A Holocene lake sediment record spanning the past 7300 years from Wishart Lake in the Turkey Lakes Watershed in the Hemi-Boreal of central Ontario, Canada, was used to evaluate the potential drivers of long-term change in diatom assemblages at this site. An analysis of diatom assemblages found that benthic and epiphytic taxa dominated the mid-Holocene (7300–4000 cal yr BP), indicating shallow, oligotrophic, circum-neutral conditions, with macrophytes present. A significant shift in diatom assemblages towards more planktonic species (mainly Cyclotella sensu lato, but also several species of Aulacoseira, and Tabellaria flocculosa) occurred ~4000 cal yr BP. This change likely reflects an increase in lake level, coincident with the onset of a more strongly positive moisture balance following the drier climates of the middle Holocene, established by numerous regional paleoclimate records. Pollen-inferred regional changes in vegetation around 4000 yrs BP, including an increase in Betula and other mesic taxa, may have also promoted changes in diatom assemblages through watershed processes mediated by the chemistry of runoff. A more recent significant change in limnological conditions is marked by further increases in Cyclotella sensu lato beginning in the late 19th century, synchronous with the Ambrosia pollen rise and increases in sediment bulk density, signaling regional and local land clearance at the time of Euro-Canadian settlement (1880 AD). In contrast to the mid-Holocene increase in planktonic diatoms, the modern increase in Cyclotella sensu lato likely indicates a response to land use and vegetation change, and erosion from the watershed, rather than a further increase in water level. The results from Wishart Lake illustrate the close connection between paleoclimate change, regional vegetation, watershed processes, and diatom assemblages and also provides insight into the controls on abundance of Cyclotella sensu lato, a diatom taxonomic group which has shown significant increases and complex dynamics in the post-industrial era in lakes spanning temperate to Arctic regions.
Citation: Beck KK, Medeiros AS, Finkelstein SA (2016) Drivers of Change in a 7300-Year Holocene Diatom Record from the Hemi-Boreal Region of Ontario, Canada. PLoS ONE 11(8): e0159937. https://doi.org/10.1371/journal.pone.0159937
Editor: Navnith K.P. Kumaran, Agharkar Research Institute, INDIA
Received: February 4, 2016; Accepted: July 11, 2016; Published: August 17, 2016
Copyright: © 2016 Beck et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data are available in the public database "Neotoma". Diatom: DatasetID = 19788; Loss-on-ignition: DatasetID = 19787; Geochronology: DatasetID = 19786. Full public access is available at: http://apps.neotomadb.org/Explorer/?datasetid=19788.
Funding: This study was funded by the Natural Sciences and Engineering Research Council of Canada (Discovery Grant RGPIN 327197-11 to SAF). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Aquatic ecosystems in the 21st century are threatened by numerous anthropogenic disturbances, such as climate warming, land use change, and atmospheric deposition; interactions between these stressors may lead to complex and unpredictable effects [1, 2]. Short-term observational datasets are rarely sufficient to parse out the effects of particular types of environmental change on species diversity or ecosystem function; therefore, paleolimnological approaches are often used to develop conceptual or quantitative models of ecosystem responses to various stressors .
While climate may be an ultimate driver of change in aquatic ecosystems, physico-chemical limnological processes are generally the proximate, mechanistic drivers. For example, climate warming in temperate and Arctic lakes affects ice phenology, lengthens the ice-free season and can produce more stable and deeper thermal stratification, thereby affecting habitat available to aquatic biota [4–6]. These processes have been hypothesized to explain major recent changes in freshwater diatom assemblages in regions with seasonal climates, notably, increases in diatoms in the Cyclotella sensu lato group, and often concomitant decreases in tychoplanktonic Aulacoseira spp. and in benthic diatom taxa .
The post-1850 AD increase in Cyclotella sensu lato in particular is widespread in many North American lakes, yet is time transgressive. In circum-arctic lakes, rises in Cyclotella sensu lato species are often detected at the end of the 19th century whereas in temperate lakes, these increases have been recorded as late as ~ 1970 AD . While increases in the length of the ice-free season explains the recent rise in Cyclotella sensu lato at many high latitude sites , changes in water clarity and nutrient flux may also be important in explaining increases in these taxa, particularly in temperate locations closer to agricultural and urbanized areas . Tracking and explaining fluctuations in abundances of Cyclotella sensu lato diatoms in the pre-industrial Holocene is necessary to better explain the recent responses of this important group of bio-indicators.
In addition to ice phenology, climate also controls lake water level as evaporative balance shifts with hydroclimatic regime . Water level, particularly in a small lake, is a key factor in determining the types of biota, notably the proportion of planktonic vs benthic algae. Holocene paleoclimate has been shown to be a fundamental control on water level in both large and small lakes [10–13] and diatoms are highly responsive to such changes . Despite Holocene paleoclimates which have remained generally humid in Eastern North America, lake level changes on the order of several meters have been documented at many sites [11, 12, 15–17]. Diatoms have been widely used as proxies for reconstructing lake level both qualitatively using diatom growth habits (planktonic to benthic ratios), and quantitatively using calculation of optimal depths along surface sediment transects or other numerical methods reviewed in ref no. .
In addition to ice phenology and water level, climate can affect lake ecosystems through vegetation dynamics. In temperate North America, palynological records show a close relationship between forest succession and Holocene paleoclimate [19–22]. Analysis of diatom assemblages in conjunction with pollen records has shown some synchronous shifts in diatom community structure, possibly in part the result of change in local catchment geochemistry brought on by vegetation change [23–27]. Soil development and vegetation succession affect lake water chemistry over both short  and long time-scales . For example, an increase in broad-leaved tree species relative to needle-leaved species can influence lake water chemistry due to differences in leaf litter composition . Further, changes in the density of vegetation in catchments can influence the contribution of snowmelt to lakes , which can alter water balances , and influence nutrient cycling . As a result, the evolution of aquatic ecosystems is influenced by catchment-mediated processes that affect lake-water chemistry, and these are at least partially influenced by climate [5, 29, 34].
Much of our knowledge of long-term changes in biogeochemistry of lakes (lake ontogeny) relies on the analysis of archives of physical, chemical, and biological indicators preserved in lake sediments. The analysis of indicators that respond both directly and indirectly to climate, such as diatoms (Chromista: Bacillariophyta) and pollen, can then be used as proxies for the prevailing environmental conditions of the past. To evaluate the relationship between forest succession, paleoclimate and diatom assemblages, we present a paleolimnological diatom record for Wishart Lake, in the Turkey Lakes Watershed, which has been a monitoring site in northwestern Ontario since 1980. The record from Wishart Lake is interpreted in the context of available regional palynological and paleoclimate data. These comparisons allow for an evaluation of the potential drivers of aquatic ecosystem change over the Holocene, including regional vegetation change, in a region that has undergone large-scale vegetation reorganization since deglaciation. In this paper we aim to answer the following questions:
- Which environmental factors influence aquatic ecology in the long-term perspective in North Temperate Zone lakes which may be subject to enhanced rates of change with climate warming?
- Does landscape change play a role in aquatic ecosystem change in ecotonal regions transitional between temperate and boreal forests?
- How have the abundances of the important bio-indicator diatoms, Cyclotella sensu lato, responded to natural environmental variability prior to significant anthropogenic impact?
Wishart Lake (47.049°N; 84.397°W) is one of four lakes in the Turkey Lakes Watershed (TWL; Fig 1), which encompasses 10.5 km2 approximately 20 km inland from Lake Superior in the Province of Ontario, Canada. The bedrock geology of the region consists mainly of Precambrian silicate greenstone (ie., metamorphosed basalt). Till deposits laid down following deglaciation are composed of felsic silt ablation material containing 0–2% calcium carbonate with a thickness on average of 1–2 m [35, 36]. The TLW region was deglaciated between 10–11 ka BP following the Algonquin interstadial .
Turkey Lakes Watershed (upper panel) and location of the study lake and nearby Upper Mallot Lake in the context of regional forest zones (lower panel). Data sources: Vegetation data extracted from ref no.  and base map layers extracted from the CanVec+ digital cartographic reference product produced by Natural Resources Canada in 2015. Both data sources are licensed under the Open Government License—Canada (http://www.nrcan.gc.ca/earth-sciences/geography/topographic-information/free-data-geogratis/licence/17285). Figures generated using ArcGIS.
Climate in the TLW is cool and continental. The on-site weather station records a mean annual, July, and January air temperatures of 4.5, 17.8, and -10.7°C respectively (1980–2011). Average annual precipitation (1982–2009) is 1197 mm/yr . The modern day vegetation consists of mixed hardwood forest (Hemi-boreal); the transition to boreal forest is approximately 30 km to the north (Fig 1). Local forest cover is dominated by Acer saccharum with mixed hardwoods and conifers present . The lake is polymictic due to summer wind mixing [35, 40]. A bathymetric map of Wishart Lake provided in ref.  shows inflow from an adjacent lake at the northwest end, a single basin with maximum depth in its centre, and a shallow channel to the outflow. Monitoring data suggest that the yearly amplitude in water level fluctuation in the TLW does not exceed 1 m, and average water level in a given year deviates from the 20–30 year mean by not more than 1 m . Physical and chemical properties of Wishart Lake are given in Table 1.
Despite its proximity to point sources of sulfate aerosols, and elevated concentration of sulfate in precipitation [36, 42], the pH of Wishart Lake has been stable near ~6.7 over the past 30 years; paleolimnological data for neighboring Batchawana Lake showed no increase in acidophilic diatoms over the past 2–3 centuries , suggesting buffering capacity in the catchment. Measured nutrient concentrations in the TLW suggest that the lakes are presently oligotrophic, and apparently limited by phosphorus (Table 1). Owing to protected status and its use as a long-term monitoring site, there has been little disturbance to forest cover within the TLW; however, significant deforestation and industrial activity have taken place in the surrounding region since the end of the 19th century.
Permission to access the Turkey Lakes Watershed for lake sediment core sampling was granted to SAF in May 2011 from the Canadian Forest Service (Natural Resources Canada). Lake cores were collected June 2011 at a water depth of 4.5 m, approximately the deepest point in the lake; an anchored coring platform was used to maintain position during core collection. A 7-m lake sediment core sequence (Core WS03) was collected in 1-m sections using a Livingstone piston corer . Woody debris at the base of the core prevented the recovery of sediments below 7 m depth. Because this coring system is not able to collect the uppermost sediments in an undisturbed fashion, a gravity corer  was used to collect replicate 26-cm long surface cores (Cores WS02 and WS04) that included the undisturbed sediment-water interface. The two surface cores were taken within <1 m of each other, both contained an undisturbed sediment-water interface, and thus were considered coeval. The Livingstone piston core sequence was collected from the same core hole as the WS02 surface core; the first drive of the Livingstone core sequence was set to commence at the base of the WS02 surface core section. Livingstone core sections were extruded in the field into PVC tubing, and wrapped in plastic wrap and then aluminum foil. The surface cores (WS02 and WS04) were immediately extruded in 1-cm increments.
An age-depth model was produced from the combination of 210Pb activity, the rise in Ambrosia pollen, and radiocarbon dating. 210Pb dating was performed by drying sediments at 60°C and grinding samples at 1-cm increments. Samples from the surface of the core to depth of 70 cm (WS04/WS03 sequences) were submitted for measurement of 210Pb activity by alpha spectroscopy at Flett Research Laboratory (Winnipeg, Manitoba) (Table 2). Five 14C dates were selected between the depths of 150–700 cm in the WS03 core sequence; the selected sediment samples were washed with distilled water through a 90-μm mesh. Plant material was selected using a stereomicroscope, and submitted to Beta Analytic Inc. (Miami, USA) for Accelerator Mass Spectrometry (AMS) dating (Table 3). Dates were calibrated using the IntCal09 calibration curve  and the program CALIB .
Po-210 is assumed to be in secular equilibrium with Pb-210.
Dates were calibrated using the INTCAL09 calibration curve and the program Calib v6 .
The Ambrosia pollen rise was used to confirm the recent chronology and the 210Pb dates. The Ambrosia pollen rise in Ontario is known to have occurred between 1830 and 1880 AD with precise ages for the Ambrosia rise assigned based on local history of forest clearance and disturbance associated with Euro-Canadian settlement . The Ambrosia rise in the TLW region is placed at ~1880 AD as this was the time of initial railway development, early mining and settlement in the local area [49, 50].
The Ambrosia peak was determined in the Wishart Lake record by processing 32 samples for pollen analysis from the top 65 cm of the record using standard procedures involving acid digestion and sieving . Frequencies of Ambrosia and non-Ambrosia pollen were identified on a light microscope at 400x magnification for a total of 200 pollen grains. Exotic Lycopodium spores were used as markers to determine concentrations of pollen per ml sediment. Two cores collected from Wishart Lake in 1980 AD  have data available on the rise in Ambrosia pollen and these were used to validate the chronology proposed here.
An age-depth model was developed using the date of core collection as an upper constraint (2011 AD), the five radiocarbon dates, the lower-most 210Pb date (50 cm depth), and the Ambrosia rise (40 cm depth), using the clam package for R . The model was based on linear interpolation between dating points, with calculations at 95% confidence ranges and 1000 iterations. Ages were calculated every 1 cm from 0 to 697 cm (Fig 2). Ages for the pre-industrial Holocene are discussed as calibrated calendar years before 1950 AD (cal yr BP); more recent ages are discussed using calendar date (ie. 1950 AD).
The model was developed using a linear regression model fit to calibrated radiocarbon dates, recent chronological data points derived from the Ambrosia pollen rise (depth 40 cm) and activity of 210Pb (lower boundary used, depth 50 cm). Radiocarbon dates (the five lowermost points) are shown with histograms corresponding to the probability distribution of calibrated dates . Figure generated using clam for R .
Point samples from the Wishart Lake core were taken at 10-cm intervals for diatom and chrysophyte cyst and scale analysis, providing an inter-sample resolution of ~100 years. Higher resolution sampling at 1-cm intervals was performed for the upper-most 100 years of the core for an inter-sample resolution of ~4–5 years. Diatoms, chrysophyte cysts and scales, were concentrated using 10% HCl, followed by 30% hydrogen peroxide . Residues of known concentration were mounted with Naphrax®. Four hundred and fifty diatom valves were counted per slide using an oil immersion DIC objective on a light microscope at 1000X magnification. Taxa were identified using Antoniades et al. , Fallu et al. , Krammer and Lange-Bertalot , Lavoie et al. , and Patrick and Reimer . Diatom taxonomy follows current nomenclature used in Algaebase (http://www.algaebase.org/). Nomenclature for Cyclotella sensu lato follows . Loss-on-Ignition (LOI) estimates of inorganic and organic matter in lake sediments were used as a general proxy for sediment provenance, including erosion in the watershed and biological productivity. Standard methods for LOI were applied, with combustion at 550°C and 950°C . LOI550 and LOI950 were measured every 5 cm, except for the upper 26 cm, where it was measured at 1-cm increments. ANOVA was used to compare the LOI550 and LOI950 among the diatom-determined zones.
Cluster analysis was used to determine boundaries between diatom assemblage zones in the stratigraphy, with the number of significant zones evaluated by the broken stick model, using the rioja package in R [62, 63]. Diatom and pollen assemblages were plotted as stratigraphies using C2 ; C2 was also used to ordinate the samples using Detrended Correspondence Analysis (DCA).
To compare our analysis of diatom and chrysophyte indicators with regional paleovegetation and paleoclimate, we extracted pollen and geochronology data for Upper Mallot Lake  from the Neotoma database (http://www.neotomadb.org/). Ages were re-modelled for the Upper Mallot Lake pollen record using the top of the core (1993 AD), the three radiocarbon dates included in the database record, and the Ambrosia rise (15 cm depth) with the clam package for R ; pollen assemblages were ordinated using DCA and axis scores were plotted.
All diatom assemblage data, geochronology results and other core data have been deposited in the publicly accessible Neotoma Database (http://apps.neotomadb.org/Explorer/?datasetid=19788). All slides have been permanently archived by the lead author.
The activity of 210Pb decays regularly with depth, with the exception of lower than expected activity in the upper 5 cm of the core (Table 2), perhaps related to dilution of 210Pb or sediment mixing. A constant rate of supply model was used to infer ages from 210Pb activities . Background 210Pb activity (0.008 Bq g-1) is reached at ~56 cm. Because of irregular 210Pb activity in the upper 5–6 cm of the core, we interpret that portion of the chronology with caution (2000–2011 AD). Five AMS radiocarbon dates were obtained (Table 3). The section of the core dated by radiocarbon has a similar rate of sediment accumulation to the rates determined using 210Pb dates and the Ambrosia rise (Fig 2).
To evaluate the proposed age model and the 210Pb dates, the Ambrosia pollen rise was identified at depth 40 cm and based on local historical records, took place approximately 1880 AD. According to the 210Pb age model, this depth corresponds to the year 1860 AD, suggesting an error on the 210Pb chronology of ~ 20 yrs. As a further test of the 210Pb dates, we also considered Ambrosia pollen counts from two Wishart Lake cores taken in 1980 AD, Core 86 and Core 87 described by Harper . In core 86, the Ambrosia rise is noted at 24 cm and in core 87, at 29 cm . By taking the difference between the findings of Harper  and this study in terms of depth of the Ambrosia rise and year of core collection, we assign the year of 1980 AD to the depth of 13 cm in the Wishart Lake core used in the present study. This is based on comparison of the mean depth for the Ambrosia rise in the 1982 cores (27 cm) with the depth of the Ambrosia rise in the 2011 core (40 cm). The difference, 13 cm, is the depth of sediment accumulated since 1980, and thus the depth of 13 cm in the WS03 core would date to approximately 1980 based on the data of Harper. The 210Pb age model proposed here gives a depth of 13 cm a date of approximately 1975 AD. These comparisons suggest errors on the 210Pb age model of 5–20 years.
Approximately 280 diatom taxa were identified from the last ~7300 years in 105 samples. The record is dominated by (tycho)planktonic taxa including (in order of decreasing abundance) Discostella stelligera Houk and Klee, Discostella pseudostelligera Houk and Klee, Aulacoseira distans Simonsen, Lindavia comensis Nakov et al 2015, Tabellaria flocculosa Kützing, and the benthic taxa Brachysira brebissonii Ross in Hartley, Achnanthidium minutissimum Czarnecki B. vitrea Ross in Hartley, Staurosirella pinnata Williams and Round, Pseudostaurosira brevistriata Williams and Round, and Nupela vitiosa Siver and Hamilton (Fig 3). Discostella stelligera and D. pseudostelligera are present through the record generally at abundances between 10–20%. Recent sediments of Wishart Lake contain even higher percentages of D. stelligera (up to 35%) and D. pseudostelligera (up to 25%). While it is somewhat surprising to record high abundances of euplanktonic diatoms in a lake with maximum depth of < 5 m, comparable abundances have been recorded for these taxa occasionally in other shallow lakes in the Great Lakes region. For example, Laird et al 2011  report D. stelligera at abundances of ~10–25% in selected Ontario lakes with depths < 5m to the northwest of the study site; further, Yang and Duthie (1995)  in a survey of surface sediments in East Lake and coastal Lake Ontario, report depth optima for D. stelligera of 1.7 m, L. comesis of 4. 9 m and L. bodanica of 3.2 m. While these taxa are clearly planktonic in habit and most often dominant in lakes with depths exceeding 8–10 m , these taxa can occasionally be found in shallower lakes, where ecological and limnological conditions permit. Three statistically significant zones were identified in the Wishart Lake diatom record. Zone 1 was split into two subzones (Zone 1a and Zone 1b). Subzones were chosen to facilitate discussion of other shifts in diatom assemblages, and are also supported by nodes in the dendrogram just below the point of significance.
Diatom species are displayed in percent abundance. Only taxa appearing at >2.5% abundance in any one sample are plotted out of a total of 280 taxa recorded. Sample points where rare taxa appear at <0.5% abundance are displayed by the + symbol. Three significant zones were delineated using cluster analysis (solid red lines, with the dashed grey line separating Zone 1 into two sub-zones). The solid green line in the upper zone indicates the Ambrosia rise. The age axis is in calibrated, calendar years BP where year -61 is the time of core collection (2011 AD). Full diatom counts are available at neotomadb.org (Dataset ID = 19788; http://apps.neotomadb.org/Explorer/?datasetid=19788). Diagram plotted using C2 .
Zone 1: 7300–4010 cal yr BP (700–470 cm)
Zone 1a (7300–6160 cal yr BP; 700–640 cm) is dominated first by Lindavia comensis, then Discostella stelligera and D. pseudostelligera; Tabellaria flocculosa and selected benthic taxa, particularly Achnanthidium minutissimum, Brachysira brebissonii, and B. vitrea, are also common. The cyst:diatom ratio is low in this zone in comparison to others, while the scale:diatom ratio is higher (Fig 3). The shift to Zone 1b (6160–4010 cal. years BP; 640–470 cm) is characterized by decreasing DCA Axis 1 and Axis 2 scores (Fig 4), an increase in Aulacoseira distans from 10 to 25%, an increase in Pseudostaurosira brevistriata, Staurosirella pinnata and other small benthic fragilarioids from <5 to 10–15% abundance, and an increase in epiphytic taxa such as Nitzschia fonticola Grunow in van Huerck, Nupela vitiosa and Pinnularia biceps Gregory. Cyclotella sensu lato spp. decline, along with T. flocculosa. The cyst:diatom ratio increases in subzone 1b, while the scale to diatom ratio declines (Fig 3).
Detrended Correspondence Analysis axis 1 (solid lines) and axis 2 (dashed lines) for Wishart Lake diatom record (lower panel) and for the past ~7200 cal. yr BP Upper Mallot Lake pollen record (upper panel; raw pollen data from ref.  and publicly archived at neotomadb.org). Red lines delineate Wishart Lake diatom zones from cluster analysis; dashed grey line indicates the subdivision of Zone 1. Analysis and diagram generated using the rioja package for R .
Zone 2: 4010–250 cal yr BP (470–63 cm)
The onset of Zone 2 is characterized by marked increases in Lindavia comensis, Discostella pseudostelligera, D. stelligera and Tabellaria flocculosa. Aulacoseira spp. that are dominated by A. distans remain stable at ~15% of the assemblage. Fragilaria sensu lato spp. are somewhat lower in abundance compared to Zone 1b. Cyst:diatom ratios remain constant throughout the zone at ~10% as well as scales:diatoms at ~5%. Both ratios are generally lower in this zone compared to the other zones in the record (Fig 3). DCA axis 1 and 2 scores stabilize with the onset of Zone 2, decreasing mid-way through ~1800 cal by BP (Fig 4).
Zone 3: 250 cal yr BP (= 1700 AD)– 2011 AD (63–0 cm)
This zone is delimited on the basis of a further significant increase in Lindavia bodanica Nakov et al 2015, L. Michiganiana Nakov et al 2015, Discostella pseudostelligera and D. stelligera (Fig 3). The most significant increase in Cyclotella sensu lato spp. in this zone is temporally coincident with the Ambrosia rise (1880 AD); Cyclotella sensu lato spp. reach values >30%, which are unprecedented in the record. The samples dating to the most recent 30 years do not show any further increase in Cyclotella sensu lato spp., rather the abundances are stable or subtly declining (Fig 5). Cyst:diatom ratios in Zone 3 increase up to ~20% around 1900 AD then decline again following 1930 AD. Scale:diatom ratios are ~10%, with a sharp peak in a single sample to almost 40% ~1935 AD. LOI550 dips in Zone 3 between 1850–1920 AD, as LOI950 increases (Fig 3). The DCA axis scores show rapid and high magnitude change, indicating that the taxonomic shifts into Zone 3 are the most marked of the record (Fig 4).
Changes in % abundance of Cyclotella sensu lato diatoms (orange curve), % Ambrosia pollen (black curve) and sediment bulk density (blue curve), are shown as a function of age in cal yr BP, with 0 yr BP = 1950 AD. Figure plotted using standard R plotting functions.
The mean value for the percent organic content of sediments as determined from percent mass lost at 550°C (LOI55) was 40% (± measurement error of 0.12%; Fig 3). There are significant increases in LOI550 from the base of the core to the surface, shown through a comparison of the LOI550 in the three diatom delineated zones using ANOVA (p < 0.05). There are two exceptions to this trend: declines in LOI550 between ~30–100 cal yr BP and ~6000–6500 cal yr BP (Fig 3). The LOI950 shows less change, with a Holocene average of 3.30% (± measurement error of 0.08%); the uppermost Zone 3 has significantly higher LOI950 (ANOVA, p < 0.05). The decreases in sediment organic matter (LOI550) and increases in inorganic carbon (LOI950) recorded in Zone 3 are also coincident with increases in sediment bulk density and the rise in Ambrosia pollen (Fig 5).
Regional Pollen Record
Pollen data from nearby Upper Mallot Lake  were obtained for comparison between regional paleovegetation and changes in the aquatic environment at Wishart Lake as tracked by diatoms (Fig 4). At the inception of the Wishart Lake diatom record (7300 cal yr BP), the Upper Mallot pollen record indicates a mixed-forest assemblage dominated by Pinus spp, with Alnus, Betula, Quercus and Ulmus present as co-dominant or minor elements. A shift in pollen assemblages is noted around 4000 cal yr BP (onset of Zone 2c), with declines in pollen of Pinus and Quercus and concomitant increase in taxa indicative of more mesic conditions (Acer saccharum, Betula, Cupressaceae, Picea). Pollen-derived transfer functions reconstruct maximum Holocene July temperatures at 6000 cal yr BP, with a gradual decline to minima during the Little Ice Age (, and data publicly available in the Neotoma Database, http://www.neotomadb.org/]).
Controls on Holocene aquatic ecosystem development
With a basal age of ~7300 cal yrs BP, the Wishart Lake sediment core did not capture the entire post-glacial record. Therefore, vegetation succession and the relationship to lake ontogeny immediately following deglaciation cannot be evaluated here. Nevertheless, the relatively high rate of sediment accumulation and robust age model permit analysis of mid- and late Holocene paleolimnology. Cluster analysis indicates significant shifts in diatom assemblages, driven mainly by fluctuations in the abundances of planktonic diatoms. A primary control on the abundance of planktonic diatoms is availability of open water habitat, determined in small, middle latitude lakes such as Wishart largely by water level. Thus, the changes in abundances of planktonic diatoms suggest that variability in lake water level through the pre-settlement part of the record is an important control on diatom assemblages.
The early portion of the Wishart Lake diatom record (Zone 1; Fig 3) is characterized by several indicators of lower water level, including lower abundances of planktonic diatoms, more benthic taxa, and a lower concentration of diatom valves. The base of the core consisted of compacted woody debris, supporting the interpretation of a low water stand around or just prior to 7300 cal yr BP. Benthic and epiphytic diatom taxa important in Zone 1 include Achnanthidium minutissimum, a broadly tolerant benthic taxon widespread in the study region, and Brachysira brebissonii, also a benthic taxon common in the study region, found preferentially in oligotrophic lakes with mildly acidic pH. Pinnularia biceps is also recorded in this zone, an epiphyte typically found in water with low mineral content . Aulacoseira distans increases through Zone 1; analyses of lake surface samples across a transect of water depths in adjacent Michigan shows A. distans to be a benthic or tychoplanktonic species, present mainly in samples with water depths < 4 m . Zone 1 assemblages as a whole indicate the availability of shallow water oligotrophic, circum-neutral to slightly acidic aquatic habitats.
Around 4000 cal yr BP, diatoms in the genera Cyclotella sensu lato, Tabellaria and Aulacoseira increase. These shifts delineate Zone 2, and suggest a possible increase in lake water level that persisted through the rest of the Holocene. Temporally coincident increases in lake level are recorded in paleolimnological studies to the west of the study site. Cyclotella sensu lato and other planktonic diatom taxa also increase after 4500 cal yr BP in three lakes in Ontario’s Experimental Lakes Area (ELA) [17, 69]. These increases were linked to rising water level (up to 8 m increase in Lake 239) through independent reconstructions of reduced effective evapotranspiration derived from pollen records from Lake 239 and a series of nearby lakes [16, 69]. While ELA sites are expected to show increased sensitivity to moisture availability because of their proximity to the prairie-boreal forest ecotone, our results confirm sensitivity of hemiboreal lakes several hundred km to the east of the prairie ecotone to shifts in moisture availability related to regional-scale Holocene climate fluctuations .
The diatom record from Wishart Lake suggests some role for water level changes in explaining long-term lake development at this site. However, given that the lake presently remains shallow and modern day water level variability is small, the record from Wishart Lake also suggests the possible influence of watershed processes on diatom assemblages. While the major shift in diatom community structure beginning around 4000 cal yr BP is marked by increases in Discostella stelligera, perhaps driven in part by regional climatically induced increases in water level, Tabellaria flocculosa also increases in Zone 2. T. flocculosa is generally interpreted as an indicator of nutrient enrichment [55, 71–74]. Lower cyst:diatom ratios and higher LOI550 in diatom Zone 2 further suggest the possibility of an increase in lake productivity, perhaps caused by a change in water quality along with water level rise after 4000 cal yr BP.
A possible driver of these changes in water quality relates to changes in paleovegetation. The pollen record from nearby Upper Mallot Lake shows a decline in Pinus pollen as Betula and other taxa more indicative of mesic conditions (Acer saccharum, Picea) increase, particularly after 4000 cal yr BP . Pollen records from Jack Lake and Nina Lake in central Ontario, located in the same ecotonal region and at similar latitude to Wishart Lake, show a similar vegetation shift from Pinus-dominated coniferous forests to mixed hardwood forest, indicated by increases in pollen of Betula and other broad-leaved taxa, around 4000 cal yr BP . This transition is also recorded at regional sites to the west including Lily Lake on Isle Royale, Michigan, a set of sites at Pictured Rocks National Lakeshore, Michigan , and Crooked Lake on the Upper Peninsula of Michigan . These changes in paleovegetation are indicative of the well documented transition from warmer, drier middle Holocene climates to wetter, cooler late Holocene Neoglacial climates recorded locally at Upper Mallot Lake  and more broadly across Eastern and Mid-western North America [68, 76–78].
The regional increase in broad-leaved hardwood tree taxa after 4000 cal yr BP likely had important watershed biogeochemical effects. As needle-leaf litter is nutrient poor relative to litter produced by broad-leaved species , altered chemical composition of runoff would favour diatom assemblages more responsive to higher nutrient concentrations [79–82]. Multi-proxy studies of the effects on lakes of the middle Holocene decline in eastern North America of Tsuga canadensis document changes to lake nutrient concentrations in response to changes in leaf litter quality driven by the replacement of Tsuga with broad-leaved trees; the Tsuga decline often coincides with an increase in diatom taxa indicative of more eutrophic conditions [25, 26], including Tabellaria flocculosa. Thus, several processes, including enhanced nutrient flux due to runoff from a catchment increasingly vegetated with broad-leaved trees, and an increase in lake level resulting from a moister climate, may have combined to produce the observed shifts around ~4000 cal yr BP at Wishart Lake.
Causes of recent changes
Zone 3 of the Wishart Lake diatom record contains assemblages unprecedented for the past 7300 years, suggesting critical effects of several facets of human activity on the system. Many Northern Hemisphere lakes show marked increases in Cyclotella spp. and declines in Aulacoseira spp. with climate warming, beginning in the late 19th century for Arctic lakes, and the 1970s for temperate lakes . In Wishart Lake, the increase in Cyclotella spp. is contemporaneous with the Ambrosia rise (Fig 5). This early date for the increase in Cyclotella sensu lato indicates a more complex response and allows for further discussion of the causes of the widespread rise in Cyclotella sensu lato diatoms in temperate lakes.
The Ambrosia rise is the result of multi-faceted landscape changes in the study region. Around 1880 AD mining became a dominant industry in nearby Sault Ste. Marie, Ontario. Many changes occurred such as land clearance, road and railway development, ore smelting and associated increases in acid rain and atmospheric deposition of nitrogen and other pollutants, and increased erosive fluxes into lakes [49, 50]. Therefore the correlation between the increase in Cyclotella sensu lato and the Ambrosia rise may not represent solely anthropogenic climate warming, particularly since syntheses of temperature records tracking anthropogenic warming do not show major temperature anomalies until into the 20th century . In a high resolution diatom record of the period following European settlement taken from a small lake proximal to Lake Michigan, Wolin and Stoermer  similarly show a distinct shift in assemblages in response to initial forest clearance, beginning around 1845 AD.
It has been suggested that recent increases in Cyclotella sensu lato may be related to enrichment by atmospheric deposition of nitrogenous compounds, as Cyclotella sensu lato abundance has been positively correlated in some studies with nutrient concentrations [6, 86, 87]. TLW has experienced minimal local changes in nutrient inputs due its status as a protected watershed with very limited settlement and it has not been extensively logged. While there has been an increase in regional atmospheric nitrogen deposition, there is no indication of nutrient enrichment at Wishart Lake based on an intensive monitoring program since the 1980s . Thus, atmospheric deposition of N is unlikely to explain the recent increase in Cyclotella sensu lato at this site.
Increases in Cyclotella sensu lato occurring around the same time as the Ambrosia rise (1880 AD) with no strong evidence for acidification, large-scale atmospheric nutrient input or rapid climate change at that time, suggest that the observed diatom assemblage changes may reflect an effect of land clearance, vegetation change and associated changes in sediment fluxes to the lake, affecting water clarity. The relationship between the Ambrosia rise and Cyclotella sensu lato in watershed lakes is not well described however a correlation between them is recorded in a number of paleoecological studies of temperate lakes in North America [43, 68, 88, 89]. More high resolution multi-proxy records from temperate lakes have potential to provide new details on the autecology of the Cyclotella sensu lato group, and new insights into the complex mechanisms through which aquatic ecosystems are affected by human activity.
Wishart Lake’s diatom community is largely driven by climate, and impacts of climate on water level, since 7300 cal yr BP. Vegetation succession driven by wetter and cooler conditions ~4000 cal yr BP, alongside potential rising lake levels caused significant increases in nutrient sensitive, planktonic taxa Cyclotella sensu lato spp. and Tabellaria flocculosa. Recent disturbance and landscape change associated with increased Ambrosia pollen correlates to unprecedented rises in Cyclotella sensu lato spp. This record shows Wishart Lake’s sensitivity to changes in the surrounding landscape caused by natural climatic variability pre-1880 AD, and by human impact thereafter. Further research in the Turkey Lakes Watershed is recommended to better understand the relationship between vegetation, aquatic communities and diatom assemblages in this changing region.
We thank the Turkey Lakes Watershed study team for logistical support and data; in particular, we thank Fred Beall and Dean Jeffries for enthusiastic support of this project. We also thank Kara Webster and Ray Semkin for helpful discussions, and Maara Packalen, Nicole Rozario and Magdalena Sobol for research assistance. We also acknowledge three anonymous reviewers for insightful comments which improved the paper.
- Conceived and designed the experiments: KKB SAF.
- Performed the experiments: KKB SAF.
- Analyzed the data: KKB SAF ASM.
- Wrote the paper: KKB SAF ASM.
- 1. Desellas AM, Paterson AM, Sweetman JN, Smol JP. Assessing the effects of multiple environmental stressors on zooplankton assemblages in Boreal Shield lakes since pre-industrial times. Journal of Limnology. 2011;70(1):41–56. WOS:000294320300007.
- 2. Catalan J, Pla-Rabes S, Wolfe AP, Smol JP, Ruhland KM, Anderson NJ, et al. Global change revealed by palaeolimnological records from remote lakes: a review. J Paleolimnol. 2013;49(3):513–35. WOS:000316886900014.
- 3. Smol JP. The power of the past: using sediments to track the effects of multiple stressors on lake ecosystems. Freshwater Biology. 2010;55:43–59. WOS:000273687100004.
- 4. Sorvari S, Korhola A. Recent diatom assemblage changes in subarctic Lake Saanajärvi, NW Finnish Lapland, and their paleoenvironmental implications. J Paleolimnol. 1998;20:205–15.
- 5. Rühland KM, Paterson AM, Smol JP. Lake diatom responses to warming: reviewing the evidence. J Paleolimnol. 2015;54(1):1–35. WOS:000355218100001.
- 6. Saros JE, Anderson NJ. The ecology of the planktonic diatom Cyclotella and its implications for global environmental change studies. Biol Rev. 2015;90(2):522–41. WOS:000352818700009. pmid:24917134
- 7. Rühland KM, Paterson AM, Smol JP. Hemispheric-scale patterns of climate-related shifts in planktonic diatoms from North America and European lakes. Global Change Biol. 2008;14:2740–54.
- 8. Smol JP, Wolfe AP, Birks HJB, Douglas MSV, Jones VJ, Korhola A, et al. Climate-driven regime shifts in the biological communities of Arctic lakes. Proceedings of the National Academy of Sciences (USA). 2005;102(12):4397–402. ISI:000227854800033.
- 9. Harrison SP. Lake levels and climatic change in eastern North America. Clim Dyn. 1989;3:157–67.
- 10. Finkelstein SA, Davis AM. Paleoenvironmental records of water level and climatic changes from the middle to late Holocene at a Lake Erie coastal wetland, Ontario, Canada. Quat Res. 2006;65:33–43.
- 11. Viau AE, Gajewski K. Holocene variations in the global hydrological cycle quantified by objective gridding of lake level databases. J Geophys Res- Atm. 2001;106(D23):31703–16. ISI:000173479100003.
- 12. Almquist H, Dieffenbacher-Krall AC, Flanagan-Brown R, Sanger D. The Holocene record of lake levels of Mansell Pond, central Maine, USA. Holocene. 2001;11(2):189–201. ISI:000167169700005.
- 13. Shuman B, Donnelly JP. The influence of seasonal precipitation and temperature regimes on lake levels in the northeastern United States during the Holocene. Quat Res. 2006;65:44–56.
- 14. Laird KR, Cumming BF. Reconstruction of Holocene lake level from diatoms, chrysophytes and organic matter in a drainage lake from the Experimental Lakes Area (northwestern Ontario, Canada). Quat Res. 2008;69(2):292–305. WOS:000254787600010.
- 15. Shuman B, Bravo J, Kaye J, Lynch JA, Newby P, Webb T III. Late Quaternary water level variations and vegetation history at Crooked Pond, southeastern Massachusetts. Quat Res. 2001;56:401–10.
- 16. Moos MT, Laird KR, Cumming B. Diatom assemblages and water depth in Lake 239 (Experimental Lakes Area, Ontario): implications for paleoclimatic studies. J Paleolimnol. 2005;34(2):217–27. ISI:000229972800006.
- 17. Karmakar M, Laird KR, Cumming BF. Diatom-based evidence of regional aridity during the mid-Holocene period in boreal lakes from northwest Ontario (Canada). Holocene. 2015;25(1):166–77. WOS:000346178000013.
- 18. Wolin JA, Stone JR. Diatoms as indicators of water-level change in freshwater lakes. In: Smol JP, Stoermer EF, editors. The Diatoms: Applications for the Environmental and Earth Sciences. 2nd edition ed. Cambridge: Cambridge University Press; 2010. p. 174–85.
- 19. Liu KB. Holocene Paleoecology of the Boreal Forest and Great-Lakes-St. Lawrence Forest in Northern Ontario. Ecol Monogr. 1990;60(2):179–212. WOS:A1990DE96200003.
- 20. McAndrews JH, Campbell ID. 6 ka mean July temperature in eastern Canada from Bartlein and Webb's (1985) pollen transfer functions: comments and illustrations. In: Telka A, editor. Proxy climate data and models of the six thousand years before present time interval: the Canadian perspective. Ottawa: Canadian Global Change Program Incidental Report Series IR93-3. The Royal Society of Canada; 1993. p. 22–5.
- 21. Shuman B, Newby P, Huang YS, Webb T. Evidence for the close climatic control of New England vegetation history. Ecology. 2004;85(5):1297–310. ISI:000221433300013.
- 22. Davis MB. Quaternary history and the stability of forest communities. In: West DC, Shugart HH, Botkin DB, editors. Forest Succession. New York: Springer; 1981. p. 132–53.
- 23. Sreenivasa MR, Duthie HC. Post glacial diatom history of Sunfish Lake, southwestern Ontario. Canadian Journal of Botany. 1973;51:1599–609.
- 24. Whitehead DR, Charles DF, Jackson ST, Smol JP, Engstrom DR. The developmental history of Adirondack (N.Y.) lakes. J Paleolimnol. 1989;2:185–206.
- 25. Huvane JK, Whitehead DR. The paleolimnology of North Pond: Watershed-lake interactions. J Paleolimnol. 1996;16(3):323–54. ISI:A1996VR64200003.
- 26. St. Jacques JM, Douglas MSV, McAndrews JH. Mid-Holocene hemlock decline and diatom communities in Van Nostrand Lake, Ontario, Canada. J Paleolimnol. 2000;23:385–97.
- 27. Moos MT, Laird KR, Cumming BF. Climate-related eutrophication of a small boreal lake in northwestern Ontario: a palaeolimnological perspective. Holocene. 2009;19(3):359–67. ISI:000265513100002.
- 28. Balasubramaniam AM, Hall RI, Wolfe BB, Sweetman JN, Wang XW. Source water inputs and catchment characteristics regulate limnological conditions of shallow subarctic lakes (Old Crow Flats, Yukon, Canada). Can J Fish Aquat Sci. 2015;72(7):1058–72. WOS:000358607900010.
- 29. Fritz SC, Anderson NJ. The relative influences of climate and catchment processes on Holocene lake development in glaciated regions. J Paleolimnol. 2013.
- 30. Prescott CE, Zabek LM, Staley CL, Kabzems R. Decomposition of broadleaf and needle litter in forests of British Columbia: influences of litter type, forest type, and litter mixtures. Can J Forest Res. 2000;30(11):1742–50. WOS:000165448300008.
- 31. Pomeroy JW, Bewley DS, Essery RLH, Hedstrom NR, Link T, Granger RJ, et al. Shrub tundra snowmelt. Hydrological Processes. 2006;20(4):923–41. WOS:000236332400020.
- 32. Turner KW, Wolfe BB, Edwards TWD, Lantz TC, Hall RI, Larocque G. Controls on water balance of shallow thermokarst lakes and their relations with catchment characteristics: a multi-year, landscape-scale assessment based on water isotope tracers and remote sensing in Old Crow Flats, Yukon (Canada). Global Change Biol. 2014;20(5):1585–603. WOS:000334361000018.
- 33. Stewart KA, Lamoureux SF. Connections between River Runoff and Limnological Conditions in Adjacent High Arctic Lakes: Cape Bounty, Melville Island, Nunavut. Arctic. 2011;64(2):169–82. WOS:000292226100003.
- 34. Medeiros AS, Taylor DJ, Couse M, Hall RI, Quinlan R, Wolfe BB. Biological and nutrient responses to catchment disturbance and warming in small lakes near the Alaskan tundra-taiga boundary. Holocene. 2014;24(10):1308–19. WOS:000342809500011.
- 35. Jeffries DS, Kelso JRM, Morrison IK. Physical, chemical and biological characteristics of the Turkey Lakes Watershed, Central Ontario, Canada. Can J Fish Aquat Sci. 1988;45:3–13. WOS:A1988Q784000001.
- 36. Beall FD, Semkin RG, Jefferies DS. Trends in the Output of First-order Basins at Turkey Lakes Watershed, 1982–96. Ecosystems. 2001;4:514–26.
- 37. Saarnisto M. The deglaciation history of the Lake Superior region and its climatic interpretations. Quat Res. 1974;4:316–39.
- 38. Brandt JP. The extent of the North American boreal zone. Environmental Reviews. 2009;17:101–61. ISI:000273768500007.
- 39. Environment Canada. Turkey Lakes Watershed Meteorological Data. Burlington, Ontario, Canada: National Water Research Institute; 1980–2009
- 40. Kelso JRM. Fish Community Structure, Biomass, and Production in the Turkey Lakes Watershed, Ontario. Can J Fish Aquat Sci. 1988;45:115–20.
- 41. White MS, Xenopoulos MA, Hogsden K, Metcalfe RA, Dillon PJ. Natural lake level fluctuation and associated concordance with water quality and aquatic communities within small lakes of the Laurentian Great Lakes region. Hydrobiologia. 2008;613:21–31.
- 42. Sirois A, Vet R, MacTavish D. Atmospheric Deposition to the Turkey Lakes Watershed: Temporal Variations and Characteristic. Ecosystems. 2001;4:503–13.
- 43. Delorme LD, Duthie HC, Esterby SR, Smith SM, Harper NS. Prehistoric Inferred pH changes in Batchawana Lake, Ontario from Sedimentary Diatom Assemblages. Archiv Fur Hydrobiologie. 1986;108(1):1–22. WOS:A1986F258700001.
- 44. Wright HE Jr. A square rod piston sampler for lake sediments. Journal of Sedimentary Petrology. 1967;37:975–6.
- 45. Glew JR. Miniature gravity corer for recovering short sediment cores. J Paleolimnol. 1991;5:285–7.
- 46. Reimer PJ, Baillie MGL, Bard E, Bayliss A, Beck JW, Blackwell PG, et al. IntCal09 and Marine09 Radiocarbon Age Calibration Curves, 0–50,000 years cal BP. Radiocarbon. 2009;51(4):1111–50.
- 47. Stuiver MP, Reimer PJ, Reimer RW. CALIB 6.0. Belfast: Queen's University, Belfast; 2013.
- 48. St. Jacques JM, Cumming BF, Smol JP. A pre-European settlement pollen-climate calibration set for Minnesota, USA: developing tools for palaeoclimatic reconstructions. Journal of Biogeography. 2008;35(2):306–24. ISI:000252396600012.
- 49. Woodside T, Green H, Evans B, Davies R, Giblin PE. A History of Sault Ste. Marie Forest District: District History Series. Ontario: Department of Lands and Forests; 1963. 121 p.
- 50. McDowall D. Steel at the Sault: Francis H. Clergue, Sir James Dunn, and the Algoma Steel Corporation 1901–1956. Toronto, Canada: University of Toronto Press; 1984. 326 p.
- 51. Fægri K, Iversen J. Textbook of Pollen Analysis. 4th ed. Chichester: John Wiley & Sons; 1989. 328 p.
- 52. Harper NS. Pollen Analysis of Sediments from Turkey Lakes Cores, Algoma Calibrated Watershed, Ontario. National Water Research Institute, editor. Burlington, Canada: Unpublished Report, Canada Centre for Inland Waters; 1982. 47 p.
- 53. Blaauw M. Methods and code for ‘classical’ age-modelling of radiocarbon sequences. Quaternary Geochronology. 2010;5:512–8.
- 54. Battarbee RW, Jones VJ, Flower RJ, Cameron NG, Bennion H, Carvalho L, et al. Diatoms. In: Smol JP, Birks HJB, Last WM, editors. Tracking Environmental Change Using Lake Sediments Volume 3: Terrestrial, Algal, and Siliceous Indicators. Dordrecht: Kluwer; 2001. p. 155–202.
- 55. Antoniades D, Hamilton PB, Douglas MSV, Smol JP. Diatoms of North America: The Freshwater Floras of Prince Patrick, Ellef Ringnes and northern Ellesmere Islands from the Canadian Arctic Archipelago. Ruggell, Liechtenstein: Gantner Verlag; 2008.
- 56. Fallu MA, Allaire N, Pienitz R. Freshwater Diatoms from nothern Quebec and Labrador (Canada). Berlin, Germany: Gebruder Borntraeger; 2000.
- 57. Krammer K, Lange-Beralot H. Bacillariophyceae. Susswasser-flora von Mitteleuropa 2. Stuttgart: Gustav Fischer; 1986–1991
- 58. Lavoie I, Hamilton PB, Compeau S, Grenier M, Dillon P. Guide d'identification des diatomées des rivières de l'Est du Canada. Québec: Presses de l'Université du Québec; 2008.
- 59. Patrick R, Reimer CW. The Diatoms of the United States Exclusive of Alaska and Hawaii 1. Philadelphia: The Academy of Natural Sciences of Philadelphia, Monograph No. 13; 1966.
- 60. Houk V, Klee R. The stelligeroid taxa of the genus Cyclotella (Kützing) Brébisson (Bacillariophyceae) and their transfer into the new genus Discostella gen. Nov. Diatom Research. 2004;19(2):203–28.
- 61. Heiri O, Lotter AF, Lemcke G. Loss on ignition as a method for estimating organic and carbonate content in sediments: Reproducibility and comparability of results. J Paleolimnol. 2001;25:101–10.
- 62. Juggins S. rioja: an R package for the analysis of Quaternary science data, Version 0.7–32012.
- 63. R Development Core Team. R: A language and environment for statistical computing. 2.15.1 ed. Vienna, Austria: R Foundation for Statistical Computing; 2012.
- 64. Juggins S. C2: software for ecological and paleoecological data analysis and visualisation. 1.6.6 ed. University of Newcastle: Department of Geography; 2003.
- 65. Appleby PG. Chronostratigraphic techniques in recent sediments. In: Last WM, Smol JP, editors. Tracking Environmental Change Using Lake Sediments Volume 1: Basin Analysis, Coring and Chronological Techniques. Dordrecht: Kluwer; 2001. p. 171–203.
- 66. Laird KR, Kingsbury MV, Lewis CFM, Cumming BF. Diatom-inferred depth models in 8 Canadian boreal lakes: inferred changes in the benthic:planktonic depth boundary and implications for assessment of past droughts. Quat Sci Rev. 2011;30:1201–17.
- 67. Yang JR, Duthie HC. Regression and weighted averaging models relating surficial sedimentary diatom assemblages to water depth in Lake Ontario. J Great Lakes Res. 1995;21(1):84–94.
- 68. Brugam RB, McKeever K, Kolesa L. A diatom-inferred water depth reconstruction for an Upper Peninsula, Michigan, lake. J Paleolimnol. 1998;20:267–76. WOS:000076951600006.
- 69. Moos MT, Cumming BF. Changes in the parkland-boreal forest boundary in northwestern Ontario over the Holocene. Quat Sci Rev. 2011;30(9–10):1232–42. WOS:000291516100016.
- 70. Viau AE, Gajewski K. Reconstructing Millennial-Scale, Regional Paleoclimates of Boreal Canada during the Holocene. Journal of Climate. 2009;22(2):316–30. WOS:000263786900009.
- 71. Bryant GA. The Auto-Ecology of Cyclotella and Melosira species in Southern Ontario Lakes [MSc]: University of Toronto; 1975.
- 72. Kingston JC. Araphid and Monoraphid Diatoms. In: Wehr J, Sheath R, Kociolek P, editors. Freshwater Algae of North America. San Diego, USA: Academic Press; 2003. p. 595–632.
- 73. Knudson BM, Kipling TH. Ecology of the epiphytic diatom Tabellaria flocculosa (Roth) Kutz. var. flocculosa in three English lakes. Journal of Ecology. 1957;45(1):93–112.
- 74. Moser KA, Smol JP, MacDonald GM. Ecology and distribution of diatoms from boreal forest lakes in Wood Buffalo National Park, Northern Alberta and the Northwest Territories, Canada. Philadelphia, PA: The Academy of Natural Sciences of Philadelphia; 2004. 59 p.
- 75. Davis MB, Douglas C, Calcote R, Cole KL, Winkler MG, Flakne R. Holocene climate in the western Great Lakes National Parks and Lakeshores: Implications for future climatic changes. Conserv Biol. 2000;14:968–83.
- 76. Liu K-B. Holocene paleoecology of the boreal forest and Great Lakes-St Lawrence forest in northern Ontario. Ecol Monogr. 1990;60(2):179–212.
- 77. Viau AE, Gajewski K, Sawada MC, Fines P. Millennial-scale temperature variations in North America during the Holocene. J Geophys Res- Atm. 2006;111(D9): ISI:000237447100001.
- 78. Yu Z, McAndrews JH, Eicher U. Middle Holocene dry climate caused by change in atmospheric circulation patterns: Evidence from lake levels and stable isotopes. Geology. 1997;25(3):251–4.
- 79. Crocker RL, Major J. Soil development in relation to vegetation and surface age at Glacier Bay, Alaska. Journal of Ecology. 1955;43(2):427-&. ISI:A1955XM34900004.
- 80. Jacobson GL, Birks HJB. Soil Development on Recent end Moraines of the Klutlan Glacier, Yukon Territory, Canada. Quat Res. 1980;14(1):87–100. WOS:A1980KF94100007.
- 81. Paré D, Bergeron Y, Camiré C. Changes in the forest floor of Canadian southern boreal forest after disturbance. Journal of Vegetation Science. 1993;4(6):811–8.
- 82. Brais S, Camiré C, Bergeron Y, Paré D. Change in nutrient availability and forest floor characteristics in relation to stand age and forest composition in the southern part of the boreal forest of northwestern Quebec. Forest Ecology and Management. 1995;76:181–9.
- 83. Rühland K, Paterson AM, Smol JP. Hemispheric-scale patterns of climate-related shifts in planktonic diatoms from North American and European lakes. Global Change Biol. 2008;14(11):2740–54. ISI:000260307100022.
- 84. Mann ME, Zhang ZH, Hughes MK, Bradley RS, Miller SK, Rutherford S, et al. Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia. Proc Natl Acad Sci U S A. 2008;105(36):13252–7. ISI:000259251700016. pmid:18765811
- 85. Wolin JA, Stoermer EF. Response of a Lake Michigan coastal lake to anthropogenic catchment disturbance. J Paleolimnol. 2005;33:73–94.
- 86. Wolfe AP, Vinebrooke RD, Michelutti N, Rivard B, Das B. Experimental calibration of lake-sediment spectral reflectance to chlorophyll a concentrations: methodology and paleolimnological validation. J Paleolimnol. 2006;36(1):91–100. WOS:000239393600007.
- 87. Werner P, Smol JP. The distribution of the diatom Cyclotella comensis in Ontario (Canada) lakes. Nova Hedwigia Beiheft. 2006;130:373–92.
- 88. Finkelstein SA, Peros MC, Davis AM. Late Holocene paleoenvironmental change in a Great Lakes coastal wetland: integrating pollen and diatom datasets. J Paleolimnol. 2005;33:1–12.
- 89. Duthie HC, Yang JR, Edwards TWD, Wolfe BB, Warner BG. Hamilton Harbour, Ontario: 8300 years of limnological and environmental change inferred from microfossil and isotopic analyses. J Paleolimnol. 1996;15(1):79–97. WOS:A1996TU74700005.