9512.net
甜梦文库
当前位置:首页 >> >>

Wetland-Water Column Exchanges of Carbon, Nitrogen, and Phosphorus in a Southern Everglades Dwarf Ma


846

Long-term changes in phytoplankton composition in seven Canadian Shield lakes in response to multiple anthropogenic stressors1
Andrew M. Paterson, Jennifer G. Winter, Kenneth H. Nicholls, Bev J. Clark, Charles W. Ramcharan, Norman D. Yan, and Keith M. Somers

Abstract: We examined long-term changes in phytoplankton composition from 1981 to 2003 in seven intensively studied lakes on the southern Canadian Shield in Ontario, Canada. Significant (P < 0.05), temporally coherent increases in the relative biovolume of colonial chrysophytes were observed in six of the seven lakes, with coincident declines in the relative biovolume of diatom algae. Variance partitioning analyses identified water chemistry variables, and the co-variation of water chemistry with physicoclimatic variables, as most important in structuring phytoplankton assemblages through time in the study lakes (variance explained: chemical variables (14%–47%, mean = 28.7%); chemistry and physicoclimatic variables (21%–30%, mean = 25.5%)). With the exception of Harp Lake, which was invaded by Bythotrephes in the early 1990s, grazing variables did not explain a significant portion of the phytoplankton variance. We hypothesize that the longterm changes in phytoplankton species composition is attributable to multiple anthropogenic stressors acting at a regional scale. Our results, coupled with paleoecological studies, indicate that increases in the relative importance of colonial chrysophytes are coincident with water chemistry changes associated with industrial activity since the mid-1900s and physical changes linked to climate indices such as the North Atlantic Oscillation. Résumé : Nous avons examiné les changements à long terme de 1981 à 2003 dans la composition du phytoplancton de sept lacs étudiés en détail dans la région sud du Bouclier canadien en Ontario, Canada. Il s’est produit des accroissements significatifs (P < 0,05) et cohérents dans le temps du biovolume relatif des colonies de chrysophytes dans six des sept lacs, avec des déclins correspondants du biovolume relatif des algues diatomées. Des analyses de partition de la variance ont identifié les variables chimiques de l’eau, ainsi que la covariation de la chimie de l’eau et des variables physico-climatiques, comme les facteurs explicatifs les plus importants de la structuration des peuplements phytoplanctoniques dans le temps dans les lacs étudiés (variance expliquée : variables chimiques (14–47 %, moyenne 28,7 %); variables chimiques et physico-climatiques (21–30 %, moyenne 25,5 %)). Sauf au lac Harp qui a été envahi par Bythotrephes au début des années 1990, les variables reliées au broutage n’expliquent pas une portion significative de la variance du phytoplancton. Nous émettons l’hypothèse selon laquelle les changements à long terme dans la composition spécifique du phytoplancton sont attribuables aux facteurs anthropiques multiples de stress qui agissent à l’échelle régionale. Nos résultats, couplés à des études paléolimnologiques, indiquent que l’accroissement de l’importance relative des colonies de chrysophytes co?ncide avec les changements dans la chimie des eaux associés à l’activité industrielle depuis le milieu des années 1900 et avec les changements physiques reliés aux indices climatiques, tels que l’oscillation nord atlantique. [Traduit par la Rédaction] Paterson et al. 861

Introduction
The biological condition of the phytoplankton community has long been recognized by lake managers and aquatic sci-

entists as an essential component of environmental assessments in Canadian Shield lakes (Dillon and Rigler 1974; Schindler 1978). Algal abundances are highly correlated with the aesthetic condition of lake ecosystems, and a num-

Received 20 April 2007. Accepted 16 January 2008. Published on the NRC Research Press Web site at cjfas.nrc.ca on 17 April 2008. J19954 A.M. Paterson,2,3 B.J. Clark, and K.M. Somers. Ontario Ministry of the Environment, Dorset Environmental Science Centre, 1026 Bellwood Acres Road, P.O. Box 39, Dorset, ON P0A 1E0, Canada. J.G. Winter.3 Ontario Ministry of the Environment, Water Monitoring Section, 125 Resources Road, Toronto, ON M9P 3V6, Canada. K.H. Nicholls. S-15 Conc. 1, R.R. 1, Sunderland, ON L0C 1H0, Canada. C.W. Ramcharan. Department of Biology, Laurentian University, 935 Ramsey Lake Road, Sudbury, ON P3E 3C6, Canada. N.D. Yan. York University, 4700 Keele Street, Toronto, ON M3J 1P3, Canada.
1

This article is part of the series “Transforming understanding of factors that regulate aquatic ecosystems on the southern Canadian Shield”. 2 Corresponding author (e-mail: andrew.paterson@ontario.ca). 3 Authors contributed equally to this manuscript.
Can. J. Fish. Aquat. Sci. 65: 846–861 (2008) doi:10.1139/F08-022 ? 2008 NRC Canada

Paterson et al.

847

ber of species have been identified as nuisance taxa (e.g., Nicholls and Gerrath 1985; Nicholls 1995). Their ubiquity, diversity, species-specific sensitivities, and ease of sampling have resulted in the widespread use of phytoplankton taxa in modern and paleoecological assessments of environmental change (European Union 2000; Smol 2002; Stevenson and Smol 2003). Phytoplankton species composition may vary across environmental gradients of temperature, light quantity and quality, micro- and macro-nutrient concentrations, lake acidity and metal concentrations, and the abundance and composition of grazers (Reynolds 1984, 1998). These gradients may be altered directly or indirectly by regional and local disturbances such as changing climate, lake acidification and recovery, exotic species invasions, and shoreline disturbance (Schindler 1998; George et al. 2004). Consequently, changes in phytoplankton composition over time have been reported in many Canadian Shield lakes (e.g., Nicholls et al. 1992; Planas et al. 2000; Findlay et al. 2001). In south-central Ontario, Canada, combinations of anthropogenic stressors are impacting aquatic ecosystems. Despite significant declines in the emission of sulphur dioxide from industrial sources since the late 1980s, the deposition of strong acids continues to exceed critical loads for lakes and watersheds (Stoddard et al. 1999; Aherne et al. 2004; Watmough et al. 2005). Consequently, long-term declines in base cation export and lake Ca concentrations are expected to continue in many lakes (Watmough and Aherne 2008). Furthermore, climate models predict increases in the severity and duration of drought in central Ontario (Smith et al. 1998), with measurable effects on regional hydrology and biogeochemical cycling in watersheds and lakes (Yan et al. 1996; Eimers et al. 2008). Recently, the introduction of exotic species such as Bythotrophes to remote lakes on the Canadian Shield has resulted in significant changes to zooplankton populations, including significant increases in the mean size of cladocerans in many lakes (Yan et al. 2001). Collectively, these and other stressors (e.g., shoreline development; Dillon and Molot 1996) will likely continue to affect lake ecosystems into the future, with unknown repercussions for lake biota. For the first time, we present the long-term phytoplankton record from seven intensively studied lakes in the Muskoka– Haliburton region of central Ontario, Canada. In contrast to those in heavily industrialized regions (e.g., Sudbury) or experimentally manipulated lakes (e.g., whole-lake experiments at the Experimental Lakes Area), lakes in this region are commonly considered to be minimally impacted reference lakes, having been exposed to decades of low to moderate environmental disturbances. Thus, the changes observed in these lakes may be representative of changes occurring in thousands of softwater lakes on the southern Canadian Shield. Despite their classification as reference ecosystems, paleoecological studies from these and other lakes in this region show that the present composition of phytoplankton is significantly different from that of preindustrial times (Hall and Smol 1996; Paterson et al. 2001; Quinlan et al. 2008), although few studies have examined the specific timing of these changes in central Ontario, Canada (Paterson et al. 2004; Quinlan et al. 2008). Specifically, we (i) test whether significant monotonic trends in phytoplankton composition have occurred in the study

lakes since 1981, (ii) determine whether the relative biovolumes of major algal groups (i.e., diatom and chrysophyte algae) have changed in a temporally coherent manner among the lakes, and (iii) quantify the unique and interactive effects of physicoclimatic, chemical, and biological (i.e., grazers) variables in explaining phytoplankton variation over time. We selected these categories of predictor variables because they encompass the current understanding of possible stressor effects on Canadian Shield lakes, including climatic effects (e.g., changes in air temperature, duration of the icefree season), chemical effects (e.g., acidification and recovery, changes in nutrient levels), and biological effects (e.g., indirect effects from invasive species). Based on evidence that there have been significant changes in many of the predictor variables in recent decades and paleoecological evidence of long-term changes in algal populations in many lakes since preindustrial times, we predict that there will be significant, long-term changes in phytoplankton community composition in these lakes. Temporal trends in other biological groups in the study lakes are discussed in detail in companion papers in this issue (Bowman et al. 2008; Rusak et al. 2008; Yan et al. 2008).

Materials and methods
Study sites The seven study lakes examined here (Blue Chalk, Chub, Crosson, Dickie, Harp, Plastic, and Red Chalk) are located in the Muskoka–Haliburton region of the Canadian Shield in south-central Ontario, Canada (44–45°N latitude, 78–79°W longitude). The region is characterized by shallow, acidic soils containing little or no carbonate-containing minerals overlying granitic (silicate) bedrock, although thicker deposits of clay, sand, or gravel occur locally (Jeffries and Snyder 1983). Mean annual air temperature and total annual precipitation are 4.5 °C and 1034 mm, respectively (Rusak et al. 2002). The study sites are primarily small (<100 ha), headwater, dimictic lakes, with the exception of Red Chalk, which is located downstream from Blue Chalk, and Dickie, which undergoes perennial mixing between periods of stratification. Chemically, the lakes differ in their acid sensitivity (Gran alkalinity, 2003, 6.8–91.5 ?equiv.·L–1) and trophic status (total phosphorus, 2003, 4.4–9.2 ?g·L–1). Many of the lakes have undergone significant temporal changes in water quality over the past three decades, including increases in pH and alkalinity (Yan et al. 2008), dissolved organic carbon (Keller et al. 2008), and road salt inputs, and decreases in total phosphorus, sulphate, and calcium concentrations (Watmough and Aherne 2008). These chemical changes are occurring coherently within and among lakes and have been related to regional stressors such as changes in deposition of strong acids (Watmough and Aherne 2008), climatic forcing (Dillon et al. 1997; Eimers et al. 2008; Keller et al. 2008), and human settlement (Dillon and Molot 1996). Harp Lake was invaded by Bythotrephes longimanus in the early 1990s, resulting in dramatic changes to the zooplankton community (Yan et al. 2002). Phytoplankton sampling and enumeration Phytoplankton samples were collected from the deep station of each lake as unfiltered, volume-weighted, euphotic? 2008 NRC Canada

848

Can. J. Fish. Aquat. Sci. Vol. 65, 2008

zone composites (approximated as 2 × Secchi depth) using a polyvinyl chloride (PVC) pump-and-hose system. Phytoplankton samples were fixed with 1 mL Lugol’s iodine solution in the field. In all lakes, the ice-free sampling frequency has changed periodically through the sampling record from weekly sampling in 1981 to biweekly (Dickie, Harp, and Plastic) or monthly (Blue Chalk, Chub, Crosson, Red Chalk) sampling presently. In four of the seven study lakes (Blue Chalk, Dickie, Harp, Plastic), there was no significant relationship between changes in sampling frequency and the main direction of variation in the group-level phytoplankton data (correlation of sampling frequency and principal component analysis (PCA) axis 1 scores of species data; P > 0.05). Moreover, in two of the three lakes where sampling frequency was significantly correlated with long-term phytoplankton changes (Blue Chalk, Crosson; P < 0.05), similar trends in phytoplankton composition were recorded in sediment cores (i.e., long-term change from diatom to chrysophyte algae; Paterson et al. 2004; C.S. Chueng, A.M. Paterson, and J.P. Smol, unpublished data). This suggests that the observed changes in phytoplankton are not an artifact of concurrent changes in sampling frequency. With the exception of Plastic Lake, no phytoplankton data are available for 1983. In the laboratory, seasonal phytoplankton subsamples were concentrated to 25 mL following settling and preserved with two drops of 37% formalin (Hopkins and Standke 1992). Subsamples were recombined at the end of each ice-free season into an annual composite sample from each lake for each year. Before amalgamation, volumes were adjusted according to a known sample concentration factor, and the final concentration factor was noted and used to determine lake-equivalent volumes (Hopkins and Standke 1992). Phytoplankton were counted using inverted microsocopy and Uterm?hl counting chambers by Ontario Ministry of the Environment (MOE) analysts, working jointly at MOE’s Algal Taxonomy Unit Laboratory (Toronto, Ontario). A minimum of 400 pieces (singly occurring cells or colonies) was counted and identification was predominantly to genus level. Generally, 100 to 150 cells or units of the most common taxa were counted for a counting error of ≤20% and to ensure sufficient precision. Estimates of cell volumes for each taxon were obtained by routine measurements of 30 to 50 cells and application of the geometric formula best fitted to the shape of the cell. A specific gravity of one was assumed for cellular mass. In this manner, cell counts were converted to wet weight biomass and expressed as biovolume (Hopkins and Standke 1992). Data analyses Taxonomic data were analysed at the group level (i.e., chlorophytes, chrysophytes (Chrysophyceae, Synurophyceae, Haptophyceae), cryptophytes, cyanophytes, diatoms, and dinoflagellates) and are expressed and plotted as a percentage of the total phytoplankton biovolume of each sample. The Euglenophyceae, Xanthophyceae, and Raphidophyceae were identified but never exceeded 1% relative abundance and so were excluded from subsequent analyses. Taxa were grouped as mixotrophic or non-mixotrophic using species lists found in Porter (1988), Holen and Boraas (1995), and Olrik (1998). To test for monotonic temporal trends in the relative (percent) biovolume of each algal group and in the percent

biovolume of mixotrophic taxa, Mann–Kendall trend tests were performed separately using a Microsoft Excel add-in developed by Libiseller and Grimvall (2002). Performing multiple Mann–Kendall tests increases the likelihood of erroneously reporting significant trends. Thus, we applied Benjamini and Hochberg’s (1995) correction procedure for the possibility of a false discovery rate (FDR). Further details of the FDR correction are provided in Yan et al. (2008). As chrysophyte and diatom algae are the dominant phytoplankton groups in the study lakes and other studies have shown significant temporal changes in the relative abundance of these groups in Canadian Shield lakes (Leavitt et al. 1999; Vinebrooke et al. 2002; Paterson et al. 2004), we tested for temporal coherence among the study lakes in the percent biovolume of chrysophytes and diatoms through time using Brien’s test (Brien et al. 1984). Beginning with the strongest pairwise correlation, additional lakes were tested sequentially for coherence based on the highest mean correlation to the first two lakes. With the addition of each lake, the resultant interlake correlation matrix was evaluated to determine if all pairwise correlations were equal (i.e., homogenous) and coherent (i.e., the grand mean of the correlation matrix was significantly different from zero). Failure of either of these conditions resulted in rejection of the new lake from the coherent subset. Variance partitioning analysis (VPA; sensu Borcard et al. 1992) was used to explore the relationship between three categories of explanatory variables (physicoclimatic, chemical, and grazers; described below) and phytoplankton assemblages from 1981 to 2001. Separate two-category and (or) three-category analyses were performed for each lake. First, detrended correspondence analysis (DCA) was used to assess the relative length of the gradient representing variation in the phytoplankton data. Based on first axes gradient lengths of less than two standard deviation units in each lake, linear ordination methods were applied in subsequent analyses (ter Braak and Prentice 1998). Second, a series of constrained and partially constrained ordinations (redundancy analysis, RDA) were performed to quantify the temporal phytoplankton variation that can be explained uniquely by each predictor category and by covariation among categories (e.g., Hall et al. 1999). Given the large number of highly correlated explanatory variables in each predictor category, the complexity of the data set was reduced in the following manner: (i) considering each predictor category separately, constrained ordinations with Monte Carlo permutation tests (999 temporally constrained permutations) were performed using Canoco (version 4.0; ter Braak and ?milauer 1998) to identify individual variables that explained significant portions of the phytoplankton variance; and (ii) a manual forwardselection procedure was applied to further reduce colinearity among the significant variables. Prescreening reduced the number of significant explanatory variables from zero to four, depending on the category and lake examined. As the percent variance explained by each category is also sensitive to the number of variables included in the analysis (Borcard et al. 1992), separate PCAs were run for each of the three categories of variables for each lake (e.g., Harp Lake, three separate PCAs with physicoclimatic variables, chemical variables, or biological variables). The first two axes of each PCA were subsequently used as input variables for each cat? 2008 NRC Canada

Paterson et al. Table 1. Sen’s slopes and p values from Mann–Kendall monotonic trend tests of the relative abundance of chrysophyte, diatom, chlorophyte, cyanophyte, cryptophyte, and dinoflagellate taxa identified during the ice-free seasons of 1981 to 2003. Statistics Chrysophytes Diatoms Chlorophytes Cyanophytes Cryptophytes Dinoflagellates Sen’s slope p value Sen’s slope p value Sen’s slope p value Sen’s slope p value Sen’s slope p value Sen’s slope p value (%·year–1) (%·year–1) (%·year–1) (%·year–1) (%·year–1) (%·year–1) BC 2.03 <0.01 –1.81 <0.01 –0.08 0.22 0.01 0.44 –0.36 0.03 0.02 0.31 CB 1.42 0.02 –0.06 0.45 –0.19 0.09 –0.13 0.14 –0.41 0.04 0.04 0.37 CN 2.87 <0.01 –1.21 0.05 –0.31 0.04 –0.25 <0.01 –0.19 0.08 0.05 0.29 DE 1.84 <0.01 –0.54 0.11 –0.13 0.26 –0.02 0.24 –0.35 0.07 –0.33 <0.01 HP 1.92 <0.01 –2.43 <0.01 0.14 0.19 0.07 0.37 0.44 0.01 0.10 0.09 PC 0.96 0.07 –0.10 0.22 –0.04 0.32 0.08 0.03 0.05 0.43 –0.68 0.01

849

RCM 3.28 <0.01 –3.57 <0.01 –0.04 0.38 –0.20 0.04 0.09 0.42 0.11 0.03

Note: Lake codes: BC, Blue Chalk; CB, Chub; CN, Crosson; DE, Dickie; HP, Harp; PC, Plastic; RCM, Red Chalk Main. Bold values indicate significance at P < 0.05 after correcting for the false discovery rate.

egory in the VPA (i.e., maximum of six predictor variables per analysis). No variables in the grazer category were significantly related to phytoplankton variation through time in any of the lakes except Harp, either uniquely or in combination, so two-category VPAs were run. A three-category VPA was subsequently run for Harp Lake that included the grazer category. We made no decision a priori regarding which variables in each predictor category should be included in the VPA analyses and relied on the aforementioned screening process to reduce the variables to an optimal subset. The physicoclimatic variables included seasonal (spring (March–May), summer (June–August), fall (September–November), winter (December–February)) and annual sums of climate indices obtained from the National Oceanic and Atmospheric Administration (NOAA) worldwide database (www.noaa.gov). Other physicoclimatic variables included daily precipitation and temperature data, presented as monthly and annual sums, ice-free season wind speed and radiation (summed), and the number of ice-free and ice-on days. Zero- and 1-year time lags of annual variables, relative to the phytoplankton data, were included in the analyses. The meteorological data used in our analysis were obtained as averages from one to four meteorological stations operated by the Ontario Ministry of the Environment and located at or within 50 km of the Dorset Environmental Science Centre (DESC). Whole-lake, volume-weighted samples for chemical analyses were collected at varying frequencies within and among lakes at weekly to monthly intervals during the ice-free season. Water samples were collected by Ministry of the Environment personnel using standard field techniques (Ingram et al. 2006) and were filtered using an 80 ?m mesh to remove large particles. Laboratory analyses were performed at the DESC laboratory using standard methods (Ontario Ministry of the Environment 1983), and parameters included pH, Gran alkalinity, conductivity, and concentrations of major anions and cations, dissolved organic carbon (DOC), metals (Mn, Al, Fe), and major and minor algal nutrients (total phosphorus, ammonium, nitrate, total Kjeldahl nitrogen, and silicate). To assess the impact of grazers on phytoplankton composition through time, we included the following annual-scale

variables: mean herbivore biomass (dry biomass of herbivorous taxa, ?g·L–1), mean total zooplankton biomass (dry biomass, ?g·L–1), taxon-specific clearance rate (mL·L–1·day–1), total community clearance rate (mL·L–1·day–1), total zooplankton density (number of animals·m–3), total annual species richness (species·year–1), and mean annual cladoceran length (mm; herbivorous taxa excluding Bythotrephes and Polyphemus). All metrics are averages over the ice-free season, although they are referred to as “annual” means in the text. Detailed methodologies regarding sample collection, preservation, processing, enumeration, and analysis of zooplankton are described in Yan et al. (2008).

Results
Significant monotonic changes in the relative biovolume of one or more phytoplankton groups were recorded in all of the study lakes. Most striking was an increase in the percent biovolume of chrysophyte algae at annual time steps from 1981 to 2003 in six of the seven study lakes (Blue Chalk, Chub, Crosson, Dickie, Harp, Red Chalk; Fig. 1; Table 1). Increases in chrysophytes were driven by increases in the percent biovolume of colonial taxa of the genera Synura, Uroglena, Dinobryon, and Chysosphaerella, although the relative importance of the genera varied among lakes (Figs. 3 and 4). Synura increased in relative biovolume in Crosson and Red Chalk lakes, Dinobryon in Blue Chalk and Harp, Uroglena in Harp, and Chrysosphaerella in all lakes, although at lower relative proportions (Fig. 4). There were corresponding increases in the relative biovolume of mixotrophic chrysophytes in five of the lakes (Table 2; Fig. 3). However, this did not necessarily result in an overall increase in mixotrophic taxa in the lakes, as significant increases in the relative biovolume of mixotrophs only occurred in two lakes when all phytoplankton groups were considered (Harp and Red Chalk; Table 2). Furthermore, the relative biovolume of mixotrophs decreased in Plastic Lake over the monitoring record as a result of a significant decline in the relative biovolume of dinoflagellate taxa (Figs. 1 and 3; Table 2). Increases in the relative abundance of chrysophytes were offset primarily by declines in the percent biovolume of dia? 2008 NRC Canada

850

Can. J. Fish. Aquat. Sci. Vol. 65, 2008

Fig. 1. The relative biovolume (%) of major algal groups from 1981 to 2003 in seven intensively studied lakes in central Ontario, Canada.

toms, although significant declines of other algal groups were also observed (diatoms in Blue Chalk, Harp, Red Chalk Main; cyanophytes in Crosson; dinoflagellates in Dickie, Plastic; Table 1). Based on our analyses, increases in chrysophytes were not significant in Plastic Lake. However, chrysophyte algae dominated the assemblage of this lake throughout the monitoring record (biovolume of chrysophytes: range = 28%–81%, mean = 54%). Our analyses were restricted to relative biovolume data. In part, this was to reduce the likelihood that temporal trends in phytoplankton were unduly influenced by short-lived monospecific blooms (e.g., Uroglena bloom in Harp Lake in 2006; Fig. 4). An examination of temporal patterns using absolute biovolume data, however, revealed similar results to
4

relative biovolume (Fig. 2; Supplementary Table S1, available online from the NRC Depository of Unpublished Data4). In five of the six lakes that showed significant increases in percent chrysophytes over time, there were also significant increases in the absolute biovolume of this group. The exception was Harp Lake, suggesting that the relative increase in chrysophyte algae in this lake was driven primarily by a significant decline in diatom biovolume from 1981 to 2003. The phytoplankton assemblages of all of the study lakes showed large interannual variation within groups (Fig. 1). In part, this may reflect the importance of single, bloom events contributing to the annual signal (Fig. 2), many of which are now caused by colonial chrysophyte algae. For example,

Supplementary data for this article are available on the journal Web site (http://cjfas.nrc.ca) or may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research Council Canada, Building M-55, 1200 Montreal Road, Ottawa, ON K1A 0R6, Canada. DUD 3726. For more information on obtaining material refer to http://cisti-icist.nrc-cnrc.gc.ca/cms/unpub_e.html.
? 2008 NRC Canada

Paterson et al. Table 2. Sen’s slopes and p values from Mann–Kendall monotonic trend tests of the relative abundance of mixotrophic taxa (chrysophytes and all groups) identified during the ice-free seasons of 1981–2003. Statistics Mixotrophs (chrysophytes) Mixotrophs (all groups) Sen’s slope (%·year–1) p value Sen’s slope (%·year–1) p value BC 0.55 <0.01 0.55 0.12 CB 0.51 0.09 –0.06 0.43 CN 0.82 0.01 0.49 0.07 DE 0.59 0.04 –0.11 0.32 HP 0.95 <0.01 0.97 <0.01 PC 0.08 0.35 –0.95 <0.01

851

RCM 1.71 <0.01 1.95 <0.01

Note: Lake codes: BC, Blue Chalk; CB, Chub; CN, Crosson; DE, Dickie; HP, Harp; PC, Plastic; RCM, Red Chalk Main. Bold values indicate significance at P < 0.05 after correcting for the false discovery rate.

Fig. 2. The absolute biovolume (mm3·L–1 × 103) of major algal groups from 1981 to 2003 in seven intensively studied lakes in central Ontario, Canada.

peaks in the percent biovolume of some groups are attributable to visible blooms in the plankton of lakes (e.g., surface bloom of the chrysophyte Uroglena, Harp Lake, 1996). Despite this variability, significant temporal coherence was observed in the percent biovolume of chrysophytes in three lakes (Blue Chalk, Red Chalk, and Harp). A repeat of this analysis following the removal of Red Chalk (which is hydrologically connected to Blue Chalk) revealed significant tem-

poral coherence in the percent biovolume of chrysophytes among five of the six remaining lakes (Blue Chalk, Harp, Crosson, Chub, and Dickie). Similarly, temporally coherent declines in the percent biovolume of diatoms were observed in five of the study lakes (Red Chalk, Blue Chalk, Harp, Crosson, and Dickie), suggesting that regional drivers may be important. This observation is supported by paleolimnological evidence that the relative abundance of colonial scaled
? 2008 NRC Canada

852

Can. J. Fish. Aquat. Sci. Vol. 65, 2008

Fig. 3. The relative biovolume (%) of (a) chrysophyte algae and colonial chrysophytes and (b) mixotrophic phytoplankton taxa and mixotrophic chrysophytes from 1981 to 2003 in seven intensively studied lakes in central Ontario, Canada. Lake codes: BC, Blue Chalk; CB, Chub; CN, Crosson; DE, Dickie; HP, Harp; PC, Plastic; RCM, Red Chalk (Main Basin). Data are plotted as three-year running means.

chrysophytes and the ratio of chrysophyte to diatom sedimentary remains have increased since preindustrial times in approximately 50 lakes examined in the study region (Paterson et al. 2004). Using VPA analysis, combinations of water chemistry, physicoclimatic, and biological variables explained significant (P < 0.05) proportions of the variation in the relative biovolume of phytoplankton groups in Blue Chalk, Dickie, Crosson, Harp, and Red Chalk lakes. The total amount of variation explained varied from 48% (Dickie) to 72% (Red Chalk). With the exception of Harp Lake, none of the bio-

logical (zooplankton) variables was significant either uniquely or in combination. In Harp Lake, mean annual cladoceran size was the most significant variable explaining temporal phytoplankton variation, although herbivore biomass and total zooplankton clearance rates were also significant. The subsets of environmental variables selected for the two- or three-category VPAs were different for each of the lakes (Table 3). Calcium concentration was the water chemistry variable explaining most of the variance in the Blue Chalk and Red Chalk phytoplankton, whereas alkalinity, conductivity, and total phosphorus concentrations were the
? 2008 NRC Canada

Paterson et al.

853

Fig. 4. The relative biovolume (%) of chrysophyte algae organized by genus, including Uroglena, Dinobryon, Synura, Chrysosphaerella, and all other genera, from 1981 to 2003 in seven intensively studied lakes in central Ontario, Canada. Lake codes: BC, Blue Chalk; CB, Chub; CN, Crosson; DE, Dickie; HP, Harp; PC, Plastic; RCM, Red Chalk (Main Basin).

? 2008 NRC Canada

854

Can. J. Fish. Aquat. Sci. Vol. 65, 2008

Table 3. Environmental variables selected to generate principal components analysis axes 1 and 2 scores used in the two- and three-category variance partitioning analysis of phytoplankton groups for Blue Chalk, Crosson, Dickie, Harp, and Red Chalk (Main Basin) lakes. Blue Chalk Water chemistry Ca DOC SO4 NH3–NH4 ns Crosson Alkalinity NO3 Si ns Dickie Conductivity NO3 Si ns Harp Ca TP Fe Herbivore biomass; zooplankton clearance rate; mean annual cladoceran size Annual temperature NAO fall Red Chalk Ca Fe

Zooplankton

ns

Climate

Spring precipitation NAO annual NAO fall

Winter temperature AO annual (1-year lag) Ice-off date (1-year lag)

Summer temperature Annual precipitation (1-year lag)

Winter temperature NAO fall

Note: Precipitation, in mm; air temperature, in °C; NAO, North Atlantic Oscillation (annual or seasonal sum); AO, Arctic Oscillation (annual or seasonal sum); ns, no variables in this category were significant in explaining variation in phytoplankton data through time, based on Monte Carlo permutation tests.

most important chemical variables in Crosson, Dickie, and Harp lakes, respectively. Nutrient concentrations and silica also explained significant proportions of the phytoplankton variance in Crosson and Dickie lakes, and iron was significant in Harp and Red Chalk lakes (Table 3). The significant climate variables selected varied among lakes and included measures of precipitation (spring and annual precipitation with a 1-year lag), air temperature (winter temperature and annual temperature), and variations of the North Atlantic Oscillation (NAO) and Arctic Oscillation (AO) signals (Table 3). The most common climatic variable identified using VPA was fall NAO, which was the most significant climatic variable in Blue Chalk, Harp, and Red Chalk lakes (Table 3), showing a significant decline from 1981 to 2000. RDA biplots of the five lakes analyzed using VPA (Blue Chalk, Dickie, Crosson, Harp, and Red Chalk lakes) revealed strong primary gradients when constrained to the significant water chemistry and physicoclimatic variables identified in the prescreening process (λ = 0.52 to 0.68). Although predictor variables differed among the lakes, years with a high percent biovolume of chrysophyte taxa were generally higher in DOC concentration (Blue Chalk), Gran alkalinity (Crosson), and (or) conductivity (Dickie) and lower in concentrations of total phosphorus (Harp), iron (Harp, Red Chalk), and silica and (or) calcium (Blue Chalk, Harp, Red Chalk) (Fig. 5). Higher percent chrysophytes were also associated with warmer winter temperatures and lower NAO values (Fig. 5). In the two-category VPAs, the shared variance explained by water chemistry and physicoclimatic variables explained 21%–30% (mean = 25.5%) of the total variance in the phytoplankton group-level data for each of the lakes (Figs. 6, 7b). These results suggest that water chemistry as influenced by climatic factors is important in the structuring of the phytoplankton communities of these lakes at the group level. Water chemistry independent of climate explained similar portions of the phytoplankton variance (14%–47%, mean = 28.7%). A smaller portion of the phytoplankton variance was explained by physicoclimatic variables, independent of water chemistry (3%–7%, mean = 5.4%).

Biological variables were included in a three-category VPA of Harp Lake (Fig. 7a). The total variance of the phytoplankton data explained by all predictor categories was 69%. The most significant predictor categories were the covariation of biological and water chemistry variables, explaining 32% of the phytoplankton variance, and the complex covariation of biological, water chemistry, and physicoclimatic variables, explaining 21% of the variation (Fig. 7b). The total variance explained by the predictor variables in Harp Lake dropped from 69% to 62% when biological variables were excluded from the VPA analysis (i.e., three-way vs. two-way VPA) (Figs. 7a, 7b).

Discussion
The strongest long-term trend in the phytoplankton we observed was a rise in the percent biovolume of colonial chrysophyte algae, increasing significantly from 1981 to 2003 in six of seven lakes. In three lakes (Blue Chalk, Harp, Red Chalk), the shift to chrysophytes coincided with significant monotonic declines in the percent biovolume of diatoms. The increase in chrysophytes was also observed in absolute biovolume in five of the lakes and thus represents a quantifiable change in the structure of the phytoplankton. A switch from epilimnetic diatoms to metalimnetic, colonial chrysophytes may represent a shift in vertical structure of algae in the water column. Large colonial chrysophytes are highly motile and, with notable exceptions such as Uroglena species that may form surface blooms, commonly form short-lived, monospecific blooms at or below the thermocline in stratified lakes (Fee et al. 1977; Nakamoto et al. 1983). As a group, chrysophytes dominate the phytoplankton assemblages of lakes characterized by low productivity, alkalinity, and conductivity and neutral to slightly acidic pH (Sandgren 1988). The relative importance of chrysophyte algae in lakes increases as ionic content decreases, although chrysophytes are broadly tolerant of concentrations of individual ions (e.g., calcium, magnesium, sodium, and chloride) (Sandgren 1988). Neither the concentration of individual ions nor total ionic strength appears to directly restrict chrysophyte
? 2008 NRC Canada

Paterson et al.

855

Fig. 5. Redundancy analysis ordination plots showing the position of site years in relation to (a) forward-selected environmental variables and (b) major phytoplankton groups. Lake codes: BC, Blue Chalk; CN, Crosson; DE, Dickie; HP, Harp; RCM, Red Chalk (Main Basin). Algal groups: Chry, chrysophytes; Diat, diatoms; Chlo, chlorophytes; Cyan, cyanophytes; Cryp, cryptophytes; Dino, dinoflagellates. Environmental variables: NAO, North Atlantic Oscillation; AO, Arctic Oscillation; Ice-off, date of ice-off; T°, air temperature; Prec, precipitation; NH4, ammonium; Ca, calcium; DOC, dissolved organic carbon; SO4, sulphate; SiO3, silicate; Alk, Gran alkalinity; NO3, nitrate; COND, conductivity; Fe, iron; TP, total phosphorus.

? 2008 NRC Canada

856 Fig. 5 (concluded).

Can. J. Fish. Aquat. Sci. Vol. 65, 2008

growth. However, these group-level ecological characteristics are derived from studies of lakes that span broad water chemistry gradients (e.g., oligotrophic to eutrophic lakes). They provide a general overview of the environmental preferences of chrysophytes relative to other algal groups but are of limited value for evaluating specific causes of the recent changes in phytoplankton composition in the study lakes. In a more detailed analysis, VPA showed similar results at an annual time scale among the study lakes. In four lakes (Blue Chalk, Crosson, Dickie, Red Chalk Main), water chemistry variables, uniquely and in covariation with physicoclimatic variables, were the most important predictors of phytoplankton variance through time. Furthermore, within each predictor category, similar or correlated variables were important in explaining variation in phytoplankton composition among lakes. These results, and the observed temporal coherence among lakes in changes in chrysophyte and diatom biovolume, suggest that regional drivers are important in structuring phytoplankton populations in these lakes. This conclusion is supported by paleoecological evidence in this geographic region. In a study of ~50 Canadian Shield lakes, Paterson et al. (2004) reported increases in the relative abundance of colonial chrysophytes since preindustrial times in more than 90% of the lakes examined, including lakes with and without pressure from shoreline development. Detailed stratigraphic analyses of two lakes (including Crosson Lake of this study) indicated that initial increases in colonial chrysophytes were coincident with the

increased industrial activity and acidic deposition observed in the mid-1900s in central Ontario (Paterson et al. 2004). Calcium (Ca) was the water chemistry variable most strongly correlated (negatively) to phytoplankton change in the three lakes with the highest increases in chrysophyte algae and (or) decreases in diatoms (Blue Chalk, Harp, and Red Chalk). However, we do not believe that chrysophytes are increasing as a direct consequence of declining Ca concentrations. Paleoecological evidence from the study lakes (C.S. Chueng, A.M Paterson, and J.P. Smol, Paleoecological Environmental Assessment and Research Laboratory, Department of Biology, Queen’s University, Kingston, Ontario, unpublished data) shows that the rise in colonial chrysophytes began in the mid-1900s and was, in fact, prominent during a period of increasing Ca concentrations in the study lakes (Watmough and Aherne 2008). The accelerated increases in chrysophytes (Paterson et al. 2004), and declines in lake Ca concentrations over the past two decades (Watmough and Aherne 2008), are indications that rapid chemical and ecological changes are now occurring in Canadian Shield lakes as a result of decades of industrial activity. A rise in importance of metalimnetic flagellates, such as colonial chrysophytes, has also been observed in experimentally and atmospherically acidified lakes on the Canadian Shield during the early stages of acidification (Schindler et al. 1991). Increases in these groups are often coincident with declines in the abundance of small, epilimnetic taxa such as
? 2008 NRC Canada

Paterson et al. Fig. 6. Results of two-category variance partitioning analysis (VPA) displaying the amount of variation in the phytoplankton data explained by the unique and shared effects of water chemistry and physicoclimatic variables in Blue Chalk, Crosson, Dickie, and Red Chalk lakes from 1981 to 2001: solid bars, variance explained by water chemistry variables; hatched bars, physicoclimatic variables; shaded bars, first-order interaction of water chemistry and physicoclimatic variables; open bars, unexplained variance.

857 Fig. 7. Results from (a) three-category and (b) two-category variance partitioning analyses (VPA) displaying the amount of variation in the phytoplankton data explained by the unique and shared effects of biological (zooplankton), water chemistry, and physicoclimatic variables in Harp Lake from 1981 to 2001. Bars: horizontally hatched bars, variance explained by biological variables; solid bars, water chemistry variables; diagonally hatched bars, physicoclimatic variables; lightly shaded bars, first-order interaction of biological and water chemistry variables; cross-hatched bars, first-order interaction of biological and physicoclimatic variables; darker shaded bars, first-order interaction of water chemistry and physicoclimatic variables; vertically hatched bars, second-order interaction of biological, water chemistry, and physicoclimatic variables; open bars, unexplained variance.

planktonic diatoms or unicellular chrysophytes (Findlay and Kasian 1986; Schindler et al. 1991; Vinebrooke et al. 2002). In other studies, these early taxonomic shifts have been attributed to variation in species- or group-level tolerances to pH (Paterson et al. 2001), indirect responses to changes in grazing pressure (Sandgren and Walton 1995), or increased transparency associated with declining concentrations of DOC (Leavitt et al. 1999). Under such conditions, deepwater phytoflagellates may have a selective advantage with a reduced exposure to harmful ultraviolet radiation (UVR), a reduction in grazing pressure, and proximity to nutrient-rich hypolimnetic waters (Leavitt et al. 1999; Vinebrooke et al. 2002; Paterson et al. 2004). However, in the three study lakes that show the greatest change in phytoplankton (Blue Chalk, Harp, Red Chalk), pH and DOC have either shown no change or increased significantly over the past two decades. Moreover, in six of the seven study lakes, grazing variables did not explain a significant portion of the phytoplankton variation through time. Thus, the aforementioned mechanisms do not provide fully satisfactory explanations. It is possible that indirect effects associated with industrial activity, such as the deposition of trace metals, may be important. Davis et al. (2006) reported a major shift from unicellular to colonial chrysophytes in two New Hampshire lakes that was correlated to increases in coal combustion products and major and trace metals in lake sediments. They suggest that pathways for metal enrichment may vary from direct erosional inputs associated with logging and watershed development to aerial inputs of trace metals in a more remote mountain lake (Davis et al. 2006). The importance of metals such as iron as a potential limiting nutrient for some

freshwater algae has been reported in other studies (references in Sandgren 1988). The high abundances of some chrysophyte taxa in highly coloured lakes, for example, may be related to readily available, chelated forms of iron or other trace metals in such lakes. The deposition of selenium (Se), a product of coal-fired power plants, may also be an important trace metal in central Ontario lakes. Wehr and Brown (1985) demonstrated that a simulated Se spike resulted in a significant increase in algal growth, using water from Dickie Lake, one of our study sites. They also reported that the bloom-forming haptophyte Chrsyochromulina breviturrita could not be maintained in culture without Se. Although long-term data do not exist for Se, and many of the trace metals are observed at levels approaching analytical detection in the study lakes, these variables may play an important role in explaining phytoplankton trends in Shield lakes and should be included in future monitoring and paleolimnological initiatives with efforts to improve the analytical precision. As suggested by the strong covariation of water chemistry and physicoclimatic variables in explaining phytoplankton variance in the study lakes, multiple stressors may also be important (Schindler 1998). The synergistic effects of acid deposition and drought, in particular, may lead to long-term
? 2008 NRC Canada

858

Can. J. Fish. Aquat. Sci. Vol. 65, 2008

declines in nutrient concentrations in lakes (e.g., total phosphorus (TP), silica) with possible measurable shifts in phytoplankton assemblages (Findlay et al. 1999). Phytoplankton that form deepwater populations and mixotrophic genera such as Dinobryon, Uroglena, and Chrysosphaerella that have increased in relative biovolume in our study lakes may have a selective advantage. Their ability to switch between autotrophy, heterotrophy, and phagotrophy in response to cell requirements and environmental conditions (Sandgren 1988), together with their ability to exist proximal to the nutrient-rich hypolimnion, may be particularly advantageous when resources are limited by drought (Findlay et al. 1999). Although we observed an increase in the percent biovolume of mixotrophic chrysophytes in five of the study lakes, the percent biovolume of mixotrophs across all algal groups only increased significantly in two lakes (Harp and Red Chalk) and decreased significantly in Plastic Lake. In part, this apparent discrepancy may be explained by recovery trends from acidification in some lakes. Mixotrophic dinoflagellates declined significantly in Dickie and Plastic lakes as pH values increased above 6, offsetting increases in mixotrophic chrysophytes and consistent with patterns observed in whole-lake experiments in other studies (Findlay and Kasian 1996; Graham et al. 2007). Independent of water chemistry changes, recent studies suggest that climate may also be a significant driver structuring algal populations in north-temperate and subarctic lakes. Physical changes associated with climate such as an increase in epilimnetic temperatures and increases in thermal stability and the duration of the ice-free season in thermally stratified systems (Magnuson et al. 1997; Sorvari et al. 2002; Rühland et al. 2003) have been related to shifts in phytoplankton composition. The NAO, in particular, has been linked to changes in the timing, composition, and intensity of freshwater phytoplankton blooms in north-temperate lakes (Gerten and Adrian 2002). In our study, however, variation in phytoplankton data that could be explained uniquely by physicoclimatic variables was small (3%–7%). In part, this is because some climatic variables (NAO fall) are changing in concert with chemical variables and are expressed as covariates in the VPA. Other climatic variables are changing in a non-monotonic fashion, despite significant linear trends in the phytoplankton of some lakes. Finally, we included time lags of 1 year for climatic variables, which may not be sufficient to detect significant changes in some ecological variables, particularly if effects are channeled through indirect pathways (e.g., effects of climate warming on grazer communities). Despite significant changes in mean annual cladoceran size in six of the seven study lakes (Yan et al. 2008), grazing variables only explained a significant portion of the phytoplankton variance in one lake at an annual time scale. Similar to other studies in oligotrophic lakes, this poor relationship may be explained by the increasing dominance of large, colonial chrysophytes in these lakes and (or) by zooplankton communities that are too sparse to exert a significant impact on phytoplankton assemblages (Ramcharan et al. 1995; Tremel et al. 2001; Christensen et al. 2006). The dominant phytoplankton genera in the study lakes (Chrysosphaerella, Dinobryon, Synura, and Uroglena; 43% to 94%

of the algal biovolume in the study lakes in 2002) form large, motile colonies that may only be edible to large Daphnia (>2 mm; Sandgren and Walton 1995) that are not common in the study lakes (Yan et al. 2008). Furthermore, many of these genera are associated with deepwater peaks in the metalimnion or upper hypolimnion of dimictic lakes (Sandgren 1988; Nicholls 1995), providing them with a spatial refugium from epilimnetic grazers. In Harp Lake, a significant portion of the phytoplankton variance was explained by three grazer variables (mean annual cladoceran size, herbivore biomass, and zooplankton community clearance rate), uniquely or in covariation with water chemistry and physicoclimatic variables. It is possible that the recent invasion by Bythotrephes has resulted in significant alterations to the food web with indirect effects on phytoplankton biomass and composition. In a detailed historical analysis of rotifer communities in Harp Lake and a reference system (Red Chalk Lake), Hovius et al. (2006) hypothesized that coherent peaks in Conchilus rotifer abundance and chlorophyll a concentrations may be the result of a release of grazer community control over phytoplankton biomass with Bythotrephes invasion. This may be further supported by a rise in the percent abundance of large, colonial chrysophyte taxa that may be less favourable to sizeselective herbivores such as Conchilus (Armengol et al. 2001). Interestingly, an a posteriori removal of grazers from the VPA analysis of Harp Lake shows patterns of explained variance that are similar to those of noninvaded lakes. Thus, it is also possible that the covariation explained by grazers and other predictor categories is spurious. With similar grouplevel changes occurring coherently in noninvaded lakes (Blue Chalk, Red Chalk), further study will be required to determine whether the VPA results reflect a true cause and effect of Bythotrephes invasion on phytoplankton species composition, biomass, and (or) size structure. Our analyses of long-term changes in the phytoplankton of seven, intensively studied lakes on the southern Canadian Shield indicate that phytoplankton composition has changed significantly over the past ~20 years. We observed a rise in the percent biovolume of chrysophytes in six of the seven lakes, accompanied by a decline in the importance of diatoms in three of these lakes. Significant temporal coherence among sites suggests that these species trends are a result of regional environmental stressors. A partitioning of the variance indicated that long-term changes in biological variables (i.e., grazers) could not explain a significant portion of the phytoplankton variation through time, with the possible exception of Harp Lake, which was invaded by the spiny water flea in the early 1990s. The largest portion of the phytoplankton variation was explained by a combination of water chemistry changes associated with industrial activity and physical changes linked to climate indices such as the North Atlantic Oscillation. Finally, it is worth noting that a relatively large portion of the phytoplankton variance remained unexplained in our models (28% to 52%), suggesting that key predictor variables were excluded from our analyses. It is also possible, however, that VPA does not adequately capture complex, nonadditive interactions between predictor variables as they influence biological assemblages (Christensen
? 2008 NRC Canada

Paterson et al.

859

et al. 2006). Future work will be required to tease apart this complexity and to identify the specific causes of these trends. This work should include a variety of modeling approaches and experimental and monitoring studies and should be conducted at multiple spatial, temporal, and taxonomic scales to improve our understanding of the factors that control phytoplankton assemblages in Canadian Shield lakes. Implications The long-term changes in phytoplankton composition in the study lakes may have serious consequences for water quality. For example, a rise in the relative abundance of bloom-forming, taste-and-odour taxa, such as Synura petersenii and Uroglena species (Nicholls and Gerrath 1985; Paterson et al. 2004), may result in an increased frequency of deleterious blooms. A rise in the importance of colonial chrysophytes and (or) mixotrophic taxa also may represent a shift in the structure and function of phytoplankton in the study lakes. As mentioned previously, large colonial phytoplankton, including members of the genera Chrysosphaerella, Dinobryon, Synura, and Uroglena, form large, motile colonies that likely are inedible to many zooplankton (Sandgren and Walton 1995). Their ability to maintain subthermocline populations in stratified lakes during the summer and their enhanced motility may also provide these taxa with a spatial refugium from epilimnetic grazers (Sandgren 1988; Nicholls 1995). These changes to the spatial and, possibly, seasonal structure of phytoplankton in Canadian Shield lakes have unknown impacts at higher trophic levels. Although we do not report on changes to absolute biovolume in this study, we have provided supplementary data4 showing that the observed changes in percent biovolume do translate into changes in algal biomass (as biovolume) at the group level. Significant changes in phytoplankton biomass may directly impact grazers, with possible implications for people who rely on inland fisheries for subsistence and sportfishing (Dillon and Molot 2005). Our results are similar to those reported in regional paleoecological studies (Dixit et al. 2002; Paterson et al. 2004), suggesting that these patterns are not an artifact of sampling or laboratory methodology, e.g., changes in field sampling frequency through the monitoring period. The paleoecological data also indicate that phytoplankton changes are widespread across the southern Canadian Shield and likely began before the collection of Dorset monitoring data (Paterson et al. 2004). Thus, interpretations of water quality and biological change should be placed into appropriate temporal contexts, with an understanding that the study lakes were significantly impacted by anthropogenic stressors prior to the onset of contemporary monitoring records (Quinlan et al. 2008).

thank past and present staff and students at DESC for field sampling support, chemical laboratory analysis, and database management. We also thank Lucja Heintsch, Lynda Nakamoto, and Sue Standke of MOE’s phytoplankton laboratory and members of the Phyto-Tec Group, who were involved with phytoplankton processing and enumeration over the monitoring period. We acknowledge Peter Dillon, former manager and current partner of DESC, for his vision and support of Dorset’s long-term monitoring program.

References
Aherne, J., Posch, M., Dillon, P.J., and Henriksen, A. 2004. Critical loads of acidity for surface waters in south-central Ontario, Canada: regional application of the first-order acidity balance (FAB) model. WASP, 4: 25–36. Armengol, X., Boronat, L., Camacho, A., and Wurtsbaugh, W.A. 2001. Grazing by a dominant rotifer Conochilus unicornis Rousselet in a mountain lake: in situ measurements with synthetic microspheres. Hydrobiologia, 446/447: 107–114. Benjamini, Y., and Hochberg, Y. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. B Methodol. 57: 289–300. Borcard, D., Legendre, P., and Drapeau, P. 1992. Partialling out the spatial component of ecological variation. Ecology, 73: 1045–1055. Bowman, M.F., Ingram, R., Reid, R.A., Somers, K.M., Yan, N.D., Paterson, A.M., Morgan, G.E., and Gunn, J.M. 2008. Temporal and spatial concordance in community composition of lake phytoplankton, zooplankton, macroinvertebrate, crayfish, and fish on the Precambrian Shield. Can. J. Fish. Aquat. Sci. 65: 919–932. Brien, C.J., Venables, W.N., James, A.T., and Mayo, O. 1984. An analysis of correlation matrices: equal correlations. Biometrika, 71: 545–554. Christensen, M.R., Graham, M.D., Vinebrooke, R.D., Findlay, D.L., Paterson, M.J., and Turner, M.A. 2006. Multiple anthropogenic stressors cause ecological surprises in boreal lakes. Global Change Biol. 12: 2316–2322. Davis, R.B., Anderson, D.S., Dixit, S.S., Appleby, P.G., and Schauffler, M. 2006. Responses of two New Hampshire (USA) lakes to human impacts in recent centuries. J. Paleolimnol. 35: 669–697. Dillon, P.J., and Molot, L.A. 1996. Long-term phosphorus budgets and an examination of a steady-state mass balance model for central Ontario lakes. Water Res. 30: 2273–2280. Dillon, P.J., and Molot, L.A. 2005. Long-term trends in catchment export and lake retention of dissolved organic carbon, dissolved organic nitrogen, total iron, and total phosphorus: the Dorset, Ontario, study, 1978–1998. J. Geophys. Res. 110: G01002, doi:10.1029/2004JG000003. Dillon, P.J., and Rigler, F.H. 1974. The phosphorus–chlorophyll relationship in lakes. Limnol. Oceanogr. 19: 767–773. Dillon, P.J., Molot, L.A., and Futter, M. 1997. The effect of El Nino-related drought on the recovery of acidified lakes. Environ. Monit. Assess. 46: 105–111. Dixit, S.S., Dixit, A.S., and Smol, J.P. 2002. Diatom and chrysophyte transfer functions and inferences of post-industrial acidification and recent recovery trends in Killarney lakes (Ontario, Canada). J. Paleolimnol. 27: 79–96. Eimers, M.C., Buttle, J., and Watmough, S.A. 2008. Influence of seasonal changes in runoff and extreme events on dissolved organic carbon concentrations in wetland- and upland-draining streams. Can. J. Fish. Aquat. Sci. 65: 796–808.
? 2008 NRC Canada

Acknowledgements
This work would not have been possible without the commitment and dedication of all technical and scientific staff at the Dorset Environmental Science Centre (DESC) and within the Ontario Ministry of the Environment (MOE). We

860 European Union. 2000. Directive 2000/60/EC of the European Parliament and the Council of 23 October 2000 on establishing a framework for community action in the field of water policy. J. Eur. Commun. L327: 1–72. Fee, E.J., Shearer, J.A., and DeClercq, D.R. 1977. In vivo chlorophyll profiles from lakes in the Experimental Lakes Area, northwestern Ontario. Department of the Environment, Freshwater Institute, Winnipeg, Manitoba, Canada, Technical Report No. 703. Findlay, D.L., and Kasian, S.E.M. 1986. Phytoplankton community responses to acidification of Lake 223, Experimental Lakes Area, northwestern Ontario. WASP, 30: 719–726. Findlay, D.L., and Kasian, S.E.M. 1996. The effect of incremental pH recovery on the Lake 223 phytoplankton community. Can. J. Fish. Aquat. Sci. 53: 856–864. Findlay, D.L., Kasian, S.E.M., Turner, M.T., and Stainton, M.P. 1999. Comparison of responses of phytoplankton and epilithon during acidification and early recovery. Freshw. Biol. 42: 159–175. Findlay, D.L., Kasian, S.E.M., Stainton, M.P., Beaty, K., and Lyng, M. 2001. Climatic influences on algal populations of boreal forest lakes in the Experimental Lakes Area. Limnol. Oceanogr. 46: 1784–1793. George, D.G., Maberly, S.C., and Hewitt, D.P. 2004. The influence of the North Atlantic Oscillation on the physical, chemical and biological characteristics of four lakes in the English Lake District. Freshw. Biol. 49: 760–774. Gerten, D., and Adrian, R. 2002. Effects of climate warming, North Atlantic Oscillation, and El Ni?o – Southern Oscillation on thermal conditions and plankton dynamics in northern hemispheric lakes. Scientific World, 2: 586–202. Graham, M.D., Vinebrooke, R.D., Keller, B., Heneberry, J., Nicholls, K.H., and Findlay, D.L. 2007. Comparative responses of phytoplankton during chemical recovery in atmospherically and experimentally acidified lakes. J. Phycol. 43: 908–923. Hall, R.I., and Smol, J.P. 1996. Paleolimnological assessment of long-term water-quality changes in south-central Ontario lakes affected by cottage development and acidification. Can. J. Fish. Aquat. Sci. 53: 1–17. Hall, R.I., Leavitt, P.R., Quinlan, R., Dixit, A.S., and Smol, J.P. 1999. Effects of agriculture, urbanization, and climate on water quality in the northern Great Plains. Limnol. Oceanogr. 44: 739–756. Holen, D.A., and Boraas, M.E. 1995. Mixotrophy in chrysophytes. In Chrysophyte algae: ecology, phylogeny and development. Edited by C.D. Sandgren, J.P. Smol, and J. Kristiansen. Cambridge University Press, Cambridge, UK. pp. 119–140. Hopkins, G.J, and Standke, S.J. 1992. Phytoplankton methods manual: with special emphasis on waterworks operation internal methods manual. Queen’s Printer for Ontario, Toronto, Ontario. ISBN 0-7729-8923-0. Hovius, J.T., Beisner, B.E., and McCann, K.S. 2006. Epilimnetic rotifer community responses to Bythotrephes longimanus invasion in Canadian Shield lakes. Limnol. Oceanogr. 51(2): 1004–1012. Ingram, R.G., Girard, R.E., Clark, B.J., Paterson, A.M., Reid, R.A., and Findeis, J.G. 2006. Dorset Environmental Science Centre: lake sampling methods. Queen’s Printer for Ontario, Toronto, Ontario. ISBN 1-4249-2049-3. Jeffries, D.S., and Snyder, W.R. 1983. Geology and geochemistry of the Muskoka–Haliburton study area. Queen’s Printer for Ontario, Toronto, Ontario, Ontario Ministry of the Environment Data Report No. DR 83/2. Keller, W., Paterson, A.M., Somers, K.M., Dillon, P.J., Heneberry, J., and Ford, A. 2008. Relationships between dissolved organic carbon concentrations, weather, and acidification in small Boreal Shield lakes. Can. J. Fish. Aquat. Sci. 65: 786–795.

Can. J. Fish. Aquat. Sci. Vol. 65, 2008 Leavitt, P.R., Findlay, D.L., Hall, R.I., and Smol, J.P. 1999. Algal responses to dissolved organic carbon loss and pH decline during whole lake acidification: evidence from paleolimnology. Limnol. Oceanogr. 44: 757–773. Libiseller, C., and Grimvall, A. 2002. Performance of partial Mann– Kendall test for trend detection in the presence of covariates. Environmetrics, 13: 71–84. Magnuson, J.J., Webster, K.E., Assel, R.A., Bowser, C.J., Dillon, P.J., Eaton, J.G., Evans, H.E., Fee, E.J., Hall, R.I., Mortsch, L.R., Schindler, D.W., and Quinn, F.H. 1997. Potential effects of climate changes on aquatic ecosystems: Laurentian Great Lakes and Precambrian Shield region. Hydrol. Proc. 11: 825–871. Nakamoto, L., Heintsch, L., and Nicholls, K. 1983. Phytoplankton of lakes in the Muskoka–Haliburton area. Queen’s Printer for Ontario, Toronto, Ontario, Ontario Ministry of the Environment Data Report No. DR 83/8. Nicholls, K.H. 1995. Chrysophyte blooms in the plankton and neuston. In Chrysophyte algae: ecology, phylogeny and development. Edited by C.D. Sandgren, J.P. Smol, and J. Kristiansen. Cambridge University Press, Cambridge, UK. pp. 181–213. Nicholls, K.H., and Gerrath, J.F. 1985. The taxonomy of Synura (Chrysophyceae) in Ontario with special reference to taste and odour in water supplies. Can. J. Bot. 63: 1482–1493. Nicholls, K.H., Nakamoto, L., and Keller, W. 1992. Phytoplankton of Sudbury area lakes (Ontario) and relationships with acidification status. Can. J. Fish. Aquat. Sci. 49: 40–57. Olrik, K. 1998. Ecology of mixotrophic flagellates with special reference to Chrysophyceae in Danish lakes. Hydrobiologia, 369/370: 329–338. Ontario Ministry of the Environment. 1983. Handbook of analytical methods for environmental samples. Vols. 1 and 2. Laboratory Services Branch, Ontario Ministry of the Environment and Energy, Sudbury, Ontario. Paterson, A.M., Cumming, B.F., Smol, J.P., and Hall, R.I. 2001. Scaled chrysophytes as indicators of water quality changes since preindustrial times in the Muskoka–Haliburton region, Ontario, Canada. Can. J. Fish. Aquat. Sci. 58: 2468–2481. Paterson, A.M., Cumming, B.F., Smol, J.P., and Hall, R.I. 2004. Marked recent increases of colonial scaled chrysophytes in boreal lakes: implications for the management of taste and odour events. Freshw. Biol. 49: 199–207. Planas, D., Desrosiers, M., Groulx, S-R., Paquet, S., and Carignan, R. 2000. Pelagic and benthic algal responses in eastern Canadian Boreal Shield lakes following harvesting and wildfires. Can. J. Fish. Aquat. Sci. 57(Suppl. 2): 136–145 Porter, K.G. 1988. Phagotrophic phytoflagellates in microbial food webs. Hydrobiologia, 159: 89–97. Quinlan, R., Hall, R.I., Paterson, A.M., Cumming, B.F., and Smol, J.P. 2008. Long-term assessments of ecological effects of anthropogenic stressors on aquatic ecosystems from paleoecological analyses: challenges to traditional perspectives of lake management. Can. J. Fish. Aquat. Sci. 65: 933–944. Ramcharan, C.W., McQueen, D.J., Demers, E., Popiel, S.A., Rocchi, A.M., Yan, N.D., Wong, A.H., and Highes, K.D. 1995. A comparative approach to determining the role of fish predation in structuring limnetic ecosystems. Arch. Hydrobiol. 133: 389–416. Reynolds, C.S. 1984. Phytoplankton periodicity: the interaction of form, function and environmental variability. Freshw. Biol. 14: 111–142. Reynolds, C.S. 1998. What factors influence the species composition of phytoplankton in lakes of different trophic status. Hydrobiologia, 369/370: 11–26.
? 2008 NRC Canada

Paterson et al. Rühland, K., Priesnitz, A., and Smol, J.P. 2003. Paleolimnological evidence from diatoms for recent environmental changes in 50 lakes across Canadian Arctic treeline. Arctic, Antarctic Alpine Res. 35: 110–123. Rusak, J.A., Yan, N.D., Somers, K.M., Cottingham, K.L., Micheli, F., Carpenter, S.R., Frost, T.M., Paterson, M.J., and McQueen, D.J. 2002. Temporal, spatial and taxonomic patterns of crustacean zooplankton variability in unmanipulated north-temperate lakes. Limnol. Oceanogr. 47: 613–625. Rusak, J.A., Yan, N.D., and Somers, K.M. 2008. Regional climatic drivers of synchronous zooplankton dynamics in north-temperate lakes. Can. J. Fish. Aquat. Sci. 65: 878–889. Sandgren, C.D. (Editor). 1988. The ecology of chrysophyte flagellates: their growth and perennation strategies as freshwater phytoplankton. In Growth and reproductive strategies of freshwater phytoplankton. Cambridge University Press, Cambridge, UK. pp. 9–104. Sandgren, C.D., and Walton, W.E. 1995. The influence of zooplankton herbivory on the biogeography of chrysophyte algae. In Chrysophyte algae: ecology, phylogeny and development. Edited by C.D. Sandgren, J.P. Smol, and J. Kristiansen. Cambridge University Press, New York. pp. 269–302. Schindler, D.W. 1978. Factors regulating phytoplankton production and standing crop in the world’s freshwaters. Limnol. Oceanogr. 23: 478–486. Schindler, D.W. 1998. A dim future for boreal waters and landscapes. Bioscience, 48: 157–164. Schindler, D.W., Frost, T.M., Mills, K.H., Chang, S.S., Davies, I.J., Findlay, D.F., Malley, D.F., Shearer, J.A., Turner, M.A., Garrison, P.J., Watras, C.J., Webster, K., Gunn, J.M., Brezonik, P.L., and Swenson, W.A. 1991. Comparisons between experimentally and atmospherically acidified lakes during stress and recovery. Proc. R. Soc. Edinb. Sect. B, 97: 193–226. Smith, J.V., Lavender, B., Auld, H., Broadhurst, D., and Bullock, T. 1998. Adapting to climate variability and change in Ontario. In The Canada Country Study: climate impacts and adaptation. Vol. IV. Environment Canada, Ontario Region, Toronto, Ontario. Smol, J.P. 2002. Pollution of lakes and rivers: a paleoenvironmental perspective. Oxford University Press Inc., New York. Sorvari, S., Korhola, A., and Thompson, R. 2002. Lake diatom response to recent Arctic warming in Finnish Lapland. Global Change Biol. 8: 153–163. Stevenson, R.J., and Smol, J.P. 2003. Use of algae in environmental assessments. In Freshwater algae of North America: ecology and classification. Edited by J.D. Wehr and R.G. Sheath. Academic Press, Amsterdam, the Netherlands. pp. 775–804.

861 Stoddard, J.L., Jeffries, D.S., Lukewille, A., Clair, T.A., Dillon, P.J., Driscoll, C.T., Forsius, M., Johannessen, M., Kahl, J.S., Kellogg, J.H., Kemp, A., Mannio, J., Monteith, D.T., Murdoch, P.S., Patrick, S., Rebsdorf, A., Skjelkvale, B.L., Stainton, M.P., Traaen, T., van Dam, H., Webster, K.E., Wieting, J., and Wilander, A. 1999. Regional trends in aquatic recovery from acidification in North America and Europe. Nature (London), 401: 575–578. ter Braak, C.J.F., and Prentice, I.C. 1998. A theory of gradient analysis. Adv. Ecol. Res. 18: 271–317. ter Braak, C.J.F., and ?milauer, P. 1998. Canoco reference manual and user’s guide to Canoco for Windows: software for canonical community ordination. Version 4. Microcomputer Power, Ithaca, N.Y. Tremel, B., Nicholls, K.H., McQueen, D.J., Ramcharan, C.W., and Pérez-Fuentetaja, A. 2001. Did phytoplankton biovolume and taxonomic composition change? Arch. Hydrobiol. Spec. Issues Adv. Limnol. 56: 187–209. Vinebrooke, R.D., Dixit, S.S., Graham, M.D., Gunn, J.M., Chen, Y-W., and Belzile, N. 2002. Whole-lake algal responses to a century of acidic industrial deposition on the Canadian Shield. Can. J. Fish. Aquat. Sci. 59: 483–493. Watmough, S.A., and Aherne, J. 2008. Estimating calcium weathering rates and future lake calcium concentrations in the Muskoka–Haliburton region of Ontario. Can. J. Fish. Aquat. Sci. 65: 821–833. Watmough, S.A., Aherne, J., and Dillon, P.J. 2005. Effect of declining lake base cation concentration on freshwater critical load calculations. Environ. Sci. Technol. 39: 3255–3260. Wehr, J.D., and Brown, L.M. 1985. Selenium requirement of a bloom-forming planktonic alga from softwater and acidified lakes. Can. J. Fish. Aquat. Sci. 42: 1783–1788. Yan, N.D., Keller, W., Scully, N.M., Lean, D.R.S., and Dillon, P.J. 1996. Increased UV-B penetration in a lake owing to droughtinduced acidification. Nature (London), 381: 141–143. Yan, N.D., Blukacz, A., Sprules, W.G., Kindy, P.K., Hackett, D., Girard, R.E., and Clark, B.J. 2001. Changes in zooplankton and the phenology of the spiny water flea, Bythotrephes, following its invasion of Harp Lake, Ontario, Canada. Can. J. Fish. Aquat. Sci. 58: 2341–2350. Yan, N.D., Girard, R., and Boudreau, S. 2002. An introduced invertebrate predator (Bythotrephes) reduces zooplankton species richness. Ecol. Lett. 5: 481–485. Yan, N.D., Somers, K.M., Girard, R.E., Paterson, A.M., Keller, W., Ramcharan, C.W., Rusak, J.A., Ingram, R., Morgan, G.E., and Gunn, J.M. 2008. Long-term trends in zooplankton of Dorset, Ontario, lakes: the probable interactive effects of changes in pH, total phosphorus, dissolved organic carbon, and predators. Can. J. Fish. Aquat. Sci. 65: 862–877.

? 2008 NRC Canada


赞助商链接

更多相关文章:
更多相关标签:

All rights reserved Powered by 甜梦文库 9512.net

copyright ©right 2010-2021。
甜梦文库内容来自网络,如有侵犯请联系客服。zhit325@126.com|网站地图