Ca-poor competitor adapted to low-calcium and brown-water lakes, and less vulnerable to invertebrate and fish predation. This widespread process, known as jellification of lakes (
Jeziorski et al. 2015), has also been shown to be associated with climate warming. It has the potential to negatively impact planktonic food webs, and the functions and services provided by boreal lakes throughout the Northern Hemisphere (
Weyhenmeyer et al. 2019).
A recent review (
Lacoul et al. 2011) assessing the biological consequences of surface water acidification in the Atlantic Provinces of Canada reported that, despite a decreasing trend in sulfate deposition since the 1990s, recovery in base cations (calcium, magnesium) and pH are not yet being reached due to acid-shock episodes in response to rapid snowmelt during early spring or rain-fall events. Similar results were reported in a large-scale study of 73 lakes in southern Québec based on the comparison of two years (1982–2017) (
Couture et al. 2021). Despite a sharp temporal 3-fold decrease in lake sulfate concentration, only a slight rise in pH and a small drop in calcium were recorded. The 2017 lake survey in southern Québec still supported a multiple forcing of spatial patterns in zooplankton community by large-scale gradients in climatic conditions, lake exposure to acidic deposition, and local-scale changes in lake water quality and morphology, as observed 35 years ago in 1982 (
Fig. 6). No major changes in zooplankton assemblages were detected over time since the 1980s. Segregation of zooplankton species with respect to acidification indicators (calcium, pH) in southern Québec in both 1982 (
Pinel-Alloul et al. 1990) and 2017 (
Couture et al. 2021) reflects the classification established by
Lacoul et al. (2011) across eastern Canada. Previous studies indicated that critical thresholds of pH (4.5 to 6) and calcium (Ca) (1.5–2 mg·L
−1) in Boreal Shield lakes might have detrimental effects on zooplankton diversity and success of key and vulnerable species such as
Daphnia (
Yan et al. 1996;
Holt and Yan 2003;
Jeziorski et al. 2008,
2015). The large-scale survey of lakes conducted in 2017 in southern Québec further highlighted the importance of calcium thresholds in supporting key zooplankton attributes such as the diversity and the abundance of crustaceans (
Couture et al. 2021). In contrast no clear pH threshold could be identified for either zooplankton diversity or abundance. A recent analysis of the multi-trophic responses of lake plankton food webs (phytoplankton, zooplankton and fish) to environmental gradients in southern Québec highlighted the complex ecosystemic responses of lakes to large-scale spatial heterogeneity in climate, lake morphometry and water quality (
Taranu et al. 2021). In the Boreal Plain and Taiga Shield ecozones in northeast Alberta, zooplankton metacommunity structure was also under the multiple control of natural abiotic (water chemistry, lake depth) and anthropogenic (industrial emissions) factors, as well as biotic factors related to predation regime and species invasion (
Bythotrephes) (
Yan et al. 2008;
Jeziorski et al. 2015;
Anas et al. 2017).
Watershed deforestation and clear-cut harvesting
Following the 1988 Yellowstone’s fires, many studies documented the impacts of natural wildfires on terrestrial and aquatic ecosystems (
Turner et al. 2003). Clear-cut logging and wildfires are the most important perturbations of the boreal forest (
Canadian Forest Service 1998), but their respective effects on lakes have remained largely unknown until the 1990s. Because of the importance of boreal forests to the Canadian economy and the need for new legislation favourable to the protection of ecosystems, the Canadian forestry industry partnered with research centers and created the Sustainable Forest Management Network (SFMN) in 1995 to address issues related to sustainable forest management practices (
University of Alberta 2020). The SFMN comparative study remains the only one that has compared the impact of logging to those of wildfires at the scale of watersheds, thus providing a framework allowing one to compare the impacts of human-made catchment-scale perturbations (logging) to those of a natural, unpredictable, uncontrollable perturbation (wildfire). However, comparing both types of perturbation on comparable terms proved difficult because wildfires generally affected greater surface areas of catchments (91% on average) relative to logging (47% on average). Since 1995, large-scale studies using comparative, BACI (Before and After Controlled Impact), and paleolimnological approaches were conducted in Québec, Ontario and Alberta (
Carignan and Steedman 2000).
Because of its trophic position between primary producers (phytoplankton) and secondary consumers (planktivorous fish), zooplankton have been considered sensitive to changes originating at either end of the trophic food chain (
McQueen et al. 1986). Since changes in nutrient availability or in the fish community structure can originate from the alteration of a lake’s catchment, zooplankton has the potential to serve as an indicator of catchment-scale perturbations (
Schindler 1987). The effects of natural fires and forest harvesting on zooplankton communities have been evaluated in the Boreal Shield lakes of Québec (
Patoine et al. 2000,
2002a,
2002b) and Ontario (
Lévesque et al. 2017), and in the Boreal Plain lakes of Alberta (
Prepas et al. 2001) (
Fig. 7). The comparison of ecosystemic responses of Boreal Shield and Boreal Plain lakes to wildfire and clearcut logging disturbances served to develop better forestry management (
Pinel-Alloul et al. 2002;
Prepas et al. 2003).
The impacts of logging and wildfires on zooplankton communities were examined in a large-scale comparative experiment set in the Canadian Boreal Shield forest, in the province of Québec (
Patoine et al. 2000). Thirty-eight (38) lakes were selected based on their similarity in terms of depth, surface area and drainage ratio. They included 20 lakes whose catchments have not been perturbed during the 80 years before the study, 9 lakes whose catchments have been logged in 1995, one year prior to the beginning of the study, and 9 lakes whose catchments have been burnt by wildfires, also one year before the first year of sampling. The impacts of logging and wildfires on water quality, plankton and fish were examined using the same set of lakes. Impacts differed quantitatively and qualitatively between logging and burning for water quality and phytoplankton community attributes. Quantitatively, the concentrations of some nutrients and chlorophyll
a were positively related to indicators of catchment forest loss. Since forest loss in the burnt group was on average greater than in the logged group, nutrient and chlorophyll
a increases were generally greater in the burnt group than in the logged group of lakes. Logging effect was characterized by higher concentrations of dissolved organic carbon (DOC), unnoticed after wildfires (
Carignan et al. 2000). The resulting decrease in light penetration in the logged group of lakes explained why phytoplankton did not increase as much in the logged group (∼+40% Chl-
a relative to the reference group in the first year) relative to the burnt group (∼+60%), despite similar increases in nutrient loadings in both logged and burnt groups (
Planas et al. 2000).
The impacts on zooplankton community structure were monitored for up to three years after wildfire and logging disturbances using different attributes: zooplankton and limnoplankton biomass (
Patoine et al. 2000), species richness and assemblages (
Patoine et al. 2002a), and crustacean biomass size structure (
Patoine et al. 2002b). Among these attributes, the biomass size structure of the crustaceans appeared to be the most sensitive, revealing a significant biovolume increase of larger-sized organisms (>1000 µm, mainly large cladocerans and cyclopoid copepods) in the “burned-lakes” group during two years following wildfires (
Patoine et al. 2002b) (Table S1
1). This was hypothetically attributed to an increased availability of calcium in the burned group. Indeed, as calcium can be a limiting element essential to the carapace of
Daphnia (
Hessen et al. 2000), its increased availability in the burned group could have favored increased growth of large species of cladocerans such as
Daphnia. Limnoplankton biomass also indicated a gradual increase in the biomass of a small size fraction (100–200 µm, mainly rotifers and algae) in the burned-lakes group during the first and second years, but not the third year after perturbations (
Patoine et al. 2000). In the logged lakes, there was no detectable variation in the biomass of rotifers, cladocerans and cyclopoid copepods, but the biomass of calanoid copepods decreased significantly (
Patoine et al. 2000). There were no significant changes in zooplankton species richness and composition one year after wildfire and logging disturbances (Patoine et al.
2000a) (Table S1
1).
Overall, the impacts of watershed perturbation by logging or wildfires seem stronger at the base of the food web (nutrients, phytoplankton), intermediate at the zooplankton community level, and weaker at the fish community. The increase in rotifer biomass in the burnt group of lakes can hypothetically be attributed to a bottom-up effect (trophic upsurge) (
Fig. 7) where increased nutrients inputs from ashes deposited in the watershed at spring lead to increased biomass of algae grazable by rotifers. In contrast, the decrease of calanoids in the logged group of lakes could result from their grazing activity being negatively impacted by decreased light availability, a consequence of increased concentrations of dissolved organic carbon observed specifically in the logged group of lakes. Differences in fish abundance and composition were not detected among the reference, logged, and burnt groups of lakes (
St-Onge and Magnan 2000).
As these results based on comparative studies might be affected by the potential confounding effect of uncontrolled variability in lake and watershed characteristics on the observed among-group differences (logging vs wildfires), further studies testing the impacts of watershed clearcut logging adopted a before-after comparison impact (BACI) design that controls for natural year-to-year changes in climate conditions (
Winkler et al. 2009;
Lévesque et al. 2017). These BACI studies conducted in the boreal forest have confirmed that long-term variations in zooplankton abundance and composition in Boreal Shield lakes were more related to variations in climatic or limnetic conditions than to short-term effects of logging activities (
Lévesque et al. 2017). In addition, comparative studies indicated that the impacts were minor when clearcut logging was limited to less than 40% of the catchment’s surface or in lakes with small drainage ratio (watershed area/lake area ratio < 4) (
Pinel-Alloul et al. 2002;
Prepas et al. 2003;
Deininger et al. 2019). Compared to the Boreal Shield lakes, the zooplankton of Boreal Plain lakes seems more sensitive to logging, as
Prepas et al. (2001) have reported significant decreases of cladocerans and calanoids in logged lakes relative to reference lakes (attributed to an increase of inedible cyanobacteria algae subsequent to increased nutrient exports from the catchment). Overall, the studies on lakes of the boreal ecozones support the Multiple Forces hypothesis showing greater influence of long-term abiotic and climate forcing than short-term effect of wildfire and clearcut logging watershed disturbances (
Fig. 7).
After the 1988 Yellowstone fires studies (
Turner et al. 2003), concerns regarding climate changes and one of its immediate consequence (increased fire frequency) have contributed to maintaining an interest in the ecological consequences of wildfires. The most recent reviews suggest that the response of lakes to wildfires may be weaker but longer-lasting than that of rivers (
McCullough et al. 2019) and that the environmental setting of the system studied (its type, size, location) plays an important role in the observed response, or lack thereof (
Robinne et al. 2020). Paleoecological studies suggest that the impact of wildfires on zooplankton species composition may only become apparent at decadal time scales, at least in lakes of the Canadian West Coast (
Bredesen et al. 2002).