Abstract

Lakes around the metal smelters of Sudbury, Ontario, Canada offer a unique opportunity to study recovery processes in stressed aquatic ecosystems. Following major reductions in the atmospheric deposition of sulphur and metal particulates, chemical and biological recovery have been observed in many lakes; however, the magnitudes and trajectories of biological recovery are variable. Crustacean zooplankton communities have proven to be particularly valuable in tracking the biological recovery patterns in Sudbury lakes. In a survey of 87 lakes, zooplankton community structure revealed strong gradients related to acidification, metal contamination, trophic status and depth. The study lakes could be divided into four groups related to these gradients: acidified, recovered, urban and reference lakes. Community composition in recovered (to pH >6.0) lakes did not differ from reference lakes for copepods and cladocerans. In contrast, urban lakes subjected to the effects of watershed development as well as atmospheric metal contamination had either copepod or cladoceran communities that differed from reference conditions. Among the acidified lakes, most lakes with pH <5.5 contained species poor copepod communities; however, substantial community recovery was evident in lakes with pH >5.5. Our study illustrates the complexity of biological recovery processes in lakes subjected to multiple environmental stressors. Although substantial zooplankton community recovery clearly occurs as lakes recover from acidification, other stressors including metal contamination and eutrophication in urban environments may greatly affect recovery processes and outcomes.

INTRODUCTION

The widespread acidification of lakes throughout acid sensitive regions in Europe, the northeastern USA and southeastern Canada has resulted in dramatic losses of aquatic biota, including fish, benthic invertebrates and zooplankton (see Havas and Rosseland, 1995 for a review). Acidification in the Sudbury, Ontario, region began in the 1920s and quickly accelerated throughout the 1960s when the Sudbury metal smelters were one of the world's largest point sources of SO2. By the 1970s, many of the lakes were severely acidified. With the implementation of large reductions in SO2 and metal emissions, water quality has improved (Keller, 1992). Numerous studies have documented both the effects of acidification (e.g. Roff and Kwiatkowski, 1977; Keller and Pitblado, 1984) and the subsequent chemical recovery on aquatic communities (e.g. Keller et al. 1992a). Zooplankton communities in particular have been shown to be excellent indicators of acidification, yielding predictable and consistent responses to changes in lake pH and other chemical parameters affected by acid deposition (e.g. Webster et al., 1992; Keller et al., 2002; Walseng et al., 2003).

Evidence of zooplankton recovery in the Sudbury area following emission reductions has been mounting (Keller et al., 1992b; Havas et al., 1995; Keller et al., 2002). In this region, a pH of 6 has been used as a threshold at which significant biological damage is known to occur (Neary et al., 1990). In a synthesis of studies of zooplankton recovery from lakes across Eastern North America and Central, and Northern Europe, Gray and Arnott (Gray and Arnott, 2009) found that zooplankton communities in all regions appear to be at least partially recovering in lakes with pH values above 6.0. While increases in pH have led to an overall recovery of zooplankton across a broad geographic scale, there exists a patchwork of partially recovered zooplankton communities in lakes across the landscape (see Keller et al., 2007 for a review). This variability may be due to physiological effects of residual metal toxicity, physical effects resulting from reductions in dissolved organic carbon (DOC) concentrations and resultant changes in UV penetration, and ecological effects of altered predator communities. All of these multiple stressors interact with the spatial heterogeneity of the landscape and with chance events to create uneven recovery (Diamond, 1975; Keller and Yan, 1998; Yan et al., 2003; Young et al., 2005). It is clear that habitat restoration through chemical recovery of lakes may not be enough to ensure rapid and full ecological recovery.

Yan et al. (Yan et al., 2004) observed a surprising pattern in the recovery of the zooplankton community in Middle Lake, which was severely contaminated by acid and metal emissions from the Sudbury smelters. They found that the cladoceran community, unlike the copepod community, had yet to recover, despite 30 years of non-acidic conditions. This delayed recovery of cladocerans was unexpected due to the quick generation time and production of hardy, desiccation-resistant resting eggs of many species. Differences in recovery rates between taxonomic groups have been shown in manipulative experiments (Frost et al., 2006), but this aspect of recovery has yet to be tested at a regional scale.

Our study contributes to the growing body of work examining zooplankton recovery in response to increases in pH (e.g. Locke et al., 1994; Yan et al., 1996, 2004; Holt et al., 2003; Frost et al. 2006). In contrast to most past studies in the Sudbury area that focused on one or a few lakes, we used a broad, synoptic survey of 87 impacted and reference lakes to determine the relative importance of a suite of different environmental parameters in affecting the species composition of zooplankton. Our first objective was to determine whether zooplankton communities had recovered to a state typical of neutral lakes once pH rose above 6. We followed the recommendations of Yan et al. (Yan et al., 1996) who found that multivariate measures of species composition provided the best indicator of recovery of zooplankton from acidification. Multivariate indicators of recovery were calculated separately for the cladoceran and copepod community.

Our second objective was, if delays in recovery could be identified, to determine whether the pattern of delayed cladoceran versus copepod recovery observed by Yan et al. (Yan et al., 2004) was a widespread phenomenon in the Sudbury area. We hypothesized that the cladoceran community would recover once lake pH had risen above 6; however, recovery would be delayed despite reaching this threshold if metal levels were sufficiently high. To achieve these objectives, we used indirect and direct gradient analyses and test site analysis (TSA) to analyze survey data on zooplankton abundance and abiotic variables collected from 87 lakes. TSA is a bio-assessment approach commonly used to assess damage and recovery in benthic invertebrate communities (Bowman and Somers, 2006) and allows for a direct assessment of the biological condition of each lake in comparison to an established reference condition.

METHOD

Study lakes

The 87 lakes chosen for study spanned a gradient of acidification (pH 4.7–7.6) and metal contamination (Cu 0.4–31.7 µg L−1 and Ni 0.75–134 µg L−1). The lakes are located on the Canadian Shield; however, they varied considerably in chemistry, morphometry and basin geology (Table I). The lakes varied slightly in elevation (182–486 m ASL) but differed considerably in size (8.1–1315.4ha) and maximum depth (8.0–90.3 m). They were located between 5 and 120 km from the Sudbury smelters.

Table I:

Average values (and standard errors) for the 18 physicochemical variables in each of the four lake groups and their proximity to the smelting complexes (>15 km or <15 km)

Variable Abbreviation (units) >15 km
 
<15 km 
  Acid Recovered Reference Urban 
Alkalinity Alk (mg L−1−0.1 (0.20) 0.9 (0.1) 4.5 (0.5) 10.6 (2.3) 
Aluminum Al (µg L−1122.1 (21.1) 36.5 (4) 18.3 (2.3) 20.5 (3) 
Dissolved inorganic carbon DIC (mg L−10.3 (0.1) 0.4 (0.03) 1.2 (0.1) 3.0 (0.6) 
Dissolved organic carbon DOC (mg L−11.7 (0.4) 3.3 (0.2) 3.7 (3.7) 3.6 (0.2) 
Conductivity Cond (uS cm) 22.1 (0.8) 23.3 (0.8) 30.1 (1.6) 159.3 (29.7) 
Copper Cu (µg L−11.7 (0.30) 1.8 (0.3) 1.6 (0.2) 11.8 (1) 
Iron Fe (µg L−142.6 (12.7) 37.2 (5.3) 36.7 (11.4) 118.4 (38.7) 
Manganese Mn (µg L−174.4 (8.6) 27.0 (4.4) 8.1 (0.7) 36.3 (10.7) 
Nickel Ni (µg L−17.2 (1.7) 5.9 (1.8) 2.4 (0.6) 62.8 (4.7) 
Total nitrogen TKN (µg L−1147.1 (14.8) 221.4 (13.9) 230.9 (15.6) 277.7 (20.2) 
pH pH 5.3 (0.1) 6.3 (0.04) 6.9 (0.04) 6.8 (0.1) 
Total phosphorus TP (µg L−12.9 (0.4) 4.5 (0.4) 5.1 (0.4) 10.7 (1.3) 
Sulfate SO4 (mg L−17.1 (0.2) 6.6 (0.3) 6.4 (0.3) 11.5 (0.9) 
Zinc Zinc (µg L−16.9 (1.0) 4.4 (0.9) 2.8 (0.8) 6.9 (0.9) 
Secchi depth Secchi (m) 10.9 (1.2) 5.3 (0.4) 5.2 (0.4) 4.2 (0.3) 
Distance from smelter Smelter (km) 55.4 (4.5) 48.2 (6.6) 55.9 (6.8) 9.9 (0.6) 
Area Area (ha) 286.3 (58.8) 300.5 (85.3) 236.1 (34.3) 123.9 (41.2) 
Maximum depth Depth (m) 24.1 (3.9) 21.3 (3.3) 18.2 (2.0) 13.3 (1.4) 
Variable Abbreviation (units) >15 km
 
<15 km 
  Acid Recovered Reference Urban 
Alkalinity Alk (mg L−1−0.1 (0.20) 0.9 (0.1) 4.5 (0.5) 10.6 (2.3) 
Aluminum Al (µg L−1122.1 (21.1) 36.5 (4) 18.3 (2.3) 20.5 (3) 
Dissolved inorganic carbon DIC (mg L−10.3 (0.1) 0.4 (0.03) 1.2 (0.1) 3.0 (0.6) 
Dissolved organic carbon DOC (mg L−11.7 (0.4) 3.3 (0.2) 3.7 (3.7) 3.6 (0.2) 
Conductivity Cond (uS cm) 22.1 (0.8) 23.3 (0.8) 30.1 (1.6) 159.3 (29.7) 
Copper Cu (µg L−11.7 (0.30) 1.8 (0.3) 1.6 (0.2) 11.8 (1) 
Iron Fe (µg L−142.6 (12.7) 37.2 (5.3) 36.7 (11.4) 118.4 (38.7) 
Manganese Mn (µg L−174.4 (8.6) 27.0 (4.4) 8.1 (0.7) 36.3 (10.7) 
Nickel Ni (µg L−17.2 (1.7) 5.9 (1.8) 2.4 (0.6) 62.8 (4.7) 
Total nitrogen TKN (µg L−1147.1 (14.8) 221.4 (13.9) 230.9 (15.6) 277.7 (20.2) 
pH pH 5.3 (0.1) 6.3 (0.04) 6.9 (0.04) 6.8 (0.1) 
Total phosphorus TP (µg L−12.9 (0.4) 4.5 (0.4) 5.1 (0.4) 10.7 (1.3) 
Sulfate SO4 (mg L−17.1 (0.2) 6.6 (0.3) 6.4 (0.3) 11.5 (0.9) 
Zinc Zinc (µg L−16.9 (1.0) 4.4 (0.9) 2.8 (0.8) 6.9 (0.9) 
Secchi depth Secchi (m) 10.9 (1.2) 5.3 (0.4) 5.2 (0.4) 4.2 (0.3) 
Distance from smelter Smelter (km) 55.4 (4.5) 48.2 (6.6) 55.9 (6.8) 9.9 (0.6) 
Area Area (ha) 286.3 (58.8) 300.5 (85.3) 236.1 (34.3) 123.9 (41.2) 
Maximum depth Depth (m) 24.1 (3.9) 21.3 (3.3) 18.2 (2.0) 13.3 (1.4) 

Means significantly different from the reference lake group are indicated in bold (ANOVA, P< 0.002).

The 87 lakes were divided into four non-overlapping groups with regard to stress from atmospheric contaminant deposition. Three of the groups were located outside the major metal deposition area. The reference group contained lakes which never acidified to a pH less than 6 (n = 25) and, therefore, represent normal conditions for this area (chemically and biologically) under minimal disturbance. The acid lake group included lakes that exhibited slow rates of chemical recovery and remained below pH 6 (n = 23). The recovered lake group included lakes which were once acidified but had returned to pH values of 6 or above (n = 17) by the time of our study. The last group, the urban lake group, contained lakes located inside the high metal deposition zone (n = 22). These lakes were above pH 6 but, due to their proximity to the smelters (<15 km), they have copper and nickel levels which exceeded Ontario's Provincial Water Quality Objectives (PWQOs) (5 µg L−1 for copper and 25 µg L−1 for nickel) (MOEE 1994). Additionally, due to their position within the heart of Sudbury, they were exposed to multiple urban stresses not found in the other lakes (e.g. nutrient inputs, road salt runoff, shoreline development).

These four lake groups were used to create two gradients, one characterizing acidification and the other metal contamination. The lakes located outside the metal deposition zone (acid lake and recovered lake groups) were compared with the reference lake group to determine patterns in recovery from acidification. The lakes located inside the metal deposition zone (urban lake group) were compared to the reference lake group to examine patterns in recovery from metal contamination. Characterization of past acidification was based on examination of long-term data on water chemistry (Keller et al., 2006).

Data collection

Environmental variables

To reduce confounding effects of seasonal changes in composition, sampling took place in 2005 during the relatively stable midsummer period (28 June to 4 August), with the order of lakes sampled at random so as not to confound sampling time and lake group. Fourteen chemical parameters were measured in addition to four physical lake descriptors: Secchi depth, distance from nearest smelter, lake area and maximum depth (Table I). Morphometric variables were obtained from various published and unpublished sources (Keller et al., 2006).

Lake water was sampled with two different methods because of different monitoring needs. Sixty-five lakes were sampled with “surface grabs” by immersing a 4 L polyethylene jug at the deep station. The other 22 lakes (urban lake group) were sampled with an integrated plastic tube sampler that was deployed through the epi- and metalimnion. A comparison of the two sampling methods for 43 lakes did not reveal any statistically significant differences in the levels of pH, SO4, Cu and Ni (Keller and Pitblado, 1986). A more recent comparison of 15 lakes revealed that 8 out of the 22 chemistry variables showed significant differences between surface and tube composite samples (Keller et al., 2006). In light of the Keller et al. (Keller et al., 2006) study, the variables aluminum, iron, manganese and zinc were expected to be slightly higher in our tube composite samples (urban lake group) in comparison to surface grab samples. However, the absolute differences would be very small and were not likely to be biologically meaningful.

Fish community assessments had been done on a number of the lakes between 2000 and 2006 and this information was used to aid in the interpretation of the results. A detailed description of these methods can be found in Valois et al. (Valois et al., 2010). These assessments were primarily done on the urban lakes (n = 22); however, some of the acid (n = 4), recovered (n = 9) and reference (n = 3) lakes had fish data as well.

Zooplankton composition

Zooplankton were sampled with a single vertical haul from 1 m above the bottom to the lake surface. Either a metered 12.5 cm diameter net with an 80 µm mesh (65 lakes) or a non-metered, 30 cm diameter tow net with an 80 µm mesh (22 lakes—urban lake group) was used. Holt and Yan (Holt and Yan, 2003) investigated the effect of net size on zooplankton community descriptions (12.5 versus 30 cm diameter), and found no significant differences in species richness or abundances between the samples from the two nets, except for the abundance of Holopedium gibberum, which was significantly higher in the larger net. The variance in abundances of H. gibberum between lakes was significantly lower than for the other species, suggesting that the variance associated with net size was not necessarily greater for this species. Nevertheless, we interpret any difference we find in this species in samples from the urban lake group cautiously.

There has been some debate surrounding the accuracy of using a single haul as representative of the zooplankton community of an entire lake. Single samples have been shown to detect 50% of the annual species pool, missing highly variable species that are present infrequently throughout the sampling period (Arnott et al., 1998, 1999). Single samples are useful for determining relationships with important drivers of zooplankton composition, such as productivity, metal toxicity and predator density (Arnott et al., 1998). However, our sampling lacked the sensitivity to detect early patterns in recovery including the immigration of new species (which initially remain at low abundances) as well as failed colonization attempts (Yan et al., 2004).

Crustacean zooplankton were identified according to Smith and Fernando (Smith and Fernando, 1978) and Balcer et al. (Balcer et al., 1984). Immature stages of Copepoda were identified to suborder. A Folsom plankton splitter was used to generate subsamples which were processed using ZEBRA2, an MSDOS-based measuring program developed by the MOE Dorset Environmental Science Center (DESC). The ZEBRA2 program was run on a computer linked to electronic calipers, a Leica MZ16 dissecting microscope and a video camera. Abundance estimates were corrected for net efficiency when possible; net efficiency averaged 83%.

Data analyses

Environmental variables were log10 transformed to satisfy the assumptions of normality and equal variance. Species data were log10(x + 1) transformed. This reduces the influence of the most dominant species and takes into account the less abundant species as well. Very rare species, found in <5% of samples, were removed from the analyses. Statistical analyses were completed using CANOCO for Windows version 4.5 (ter Braak and Smilauer, 2002) with a significance level of α = 0.05.

Zooplankton structure and lake groups

Before applying ordination methods, principal components analysis (PCA) and redundancy analysis (RDA), the zooplankton data were checked for suitability. A preliminary detrended correspondence analysis indicated that the zooplankton data set exhibited relatively short gradients, less than two standard deviations, implying that RDA, which assumes a linear response of species to underlying gradients, was suitable for analyzing the data alone (ter Braak, 1995). We used PCA to search for overall patterns in both the environmental data and the zooplankton assemblages. A PCA on a correlation matrix with centering and standardization of the data was performed to detect the major environmental and biological gradients in the 87 lake data set.

Identifying delays in recovery

A common definition of recovery is a change in species composition to one typical of non-acidified or reference lakes (Holt and Yan, 2003). As no pre-stress data exist for lakes in this area, we used a reference condition approach in which non-acidified lakes were used to determine what communities would be expected in this area in the absence of acidification. The reference lakes were selected to contrast with the impacted lakes in terms of pH and metal concentrations; however, they also differed in other variables indirectly related to acidification (e.g. Al, DOC, conductivity, TKN and water clarity). Urban lakes were more productive and smaller than the reference lakes and so differences in zooplankton communities between these lake groups may also reflect inherent differences in productivity and size. We examined whether any of the impacted lakes were significantly outside of the normal range (reference condition) for selected metrics. We used PCA to calculate our multivariate metrics of recovery.

To determine whether the lakes with pH above 6 (recovered and urban groups) had similar community composition to the reference lakes, each lake was compared to the reference group using TSA. The TSA method has been used previously to evaluate the recovery of stream and lake benthic invertebrate communities from acidification (Bowman et al., 2006; Szkokan-Emilson et al., 2009) as well as impacts of mining activity on sediment quality, benthic algae, benthic invertebrates and fish communities, and slimy sculpin (Cottus cognatus) condition (Bowman et al., 2010). A TSA uses a generalized distance (D2) to quantify the difference between a potentially impacted site (test site) and a reference site (Bowman and Somers, 2006). Kilgour et al. (Kilgour et al., 1998) proposed that the difference between the mean of the test site and the reference sites should be significantly greater than the normal range of variation among the reference sites. This normal range is defined as the confidence region enclosing 95% of the reference sites. A non-central interval test is used to generate a probability that each metric, in this case the position defined by PCA1 and PCA2, is inside the normal range. The non-central P-value measures if the difference between the test and reference lakes is significant, categorizing each lake as significantly impaired (P < 0.05), potentially impaired (0.05 < P < 0.95) or in reference condition (P > 0.95). Using the first two axes of the ordination, two TSAs were run, one for cladoceran species and the other for copepods. TSAs were performed on cladocerans and copepods separately to determine if one group recovered before the other. TSA calculations were performed using Microsoft Excel 2003 (Microsoft Corp.). Four urban lakes were sampled twice; Clearwater, Crooked, Silver, and Tilton Lakes. Analyses were re-run with these samples to determine whether results could be attributed to sampling error. The results of the TSA did not change.

Physicochemical gradients

RDA was used to relate cladoceran and copepod community structure to the underlying environmental gradients in order to detect differences among these two groups in terms of their response to acidification and metal contamination. Statistical tests of significance were carried out by Monte Carlo permutation tests with 499 unrestricted permutations. In the Monte Carlo test, the distribution of the test statistics under the null hypothesis is generated by random permutations of cases in the environmental data (Gittens, 1985).

RESULTS

Among the 87 study lakes, a total of 36 crustacean zooplankton species were collected, comprised of 24 cladocerans and 12 copepods. Seven species were found in more than half of the lakes and included the cladocerans Bosmina freyi, H. gibberum, Eubosmina longispina, Diaphanosoma birgei and Daphnia mendotae as well as the copepods, Leptodiaptomus minutus and Mesocyclops edax. In all samples, immature copepods dominated by numbers forming up to 90% of the zooplankton community in some lakes. Total zooplankton abundance ranged from 4184 to 124 112 individuals m3. The relative occurrence of each species in each of the lake groups is summarized in Table II.

Table II:

Summary of the average relative occurrence (%) of common zooplankton (those found in >10% of samples) in the four lake groups (acid, recovered, reference, urban)

 >15 km
 
<15 km
 
Species Acid Recovered Reference Urban 
Cladocerans 
 Bosmina freyi 78 100 96 93 
 Ceriodaphnia lacustris 33 
 Chydorus sphaericus 17 16 
 Daphnia ambigua 13 35 12 
 Daphnia catawba 57 41 24 
 Daphnia dubia 36 
 Daphnia longiremis 12 40 
 Daphnia mendotae 13 35 84 57 
 Daphnia retrocurva 17 18 44 13 
 Daphnia birgei 35 47 80 83 
 Eubosmina longispina 83 82 68 
 Holopedium gibberum 83 100 84 53 
Copepods 
 Cyclops scutifer 35 53 44 13 
 Diacyclops biscuspidatus thomasi 17 59 84 40 
 Epichura lacustris 47 64 20 
 Leptodiaptomus minutus 100 100 92 93 
 Mesocyclops edax 65 76 92 60 
 Skistodiaptomus oregonensis 18 44 43 
 Tropocyclops extensus 13 53 44 67 
 >15 km
 
<15 km
 
Species Acid Recovered Reference Urban 
Cladocerans 
 Bosmina freyi 78 100 96 93 
 Ceriodaphnia lacustris 33 
 Chydorus sphaericus 17 16 
 Daphnia ambigua 13 35 12 
 Daphnia catawba 57 41 24 
 Daphnia dubia 36 
 Daphnia longiremis 12 40 
 Daphnia mendotae 13 35 84 57 
 Daphnia retrocurva 17 18 44 13 
 Daphnia birgei 35 47 80 83 
 Eubosmina longispina 83 82 68 
 Holopedium gibberum 83 100 84 53 
Copepods 
 Cyclops scutifer 35 53 44 13 
 Diacyclops biscuspidatus thomasi 17 59 84 40 
 Epichura lacustris 47 64 20 
 Leptodiaptomus minutus 100 100 92 93 
 Mesocyclops edax 65 76 92 60 
 Skistodiaptomus oregonensis 18 44 43 
 Tropocyclops extensus 13 53 44 67 

In general, the relationships between zooplankton species and lake groups were well characterized by the first two axes of PCA plots for both cladocerans and copepods. The first two axes of a PCA of cladoceran species abundance explained 41.7% of the variation among lakes (Fig. 1) and there were substantial differences in the loadings of the various species on the first two axes. The first axis primarily separated many of the urban lakes, which contained high abundances of D. mendotae, Diaphanosoma, B. freyi and Cerodaphnia lacustris from all other lakes. In contrast to the recovered lake group, the acid lakes were characterized by higher abundances of E. longispina, H. gibberum, D. ambigua and D. catawba. Many of the reference lakes loaded positively on the second axis. These lakes were characterized by D. longiremis, D. dubia and D. retrocurva. Some of the most severely acidic (pH <5.3) lakes and the urban lakes loaded negatively on this axis with a noticeable absence of many Daphnia species.

Fig. 1.

Principal component analysis biplot on (a) cladoceran zooplankton and (b) copepod zooplankton in the 87 study lakes. Lake groups are indicated by the following symbols: acid lakes (triangle), recovered lakes (square), reference lakes (diamond) and urban lakes (circle).

Fig. 1.

Principal component analysis biplot on (a) cladoceran zooplankton and (b) copepod zooplankton in the 87 study lakes. Lake groups are indicated by the following symbols: acid lakes (triangle), recovered lakes (square), reference lakes (diamond) and urban lakes (circle).

The first two axes of a PCA of copepod species abundance explained 48.8% of the variation among lakes (Fig. 1). The first component, explaining 34.3% of the variance, was characterized by negative loadings of Epischura lacustris, Skistodiaptomus oregonensis, Diacyclops biscuspidatus thomasi and immature cyclopoids. Many of the acid lakes loaded positively on this axis and were characterized by low abundances or absences of most copepods except L. minutus. The second component, explaining 14.5% of the variance, was characterized by negative loadings of L. minutus, M. edax and immature calanoids. The most common copepods in the reference lake group included M. edax and D. b. thomasi. The recovering lake group, while similar to the reference group, had a decreased occurrence of D. b. thomasi. The urban lakes were characterized by higher abundances of S. oregonensis and a noticeable decrease in Cyclops scutifer.

Test site analysis

To illustrate the difference in PCA lake scores between each lake group and the reference lakes, the first two axes were plotted, and the 95% confidence ellipses were used to show the variability within each group (Fig. 2). The reference lakes were much more variable in their cladoceran communities in comparison to their copepod communities, primarily along axis two. The reference lakes varied in their daphniid communities, with most lakes containing a number of species especially D. longiremis and D. dubia, while some were primarily dominated by D. mendotae. Using the axis scores and the TSA method, we found that the majority of recovered lakes were not impaired. Two lakes, Annie and White Oak Lakes, were found to have significantly impaired copepod communities, lacking most cyclopoid copepod species.

Fig. 2.

Biplot of axis scores (PCA1 and PCA2) for the cladoceran and copepod communities in the reference lakes (closed symbols) relative to the test lakes (open symbols). Ellipses represent 95% confidence regions for the data values for each lake group (acid, recovered, urban—dotted lines) compared to the reference lake group (solid lines).

Fig. 2.

Biplot of axis scores (PCA1 and PCA2) for the cladoceran and copepod communities in the reference lakes (closed symbols) relative to the test lakes (open symbols). Ellipses represent 95% confidence regions for the data values for each lake group (acid, recovered, urban—dotted lines) compared to the reference lake group (solid lines).

Surprisingly, the cladoceran community of the acid lake group was not significantly impaired with more than half of the acid lakes in reference condition (some lakes were found to be potentially impaired). In 13 of the acid lakes, copepod community composition was significantly impaired, with many lakes lacking T. extensus, D. b. thomasi, S. oregonensis and Epischura lacustris (Table II). The difference between the acid lakes and the reference lakes increased with decreasing pH (Fig. 3), with most lakes with a pH <5.5 containing impaired copepod communities. Unlike the recovered lakes, TSA analysis revealed that most urban lakes differed significantly from reference, and either cladoceran or copepod communities differed. Only one lake, Johnny Lake, a fishless lake, was found to be significantly impaired for both copepods and cladocerans. Copepod communities in these lakes were highly variable (Fig. 2). Urban lakes with impaired copepod communities were characterized by low pH and low DOC levels in addition to elevated metal concentrations. These lakes contained few Daphnia species other than D. mendotae and lacked (or contained low abundances) of H. gibberum and characterized by elevated pH and conductivity. There was no relationship between increasing metal concentrations (copper is shown) and distance from reference condition for the cladoceran community (r = −0.07, P > 0.05), although distance from reference condition appeared to increase for copepods (r = 0.35, P = 0.07) (Fig. 4). The four lakes with additional samples; Clearwater, Crooked, Silver, and Tilton Lakes had similar results when the TSA was re-run, with significant impairment found in the copepod but not the cladoceran community.

Fig. 3.

Relationship between pH and each lake distance from reference condition in the acid lakes group. Distance (D2) is a generalized Euclidean distance calculated using the first two axes of a PCA of cladoceran and copepod species abundance. Lakes falling above the horizontal line are significantly different from reference.

Fig. 3.

Relationship between pH and each lake distance from reference condition in the acid lakes group. Distance (D2) is a generalized Euclidean distance calculated using the first two axes of a PCA of cladoceran and copepod species abundance. Lakes falling above the horizontal line are significantly different from reference.

Fig. 4.

Relationship between copper concentrations (log-transformed) and each lake distance from reference condition in the urban lakes group. Distance (D2) is a generalized Euclidean distance calculated using the first two axes of a PCA of cladoceran and copepod species abundance. Lakes falling above the horizontal line are significantly different from reference.

Fig. 4.

Relationship between copper concentrations (log-transformed) and each lake distance from reference condition in the urban lakes group. Distance (D2) is a generalized Euclidean distance calculated using the first two axes of a PCA of cladoceran and copepod species abundance. Lakes falling above the horizontal line are significantly different from reference.

Zooplankton responses to environmental gradients

All 18 physical and chemical variables had significant marginal effects with the abundance of cladocerans. Forward selection indicated that five variables were significant; alkalinity, copper, manganese, conductivity and aluminum, explaining 29% of the variation (Table III). The most important variable was alkalinity, which alone explained 15% of the variation. An RDA of copepod species indicated that 16 variables exhibited significant marginal effects, explaining a total of 49.5% of the variation. Forward selection identified nine significant variables; pH, area, DOC, conductivity, depth, area, aluminum, iron and DIC which together explained 40.5% of the variation (Table III). Cyclopoid copepods such as D. b. thomasi and M. edax were strongly, negatively correlated with pH. The abundance of D. b. thomasi was negatively correlated with metals (copper and nickel). The first two axes were significant (P = 0.002) in the Monte Carlo permutation test with 499 permutations for both the cladoceran and copepod analyses.

Table III:

Marginal and conditional effects of forwardly selected environmental variables produced by RDA for the cladoceran and copepod community data

 Cladocerans
 
Copepods
 
 Marginal Conditional
 
Marginal Conditional
 
Variables P-value % explained P-value P-value % explained P-value 
Alk 0.002 15 0.002 0.002 0.666 
Al 0.002 1.5 0.18 0.002 0.016 
DOC 0.002 0.38 0.002 3.5 0.002 
DIC 0.002 0.5 0.84 0.002 0.010 
Cond 0.002 0.002 0.002 2.5 0.014 
Cu 0.002 0.002 0.002 0.002 
Fe 0.008 1.5 0.120 0.008 0.036 
Mn 0.002 0.002 0.002 0.5 0.856 
Ni 0.002 0.360 0.002 0.5 0.930 
TKN 0.002 1.5 0.100 0.002 1.5 0.224 
pH 0.002 0.340 0.002 16 0.002 
TP 0.002 0.5 0.850 0.002 1.5 0.118 
SO4 0.002 0.300 0.008 1.5 0.230 
Zn 0.002 0.590 0.002 0.5 0.736 
Secchi 0.002 0.002 0.002 1.5 0.084 
Smelter 0.002 0.270 0.002 0.5 0.748 
Area 0.002 1.5 0.070 0.004 2.5 0.016 
Depth 0.002 1.5 0.070 0.002 0.002 
 Cladocerans
 
Copepods
 
 Marginal Conditional
 
Marginal Conditional
 
Variables P-value % explained P-value P-value % explained P-value 
Alk 0.002 15 0.002 0.002 0.666 
Al 0.002 1.5 0.18 0.002 0.016 
DOC 0.002 0.38 0.002 3.5 0.002 
DIC 0.002 0.5 0.84 0.002 0.010 
Cond 0.002 0.002 0.002 2.5 0.014 
Cu 0.002 0.002 0.002 0.002 
Fe 0.008 1.5 0.120 0.008 0.036 
Mn 0.002 0.002 0.002 0.5 0.856 
Ni 0.002 0.360 0.002 0.5 0.930 
TKN 0.002 1.5 0.100 0.002 1.5 0.224 
pH 0.002 0.340 0.002 16 0.002 
TP 0.002 0.5 0.850 0.002 1.5 0.118 
SO4 0.002 0.300 0.008 1.5 0.230 
Zn 0.002 0.590 0.002 0.5 0.736 
Secchi 0.002 0.002 0.002 1.5 0.084 
Smelter 0.002 0.270 0.002 0.5 0.748 
Area 0.002 1.5 0.070 0.004 2.5 0.016 
Depth 0.002 1.5 0.070 0.002 0.002 

The gradients identified in this analysis were consistent with the general environmental factors on which the lakes had been originally classified. The first axis was primarily an axis of acidification variables separating high pH, low aluminum reference lakes from the high aluminum, low DOC and acid lakes. The recovered lakes fell somewhere between these two extremes. These lakes contained low abundances of Diaphanosoma birgei and many Daphnia species, especially D. mendotae. On the other hand, the second RDA axis can be defined as a gradient of trophic status, metals and depth, separating the shallow, metal contaminated, high nutrient lakes within the city from the low nutrient, deeper lakes outside the urban zone. Many of the urban lakes also resembled the acid lakes. Although D. mendotae was quite common in many of the urban lakes, other Daphnia species were lacking.

DISCUSSION

Biological recovery from the acidification of lakes can follow unpredictable trajectories. Although zooplankton communities in many Sudbury area lakes are recovering as water chemistry improves, incomplete and divergent trajectories have been observed (e.g. Keller et al., 2002; Yan et al., 2004). In a review of studies from regions severely affected by acidification, Gray and Arnott (Gray and Arnott, 2009) found that the importance of water quality, dispersal and community-level impediments in influencing zooplankton recovery may differ among regions. By documenting recovery patterns in relation to water quality improvements at a regional level, we have the opportunity to increase our understanding of the processes regulating limnetic community structure and biodiversity.

Appropriate interpretation of community data relies on the assumption that sampling methods provided an adequate estimate of community composition and that any sampling biases were at least consistent across lakes (Magalhaes et al., 2002). Previous zooplankton studies have shown that a potential drawback to synoptic surveys is the small per-lake sampling effort, which is suitable for common species but tends to miss rare species (Arnott et al., 1999). Accounting for spatial heterogeniety may add to the amount of explained variation but would also require a huge increase in sampling effort. On the other hand, in our study, the substantial number of lakes within each lake group did mitigate against missing important patterns. Moreover, some of the best-known long-term, comparative studies of zooplankton communities have used data collected with a variety of methods without any apparent biases caused by sampling gear (e.g. Shurin et al., 2000).

Through the use of a reference condition approach, our findings support the hypothesis of Gray and Arnott (Gray and Arnott, 2009) that zooplankton communities do recover when lake pH rises above 6, at least in systems uncontaminated with metals. The TSA results revealed that the recovered lake group was not significantly different from reference condition for the cladoceran community, as well as for the copepod community in all but two lakes (Annie and White Oak Lakes). White Oak Lake only reached the recovery threshold within the last year, possibly indicating a delay between chemical and biological recovery. Time lags have been observed in many other studies (Havas et al., 1995; Snucins et al., 2001; Monteith et al., 2005). Snucins (Snucins, 2003) found that sensitive benthic invertebrates took 4–8 years to re-colonize two recovering lakes after the estimated pH threshold for these species was reached. Moreover, the time before successful reestablishment throughout the lake was much longer, from 11 to 22 years. Annie Lake, on the other hand, has been at pH > 6 for 7 years, and the reasons for the absence of C. scutifer and low abundance of immature cyclopoids are unknown.

Surprisingly, many of the acid lakes appear to have recovered, particularly lakes above pH 5.5. Clearly, substantial improvements in zooplankton communities can occur before a level of pH 6.0 is reached. In the most severely acidified lakes (pH <5.5), it was the copepod and not the cladoceran community which deviated significantly from reference. In these lakes, cyclopoid copepods were absent or quite reduced in numbers. Different recovery rates for different taxonomic groups have also been found in other studies. In Swan Lake (Sudbury, Ontario), recovery was most rapid in the algae and rotifer communities, with slow and incomplete recovery in crustacean zooplankton (Arnott et al., 2001). In a comparison of three taxonomic groups (cladocerans, copepods and rotifers), copepod recovery was slowest in Little Rock Lake (Wisconsin, USA) (Frost et al., 2006). Metal toxicity is of particular concern in lakes closest to the smelters where copper and nickel concentration are often very high. Zooplankton communities in many of the urban lakes were found to be impaired despite years of non-acidic conditions, which may be attributable to elevated metal concentrations. Impairments were seen in both the cladoceran and copepod communities. Lakes with impaired cladoceran communities were lacking or contained very small populations of H. gibberum, a common species for Canadian Shield lakes. The affected lakes were often shallow, with elevated pH and conductivity, suggesting that the delays in cladoceran recovery we found were caused by increased urban stresses. The absence of H. gibberum from Middle Lake observed by Yan et al. (Yan et al., 2004) was most likely the result of urban stresses and not metal levels as this species was often quite abundant in many of the metal-contaminated lakes nearby. To better assess recovery in these lakes would require the use of reference lakes subjected to similar stressors (e.g. shoreline modifications, with inputs of fertilizer and road salts).

Urban lakes with impaired copepod communities had elevated metal levels as well as low concentrations of DOC and ions, which can increase metal toxicity through increased bioavailability (Borgmann et al., 2005). Copepod recovery may have been limited by residual metal toxicity. These lakes lacked many large calanoids (e.g. Epischura lacustris, S. oregonensis) which agrees with the results of Roch et al. (Roch et al., 1985) who found that some calanoid copepods, especially the juvenile stages, were particularly sensitive to metal contamination. A striking similarity in the copepod communities of both acidic and highly metal contaminated lakes was found, with both lake types dominated by L. minutus. This small calanoid appears to be well-adapted to both metal and acid stress (Derry et al., 2010), although some studies have found L. minutus to decrease under severe acid stress (e.g. Fischer et al., 2001).

In addition to the chemical factors considered here, biological factors may also be important in determining zooplankton community recovery (e.g. Valois et al., 2010). It appears that the recovery of zooplankton communities cannot necessarily be expected until the return of the pre-acidification predator community has occurred (Yan et al., 1991; Marmorek and Korman, 1993; Keller et al., 2002). Binks et al. (Binks et al., 2005) suggested that high abundances of macro-invertebrate predators may be contributing to the failed recolonization of copepod communities in lakes lacking planktivorous fish, and that the recovery of the fish community may be necessary to facilitate the reestablishment of zooplankton in these systems. Of the non-acidic lakes, Johnny Lake was the only lake known to be fishless and thus should have supported a high abundance of invertebrate predators, e.g. Chaoborus larvae. This lake was found to be significantly impaired for both the cladoceran and copepod community, lacking large cladocerans (Daphnia, H. gibberum) and with low abundances of calanoid copepods. The absence of large cladocerans is very unusual and not likely the result of high predation pressure by Chaoborus, which generally exclude or reduce small zooplankton (Dodson, 1974; Vanni, 1988; Arnott and Vanni, 1993). Calanoid copepods have been shown to be vulnerable to invertebrate predators (Neill, 1981; Nyberg, 1984) and the conspicuous reduction in L. minutus may be a result of predation effects. Some of the severely acidified lakes were expected to be fishless and we were unable to determine what role predation by invertebrate predators may have played in zooplankton recovery in these lakes.

Fish are also known to have strong direct effects on zooplankton communities. Although the majority of the lakes in this study contained fish, many of the urban lakes had unusual fish communities dominated by yellow perch (Perca flavescens) (Gunn and Keller, 1990). In these lakes, perch may have relied heavily on zooplankton as a result of a benthic community that was impaired by sediment toxicity (Luek et al., 2010). Valois et al. (Valois et al., 2010) attributed the dominance of H. gibberum in some of these lakes to the ability of this species to out-compete D. mendotae at high levels of planktivory. We had information on fish communities for 9 of the 12 lakes with significantly impaired cladoceran communities (Valois et al., 2010). Seven of these lakes contained relatively diverse fish communities, with well-established piscivore populations (smallmouth bass, brook trout, lake trout or northern pike) and low abundances of yellow perch. Little Raft Lake contained a high abundance of yellow perch; however, the lack of H. gibberum in this lake was more likely a result of its small size (19ha) and higher productivity (TP = 13 µg/L). It does not appear that fish planktivory played a dominant role in delaying recovery in these lakes. In comparison, most of the lakes with significantly impaired copepod communities were also dominated by yellow perch. This was most likely a result of elevated metal toxicity, which appears to exclude both piscivorous fish and cyclopoid copepods from establishing in these lakes (Valois et al., 2010).

Multiple stressors can result in complex, interactive effects on lake communities (Folt et al., 1999; Christensen et al., 2006). Stressors, such as shoreline development (Keller et al., 1999), and changes in UV penetration (Christensen et al., 2006) and calcium concentrations (Smol, 2010) have been shown to act antagonistically (shoreline development) or synergistically (increased UV and decreased calcium concentrations) on plankton communities recovering from acid stress. Although it appears that changes in nutrient concentrations resulting from urbanization and alterations in fish community structure are contributing to the lack of recovery in some of the urban lakes, the magnitude and consequences of these interactions are unknown. Multifactorial experiments (e.g. Strecker and Arnott, 2005; Christensen et al., 2006) are needed to identify the nature (synergy versus antagonism) of these potentially interactive effects.

Summary

The fundamental objective of emission reductions and protective legislation is the restoration of ecosystem structure and function, often measured as a return in the community to either its pre-disturbance state or to one similar to un-impacted communities in the area. However, assessing community recovery is difficult. Barriers which may stall or prevent the return of the pre-disturbance communities once the stress is removed may be unevenly distributed across damaged sites. Moreover, confounding effects of multiple stressors may obscure recovery patterns. In this study, we found that the zooplankton communities often recovered once lake pH reached 5.5; however, delays were evident in lakes with elevated metal concentrations. The delayed recovery of cladocerans, observed by Yan et al. (2004), was not localized to Middle Lake, or even solely to the cladoceran community, but was common to both cladocerans and copepods in many of the urban lakes.

Although clear relationships could be established between alterations in the copepod community and effects of acidity and elevated metal levels, these effects were not as clear for the cladoceran community. Given their larger size, cladocerans are often much more susceptible to fish predation (e.g. Vinebrooke et al., 2001; Nilssen and Waevergan, 2002) which was not accounted in this study. The results of this study add to the growing body of evidence that recovery of the zooplankton community occurs fairly well given sufficient water quality improvements, with evidence of recovery in lakes with pH >5.5. However, untangling the complex effect of urbanization, including its effect on metal toxicity, and the effect of altered predator communities will require the use of more appropriate reference lakes, toxicity testing and in situ experiments (i.e. mesocosms). Clearly, further studies are needed to increase our understanding of the roles and interactions of multiple stressors (metal toxicity, shoreline development, fish community composition) in affecting zooplankton recovery from acidification.

FUNDING

Research funding was provided by NSERC Discovery and Collaborative Research Development grant to W.B.K. and C.W.R. with industrial support from Vale INCO and Xstrata. The Laurentian University Research Fund provided additional support.

ACKNOWLEDGEMENTS

We would like thank the staff of the Cooperative Freshwater Ecology Unit with special acknowledgements to Jocelyne Heneberry, Shannon MacPhee and George Morgan. Erik Szkokan-Emilson and Michelle Bowman provided essential technical assistance when working with TSA. We also thank Lynne Witty for assistance with zooplankton taxonomy.

REFERENCES

Arnott
S. E.
Magnuson
J. J.
Yan
N. D.
Crustacean zooplankton species richness: single- and multiple-year estimates
Can. J. Fish. Aquat. Sci.
 , 
1998
, vol. 
55
 (pg. 
1573
-
1582
)
Arnott
S. E.
Vanni
M. J.
Zooplankton assemblages in fishless bog lakes: influence of biotic and abiotic factors
Ecology
 , 
1993
, vol. 
74
 (pg. 
2361
-
2380
)
Arnott
S. E.
Yan
N.
Keller
W.
, et al.  . 
The influence of drought-induced acidification on the recovery of plankton in Swan Lake (Canada)
Ecol. Appl.
 , 
2001
, vol. 
11
 (pg. 
747
-
763
)
Arnott
S. E.
Yan
N. D.
Magnuson
J. J.
, et al.  . 
Interannual variability and species turnover of crustacean zooplankton in Shield lakes
Can. J. Fish. Aquat. Sci.
 , 
1999
, vol. 
56
 (pg. 
162
-
172
)
Balcer
M. D.
Korda
N. L.
Dodson
S. I.
Zooplankton of the Great Lakes: A Guide to the Identification and Ecology of the Common Crustacean Species
 , 
1984
Madison, WI
University of Wisconsin Press
Binks
J. A.
Arnott
S. E.
Sprules
W. G.
Local factors and colonist dispersal influence crustacean zooplankton recovery from cultural acidification
Ecol. Appl.
 , 
2005
, vol. 
15
 (pg. 
2025
-
2036
)
Borgmann
U.
Grapentine
L.
Norwood
W. P.
, et al.  . 
Sediment toxicity testing with the freshwater amphipod Hyalella azteca: relevance and application
Chemosphere
 , 
2005
, vol. 
61
 (pg. 
1740
-
1743
)
Bowman
M. F.
Somers
K. M.
Evaluating a novel test site analysis (TSA) bioassessment approach
J. North Am. Benthol. Soc.
 , 
2006
, vol. 
25
 (pg. 
712
-
727
)
Bowman
M. F.
Somers
K. M.
Reid
R. A.
, et al.  . 
Temporal response of stream benthic macroinvertebrate communities to the synergistic effects of anthropogenic acidification and natural drought events
Freshwater Biol.
 , 
2006
, vol. 
51
 (pg. 
768
-
782
)
Bowman
M. F.
Spencer
P.
Dube
M.
, et al.  . 
Regional reference variation provides ecologically meaningful protection criteria for northern World Heritage Site
Integr. Environ. Manage. Assess.
 , 
2010
, vol. 
6
 (pg. 
12
-
27
)
Christensen
M. J.
Graham
M. D.
Vinebrook
R. D.
, et al.  . 
Multiple anthropogenic stressors cause ecological surprises in boreal lake
Global Change Biol.
 , 
2006
, vol. 
12
 (pg. 
2316
-
2322
)
Derry
A. M.
Arnott
S. E.
Boag
P. T.
Evolutionary shifts in copepod acid tolerance in an acid-recovering lake indicated by resurrected resting eggs
Evolutionary Ecology
 , 
2010
, vol. 
24
 (pg. 
133
-
145
)
Diamond
J. M.
Cody
M. L.
Diamond
J. M.
Assembly of species communities
Ecology and Evolution of Communities
 , 
1975
Cambridge, MA
Belknap
(pg. 
342
-
444
)
Dodson
S. I.
Adaptive change in plankton morphology in response to size-selective predation: a new hypothesis of cyclomorphosis
Limnol. Oceanogr.
 , 
1974
, vol. 
19
 (pg. 
721
-
729
)
Fischer
J. M.
Frost
T. M.
Ives
A. R.
Compensatory dynamics in zooplankton community responses to acidification: measurement and mechanisms
Ecol. Appl.
 , 
2001
, vol. 
11
 (pg. 
1060
-
1072
)
Folt
C. L.
Chen
C. Y.
Moore
M. V.
, et al.  . 
Synergism and antagonism among multiple stressors
Limnol. Oceanogr.
 , 
1999
, vol. 
44
 (pg. 
864
-
877
)
Frost
T. M.
Fischer
J. M.
Klug
J. L.
, et al.  . 
Trajectories of zooplankton recovery in the Little Rock Lake whole-lake acidification experiment
Ecol. Appl.
 , 
2006
, vol. 
16
 (pg. 
353
-
367
)
Gittens
R.
Canonical Analysis: A Review of Applications in Ecology
 , 
1985
New York, NY
Springer
Gray
D. K.
Arnott
S. E.
Recovery of acid damaged zooplankton communities: measurement, extent, and limiting factors
Environ. Rev.
 , 
2009
, vol. 
17
 (pg. 
81
-
99
)
Gunn
J. M.
Keller
W.
Biological recovery of an acid lake after reductions in industrial emissions of sulfur
Nature
 , 
1990
, vol. 
345
 (pg. 
431
-
433
)
Havas
M.
Rosseland
B. O.
Response of zooplankton, benthos, and fish to acidification: an overview
Water Air Soil Pollut.
 , 
1995
, vol. 
85
 (pg. 
51
-
62
)
Havas
M.
Woodfine
D. G.
Lutz
P.
, et al.  . 
Biological recovery of two previously acidified, metal-contaminated lakes near Sudbury Ontario, Canada
Water Air Soil Pollut.
 , 
1995
, vol. 
85
 (pg. 
791
-
796
)
Holt
C.
Yan
N. D.
Recovery of crustacean zooplankton communities from acidification in Killarney Park, Ontario, 1971–2000: pH 6 as a recovery goal
Ambio
 , 
2003
, vol. 
32
 (pg. 
203
-
207
)
Holt
C. A.
Yan
N. D.
Somers
K. M.
pH 6 as the threshold to use in critical load modeling for zooplankton community change with acidification in lakes of south-central Ontario: accounting for morphometry and geography
Can. J. Fish. Aquat. Sci.
 , 
2003
, vol. 
60
 (pg. 
151
-
158
)
Keller
W.
Introduction and overview to aquatic acidification studies in the Sudbury, Ontario, Canada area
Can. J. Fish. Aquat. Sci.
 , 
1992
, vol. 
49
 
Suppl. 1
(pg. 
3
-
7
)
Keller
W.
Gunn
J. M.
Yan
N. D.
Evidence of biological recovery in acid stressed lakes near Sudbury, Canada
Environ. Pollut.
 , 
1992
, vol. 
78
 (pg. 
79
-
85
)
Keller
W.
Heneberry
J. H.
Gunn
J. M.
Effects of emission reductions from the Sudbury smelters on the recovery of acid- and metal-damaged lakes
J. Aquat. Eco. Stress Recov.
 , 
1999
, vol. 
6
 (pg. 
189
-
198
)
Keller
W.
Heneberry
J. H.
McLachlan
E.
, et al.  . 
Data Report: 25 Years of Extensive Monitoring of Acidified Lakes in the Sudbury Area, 1981 to 2005
 , 
2006
Sudbury, ON
Cooperative Freshwater Ecology Unit
Keller
W.
Pitblado
J. R.
Crustacean plankton in northeastern Ontario lakes subjected to acidic deposition
Water Air Soil Pollut.
 , 
1984
, vol. 
23
 (pg. 
1573
-
2932
)
Keller
W.
Pitblado
J. R.
Water quality changes in Sudbury area lakes: a comparison of synoptic surveys in 1974–1976 and 1981–1983
Water Air Soil Pollut.
 , 
1986
, vol. 
29
 (pg. 
285
-
296
)
Keller
W.
Pitblado
J. R.
Carbone
J.
Chemical responses of acidic lakes in the Sudbury, Ontario area to reduced smelter emissions, 1981–89
Can. J. Fish. Aquat. Sci.
 , 
1992
, vol. 
49
 (pg. 
25
-
32
)
Keller
W.
Yan
N. D.
Biological recovery from lake acidification: zooplankton communities as a model of patterns and processes
Restor. Ecol.
 , 
1998
, vol. 
6
 (pg. 
364
-
375
)
Keller
W.
Yan
N. D.
Gunn
J. M.
, et al.  . 
Recovery of acidified lakes: lessons from Sudbury, Ontario, Canada
Water Air Soil Pollut. Focus
 , 
2007
, vol. 
7
 (pg. 
317
-
322
)
Keller
W.
Yan
N. D.
Somers
K. M.
, et al.  . 
Crustacean zooplankton in lakes recovering from acidification
Can. J. Fish. Aquat. Sci.
 , 
2002
, vol. 
59
 (pg. 
726
-
735
)
Kilgour
B. W.
Somers
K. M.
Matthews
D. E.
Using the normal range as a criterion for ecological significance in environmental monitoring and assessment
Ecoscience
 , 
1998
, vol. 
5
 (pg. 
542
-
550
)
Locke
A.
Sprules
W. G.
Keller
W.
, et al.  . 
Zooplankton communities and water chemistry of Sudbury area lakes: changes related to pH recovery
Can. J. Fish. Aquat. Sci.
 , 
1994
, vol. 
51
 (pg. 
991
-
1003
)
Luek
A.
Morgan
G. E.
Wissel
B.
, et al.  . 
Rapid and unexpected effects of piscivore introduction on trophic position and diet of perch (Perca flavescens) in lakes recovering from acidification and metal contamination
Freshwater Biol.
 , 
2010
, vol. 
55
 (pg. 
1616
-
1627
)
Magalhaes
M. F.
Batalha
D. C.
Collares-Pereira
M. J.
Gradients in stream fish assemblages across a Mediterranean landscape: contributions of environmental factors and spatial structure
Freshwater Biol.
 , 
2002
, vol. 
47
 (pg. 
1015
-
1031
)
Marmorek
D. R.
Korman
J.
The use of zooplankton in a biomonitoring program to detect lake acidification and recovery
Water, Air, Soil Pollut.
 , 
1993
, vol. 
69
 (pg. 
223
-
241
)
Monteith
D. T.
Hildrew
A. G.
Flower
R. J.
, et al.  . 
Biological responses to the chemical recovery of acidified fresh waters in the UK
Environ. Pollut.
 , 
2005
, vol. 
137
 (pg. 
83
-
101
)
Neary
B. P.
Dillon
P. J.
Munro
J. R.
, et al.  . 
The Acidification of Ontario lakes: An Assessment of their Sensitivity and Current Status with Respect to Biological Damage
 , 
1990
Dorset, ON, Canada
Technical Report of the Ontario Ministry of Environment and Energy
Neill
W. E.
Impact of Chaoborus predation upon the structure and dynamics of a crustacean zooplankton community
Oecologia
 , 
1981
, vol. 
48
 (pg. 
164
-
177
)
Nilssen
J. P.
Waevergan
S. V.
Intensive fish predation: an obstacle to biological recovery following liming of acidified lakes?
J. Aquat. Ecosyst. Stress Recov.
 , 
2002
, vol. 
9
 (pg. 
73
-
84
)
Nyberg
P.
Impacts of Chaoborus predation on planktonic crustacean communities in some acidified and limed forest lakes in Sweden
Inst. Freshwater Res. Drottningholm
 , 
1984
, vol. 
61
 (pg. 
154
-
166
)
Roch
M.
Nordin
R. N.
Austin
A.
, et al.  . 
The effects of heavy metal contamination on the aquatic biota of Buttle Lake and the Campbell River drainage (Canada)
Arch. Environ. Contam. Toxicol.
 , 
1985
, vol. 
14
 (pg. 
347
-
362
)
Roff
J. C.
Kwiatkowski
R. E.
Zooplankton and zoobenthos communities of selected northern Ontario lakes of different acidities
Can. J. Zool.
 , 
1977
, vol. 
55
 (pg. 
899
-
911
)
Shurin
J. B.
Havel
J. E.
Leibold
M. A.
, et al.  . 
Local and regional zooplankton species richness: a scale-independent test for saturation
Ecology
 , 
2000
, vol. 
81
 (pg. 
3062
-
3073
)
Smith
K.
Fernando
C. H.
A guide to the freshwater calanoid and cyclopoid copepod Crustacea of Ontario
Univ. Waterloo Biol. Ser.
 , 
1978
, vol. 
18
 (pg. 
1
-
74
)
Smol
J. P.
The power of the past: using sediments to track the effects of multiple stressors on lake ecosystems
Freshwater Biol.
 , 
2010
, vol. 
55
 
Suppl. 1
(pg. 
43
-
59
)
Snucins
E.
Recolonization of acid-damaged lakes by the benthic invertebrates Stenacron interpunctatum, Stenonema femoratum and Hyalella azteca
Ambio
 , 
2003
, vol. 
32
 (pg. 
225
-
227
)
Snucins
E.
Gunn
J.
Keller
W.
, et al.  . 
Effects of regional reductions in sulphur deposition on the chemical and biological recovery of lakes within Killarney park, Ontario, Canada
Environ. Monit. Assess.
 , 
2001
, vol. 
67
 (pg. 
179
-
194
)
Strecker
A. L.
Arnott
S. E.
Impact of Bythotrephes invasion on zooplankton communities in acid-damaged and recovered lakes on the Boreal Shield
Can. J. Fish. Aquat. Sci.
 , 
2005
, vol. 
62
 (pg. 
2450
-
2462
)
Szkokan-Emilson
E. J.
Wesolek
B. E.
Gunn
J. M.
, et al.  . 
Recovery of benthic invertebrate communities from acidification in Killarney Park lakes
Environ. Monit. Assess.
 , 
2009
, vol. 
166
 (pg. 
293
-
302
)
ter Braak
C. J. F.
Jongman
R. H. G.
ter Braak
C. J. F.
van Tongeren
O. F. R.
Ordinations
Data Analysis in Community and Landscape Ecology
 , 
1995
2nd edn
New York, NY
Cambridge University Press
(pg. 
91
-
173
)
ter Braak
C. J. F.
Smilauer
P.
CANOCO Reference manual and CanoDraw for Windows User's Guide: Software for Canonical Community Ordination (version 4.5)
 , 
2002
Ithaca, NY
Microcomputer Power
Valois
A. E.
Ramcharan
C. W.
Keller
W.
Abiotic and biotic processes in lakes recovering from acidification: the relative roles of metal toxicity and fish predation as barriers to zooplankton re-establishment
Freshwater Biol.
 , 
2010
, vol. 
55
 (pg. 
2585
-
2597
)
Vanni
M. J.
Freshwater zooplankton community structure: introduction of large invertebrate predators and large herbivores to a small species community
Can. J. Fish. Aquat. Sci.
 , 
1988
, vol. 
45
 (pg. 
1758
-
1770
)
Vinebrooke
R. D.
Turner
M. A.
Kidd
K. A.
, et al.  . 
Truncated foodweb effects of omnivorous minnows in a recovering acidified lake
J. North Am. Benthol. Soc.
 , 
2001
, vol. 
20
 (pg. 
629
-
642
)
Walseng
B.
Yan
N. D.
Schartau
A. K.
Littoral microcrustacean (cladocera and Copepoda) indicators of acidification in Canadian Shield lakes
Ambio
 , 
2003
, vol. 
32
 (pg. 
208
-
213
)
Webster
K. E.
Frost
T. M.
Watras
C. J.
, et al.  . 
Complex biological responses to the experimental acidification of Little Rock Lake, Wisconsin, USA
Environ. Pollut.
 , 
1992
, vol. 
78
 (pg. 
73
-
78
)
Yan
N. D.
Girard
R. E.
Heneberry
J. H.
, et al.  . 
Recovery of copepod, but not cladoceran, zooplankton from severe and chronic effects of multiple stressors
Ecol. Lett.
 , 
2004
, vol. 
7
 (pg. 
452
-
460
)
Yan
N. D.
Keller
W.
MacIsaac
H. J.
, et al.  . 
Regulation of zooplankton community structure of an acidified lake by Chaoborus
Ecol. Appl.
 , 
1991
, vol. 
1
 (pg. 
52
-
65
)
Yan
N. D.
Keller
W.
Somers
K. M.
, et al.  . 
Recovery of crustacean zooplankton communities from acid and metal contamination: comparing manipulated and reference lakes
Can. J. Fish. Aquat. Sci.
 , 
1996
, vol. 
53
 (pg. 
1301
-
1327
)
Yan
N. D.
Leung
B.
Keller
W.
, et al.  . 
Developing conceptual frameworks for the recovery of aquatic biota from acidification
Ambio
 , 
2003
, vol. 
32
 (pg. 
165
-
169
)
Young
T. P.
Peterson
D. A.
Clary
J. J.
The ecology of restoration: historical links, emerging issues and unexplored realms
Ecol. Lett.
 , 
2005
, vol. 
8
 (pg. 
662
-
673
)

Author notes

Corresponding editor: Beatrix E. Beisner