A simple meta-model was used to examine how climate warming and stresses due to other human activities might affect the productive capacity of fisheries in all of Canada’s lakes. Recent estimates of lake resource characteristics by secondary watershed and area size-class provided the basis for the model. Potential fishery productivity was estimated using a variant of the Schlesinger and Regier (1982) model which had lake mean depth, total dissolved solids concentration, and mean annual air temperature as inputs. A business-as-usual climate change scenario (SRES A2) was used to estimate worst case future temperature increases (4.5–8.3°C by the 2080s). The stress index from Chu et al. (2003) was used as a proxy for the impact on fisheries of other human activities. Projected populations for the SRES A2 scenario were used to scale future stress index levels. Potential biotic responses to warming were represented in two ways; the first as potential biotic displacement of currently dominant species when temperature rose beyond their preferred range and the second as potential biotic adaptation of other species, particularly in species rich areas, replacing displaced species. Potential productive capacity of fisheries in all Canadian lakes was 361,000 tonnes for the baseline climate norms period of 1961–1990. Climate warming increased productivity by 80.7% in the 2080s but stress reduced levels by 19.4% in the norms period and held the increase to 10.3% in the 2080s. Biotic displacement alone resulted in large decreases in productivity, by 65.2% in the 2080s and, when stress was added, by 79.5%. Biotic adaptation largely offset displacement. Applying stress and both biotic responses productivity was reduced by 31.4% in the 2080s from the unstressed norms baseline or 12% with stress added. Further investigations are needed to better establish the likely extent of stress impacts and potential biotic responses to climate warming in Canada’s lakes.
Canada, like the rest of the Earth, is undergoing substantial change as a result of cumulative human activities (Costanza et al., 2007). Freshwater ecosystems are especially vulnerable to degradation and destruction (Hails et al., 2006) and Canada holds the largest national share (37%) of global surface freshwater resources, primarily as lakes, an estimated 910,000 lakes with an area ≥ 0.1 km2 (10 ha) and 562 lakes with an area ≥ 100 km2(Minns et al., 2008). Besides the impacts arising from existing human activities, freshwater ecosystems are highly vulnerable to the effects arising from projected climate (Shuter et al., 2002). Canada’s lakes provide many ecological and human services (Postel and Carpenter, 1997) and support substantial commercial, recreational, aboriginal, and subsistence fisheries. As temperature is a primary environmental factor for freshwater fishes (Magnuson et al., 1979), climate warming may have important impacts on individual species and species assemblages (Shuter et al., 2002). Individual species may be displaced as environmental temperatures shift away from preferred ranges, causing decreases in productivity and distribution. Meanwhile, assemblages may facilitate an overall adaptation as new temperature conditions favour some species while allowing the spread of others. Given the rising pace of human activities and climate change is accelerating, their potential impacts on the potential fishery productivity, or productive capacity, of lakes need to be assessed with attention to the degree to which biotic displacement and adaptation might exacerbate or ameliorate those impacts.
The goals of this paper are to: 1) Estimate the productive capacity of Canadian lakes fisheries under a baseline climate (norms of the recent past, 1961–1990) and under projected future scenarios; 2) Use recent estimates of the degree of human-induced stress (Chu et al., 2003) as an index of damage and discount the productivity estimates accordingly using projected human population levels as a guide to future damage levels; and 3) Adjust the productivity estimates for potential biotic response of individual species (displacement) and species assemblages (adaptation) to temperature increases. A meta-modelling approach is used to represent the key processes that might be expected to impact the future fishery productivity of Canadian lakes as detailed, validated quantitative models of those processes have not been developed. Meta-models present a higher level of abstraction than models and their predictions should be considered as indicators of likely responses and not as exact forecasts.
Materials and methods
The meta-model for making projections of fishery productivity for all of Canada’s lakes has several components: 1) Measuring the characteristics of Canada’s lakes; 2) Mapping recent and projected air temperatures; 3) Selecting a model for predicting potential fish productivity in lakes; 4) Representing the stress effects of aggregate human activities on potential fishery productivity; and 5) Devising models to represent potential biotic displacement and adaptation effects as the climate warms. The first three components are grounded in established models and data. The last two components represent hypotheses about the impact of human activities and responses of biodiversity and render the whole a meta-model.
Lake resources in Canada
Minns et al. (2008) described the implementation of a framework for estimating key lake characteristics for all lakes across Canada by secondary watershed (Figure 1a; 164 watersheds) and in a semi-logarithmic series of ten lake area size classes with cut-offs at 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 50, and 100 km2. This framework called CLAM, Canadian Lakes Assessment Model, provides a basis for assessing the large-scale and cumulative consequences of environmental and other changes in Canada’s lakes. By watershed and lake size-class, the lake characteristics include number of lakes, maximum and mean depth, Secchi depth, total dissolved solids (TDS, mg l−1), and pH. Lakes with areas ≥ 100 km2 are included individually. While Minns et al. (2008) focused on lakes containing lake trout, Salvelinus namaycush, the regression models of lake characteristics with coefficients for lake size and watershed were used with the estimated proportions of lakes with lake trout present to estimate the mean lake characteristics in each watershed – size class category. Minns et al. (2008) also showed the secondary watersheds can be grouped by ecozone (Ecological Stratification Working Group, 1996) (Figure 1b). Ecozones are designated using a mixture of climatic, physiographic and vegetation characteristics of the landscape and map easily onto the secondary watershed spatial organization. While future climate change may significantly redraw the map of Canadian ecozones, the current ecozones provide a convenient, compact basis for examining broad regional patterns in ecosystem resources.
Climate norms and future scenarios
Mean annual air temperature (MAAT) in each secondary watershed was estimated by overlaying the 0.5° of latitude and longitude grid of 1961–90 norms climate data (New et al., 1999) on the secondary watersheds and computing watershed means. The norms data were obtained from the IPCC Data Distribution Centre (www.ipcc-ddc.cru.uea.ac.uk).
The IPPC Data Distribution centre also provides future climate scenario data on a 3.75° grid for various global climate models (GCMs). Predicted increases in MAAT over the simulated 1961–1990 norms period were assembled for the SRES A2 scenario which represents a business-as-usual scenario based on the Canadian GCM2 (Canadian Institute of Climate Studies, 2002). This scenario was selected as concerted global action to bring greenhouse gas emissions under control has yet to be initiated. Increases for three time stanzas were assembled: a) 2020s – average of 2011 to 2040, b) 2050s—2041 to 2070, and c) 2080s—2071 to 2100. Projected temperature increases over the simulated 1961–1990 norms period were assembled and used in conjunction with the observed norms values to estimate future temperatures by secondary watershed.
Productive capacity of lake fishery resources
Numerous efforts have been made to predict the potential fish yield in lakes (Leach et al., 1987), notably using Ryder’s (1965) morpho-edaphic index (MEI), which is the ratio of total dissolved solids (TDS mg L−1) and mean lake depth (metres). Here, a new version of Schlesinger and Regier’s (1982; S&R hereafter) sustained yield model:
The quadratic regression model was applied, by secondary watershed (SWS) and lake size class, using mean lake attributes for area, mean depth, and TDS and a correction for transformation bias. It was assumed that MEI, and hence TDS and mean depth, remain unchanged in future scenarios. Mean annual air temperatures for the 1961–1990 norms period, and for three future time stanzas under the A2, business-as-usual, scenario were used to estimate productivity for each time stanza. The estimated yield per unit area was multiplied by area and lake number in each SWS-lake size class combination. The resulting total yield estimates, a measure of productivity, were aggregated by ecozone.
Stress due to human activities
Impacts on aquatic resources due to other human activities were estimated using the composite stress index developed by Chu et al. (2003) for Canada’s freshwater resources by tertiary watershed. The stress index is a composite measure of a wide range of human development metrics including dwelling and road densities, land use proportions, and waste discharges. That index was used to infer cumulative impacts on lake resources to date, i.e., in the 1961–1990 norms period. The tertiary watershed stress index values (SI) were averaged by secondary watershed to match the spatial resolution of the lake resource data. By secondary watershed, the fishery productivity estimates (Y) were discounted using the stress index:
To predict how SI might change in each watershed using the Chu et al. (2003) approach would have required very detailed socioeconomic projections that were far beyond the scope of this study. Instead a simple geometric model was used to increase SI values in relation to projected overall increases in human population. To estimate the impact of expected future human population levels on stress levels in secondary watershed, the stress model also assumed that already highly stressed areas will experience greater increases in stress compared to areas that currently have low stress levels, using the following equation:Figure 2).
The choice of 0.5 as the pivot point was arbitrary and kept the geometric model as simple as possible. This model assumes that stress levels watersheds with low index values in the norm period will not increase very much and that stress levels cannot exceed 1. Given the maximum 1961–1990 norms SI was 0.49 (Table 2), the model implies an almost complete loss of fisheries production in the watershed with that value. The IPCC SRES A2 climate scenarios described above include human population projections (Lutz, 1996). Canada Census data was assembled to estimate the population in the 1961–1990 norms period (http://www12.statcan.ca/english/census/index.cfm). The ratio of the projected mean population in the future time periods to the mean 1961–1990 population was used to estimate the maximum stress index increases for each scenario assuming a maximum increase of 0.5 into the 2080s. The projected stress-discounted fishery productivities (equations 2 and 3) were computed by secondary watershed and summarized by ecozone.
Responses to increasing temperatures can occur at the species and assemblage levels (Shuter et al., 2002). At a species level, individual fish species generally have a preferred temperature with a typical tolerance of ± 2°C, giving a tolerance zone of 4°C (Magnuson et al., 1979). Increases in MAAT are assumed to lead to chronic increases in water temperatures and exposures. Wehrly et al. (2007) and others have shown that fish can tolerate short-term exposures. Beyond that tolerance zone, growth often decreases substantially and with large displacements the species will be extirpated. The individual species effect is termed here biotic displacement. At an assemblage level species have a range of thermal preferences. As biotic displacements occur, the productivity of some species may be expected to decrease while that of others increases. In Canada, coldwater species constitute a large portion of the assemblages found in the northern regions while cool- and warm-water species are more dominant in the southern regions where species numbers are greater overall. Under climate change, the ability of local fish communities to adapt to change might be expected to depend on the extent of the temperature change and the richness of the species pool as a source of replacement productivity. The assemblage response was termed here biotic adaptation.
These species and assemblage effects are represented here in a simple two part model. The first part of the model assumed current fish productivity is dominated locally by species with thermal preferences matched to local norms climate conditions. As air temperature increases over the norms baseline get larger, biotic displacement causes the proportion of potential yield realized (α) to decrease:Figure 3a). This mimicked the pattern of an individual species response to departures from its preferred temperature range (Magnuson et al. 1979).
The second part of the biotic model assumes that the extent to which the first part occurs depends on species richness. Chu et al. (2003) showed how fish species richness by tertiary watershed in Canada increases with mean annual air temperature and a similar pattern applies for secondary watersheds. The proportion of the potential productivity that adjusts to temperature increases (γ) was modelled here as a function of fish species richness (S) in the secondary watershed(Figure 3b):
The effects of biotic responses on potential fish yields were examined in stages by application of estimated adjustments by secondary watershed using equations 1, 2, 3, 4, and 6. First, the effect of biotic displacement is computed with (Y′· α) and without (Y· α) stress. Second, the effect of biotic adaptation is computed with (Y′· β) and without (Y· β) stress. The resulting production estimates were aggregated by ecozone.
Canadian lake resources
Lakes with areas < 100 km2 account for the vast majority of lakes numbers in Canada but lakes ≥ 100 km2 account for 44.5% of the total lake area which exceeds one million km2 (Table 1). Lakes are most abundant in the boreal and taiga shield ecozones followed by the southern and northern arctic ecozones. Similar distribution patterns are seen for small and large lakes.
Minns et al. (2008) showed that TDS did not vary significantly with lake size but did vary considerably among secondary watersheds with ecozone being a good predictor of observed levels. By secondary watershed, TDS values ranged from 1.9 to 2468.3 with a mean of 142.8 mg L−1 (Table 2). Ecozone means show how levels of TDS vary regionally from a low of 11.7 in the southern arctic to 807.4 on the prairies which include a saline lake region. Values in shield and maritime areas are generally lower than in other ecozone types. The secondary watershed summary of the tertiary watershed fish species occurrences reported in Chu et al. (2003) showed species richness varying from 2 to 138 species with a mean of 30 (Table 2). Across ecozones, mean species richness generally increased with increasing MAAT although the maritime ecozones had somewhat lower values whilst the mixedwood plains, which includes the lower Great Lakes, had a much higher value reflecting the major post-glacial invasion route for fish species into the Canadian interior (Mandrak and Crossman, 1992).
Mean annual air temperature in the norms period and projected increases
Across the 164 secondary watersheds, mean annual air temperature (MAAT, 1961–1990 norms period) ranged from −20.4 to 7.8 with a mean of −2.3°C (Table 2). Projected temperature increases (°C) varied somewhat among ecozones (Table 2) ranging from 1.3 to 2.6 by the 2020s, 2.9 to 5.2 by the 2050s, and 4.5 to 8.3 by the 2080s with the range increasing over time.
Potential lake fishery productivity
Reanalysis of total fish yield data from Schlesinger and Regier (1982) resulted in a regression model with mean annual air temperature (MAAT) and the modified morpho-edaphic index MEI (TDS/Min (Mean depth, 25.0) in a quadratic form:Figure 4). This new model was applied (with a mean error square bias correction of 1.245 was used when de-transforming the regression equation) to estimate total productive capacity. Sustained yield assumed to be equivalent to potential productivity.
The potential productivity of Canadian lakes was estimated at 361,000 tonnes. y−1 (Table 1). The majority came from lakes with areas ≥ 100.0 km2 (Table 1). Four ecozones accounted for most of the yield with the highest on the boreal shield (27%) followed by boreal plains (19%), mixedwood plains (19%) and taiga shield (12%). Yield from small lakes was greatest in the boreal and taiga shield ecozones where it exceeded the contribution of large lakes.
For the baseline norms period the secondary watershed stress index levels derived from Chu et al. (2003) ranged from 0.00 to 0.49 with a mean of 0.14 (Table 2). Lower values occurred in the northerly, colder ecozones and the highest value was found in the mixedwood plain ecozone where a large portion of Canada’s population lives in the urban sprawls of southern Ontario.
The population of Canada increased from 18.2 to 28.0 M from 1961 to 1991 and averaged 23.2 M for the climate norms period. Under the IPCC A2 scenario, Canada’s population is projected to increase from 30.2 M in 2000 to 40.7M in 2050 and to 60.9 M in 2100. The average populations projected for the three scenario periods, 2020s, 2050s, and 2080s, were 36.0, 42.5, and 53.6 M respectively. Assuming the 2080s population level will cause a maximum stress increase (Δ SI2080) of 0.5, raising the middle stress level from 0.5 to 1.0, maximum stress level increases of 0.211 and 0.318 were estimated for the 2020s and 2050s scenarios respectively. The mean stress index values continue to be higher in the southern, warmer ecozones of Canada where much of the human population is concentrated (Table 3). In the mixedwood plains region the mean stress index would rise from 0.38 in the norms period to 0.54 in the 2020s, 0.62 in the 2050s and 0.76 in the 2080s. In the colder ecozones only minor increases in the stress index were inferred.
Biotic displacement and adaptation
If only species presently dominating in different regions of Canada are considered, the modelled biotic displacement effects produce substantial losses of productivity potential as the magnitude of the MAAT increases grows (Table 3). By the 2020s, the proportion of productivity realized (α) ranged from 0.85 to 0.99. The levels would decline to 0.26 to 0.77 by the 2050s and to 0.05 to 0.38 by the 2080s. The species richness adjustment (γ) varied from 0.07 to 0.85, indicating the proportion of adaptation (Table 3) with the ecozone means ranging from 0.20 in the northern arctic to 0.82 in the mixedwood plains. Combining biotic displacement and adaptation moderates the loss of productivity due to an inferred inability of the current assemblage to adapt to climate change. The proportion of productivity realised (β) ranged from 0.90 to 0.99 in the 2020s, from 0.38 to 0.95 in the 2050s, and from 0.17 to 0.88 in the 2080s (Table 3). The ecozone means had narrower ranges of values.
Potential fishery productivity under climate change with various model options
Various combinations of the stress index, and the biotic displacement and adaptation indices were applied to the baseline estimates to assess their impact on potential productivity (Table 4). If temperature were the only changing determinant of future fish productivity, the climate warming projected in the A2 scenario led to productivity increases of 22.0, 46.1, and 80.7 % over the norms baseline in the 2020s, 2050s, and 2080s respectively. Stress due to human activities, driven both by population growth and increasing per capita resource use, caused the projected productivity to be reduced in the near-term and rise later in the 21st century as temperature-linked increases overtake stress-linked decreases. If the increased stress occurred without climate change, catastrophic decreases in productivity occur. If only biotic displacement occurs alongside warming, productivity decreases later in the century after an increase in the 2020s. By the 2080s, productivity declines by 63.2%. When increased stress and biotic displacement are both applied with warming, productivity is lowered by 42.9% in the 2050s and by 79.5% in the 2080s. When biotic adaptation was added to displacement without stress, productivity increased again but by more moderate amounts: 19.3, 21.2, and 22.2% in the 2020s, 2050s, and 2080s respectively. When all factors are applied productivity was lowered and stayed so into the 2080s with a decrease of 31.4%.
When the projected watershed fishery productivities in the 2080s were compared with the stress-adjusted estimates for the norms period, the percentage change had a humpbacked relationship with the norms MAAT (Figure 5). The greater losses in the colder far northern areas mostly arise from the biotic displacement effect with low species richness reducing the scope for biotic adaptation. The greater losses in the warmest southern areas are due to the rising stress effects associated with human population growth and activities overtaking the warming effects. The positive values were centred on northwest Ontario where a combination of lower baseline stress levels and higher species richness allow the greatest productivity increases (Figure 6). The greater losses in the Atlantic maritime regions were due in part to the lower species diversity there compared to other southern areas. Increased stress was the main cause of the greatest decreases in southern Ontario and central Alberta while the greatest losses in the Arctic regions arose from the very low species diversity.
The results obtained by meta-modelling here provide a preliminary indication of how climate change and other cumulative human-induced impacts might affect the productive capacity of fisheries in Canada’s many lakes. The meta-model used here was a mixture of quantitative and qualitative components. The quantitative components were mainly grounded in data and empirical model of data and included the extent of lake resources, the fishery yield model, and the climate norms and projections. The qualitative components were well-grounded in observations and understanding but lacked well-established empirical or process models to support them. The qualitative components included the stress index, and the two parts of the biotic response models. First, each component of the meta-model is scrutinized. Next, the results are placed in context with respect to actual fisheries yields from Canada’s freshwaters and to evidence of large-scale impacts of climate warming. Finally, future research needs are considered whereby better estimates of the climate warming impacts might be obtained.
The basic fishery productivity model is, admittedly, based on relatively simple variables but MEI-based models have been widely used around the world (Kerr and Ryder, 1988). Total dissolved solids concentration can be viewed as a surrogate for more precise variables such as ionic strength which at very low and very high levels can affect osmo-regulation in freshwater fish and nutrients like phosphorus which is the primary determinant of biological productivity in temperate freshwaters (cf Dillon et al., 2004). Climate strongly influences aquatic temperatures which drive biological processes from biochemical to ecosystem levels of organization (Regier et al., 1990).
The stress index (SI) based on Chu et al. (2003) was assumed to represent the aggregate effect of all human activities on fisheries productivity. If the Chu et al. (2003) approach had been used to develop future SI values using detailed projections linked to all it underlying components there can be little doubt that, without major changes in human attention to environmental protection, SI values would have increased as population increased with greater stress level increases occurring in urban areas. The geometric model used (equation 3) provided a qualitative mimic of those socioeconomic processes. Water quality problems (e.g., eutrophication, contaminants, and toxicants) and lower habitat quality conditions (e.g. infilling, fragmentation of watersheds, dewatering, destruction of wetlands) more typically occur in areas with higher human population densities (e.g. urban regions and road infrastructure) or intensive human activities (e.g. intensive agriculture, mining, and electrical power generation), compared to areas with very low densities (Arlinghaus et al., 2002). Freshwater species-at-risk are more likely to occur in such areas more affected by human activities as are invasive species (Chu et al., 2005; Cambray, 2003). Preferred and valued fishery species are more often the casualties of such stress causing the loss of important ecosystem services (Holmlund and Hammer, 1999). Equation (3) provided a qualitative proxy for this complex array of impacts engendered by further population growth and increased resource utilization.
Exploitation effects were not specifically included in the derivation of the stress index (Chu et al., 2003) but freshwater fisheries are generally considered to be over-fished in many instances (Post et al., 2002) and studies in western Canada have shown how exploitation pressure is much greater close to dense population areas (Post et al., 2008). The widespread use of stocking in southern Canadian lakes to shore up productivity is further evidence of the cumulative negative effects of concentrated human activities (Cambray, 2003; Kerr, 2006).
If the stress index values are indicative of cumulative effects and directly represent impact levels, those heavily developed ecozones are already in the range where significant impacts on productive capacity can be expected. Urbanization of watersheds degrades water quality, habitat, and biotic communities as Stanfield and Kilgour (2006) showed for S. Ontario streams. Substantial cumulative impacts are already apparent in and near the Canada’s major conurbations like Montreal, Toronto, Calgary-Edmonton, and Vancouver. The projected doubling of Canada’s population with much of the increase occurring in already densely-populated areas must inevitably lead to further significant loss and damage of natural ecosystems.
The biotic response models though only loosely grounded in data and empirical models did provide a means of indicating the full range of possible effects. If only biotic displacement occurs (equation 4) dramatic losses must be expected. If biotic adaptation occurs through species substitution, whether from among native or invading species, productivity losses will be minimized. The levels of biotic displacement and adaptation will depend on the vulnerability of the dominant species in each area, whether other species already present will replace those decreasing, and how rapidly additional fish species can expand their distributions into more northerly areas as climate warming takes place. This study assessed change in relation to current fish assemblages and did not account for shifts in assemblage size and composition that are likely as southern species expand northwards (Dextrase and Mandrak, 2006). How much individual species will be affected by climate warming is not entirely known but clearly projected warming will, for many cold-water species, mean that suitable habitats are no longer available or reduced. Mackenzie-Grieve and Post (2006) showed that many lakes will lose large parts or all of the critical late summer habitat space suitable for lake trout, a typical cold water species. In the northern part of their range lake trout often occupy shallow lakes as temperatures are low enough but as temperatures increase those smaller shallower lakes will not longer be suitable, lacking deep cool areas. Walleye, a typical coolwater species, will benefit in the northerly portion of their distribution in Canada but decline on the southern edge of the range in Southern Ontario (Shuter et al., 2002). Sharma et al. (2007) showed how the distribution of the warm-water smallmouth bass, Micropterus dolomeui, might be greatly increased with climate warming. Shuter and Post (1990) showed how the northern distributional limits of yellow perch and smallmouth bass were determined through thermal constraints on key life history parameters. In Canada’s inland freshwaters (boreal, taiga, and arctic ecozones) coldwater salmonids and coregonids make up the bulk of the northern fish species diversity, paralleling the important anadromous salmonid fishery stocks in the Atlantic and Pacific river drainages. Warmwater fishes account for much of the higher species richness in the southern ecozones with the potential to expand northwards with climate warming (Rahel and Olden, 2008).
Refining this meta-model into a more detailed data-based model will be challenging. Making and validating projections for the future is difficult. However, models are an essential feature of science-based decision-making and policy analysis. For example, in the 1970s the Club of Rome’s report ‘The Limits to Growth’ stirred up controversy and much criticism of it simplistic modelling but a recent assessment by Turner (2008) has shown that the 1970s forecasts were broadly consistent with trends observed up to 2000. More recently, Nelson et al (2009) have produced detailed data-based models of how future urbanization and climate change may impact Piedmont headwater streams in the U.S. Urbanization and climate change alone or together led to depressed growth and reproduction in many stream species, especially among important fishery species. In the meta-model presented here the lake resource estimates and the fishery production models are well-founded on data while the stress and biotic response models represent reasonable extrapolations from observed ecological patterns in freshwater fishes.
Reported Canadian freshwater fishery harvests, circa 2004-5, were approximately 90,000 metric tonnes yr−1 (www.dfo-mpo.gc.ca/communic/statistics/.) with 41,000 tonnes coming from commercial fisheries concentrated in the Great Lakes and in many larger lakes across the Canadian Shield from Ontario to Alberta. The recreational fishery accounts for the other 49,000 tonnes (assuming the average freshwater fish caught weighs 0.25 kg and that 20% of released fish suffer hook mortality). Thus, in the climate norms period, the reported yield accounts for 25% of the estimated productive capacity, or 31% of the stress-adjusted value. As much of the fish production occurs in remoter areas on the boreal and taiga shield and to a lesser extent in the arctic ecozones where fishing pressure is low and distant from human population centres, the low proportion is not surprising. Cold- and cool-water species dominate the catches. Walleye, or sander, a coolwater species, is the dominant commercial and recreational species. Coregonids like lake whitefish make up a large portion of the commercial catches and salmonids like lake trout are important given their value despite a lower contribution to harvest. Lake trout are the flagship recreational fishery across the boreal ecozone (Gunn et al., 2003). Where biotic adaptation occurs under climate change, either via species already present or via new invading species, the replacement species are unlikely to provide the same fishery opportunities, especially for commercial and subsistence fishers.
These analyses indicate the magnitude of Canada’s lake fisheries resources and the extent to which they may be threatened by continued expansion of direct human pressures and climate change. These analyses probably overstate the potential for increased productivity due to warming as they do not account for the negative impacts that dominant species will experience as conditions shift away from optimal or preferred in the areas where they have been most productive or as northward invaders replace them.
This basic approach to estimating productive capacity and how anthropogenic stress can reduce or alter the productive capacity could easily be elaborated with greater spatial resolution and with species- or community-specific breakdowns. Methods for quantitatively characterizing the effects of human activities on fisheries and the biotic responses in the fish communities should be developed to replace the more qualitative components of the meta-model presented here. The cumulative effects of the various components of ecosystem stress could also be estimated in more detail, e.g., reduced pH levels due to acidic deposition in the south-east region of Canada (Minns et al 2008) has undoubtedly reduced productive capacity.
Duinker and Grieg (2006) have highlighted the deficiencies of current approaches to cumulative impact assessment in Canada. The pervasive impacts that will arise with climate change will compound the strains already apparent in many ecosystems making it all the more urgent that sounder means for forecasting and managing into the future be developed. Holman and Harman (2008) have demonstrated a participatory impact assessment model for use in the United Kingdom and the Millennium Assessment reports (cf Bennett et al., 2003) demonstrated that global assessments were feasible. As the gathering environmental threats to humanity escalate greater energies must be applied to equipping society with the means to make informed decisions. The meta-model presented here demonstrates how the joint effects of projected climate warming and continued expansion of human activities might combine to affect the potential productivity, or productive capacity, of Canada’s lake fisheries. If current trends in human activities and pollution continue unchecked, substantial deleterious changes to Canada’s lake ecosystems may be unavoidable.
I extend my thanks to Mohi Munawar for encouraging me to write this paper, and to Brian Shuter and Cindy Chu for helpful discussions and feedback during the formative stages.