Increased algal blooms are a threat to aquatic ecosystems worldwide, although the combined effects of multiple stressors make it difficult to determine the underlying causes. We explore whether changes in trophic interactions in response to declining calcium (Ca) concentrations, a water quality issue only recently recognized in Europe and North America, can be linked with unexplained bloom production. Using a palaeolimnological approach analysing the remains of Cladocera (herbivorous grazers) and visual reflectance spectroscopically inferred chlorophyll a from the sediments of a Nova Scotia (Canada) lake, we show that a keystone grazer, Daphnia, declined in the early 1990s and was replaced by a less effective grazer, Bosmina, while inferred chlorophyll a levels tripled at constant total phosphorus (TP) concentrations. The decline in Daphnia cannot be attributed to changes in pH, thermal stratification or predation, but instead is linked to declining lakewater [Ca]. The consistency in the timing of changes in Daphnia and inferred chlorophyll a suggests top-down control on algal production, providing, to our knowledge, the first evidence of a link between lakewater [Ca] decline and elevated algal production mediated through the effects of [Ca] decline on Daphnia. [Ca] decline has severe implications for whole-lake food webs, and presents yet another mechanism for potential increases in algal blooms.
Lakewater calcium (Ca) decline has emerged as a novel and a widespread threat to softwater lakes in eastern North America and Europe (and probably elsewhere), where decades of high acidic deposition and sustained logging cycles have depleted Ca pools in catchment soils, the main source of Ca to lakes . Increases in Ca leaching from soils have not been accompanied by compensatory increases in bedrock weathering rates, resulting in significant, long-term declines in surface water Ca levels . The biological implications of lakewater Ca decline are predicted to be extensive, because Ca is an essential nutritional element for many different types of aquatic biota . In particular, many freshwater crustaceans such as Daphnia (Daphniidae, Anomopoda, Crustacea) require relatively large amounts of aqueous Ca for the development of Ca-rich exoskeletons, which must be regenerated following each moult .
Daphnia are an important component of the zooplankton community in many freshwater lakes. Like many crustaceans, they have relatively high amounts of Ca in their exoskeletons, and rely mainly on the uptake of aqueous Ca to fulfil their Ca requirements . Only trace amounts of Ca are retained during moulting . In low Ca environments, some daphniids must, therefore, expend large amounts of energy obtaining Ca from the water column, and are at a competitive disadvantage relative to smaller zooplankton species with lower Ca requirements . In fact, large declines in Daphnia populations, coincident with increases in less Ca-rich Bosmina, have been documented in softwater lakes in Ontario and Nova Scotia, Canada, and in Adirondack Park (New York, USA) when Ca levels fell below approximately 2 mg l−1 .
Substantial reductions in daphniid populations have the potential to dramatically affect lake food webs, as Daphnia are important herbivores in many freshwater systems. Owing to their larger body size, Daphnia are able to filter food particles at a much faster rate and graze a wider size-range of algae compared with Bosmina , a taxon that can withstand lower Ca levels . As a result, in lakes with daphniid-dominated zooplankton assemblages, algal concentrations are often much lower than in bosminid-dominated lakes with comparable total phosphorus (TP) levels [8,9]. Species replacement of Daphnia by Bosmina, then, can potentially release phytoplankton from grazing pressure and lead to increased algal production. This presents problems for lake management programmes, as the link between TP and algal production becomes partly decoupled, limiting the use of nutrient-productivity model predictions . Prior to this investigation, however, evidence of a link between lakewater [Ca] decline and increased algal production had not been documented.
In many lakes of Kings County, Nova Scotia (Canada), chlorophyll a (a common proxy for total primary production) has been steadily rising independently of TP since lake monitoring began in 1997 . This includes Lake George, a popular recreational lake in the area where chlorophyll a levels are now approaching water quality guidelines (greater than 2.5 µg l−1) set by the Kings County Municipal Plan . A previous diatom-based palaeolimnological study of Lake George did not record any diatom-inferred changes in TP or pH over the last approximately 100 years of the lake's history that can account for this increase in algal production, but do show changes in the diatom assemblage at approximately 1970 that are consistent with the timing of recent anthropogenic climate warming in this region . The inferred timing of diatom assemblage changes related to warming, however, occurs too early to be considered a primary driver of recently documented increases in algal production in Lake George. Instead, these lakes are located on Ca-poor granitic bedrock in southwest Nova Scotia, where declines in lakewater [Ca] to very low levels have been documented . We hypothesize, then, that in the absence of nutrient enrichment, elevated chlorophyll a in Kings County lakes may be related to top-down controls that were exacerbated by the effects of climate warming, and specifically that declines in Daphnia abundances, which could release phytoplankton from grazing pressure, may be occurring in these lakes related to [Ca] decline.
Since zooplankton monitoring data from the Maritimes are scarce (as they are in almost all regions), and the onset of environmental stressors such as acid deposition, Ca decline and shoreline development precede the beginning of the Kings County Lake Monitoring Programme, we use palaeolimnological techniques to reconstruct trends in zooplankton assemblages, and assess how trends in chlorophyll a and zooplankton species composition might be linked. We tracked changes in subfossil Daphnia and Bosmina (and other cladocerans) relative abundances, as well as sedimentary visual reflectance spectroscopically inferred chlorophyll a  levels in Lake George (figure 1), to assess whether changes in Daphnia abundances have occurred in the lake's history that coincide with increases in algal production. This method measures both chlorophyll a and its post-depositional degradation products, and is therefore unaffected by diagenetic processes. To assess the role of changing predation regimes in any observed cladoceran community changes, we also measured Bosmina mucro length, as changes in mucro length recorded from lake sediments can be linked to changes in invertebrate and fish predation [15,16].
2. Material and methods
Lake George (44°55′ N, 64°42′ W; figure 1) is a weakly stratified, circumneutral (pH = 6–7), oligo-mesotrophic (unfiltered TP = 5–20 µg l−1) lake located on the South Mountain in Kings County, Nova Scotia . It has a mean depth of 9.0 m and a surface area of 153 ha, and drains into the Gaspereau River . Its catchment is mid-Devonian granitic bedrock, and is comprised of mixed coniferous–deciduous forests . It has a moderately developed shoreline with private cottages, homes, a trailer park, a public access beach and waterfront for the Lake George Provincial Park. Water chemistry for Lake George has been monitored monthly during the ice-free months since 1997 by the Kings County Lake Monitoring Programme , with some isolated measurements also available from 1993. Over the period of the monitoring record, a significant increase in chlorophyll a has been observed that cannot be explained by an increase in TP (figure 2a) .
A surface-sediment core was obtained from the deepest basin of Lake George in July 2005 using a Glew gravity corer with an internal diameter of 7.6 cm equipped with a 50 cm long Lexan core tube [18,19]. The core was sectioned into 0.5 cm intervals using a Glew vertical extruder , and then transferred to individual Whirl-Pak sample bags, and stored at 4°C at the Paleoecological Environmental Assessment and Research Lab (PEARL) at Queen's University in Kingston, Ontario. An additional water sample was taken in June 2010 following the protocols of the Ontario Ministry of the Environment's Lake Partner Programme and analysed for Ca at the Dorset Environmental Science Centre (Dorset, Ontario).
The sediment core was dated using excess 210Pb activities and 137Cs, in an Ortec germanium crystal well detector following the procedures outlined by Schelske et al.  and Appleby . Details of the dating of the Lake George sediment core can be found in Thienpont et al. . The chronology was developed using the constant rate of supply programme developed by Binford .
For analysis of cladoceran remains, sediments were deflocculated in 10 per cent potassium hydroxide solution for approximately 30 min, then passed through a 37 µm sieve . The material retained on the sieve was transferred to a vial using deionized water, and stained with safranin-glycerol solution. A minimum of 100 individuals was enumerated per interval . Daphnia were grouped into the Daphnia pulex complex based on the presence of stout spines on the middle comb of the post-abdominal claw, and in Lake George, this complex may consist of Daphnia pulicaria or Daphnia catawba (no remains belonging to the Daphnia longispina complex were identified). A total of 50 Bosmina remains were measured per interval using techniques outlined in Korosi et al. . In order to estimate sedimentary chlorophyll a content, selected intervals were freeze-dried and sieved (125 µm), and then analysed for spectral reflectance using a FOSS NIRSystems Model 6500 rapid content analyser [27,28]. Since this technique detects both chlorophyll a and its derivatives (pheophytin a and pheophorbide a), which absorb in similar regions of the electromagnetic spectrum, visual reflectance spectroscopically (VRS)-inferred chlorophyll a are accurately tracking overall trends in lake production, and not simply diagenetic processes [14,28]. A stratigraphy was prepared, and a constrained incremental sum of squares (CONISS) cluster analysis was performed on the cladoceran relative abundance data (chlorophyll a and Bosmina measurements were not included) using the computer program TGView v. 2.0.2 . Pairwise Pearson correlations of select variables were conducted using the stats package for the R software environment .
3. Results and discussion
Chlorophyll a levels in the water column of Lake George have been steadily and inexplicably rising over 10 years of lake monitoring, indicating that phytoplankton production has increased (figure 2a) . We used palaeolimnological techniques to determine whether these modern trends are a symptom of significant ecological changes occurring in the lake, or just natural variability in lake productivity. Similar to the modern limnological monitoring record, we observed an increase in VRS-inferred chlorophyll a (and its diagenetic products) in recent years, tripling in value between approximately 1990 and 2005, the year the sediment core was taken (figures 2b and 3). Prior to this, chlorophyll a levels had remained relatively unchanged over the last century. Therefore, we show that the elevated chlorophyll a values recorded by the Kings County Lake Monitoring Programme began in the early 1990s, and are unprecedented over approximately 110 years of the lake's history. No significant changes in monitored lakewater or diatom-inferred TP and pH have occurred in Lake George that can account for this recent increase (figures 2a and 3) [11,12]. Instrumental temperature records from Halifax, Nova Scotia, have recorded an increase of 0.8°C in mean summer temperature since 1948, and recent changes observed in the diatom sedimentary assemblage are consistent with limnological changes related to climate warming . While increased thermal stratification and growing season could account for increased lake productivity, the timing of diatom-inferred changes in lake thermal properties in Lake George are not consistent with the timing of changes in chlorophyll a, occurring more than a decade prior to the beginning of the rise in chlorophyll a (diatom principal components analysis axis 1 score on figure 3) . Therefore, anthropogenic climate change alone cannot account for increased algal production in Lake George.
We hypothesized that changes in chlorophyll a might be linked to ‘top-down’ grazing control by Daphnia rather than ‘bottom-up’ influences, as the abundance of large-bodied Daphnia has been identified as an important predictor of lake production in previous studies [10,31,32]. In theory, a zooplankton assemblage dominated by Daphnia can graze a broader size-range of algae, filter food particles faster (filter-feeding rate is a function of body length), and recycle nutrients at a slower rate than a zooplankton assemblage of equal biomass dominated by Bosmina, ultimately leading to reduced algal production . We observed considerable changes in the pelagic cladoceran community of Lake George, which consisted almost entirely of Bosmina and Daphnia (figure 3). Prior to 1950, species of the D. pulex complex were present in the lake at relative abundances of 5–10% on average (figure 3). Above a core depth of 6 cm (approx. 1950), we observed a decrease in Bosmina relative abundance and an increase in Daphnia relative abundance to approximately 15–20%, nearly double background abundances (figure 3). Above a core depth of 2 cm (approx. 1990), however, Daphnia and Bosmina relative abundances returned to pre-industrial levels (figure 3). The most recent reduction in Daphnia relative abundance occurs immediately prior to the increase in VRS-inferred chlorophyll a (figure 4), and there is a significant negative relationship between Daphnia abundance and VRS-inferred chlorophyll a since 1960 (t10 = −2.8, p = 0.02). This supports our hypothesis of top-down control on algal populations, with Daphnia as more effective grazers on algae than smaller-bodied Bosmina. This is consistent with previous field and laboratory studies, which show that chlorophyll a yields at constant TP are substantially lower when Daphnia are abundant [8,9,34]. We conclude, then, that a decrease in grazing pressure by Daphnia, probably exacerbated by recent changes in lake thermal properties , have led to anomalous levels of algal production in Lake George without increased nutrient enrichment.
Similar to the chlorophyll a trends (whether they be VRS-inferred or monitored from the water column; figure 2b), changes in Daphnia cannot be explained by changing pH, TP or lake thermal structure in our lake, as discussed above. Owing to their larger body size relative to many other species of zooplankton, Daphnia are one of the preferred prey items for planktivorous fishes, and changes in predation intensity by planktivorous fishes can affect daphniid abundances . Reduced predation by planktivorous fishes might be responsible for the increased relative abundance of Daphnia observed at a core depth of 6 cm, although there is no corresponding increase in Bosmina mucro length (figure 3). Increases in the length of the mucro are an anti-predation strategy when predation by grasping copepod (and other invertebrate) predators is high, as a long mucro will make them more difficult to handle, providing them with increased opportunities of escape . Predatory copepods, similar to Daphnia, are heavily preyed upon by planktivorous fishes owing to their large body size, and therefore increased mucro length can also indirectly infer reduced planktivorous fish predation . Moreover, the increase in Daphnia relative abundance does not match known fish stocking records. Stocking of brook trout yearlings and fingerlings began in 1976, after Daphnia populations reached 15–20% abundance (figure 4) .
A strong increase in Bosmina mucro length is observed, however, above a core depth of 2 cm, with the average length reaching 100 µm in the uppermost sediments (figure 3). This matches well with the first-known record of smallmouth bass in Lake George in 1997 , a piscivorous fish, which often reduces the abundance of planktivorous fishes . Fish predation, therefore, cannot explain the decrease in Daphnia observed above 2 cm. Furthermore, significant increases in dissolved organic carbon (DOC) have been measured in Lake George since 1993 , which impairs the visibility of planktivorous fishes and should provide further release for Daphnia from grazing pressure .
Lake George is located in a region of Nova Scotia that receives acidic deposition from long-range sources , and a few lakes in Kings County have acidified . The timing of the increase in Daphnia abundances corresponds to the height of acid deposition before emission restrictions were implemented . When acid deposition rates are high, Ca ions have been shown to leach into lakes at an increased rate, as hydrogen ions would replace base cations at soil-binding sites, leaving them free to wash into lakes and rivers . Lake George could probably have experienced elevated lakewater Ca levels in the 1950–1960s that may account for the observed increase in Daphnia relative abundances (figure 3). Eventually, following decades of increased leaching, soils become depleted in Ca, leading to long-term lakewater [Ca] declines that negatively impact Daphnia . This has been demonstrated in nearby Little Wiles Lake (Bridgewater, Nova Scotia), where the loss of Daphnia in recently deposited lake sediments (similar to the trend observed in Lake George) was attributed to declining lakewater Ca levels . Similarly, in 42 softwater, oligotrophic lakes in south-central Ontario, declines in Daphnia relative abundances and coincident increases in Ca-poor Bosmina since pre-industrial times were observed in lakes where modern Ca levels were less than 2 mg l−1, whereas in lakes with [Ca] greater than 2 mg l−1, Daphnia relative abundance increased owing to factors such as increasing DOC and the spread of smallmouth bass , changes that have also occurred in Lake George.
Unlike Little Wiles Lake, there are no long-term monitoring data on [Ca] for Lake George. At the time of sediment collection, alkalinity in Lake George was low (2.8 mg CaCO3 l−1), and [Ca] sampled from Lake George in June 2010 (5 years after the sediment core used in this study was collected) was 1.14 mg l−1, well below levels that have been shown to be detrimental to daphniid populations . We conclude, based on the geology of the region and its history of acid deposition, that Lake George has also experienced declines in base cations similar to other lakes in southwest Nova Scotia that have been monitored for Ca , and that this decline in lakewater [Ca] is responsible for the decrease in Daphnia populations observed at a core depth of 2 cm. There can be considerable inter-annual variability in lakewater [Ca] levels related to climate variables that control the washing of ions into lakes , and the spike in Daphnia observed at 0.5 cm may be reflecting years when [Ca] was higher. Local climate records from nearby Greenwood, Nova Scotia, show that 2003, the approximate 210Pb-inferred year the Daphnia spike occurred, was a wetter year than the decadal average , lending some support to this hypothesis. Interestingly, the long-term monitoring record for Lake George shows that lakewater chlorophyll a was also lower in 2003 (figure 2) when the Daphnia were more abundant, which was not detected by VRS in the sediment record, probably because it integrates material from several years.
We document a link between declines in Daphnia and elevated VRS-inferred chlorophyll a (a trend that matches closely with monitored lakewater chlorophyll a), where the decline in Daphnia cannot be attributed to changes in predation, TP, pH or lake thermal structure, but instead to a new environmental stressor: lakewater [Ca] decline. The decline in Daphnia is especially striking, as many of the known limnological changes that occurred in recent years in Lake George (e.g. increasing DOC, smallmouth bass introduction) should have favoured an increase in Daphnia abundances in the absence of [Ca] decline as a stressor. Indeed, Daphnia did increase as predicted in approximately 1950 in Lake George, until a sudden reversal occurred in approximately 1990. The reversion of Daphnia and Bosmina back to pre-industrial relative abundances, combined with warmer surface water temperatures and a longer growing season, probably resulted in a striking rise in algal production as inferred from both sedimentary and measured lakewater chlorophyll a that is unprecedented over the last century.
The negative impacts of [Ca] decline on Daphnia in softwater lakes have recently been well established, and we provide, to our knowledge, the first evidence of a link between lakewater [Ca] decline and unusually high levels of algal production, implicating [Ca] decline as an additional mechanism for reductions in water quality related to increased algal production. This has become apparent in Lake George, where both modern limnological sampling and palaeolimnological data detect increasing chlorophyll a levels in the absence of nutrient enrichment, and complaints of episodic algal blooms may become more frequent in response to the combined effects of reduced grazing control by Daphnia, potential future nutrient enrichment from continued watershed development and anthropogenic climatic warming.
It is becoming increasingly clear that lakes are being affected by multiple stressors. As [Ca] decline in lakes can negatively affect keystone herbivores, we show that, through trophic linkages, this newly identified anthropogenic stressor may have complex, and potentially negative consequences for whole lake ecosystems.
This project was funded by Natural Science and Engineering Research Council (NSERC) awards to J.P.S. and J.B.K. We are grateful to Ian Watson, Leanne Jennings and the volunteers of the Kings County Lake Monitoring Programme for providing water samples as well as chemical and physical data on Lake George, and to Ian Spooner and Michael MacIntyre for helping provide historical information. Anna DeSellas and the Dorset Environmental Science Centre provided the Ca data. We thank B. Ginn, A. Paul, A. Coombs and K. Lauersen for collecting the sediment core from Lake George, and J. Kurek, A. Jeziorski, A. DeSellas, and two anonymous reviewers for providing comments that improved the quality of the manuscript.
- Received July 6, 2011.
- Accepted September 8, 2011.
- This journal is © 2011 The Royal Society