Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

Contents lists available at ScienceDirect

Journal of Great Lakes Research

j ourna l homepage: www.e lsev ie r .com/ locate / jg l r

Decadal regulation of phytoplankton abundance and water clarity in alarge continental reservoir by climatic, hydrologic and trophic processes

R.J. Vogt a,⁎, S. Sharma b, P.R. Leavitt a

a Limnology Laboratory, Department of Biology, University of Regina, Regina, Saskatchewan, Canadab Department of Biology, York University, Toronto, Ontario, Canada

⁎ Corresponding author at: Department of BiologicPeterborough, Ontario, Canada. Tel.: +1 705 748 1011 (6

E-mail address: RichVogt@gmail.com (R.J. Vogt).

http://dx.doi.org/10.1016/j.jglr.2014.11.0070380-1330/© 2014 International Association for Great Lak

a b s t r a c t

a r t i c l e i n f o

Article history:Received 13 July 2014Accepted 2 September 2014Available online 4 December 2014

Communicated by Rebecca North

Index words:ReservoirWater qualityTransparencyClimateMultiple stressorsChl a

Large continental reservoirs are impacted by multiple stressors including climate change, watershed develop-ment, food-web alteration, water extraction, hydroelectric-power generation, and industrial aquaculture. Suchcomplexity makes it difficult to identify the hierarchical relationships among regulatory processes, forecasteffects of environmental change, or develop adaptive management strategies. Here we present a regression-based analysis of 19 years of limnological change in a representative main-stem reservoir within centralNorth America to suggest that phytoplankton abundance and water clarity are regulated by hierarchical but in-dependent mechanisms. The Qu'Appelle arm of Lake Diefenbaker, Canada, was monitored during summers(May–August) of 1995–2013 to evaluate the unique and interactive effects of continental climate systems,regional meteorology, river hydrology and limnological characteristics on phytoplankton abundance and waterclarity. Regression models explained 48–52% of historical variation despite the absence of pronounced temporalpatterns in monthly or summer phytoplankton abundance and water clarity. Phytoplankton abundance (mainlydiatoms, flagellates) was correlated positively with soluble reactive phosphorus concentrations and inverselywith the density of large herbivores, factors which were themselves correlated to variation in chemical conditions(oxygen, dissolved inorganic carbon) and dissolved organic carbon content, respectively. In contrast, water clarityvaried directly as a function of climate systems (producing warm, dry winters) and inversely with river flow. Weconclude that anticipated climate change (warmer, less runoff) will improve water clarity in lacustrine regions oflarge prairie reservoirs on decadal scales by reducing inorganic turbidity, while nutrient fluxes associated witheconomic development may independently regulate algal abundance.

© 2014 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.

Introduction

As with natural lakes, water quality in large continental reservoirs isimpacted by climate change (Chen et al., 2007; Marce et al., 2010;Sukenik et al., 2012), watershed development (Hall et al., 1999a;Jackson et al., 2001; Leigh et al., 2010) and food-web alteration (Postet al., 2002), all of which interact and vary in intensity through time. Inaddition, reservoirs may be managed for water extraction (Baldwinet al., 2008; Joeckel andDiffendal, 2004), hydroelectric power generation(Soumis et al., 2004) and industrial aquaculture (Figueredo and Giani,2005; Guo and Li, 2002), while ontogenetic changes in internal nutrientsupply can cause a decadal-scale upsurge in algal production (Hall et al.,1999a; Thornton et al., 1990). Such complex interactions betweenmulti-ple stressors makes it difficult to identify the hierarchical relationshipsamong regulatory processes, forecast effects of environmental change,or develop adaptive management strategies to balance economic andenvironmental issues (Crossman et al., 2013; Marce et al., 2010).

al Sciences, Trent University,129)

es Research. Published by Elsevier B

Climate variability and directional warming are expected to havepronounced effects on water resources in the northern Great Plains(Barrow, 2009). Present-day climate variability arises from the interac-tion of threemajor climate systems and three air masses (Arctic, Pacific,Gulf of Mexico) that supply moisture into the continental interior(Bonsal and Shabbar, 2008; Bryson and Hare, 1974). Variation in theNorth Atlantic Oscillation duringwinter (NAOw) appears to regulate cy-clonic activity over the Prairies early each year (Hurrell, 1995; Wanget al., 2006) and is associated with changes in timing of ice melt, algalproduction and development of the clear water phase during spring(Dröscher et al., 2009;Weyhenmeyer et al., 1999). Similarly, the PacificDecadal Oscillation (PDO) (Mantua et al., 1997) and El Niño–SouthernOscillation (ENSO) (Trenberth and Hurrell, 1994) regulate influx of Pa-cific precipitation to the Prairies and runoff from western mountains(Shabbar et al., 2011; St. Jacques et al., 2010) and can interact to produceexceptionally warm, dry conditions in winter and spring (Mantua et al.,1997;McCabe et al., 2004) andwhich, in turn, alter plankton phenology(McGowan et al., 2005b; Winder and Schindler, 2004). Finally, generalcirculationmodels predict that themean annual temperature of the Ca-nadian Prairies will increase ~4 °C by 2050 (Barrow, 2009; Lapp et al.,2012), leading to higher hydrological variability and intensification of

.V. All rights reserved.

82 R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

both droughts (Lapp et al., 2013; van der Kamp et al., 2008) and pluvialperiods (Winter and Rosenberry, 1998). These events will further leadto more variable water chemistry (Pham et al., 2009; Starks et al.,2014), altered regional hydrology (Pomeroy et al., 2007; Schindlerand Donahue, 2006) and changes in heat budgets that regulate plank-tonic production (Johnk et al., 2008; Paerl and Otten, 2013; Winderand Sommer, 2012), community composition (Cantin et al., 2011;Huisman et al., 2004; Magnuson et al., 1997), and outbreaks ofpotentially-toxic cyanobacteria (Donald et al., 2011; Sukenik et al.,2012).

Large reservoirs in continental interiors are often found in fertilecatchments subject to extensive anthropogenic influence including agri-culture (nutrients, erosion, pesticides, drainage), urbanization (microbes,pharmaceuticals, metals, nutrients), and hydrologic modifications(Bolgrien et al., 2009; Fernandez et al., 2014; Hall et al., 1999a). For exam-ple, ancestral European settlement of the Canadian Prairies since 1880has transformed the grasslands due to mechanized cultivation (since1920), fertilization (since the 1940s) and localized urban development(Bunting et al., 2011; Hall et al., 1999b; Leavitt et al., 2006). Over 70% ofsurface cover has been converted to grain and livestock farming whilepermanent surface water cover has declined over 50% to 5–8% of landarea (Bunting et al., 2011; Hall et al., 1999b). Elevated fluxes of nutrientsand suspended solids have increased algal productivity, reduced waterclarity, and intensified blooms of cyanobacteria in natural lakes (Leavittet al., 2006; Orihel et al., 2012) although less is known of the responsesof regional reservoirs (McGowan et al., 2005a; North et al., 2014).Large-scale elimination of wetlands by farming has changed runoff ofsnow and rain (Pomeroy et al., 2007; van der Kamp et al., 2008), an im-portant control on the fluxes of nutrients and other solutes to aquaticecosystems in the grasslands (Pham et al., 2008; Starks et al., 2014). Sim-ilarly, reservoirs in continental interiors are often subject to regulated butextensive (N5 m) variation in lake water-level reflecting the highly sea-sonal nature of local snowmelt (March–April) (Akinremi et al., 1999;Fang and Pomeroy, 2007), the importance of snowmelt from the RockyMountains to the hydrological budgets of main-stem reservoirs (June)(Saskatchewan Water Security Agency, SWSA, 2012), and the use oflarge reservoirs for industrial, urban, and agricultural applications (Hallet al., 1999a).

Complex, hierarchical, and temporally-variable interactions be-tween multiple stressors complicate development of a mechanistic un-derstanding of how humans and climate interact to regulate waterquality and ecosystem structure in reservoirs (Crossman et al., 2013;Hart and Calhoun, 2010). To address this issue, we analyzed a 19-yeartime series from a large lacustrine embayment (Qu'Appelle arm) ofthe Lake Diefenbaker reservoir to quantify how environmental variabil-ity associated with climatic regimes, regional meteorology, hydrologicvariability and intrinsic limnological properties related to human activ-ity (nutrient content, food-web structure, etc.) influences phytoplank-ton abundance and water clarity in continental reservoirs. Such long-term ecological research (LTER) is of particular use in developing robustpredictive models for changes in biological, chemical, and physicalproperties of aquatic ecosystems, and allowing scientists and managersto develop management strategies to adapt to future environmentalchange (Adrian et al., 2009; Chen et al., 2007). The empiricalmodels de-veloped here serve as an important first step to identifying targets forpotential future management action both in this region and in similarsub-humid ecozones around the world (e.g., southern South America,central Asia, central China, Australia).

Methods

Site description

Lake Diefenbaker is a 220-km long reservoir located ~554 m abovesea level (a.s.l.) on the South Saskatchewan River, Canada, and whichserves as headwater to Qu'Appelle River drainage basin (Hall et al.,

1999a; Hecker et al., 2012). The reservoir's catchmentwhich liesmainlywithin Alberta, Saskatchewan, and upstate Montana was originallymixed-grass Prairie, but it is presently under predominantly (~65%) ag-ricultural land use (cereal crops, pasture, livestock) (Hall et al., 1999a).The reservoir lies within a broad glacial outwash valley formed by themeltwaters of the Wisconsin ice sheet (Christiansen, 1960) and isunderlain by cretaceous marine sedimentary bedrock (Bearpaw forma-tion) with a thick overburden of glacial till and predominantly dark-brown chernozemic soils (Acton et al., 1998). There are no majorurban centers in the immediate vicinity of Lake Diefenbaker, althoughCalgary (pop. 990,000), Red Deer (pop. 97,100) and Medicine Hat(pop. 61,200) lie 200–400 km upstream in the upper catchment withinAlberta while smaller Swift Current, Saskatchewan (pop. 15,500) islocated ~50 km south of the reservoir.

Regional climate is classified as sub-humid continental, exhibits amoisture deficit of 30–60 cm/year, and is characterized by short warmsummers (mean 19 °C in July), cold winters (mean−16 °C in January),and lowmean annual temperatures (~1 °C)with high seasonal variabil-ity.Most regional runoff occurs in spring during a brief snowmelt periodin late March or early April (Pham et al., 2009; Pomeroy et al., 2007)although monthly inflows to Lake Diefenbaker are highest in Juneand early July when melt-waters arrive from the Rocky Mountains(Akinremi et al., 1999; Fang and Pomeroy, 2007; SWSA, 2012). Reservoirhydrology varies widely among years with a four-fold difference be-tween minimum and maximum annual inflows since 1995 and a N6-mseasonal variation in lake level (Hall et al., 1999a; Quiñones-Riveraet al., 2015).

Lake Diefenbaker is a large (area = 394 km2), deep (Zmax ~62 m)basin with a relatively short residence time (~1.3 years) characteristicof large riverine reservoirs in central North America (Bolgrien et al.,2009; Hall et al., 1999a). Surface waters in the Qu'Appelle arm (area~50 km2; Zmax ~23 m) are cool and well oxygenated (Dröscher et al.,2009; Sadeghian et al., 2015) with low nutrient content and phyto-plankton abundance relative to other regional lakes (Leavitt et al.,2006; Patoine et al., 2006). In general, the embayment warms slowlyand is polymictic although in some years limited thermal stratificationis observed during late-July or August (Dröscher et al., 2009; Hudsonand Vandergucht, 2015). Phytoplankton are characteristic of well-mixed mesotrophic lakes with abundant diatoms in spring giving wayto later-summer assemblages of flagellates and occasionalcyanobacteria (Hecker et al., 2012; McGowan et al., 2005b; Patoineet al., 2006). The zooplankton community is composed mainly ofcyclopoid (Diacyclops thomasii) and calanoid (Leptodiaptomus siciloides)copepods, but also includes large (Daphnia pulex, Daphnia galeatamendotae, Diaphanosoma birgei, Holopedium gibberum) and small(Bosmina longirostrus, Daphnia retrocurva) cladocerans at lowerdensities (Patoine et al., 2006). Fish assemblages include 25 nativeand stocked species, such as walleye (Sander vitreus), northern pike(Esox lucius), rainbow trout (Oncorhynchus mykiss), lake char (Salvelinusnamaycush), yellow perch (Perca flavescens), cisco (Coregonus artedi),bigmouth buffalo (Ictiobus cyprinellus), and white sucker (Catostomuscommersonii) (for full list, see SWSA, 2012).

Paleolimnological analyses demonstrate that the downstream lacus-trine portions of Lake Diefenbaker have experienced three intervals ofcontrasting water quality since valley inundation in 1967 (Hall et al.,1999a; Lucas et al., 2015; Tse et al., 2015). First, lake production in-creased ca. 1968–1975 when diatoms were characteristic of eutrophicconditions (Stephanodiscus niagarae, Stephanodiscus parvus) and colo-nial and N2-fixing cyanobacteria were present. Second, reservoir pro-duction declined ca. 1975–1986, as mesotrophic diatoms (Tabellariaflocculosa str. IIIp, Fragilaria crotonensis, Aulacoseira ambigua) replacedeutrophic species (Hall et al., 1999a; Lucas et al., 2015). Finally, the tro-phic status of Lake Diefenbaker increased again after ca. 1986, with ele-vated abundance of productive diatoms (A. ambigua), chlorophytes andcolonial cyanobacteria (Hall et al., 1999a; Tse et al., 2015). During this lat-ter transition, water-column chlorophyll (Chl a) increased from ~1 μg/L

83R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

in 1984 to ~5 μg/L in 1994 (Environment Canada, 1988). Recent publica-tions suggest that modern phytoplankton abundance has remained ele-vated (McGowan et al., 2005b; Patoine et al., 2006) and that colonialcyanobacteria may be problematic (Hecker et al., 2012), although therehas been no comprehensive evaluation of water quality change sincethe mid-1990s. Whole-basin carbon budgets demonstrate that Lake Die-fenbaker has been net heterotrophic (respiration N primary production)during the past 20 years (but see Quiñones-Rivera et al., 2015), althoughCO2 sequestration has increased since ca. 2003 in this and other basins(Finlay et al., 2009, 2010).

Climate data

Time series of climate variables (1995–2010) were assembled frompublic archives, including the US National Oceanic and Atmospheric Ad-ministration (NAO), the Australian National Climate Centre (SouthernOscillation Index, SOI), and the University of Washington Joint Institutefor the Study of the Atmosphere and Ocean (PDO). Additionally, we es-timated the known synergistic interaction between El Niño phases ofENSO and positive phases of the PDO as the product of their respectiveindices (Bonsal and Shabbar, 2008; Shabbar and Yu, 2012). The calendarday of year (DOY) of spring ice melt from Lake Diefenbaker wasobtained from Vogt et al. (2011).

Meteorology

Meteorological measurements (1995–2013) within the catchmentwere obtained from the Environment CanadaNational Climate Archives(http://www.climate.weatheroffice.gc.ca/ climateData; April 2014) forthe Outlook, Saskatchewan, weather station situated ~28 km north ofLake Diefenbaker (51°29′N, 107°03′W). Mean monthly air temperature(°C), precipitation (mm) and daily wind speed (km/h) were computedfor an ice-free (‘summer’) period that overlapped with limnologicalsampling (01May–31 August). Snow accumulation (mm)was estimat-ed as the sum of monthly precipitation from January–March in eachcalendar year.

Hydrology

Hydrologic data (1995–2010) were collected by the SaskatchewanWater Security Agency (SWSA) and used to assess the monthly ortotal summer inflow to Lake Diefenbaker (m3/d). Inflowwas estimatedusing standard governmental procedures, based on data from dischargegauges on streams, measurements of lake levels, and estimates of netevaporation (Saskatchewan Water Security Agency, SWSA, 2012 andunpublished).

Limnological characteristics

The Qu'Appelle arm of Lake Diefenbaker is one of two large lacustrineembayments in the reservoir and was sampled bi-weekly during sum-mer 1995–2013 using standard protocols as part of a regional long-term ecological monitoring program (QU-LTER) (Finlay et al., 2009;McGowan et al., 2005b; Patoine et al., 2006). Briefly, depth-integratedsamples were collected at midday at a standard geo-positioned station(51°01.260′N, 106°29.840′W) by pooling casts of a 2-L Van Dorn waterbottle deployed at 1-m intervals to 15 m depth. Integrated sampleswere filtered through a 0.45-μm pore membrane filter and analyzed atthe University of AlbertaWater Chemistry Laboratory for concentrationsof soluble reactive phosphorus (SRP, μg P/L) and total dissolved phos-phorus (TDP, μg P/L), as well as total dissolved nitrogen (TDN), ammoni-um (NH4

+), and the sum of NO22− and NO3

2− (all μg N/L) (Patoine et al.,2006). Total inorganic (TIC) and dissolved organic carbon (DOC) concen-trations (mg C/L) in filtered samples were quantified using a Shimadzumodel 5000A (Shimadzu, Kyoto, Japan) total carbon analyzer followingFinlay et al. (2009). Water temperature (Twater; °C) and oxygen content

(mg O2/L) were computed as average values of depth profiles collectedevery 1 m using a YSI model 85 meter (YSI, Yellow Springs, Ohio,USA). Surface water pH was measured at 0.5-m depth using a calibratedhandheld probe, while water clarity was measured using a 20-cm diam-eter Secchi disk. Zooplankton densities were estimated bi-weekly, May–August for 1995–2012, from vertical tows of a 20-cm diameter Wiscon-sin net (243-μm mesh) at the standard sampling station. Invertebratesamples were preserved and enumerated according to Patoine et al.(2006). Individual species densities (ind./L) were summed as total zoo-plankton, copepods, large cladocerans and small cladocerans (seeabove for species). Rotifers were not enumerated.

Phytoplankton abundance and community composition in the Qu′Appelle arm was estimated 1995–2012 using standard trichromaticand high performance liquid chromatographic (HPLC) analysis of carot-enoid and chlorophyll pigments (Leavitt and Hodgson, 2001; Vogt et al.,2011). Briefly, particulate organic matter (POM) from depth-integratedwater samples (see above) was collected on GF/C glass fiber filters(nominal pore 1.2-μm) and extracted using either pure acetone(trichromatic) or a mixture of acetone:methanol:water (80:15:5, byvolume) (HPLC). Trichromatic estimates of chlorophyll a (Chl a) con-centrations were used as an index of total phytoplankton abundance(μg Chl/L). In contrast, carotenoids, chlorophylls and their derivativeswere isolated and quantified (nmol pigment/L) using an Agilentmodel 1100 HPLC system (Agilent Technologies Inc., Mississauga,Canada) fitted with photodiode-array and fluorescence detectors andcalibrated using authentic standards (Leavitt and Hodgson, 2001).Historical changes in phytoplankton community composition wereestimated using pigment biomarkers from siliceous algae (mainlydiatoms and chrysophytes; fucoxanthin), cryptophytes (alloxanthin),dinoflagellates (peridinin), chlorophytes (Chl b), colonial cyanobacteria(myxoxanthophyll), and N2-fixing cyanobacteria (aphanizophyll). Asshown by Donald et al. (2013), trichromatic- and HPLC-based analysesprovide equivalent estimates of gross phytoplankton abundance andcomposition as do enumerations of cellular biovolume using lightmicroscopy.

Data analysis

For modeling purposes, historical changes in water quality wereassessed using estimates of total phytoplankton abundance (Chl a)and water clarity (Secchi depth) during 1995–2010, the interval inwhich complete predictor data were available. First, linear regressionmodels were developed for both explanatory and response variablesto assess the mode of variation of each time series. Time series weredetrended using first-difference procedures if temporal autocorrelationwas detected. Second, multiple regression models were developed forboth water quality indices using forward selection of predictors withtwo stopping criteria (α and R2

adj; 9999 permutations) from the fullsuite of climatic,meteorological, hydrological and limnological variables(Blanchet et al., 2008). As our goal was to identify potential regulatorymechanisms rather than develop the most parsimonious model, finalmodel composition was not based solely on Akaike's Information Crite-rion (AICc), although typically the final model also exhibited the lowestAICc (analysis not shown). The relative contribution of each predictorwas quantified using a type III sum of squares analysis of variance(ANOVA) in which variation explained by each predictor was summedinto classes representing climate, meteorology, hydrology and internallake characteristics. Subsequently, regression models were developedfor the statistically significant predictors of water quality and clarity,thereby allowing us to evaluate the hierarchical relationships amongpotential regulatory variables. This approach avoids over-fitting withcomplexmodels of poor predictive capabilities (Sharma et al., 2013). Fi-nally, regression analyseswere repeated using time serieswithmonthlyresolution to assess seasonal variation in potential regulatory processes.Variables were transformed (log10) as necessary to produce normal dis-tributions and co-linear parameters were excluded from final models.

84 R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

All data manipulation and statistical analyses were performed in theR-language environment using the lm (linear regression), acf (auto-correlation) and forward.sel (packfor) (forward selection) functions(R Development Core Team, 2014).

Results

Water quality time-series

Linear regression analysis revealed no significant linear trends intotal phytoplankton abundance or water clarity (Fig. 1a, Table 2).When highly productive 1995 was not considered, a small but signifi-cant increase in total phytoplankton abundancewas recorded for the in-terval 1996–2012 (R2adj = 0.17, p=0.05) (Fig. 1a). Regression analysisof time series with monthly resolution showed the increase in phyto-plankton abundance was restricted to July (R2adj = 0.26, p = 0.019)only if 1995 was eliminated, and was not observed in any analysesof data derived solely from May, June or August (all R2

adj = 0.00,p N 0.34). In contrast, there were no significant linear trends in waterclarity recorded in any month (R2adj = 0.00–0.07, p N 0.15), though

Tota

l Phy

topl

ankt

on (µ

g C

hl L

-1)

Wat

er C

larit

y (m

)

0.0

5.0

10.0

15.0

20.0

25.0

30.0

1995 1997 1999 2001 2003

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

1995 1997 1999 2001 200

May June

Tota

l Phy

topl

ankt

on (µ

g C

hl L

-1)

A)

B)

C)

Wat

er C

larit

y (m

)

0.0

5.0

10.0

15.0

20.0

1995 1997 1999 2001 2003

Total Phytoplankt

Fig. 1. Historical changes in mean (±standard error) total phytoplankton abundance (as μg CCanada, based on annual (May–August) (A) andmonthlymonitoring (B, C) during 1995–2013.increase in total algae for the interval 1996–2013.

the greatest water clarity occurred during August and the least duringJune in most years (Fig. 1c).

Phytoplankton community composition

HPLC analysis of phytoplankton community composition revealedthat siliceous algae (diatoms and chrysophytes) composed 50–80%total algal biomass in the Qu'Appelle arm of Lake Diefenbaker in eachyear (Fig. 2). In general, relative abundance of these taxa peaked duringMay, before being partly replaced by cryptophytes (most years), dino-flagellates (9 years) and colonial cyanobacteria (14 years). Overall,diazotrophic cyanobacteria were abundant early (1996–1999) andlate (2007–2012) in the time series, whereas non-N2-fixing colonialcyanobacteriaweremore common in other years (Fig. 2). Cyanobacteriawere uncommon in 2001, 2004 and 2006, whereas this group exhibitedits maximum abundance within the QU-LTER during 2012.

Environmental time series

Regression-based analysis of time series of physico-chemical, cli-matic, meteorological and hydrological parameters (Table 1) revealed

2005 2007 2009 2011 2013

3 2005 2007 2009 2011 2013

July August

2005 2007 2009 2011 2013

on Water Clarity

hl a/L) and water clarity (as Secchi depth, m) in the Qu'Appelle arm of Lake Diefenbaker,Therewere no statistically significant trends in any time series, though therewas amodest

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA MJ JA

N2-fixing cyanob. colonial cyanob. chlorophytes siliceous algae cryptophytes dinoflagellates1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Rel

ativ

e ab

unda

nce

(% p

igm

ent s

um)

Fig. 2.Meanmonthly (May–August) relative abundance of themajor phytoplankton groups in the Qu'Appelle arm of Lake Diefenbaker, Canada (1995–2012). Absolute and relative abun-dance of phytoplankton groups was determined by high performance liquid chromatography of their diagnostic pigments. Phytoplankton groups (and pigments) include N2-fixingcyanobacteria (aphanizophyll), colonial cyanobacteria (myxoxanthophyll), chlorophytes (chlorophyll b), siliceous algae (diatoms and chrysophytes; fucoxanthin), cryptophytes(alloxanthin), and dinoflagellates (peridinin).

85R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

few significant trends during the past 20 years. Total zooplankton den-sities declined significantly and substantially since the mid-1990s(R2adj = 0.38 p = 0.004), with a particularly steady rate of decreasesince 2000 (R2adj = 0.78 p b 0.001) (Fig. 3a). This trend was driven byconcomitant reductions in mean summer densities of large-bodied cla-docerans (R2adj = 0.69, p b 0.001) and copepods (R2adj = 0.47, p =0.002), but not small cladocerans (R2adj= 0.00, p=0.46) (Fig. 3b). Sim-ilarly, the pH of Lake Diefenbaker surface waters increased nearly1.0 unit since 1995 (R2adj = 0.37, p=0.008), similar to patterns record-ed in other prairie lakes (Finlay et al., 2009; Vogt et al., 2011). Of the 20remaining variables, only concentrations of oxygen (increase; R2adj =0.31, p = 0.015) and NH4

+ (decline; R2adj = 0.37, p = 0.007) exhibitedsignificant directional changes during the monitoring period.

Models of summer water quality

Intrinsic limnological characteristicswere the strongest predictors oftotal phytoplankton abundance during 1995–2010, accounting for 48%of historical variation in concentrations of Chl a (Fig. 4a). Specifically,mean phytoplankton abundance during summer was correlated posi-tively to SRP concentration (29% variation explained), and negativelyto total zooplankton abundance (19% variation) (Fig. 4b, Table 2). Anal-ysis of models based on monthly data (not shown) suggested that SRPwas a significant predictor (p b 0.05) of phytoplankton abundance late

Table 1Physico-chemical, morphometric, climatic, hydrological and meteorological variables for the Qcalculated for May-August during 1995–2012. Ranges are standard deviations of themean. Varicarbon (DOC), total inorganic carbon (TIC), depth-integratedwater temperature (Twater), oxygenOscillation (ENSO), and North Atlantic Oscillation (NAO). DOY = calendar day of year.

Physico-chemistry Morphology

SRP (μg/L) 10.5 ± 12.3 Lat (°N)–long (°W) 51.12TDP (μg/L) 18.9 ± 18.1 Elevation (m) 552NO3 (μg/L) 167.2 ± 93.7 Lake area (km2) 394NH4 (μg/L) 18.6 ± 20.9 Mean depth (m) 33DOC (mg/L) 6.8 ± 3.2 Max depth (m) 62TIC (mg/L) 33.6 ± 4.9 Volume (m3) 9.40 ×pH 8.7 ± 0.5 Water residence time (year) 1.3Twater (°C) 13.2 ± 1.0 Gross drainage area (km2) 1.36 ×O2 (mg/L) 9.0 ± 1.6 Effective drainage area (km2) 8.69 ×Total zoopl. (ind/L) 23.8 ± 14.5Chlorophyll a (μg/L) 5.4 ± 2.5Secchi depth (m) 3.4 ± 0.4

in summer (July–August), while effects of zooplankton were restrictedto July (analysis not shown). In contrast, changes in mean summerwater clarity (52% variation explained) were predicted by hydrologicinflow (37% variation) and the prevalent climate system (ENSO-PDO;15% variation), but not by limnological or meteorological parameters(Fig. 4b, Table 2).

Additional regression models suggested that the controls of waterquality (SRP, zooplankton)were themselves under independent regula-tion. For example, water-column concentrations of SRP were correlatednegatively to oxygen content and positively to concentrations of inor-ganic C (R2adj = 0.48 p=0.006), whereas total zooplankton abundancewas correlated negatively to concentrations of DOC (R2adj = 0.21 p b

0.05) (Table 2). Regressionmodels were not developed for the principalpredictors of water clarity (hydrologic inflow, ENSO-PDO), becausethose factors are regulated by processes outside the spatial domain ofour data collection (snowmelt in Rocky Mountains and Pacific sea-surface temperature, respectively) (Lapp et al., 2013; St. Jacques et al.,2010).

Discussion

Analysis of a comprehensive suite of limnological and environmen-tal variables from the lacustrine zone of a major continental reservoir(Table 1) demonstrated that phytoplankton abundance, community

u'Appelle arm of Lake Diefenbaker. Mean physico-chemical and meteorological values areables include soluble reactive phosphorus (SRP), total dissolved P (TDP), dissolved organicconcentration (O2), and indices of the Pacific Decadal Oscillation (PDO), El Niño–Southern

Climate, hydrology, meteorology

–106.63 Ice out date (DOY) 120.7 ± 8.8PDO 0.03 ± 0.74ENSO 0.54 ± 6.24NAO −4.82 ± 12.79Annual inflow (m3) 7.0 ± 2.6 × 109

109 Summer temperature (°C) 15.9 ± 1.0Mean monthly precipitation (May–Aug) (mm) 50.5 ± 17

105 Mean monthly precipitation (Jan–Dec) (mm) 29.5 ± 8.6104 Winter snowfall (mm) 133 ± 55

Wind speed (km/h) 11.9 ± 3.7

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

50.00

1995 1998 2001 2004 2007 2010 20130.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

1995 1998 2001 2004 2007 2010 2013

Tota

l Zoo

plan

kton

(ind

L-1

)

Year

Zoop

lank

ton

(indL

-1)

A) B)

CopepodsLarge CladoceransSmall Cladocerans

Fig. 3. Mean summer density of (A) total zooplankton density (ind./L) and (B) copepods (Leptodiaptomus siciloides, Diacyclops thomasii), large cladocerans (Daphnia spp.) and smallcladocerans collected from the Qu'Appelle arm of Lake Diefenbaker, Canada, during May–August 1995–2012.

86 R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

composition and water clarity have been resilient to climatic and an-thropogenic stressors during the past 20 years (Figs. 1, 2). All three pa-rameters show little evidence of directional change during 1995–2013,consistent with recent paleolimnological quantification of phytoplank-ton abundance in lacustrine regions since 1967 (Tse et al., 2015) and30 years of Landsat imagery (Yip et al., 2015), but in contrast to contin-ued increases in regional population, agricultural production and cli-mate warming, and to declining river discharge (Bunting et al., 2011;Leavitt et al., 2006; St. Jacques et al., 2010). Instead, half of observedinter-annual variation in mean-summer phytoplankton abundance(mainly diatoms,flagellates)was explained by trophic processes (nutri-ent availability, herbivory) (Table 2, Fig. 3), whereas a similar propor-tion of changes in water clarity were explained mainly by physicalcontrols (ENSO-PDO, river inflow) (Table 2). Given that the food-webstructure and hydrological cycle of Lake Diefenbaker are similar tothose observed for many of the largest reservoirs in the United States(Bolgrien et al., 2009; Bowersox et al., 2014; Lynott et al., 1995),we sug-gest that our findings provide a useful model for lacustrine regulatoryprocesses in reservoirs in much of central and western North America.Specifically, we forecast that future declines in river inflow frommoun-tain snowfields (Lapp et al., 2013; St. Jacques et al., 2010) should reducethe inorganic turbidity, whereas continued economic development thatelevates nutrient supply will likely favor increased phytoplanktonabundance and possibly cyanobacterial outbreaks in lacustrine embay-ments (Donald et al., 2011; Leigh et al., 2010). Consequently, selectionof the appropriate management strategy will depend on whether suchreservoirs are exploited principally for industrial, recreational ordomestic uses.

Phytoplankton abundance

Moderate phytoplankton abundance (Fig. 1) and diatom-rich com-munity composition (Fig. 2) are consistent with observations in otherlarge temperate reservoirs in the Great Plains of North America(Bolgrien et al., 2009; Joeckel and Diffendal, 2004). In these largelakes, hydrologic and nutrient influx (Finlay et al., 2009; Patoine et al.,2006) is controlled mainly by snowmelt in westerly mountain rangesduring early summer (Lapp et al., 2013; St. Jacques et al., 2010), ratherthan local snowmelt during March and April (Fang and Pomeroy,2007; Pham et al., 2009; Quiñones-Rivera et al., 2015). Typically,water from mountain ranges is cold, has lower nutrient content thanseen in natural prairie lakes (Pham et al., 2008, 2009), and exhibitshigh turbidity associatedwith suspended inorganic particulates, leading

to low algal production (Dubourg et al., 2015). Diatoms predominatewhen turbulence related to inflow is greatest (spring, early summer)(Hall et al., 1999a), while flagellates (cryptophytes, chlorophytes, dino-flagellates) are increasingly common as inflow slows, turbulence de-clines, and thermal stratification may occur (Abirhire et al., 2015;Dröscher et al., 2009; McGowan et al., 2005b). These patterns are inmarked contrast to many natural lakes in the prairie ecozone (Donaldet al., 2013; Leavitt et al., 2006; Pham et al., 2008), as well as somelarge reservoirs from continental interiors of China (Lv et al., 2014),Australia (Bowling et al., 2013), and South America (Fernandez et al.,2014), where high turbidity reflects dense blooms of cyanobacteria aris-ing from excessive supply of P and N (Leigh et al., 2010; Sukenik et al.,2012). Differences in production between the large reservoirs of westernNorth America and natural prairie lakes may reflect differences in thepredominant source of nutrients to the water-column of artificial (al-lochthonous, riverine) and natural aquatic ecosystems (autochthonous,sedimentary) (e.g., Leavitt et al., 2006; Abirhire et al., 2015; Patoineet al., 2006).

Absence of decadal-scale trends in water quality is consistent withobservations from other large prairie lakes (Bunting et al., 2011;Patoine et al., 2006), Landsat analysis of this reservoir (Yip et al.,2015), and recent paleolimnological analysis of phytoplankton fossilsin Lake Diefenbaker sediments (Yip et al., 2015). Although phytoplank-ton abundance has increased N300% above baseline in many large prai-rie lakes, most cultural eutrophication occurred by 1990 mainly due tointensification of grain and livestock production during the 20th centu-ry (Bunting et al., 2011;Hall et al., 1999a). Similarly, although thedown-stream reaches of Lake Diefenbaker experienced a small trophic surgefor several years following inundation of the valley in 1967 (Hall et al.,1999a; Lucas et al., 2015), phytoplankton abundance declined beforeexpanding again to a stable plateau in the late 1980s and early 1990s(Tse et al., 2015). Although speculative, recent research suggests thatmodernwater quality conditionsmay be the result of an interaction be-tween long-term land-use patterns and a sudden persistent change inclimatic forcing within the northern Great Plains (McCullough et al.,2012).

Multiple regression models suggest that nutrient supply fromreservoir sediments during summer may be an important control ofdecadal variability in total phytoplankton abundance. Specifically,depth-integrated estimates of phytoplankton abundancewere correlat-ed positively with that of SRP, whereas the latter was correlatednegatively with water-column oxygen level and positively with DICcontent (Table 2). Bottle bioassays conducted biweekly during summers

A) Water Quality Models

% V

aria

tion

Exp

lain

ed

100

90

80

70

60

50

40

30

20

10

0Total Phytoplankton Water Clarity

R2 ad

j= 0

.48

R2 ad

j= 0

.52

Internal

Hydrology

Climate

TotalPhytoplankton Water Clarity

B) Specific Predictor Variables

SRP Zoop. PDO-ENSO Inflow

-70%

+30%

-40%

+60%

Fig. 4. (A) Proportion of total explained variation (open bars) and sources of explained variation (filled bars) in total phytoplankton abundance (as μg Chl a/L) and water clarity (as Secchidepth,m) as identified inmultiple regression analysis of mean summer (May-August) conditions in theQu'Appelle arm of Lake Diefenbaker, Canada, 1995–2010.Model performancewasevaluated by adjusted coefficient of determination (R2adj). Significant (p b 0.05) explanatory variables were selected by forward selectionmultiple regression based on 9999 permutationsandwere classified into categories associatedwith variation in climate systems (black), river hydrology (dotted), and internal lake characteristics (waves). (B) Potentially-causal relation-ships among water quality variables and independent predictors were identified from multiple regression models (Table 2). Predictors include soluble reactive phosphorus (SRP), meantotal zooplankton density (Zoop.), indices of interactions between the Pacific Decadal Oscillation (PDO) and El Niño-Southern Oscillation (ENSO), and riverine inflow to Lake Diefenbaker(Inflow). Numbers and signs reflect the proportion of explained variation inwater quality parameters (ovals) attributable to variation in predictor variables (squares) due to inhibitory (-)or stimulatory (+) interactions.

87R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

since 1994 as part of theQU-LTER program reveal thatwhilemost down-streamQu'Appelle lakes exhibit pronounced algal limitation by supply ofN (Donald et al., 2011; Leavitt et al., 2006), instantaneous algal growth inthe lacustrine portion of Lake Diefenbaker is regulated mainly by P sup-ply (Dubourg et al., 2015; Hecker et al., 2012; Quiñones-Rivera et al.,2015). Further, diazotrophic cyanobacteria are rare (Fig. 2) (Abirhireet al., 2015;McGowan et al., 2005b) and there is no net fixation of atmo-spheric N2 inmost years (Patoine et al., 2006). Instead, Lake Diefenbakeris a substantial net sink for P flux through the South Saskatchewan Riverdrainage basin (North et al., 2015). As shown by Nürnberg (1998),water-column P concentrations are thought to be influenced by the de-velopment of deepwater hypoxia (but see Hupfer and Lewandowski,2008), mainly due to declines in the effectiveness of the oxidizedmicrozone as a P trap with loss of O2 (reviewed in Wetzel, 2001).Although not directly measured in this study, the positive correlationbetween SRP and DIC concentrations is also consistent with the effectsof organic matter decomposition in sediments both reducing oxygentensions and increasing water-column DIC levels, as respiratory CO2

reacts with insoluble sedimentary carbonate to produce two dissolvedbicarbonate molecules (Wetzel, 2001). Consistent with this hypothesis,sediments of the Qu'Appelle arm of Lake Diefenbaker are over 50%inorganic carbonate by mass (Finlay et al., 2010) and appear to be asignificant source of P to the water column (North et al., 2015).

Internal nutrient loading is also hypothesized to be responsible fordecade-long surges in planktonic production in reservoirs. In this case,flooding of dry river valleys leaches soluble nutrients from formerlyterrestrial soils, erodes newly formed shorelines, and decomposesinundated vegetation (reviewed in Thornton et al., 1990). Oxygen con-sumption during vegetation decomposition favors deepwater anoxia,further increasing nutrient flux from sediments (see above). However,continued erosion of shorelines also buries submerged organic sub-strates, reduces oxygen demand and nutrient release, and reduces pro-duction by benthic communities in favor of planktonic assemblages,leading to a decline in lake production (Hall et al., 1999a; Hewlettet al., 2015; Thornton et al., 1990). Although this ontogenetic patterncan be modified by differences in the original habitat type (terrestrial,

Table 2Summary of regression models predicting total phytoplankton abundance (as μg Chl/L), water clarity (Secchi depth, m), soluble reactive phosphorus concentration (μg P/L) and totalzooplankton density (individuals/L). Models are based on 16 years of monitoring in the Qu'Appelle arm of Lake Diefenbaker, Canada (May–August, 1995–2010). Model performance issummarized using an adjusted coefficient of determination (R2adj), and models were significant with a probability level of p. Multiple regression models include predictor variables foreach water quality index selected using a Monte Carlo forward selection. Predictors selected include river inflow (m3/month), indices of Pacific Decadal Oscillation (PDO) and El Niño–Southern Oscillation (ENSO), oxygen concentration (mg O2/L), inorganic carbon concentration (TIC, mg C/L), and dissolved organic C content (mg C/L).

Water quality variable Linear regression Multiple regression

R2adj p Slope R2adj p Predictor Coefficient % total variationexplained

% explainedvariation

Total phytoplankton(μg Chl/L)

0.02 0.09 0.60 SRP − total zooplankton 0.48 0.005 SRP 1.67 29.0 60.4Total zoop −2.29 19.0 39.6Residual 52.0Intercept 8.97

Water clarity(m)

0.00 0.57 0.00 −Inflow + PDO–ENSO 0.52 0.003 Inflow −2.74 × 10−8 36.6 70.5PDO–ENSO 7.5 × 10−3 15.3 29.5Residual 48.1Intercept 1.68

Soluble reactive phosphorus(μg P/L)

0.001 .33 −0.05 −O2 + TIC 0.48 0.006 O2 −0.42 34.6 72.6TIC 0.09 13.0 27.4Residual 52.4Intercept 2.95

Total zooplankton(ind./L)

0.38 0.004 −0.07 −DOC 0.21 0.04 DOC −0.09 21.0 100.0Residual 79.0Intercept 3.66

88 R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

aquatic), area of flooded land, degree of shoreline development, valleygeometry, water turbidity and magnitude of water level fluctuations(Hecky and Guildford, 1984; Thornton et al., 1990), studies generallyagree that autothchonous nutrient sources can be important controlsof decadal-scale development of reservoir productivity independent ofallochthonous sources (Thornton et al., 1990). Better estimation of theimportance of internal nutrient sources is essential for Lake Diefenbakerand other large reservoirs in central North America (North et al., 2015),as these water bodies supply much of the potable water to urban cen-tres in the continental interior (e.g., Saskatchewan Water SecurityAgency, SWSA, 2012).

Recent changes in the grazing regime in Lake Diefenbaker may havealso had an important effect on phytoplankton abundance during thepast 20 years (Table 2, Fig. 3).While quantitative estimates of fish abun-dances are not available, themarked and persistent decline in density oflarge zooplankton (Fig. 3) coincides with a documented escape of~360,000 rainbow trout from a local fish farm in the year 2000(Saskatchewan Water Security Agency, SWSA, 2012; SaskatchewanEnvironment, unpublished data). This event was one of a series ofunintended releases during the past 15 years. Rainbow trout have avaried diet in large reservoirs of central North America, but preferlarge-bodied zooplankton and other invertebrates when fish are up to46 cm in length (Lynott et al., 1995). Consistent with the effects ofsize-selective predation by visually-orienting fish, densities of large-bodied Daphnia and copepods declined nearly 75% between 2000 and2012 (Fig. 3a), whereas the abundance of small cladocerans was un-changed (Fig. 3b). Similarly, research on other large western lakes(e.g., L. Tahoe) reveals that stocked rainbow trout prefer D. thomasii(Elser et al., 1995), one of the two predominant copepods in easternLake Diefenbaker. As noted by Vogt et al. (2011), D. thomasii andthe equally large-bodied Leptodiaptomus siciloides are the biomass-dominant taxa in the Qu'Appelle arm and other regional lakes, andboth taxa are strongly omnivorous (Patoine et al., 2006). Finally, analy-sis of time series with monthly resolution demonstrates that phyto-plankton abundance increased only during July, a month in whichedible flagellates are most common (Fig. 2), densities of large herbi-vores are normally greatest (Dröscher et al., 2008, 2009), and relaxationof predation by fish would be expected to have large effects on phyto-plankton abundance. Thus while variation in invertebrate grazing ap-peared to explain less variation (19%) in phytoplankton abundancethan does inter-annual variation in SRP supply (29%), our regression

analysis and supporting evidence suggest that decadal-scale variationin phytoplankton abundance is under mainly trophic control.

Water clarity

Multiple regression analysis shows that inter-annual variation inwater clarity is regulated mainly by changes in climatic conditionsthat regulate river discharge and inorganic turbidity (Table 2). Specifi-cally, the Qu'Appelle arm of Lake Diefenbaker was more transparent inyears with dry winters (ENSO-PDO) and low discharge of the SouthSaskatchewan River (inflow). Runoff to reservoirs in the NorthernGreat Plains is derived from two main water sources; local snowmeltduring March–April (Fang and Pomeroy, 2007; Pham et al., 2009), andinflux of alpine meltwaters via east-flowing rivers in June (e.g., SouthSaskatchewan, Missouri) (Quiñones-Rivera et al., 2015; St. Jacqueset al., 2010). While the former combines with regional groundwatersources to sustain natural lakes in the Canadian Prairies (Pham et al.,2009; van der Kamp et al., 2008), discharge of large prairie rivers thatregulate reservoir hydrology is controlled mainly by release in alpinesnow packs and upstream hydrologic management (Lapp et al., 2013).

Timing and magnitude of release of water from mountain sourcesappears to be controlled in part by variation in the PDO (Lapp et al.,2012, 2013; Schindler and Donahue, 2006), with anomalously low dis-charge occurring when ENSO and PDO systems reinforce each other tosplit the winter jet stream and induce warm dry winters in westernCanada and northern USA (Mantua et al., 1997; McCabe et al., 2004;Shabbar et al., 2011). In general, years with low river discharge resultin lower turbidity associated with suspended solids (Thornton et al.,1990), because of both lower influx and reduced erosion of reservoirshorelines (Coakley and Hamblin, 1969; Hewlett et al., 2015).Interestingly, although both ENSO and PDO systems are known to regu-late spring climate, trophic interactions, phytoplankton compositionand primary production (Dröscher et al., 2009; Weyhenmeyer et al.,1999), our regression analysis revealed no evidence of direct effects ofclimate, hydrology or regional meteorological condition on mean sum-mer phytoplankton abundance or composition within the Qu'Appellearm (Table 2). Instead, we conclude that water clarity and phyto-plankton abundance in large lacustrine reaches are under parallelbut largely independent regulatory mechanisms (physical and trophic,respectively).

89R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

Forecasts of reservoir response to climate and anthropogenic stressors

General circulationmodels and downscaled regional climatemodelspredict that the North American grasslands will warm ~4 °C within 40years (Barrow, 2009). Such effectsmay be particularly noticeable in cen-tral Canada,whereminor changes inmean annual temperatures (~1 °C)change the state of water (solid, liquid). Although precipitation on thenorthern Great Plains is expected to increase and becomemore variablein coming decades, elevated evaporation associated with atmosphericwarming is expected to offset increased receipt of snow and rain, pro-ducing amore arid climate (Lapp et al., 2012).We infer that river inflowto large prairie reservoirs (Diefenbaker, Oahe, Sakakawea, Fort Peck)will remain a key regulatory mechanism in the future, given that baseflow in prairie rivers has already declined nearly 50% since 1900 dueto climate change and industrial use (St. Jacques et al., 2010), futurePDO systems may be weaker than at present (Lapp et al., 2012), andthe snowmelt in the Rockies is expected to occur earlier and decline inmagnitude (St. Jacques et al., 2010). Based on our analyses, we antici-pate that these declineswill favor less turbidwater suitable for domesticand agricultural consumption, particularly in downstream lacustrineregions.

Continued strong urban and industrial growth in much of the Cana-dian Prairies, combinedwith expected economic recovery in the centralUnited States, may increase nutrient influx to regional reservoirs. Forexample, application of agricultural fertilizers has risen steadily since1960 and is expected to nearly double again by 2050 (Bogard et al.,2012; Donald et al., 2011). Similarly, stronger commodity prices(grain, oil, potash), combined with plentiful regional reserves, suggestthat anthropogenic nutrient influx to large reservoirs will continue toincrease due to future economic development. While these nutrientsources do not appear to have degraded water quality in eastern LakeDiefenbaker during the past 20 years (Fig. 1, Table 2) (Tse et al., 2015;Yip et al., 2015), modern and paleolimnological analysis of large region-al water bodies suggest that anthropogenic effects may develop slowly,before reaching a tipping point beyond which water quality changessuddenly (Bunting et al., 2011). Further research is needed to determinewhether increased phytoplankton production after 1990 reflects such athreshold effect (Bunting et al., 2011), the result of a trophic upsurgeduring reservoir ontogeny (Hall et al., 1999a; Thornton et al., 1990), ora sudden change in regional climatic forcing mechanisms(McCullough et al., 2012).

Regression analysis of nearly two decades of limnological monitor-ing suggests that optimal management strategy for lacustrine reachesof large reservoirs within central North Americamay depend on the rel-ative importance of industrial, recreational and domestic uses of theecosystem, as well as whether the reservoir is managed to optimizephytoplankton abundance or water clarity. Longer reservoir residencetimes will favor less turbid water suitable for consumption by livestockand humans, although this strategy will likely reduce generation of hy-droelectric power (SaskatchewanWater Security Agency, SWSA, 2012)and may favor potentially-toxic blooms of cyanobacteria (Fig. 2) be-cause of reduced dilution of internal nutrient sources and increasedwater clarity (Dubourg et al., 2015). Similarly,management for industri-al aquaculture will likely increase nutrient supply (Figueredo and Giani,2005; Guo and Li, 2002), while reducing herbivore control of phyto-plankton (Fig. 3, Table 2). Finally, managersmust recognize that specificstrategies are likely to interactwith continued economic growth and cli-matic variability to produce complex, hierarchical pathways of ecosys-tem regulation (Leavitt et al., 2009).

Acknowledgments

We thank the members of the Limnology Laboratory for assistancewith data collection since 1994, and Curtis Halborg and Bart Oegemaat the Saskatchewan Water Security Agency for estimates of riverdischarge. This work was supported by the Natural Sciences and

Engineering Research Council of Canada Discovery Grants (NSERC),the Canada Research Chair Program, Canada Foundation for Innovation,the Province of Saskatchewan, and the University of Regina, YorkUniversity and Trent University. We thank two reviewers and editorRebecca North for valuable comments that improved this paper.

References

Abirhire, O., North, R.L., Hunter, K., Vandergucht, D., Sereda, J., et al., 2015. Environmentalfactors influencing phytoplankton in Lake Diefenbaker, Saskatchewan, Canada. J.Great Lakes Res. 41 (Supplement 2), 118–128.

Acton, D.F., Padbury, G.A., Stushnoff, C.T., 1998. The Ecoregions of Saskatchewan.Canadian Plains Research Centre, Regina, Canada.

Adrian, R., O'Reilly, C.M., Zagarese, H., Baines, S.B., et al., 2009. Lakes as sentinels of climatechange. Limnol. Oceanogr. 54, 2283–2297.

Akinremi, O.O., McGinn, S.M., Cutforth, H.W., 1999. Precipitation trends on the CanadianPrairies. J. Clim. 12, 2996–3003.

Baldwin, D.S., Gigney, H., Wilson, J.S., Watson, G., Boulding, A.N., 2008. Drivers of waterquality in a large storage reservoir during a period of extreme drought. Water Res.42, 4711–4724.

Barrow, E., 2009. Climate scenarios for Saskatchewan. Prairie Adaptation ResearchCollaborative Report, Regina, Saskatchewan.

Blanchet, F.G., Legendre, P., Borcard, D., 2008. Forward selection of explanatory variables.Ecology 89, 2623–2632.

Bogard, M.J., Donald, D.B., Finlay, K., Leavitt, P.R., 2012. Distribution and regulation of ureain lakes of central North America. Freshwat. Biol. 57, 1277–1292.

Bolgrien, D.W., Scharold, J.V., Angradi, T.R., Corry, T.D., Schwieger, E.W., Kelly, J.R., 2009.Trophic status of three large Missouri reservoirs. Lake Reserv. Manag. 25, 176–190.

Bonsal, B., Shabbar, A., 2008. Impacts of large-scale circulation variability on lowstreamflows over Canada: a review. Can. Water Resour. J. 33, 137–154.

Bowersox, B.J., Scarnecchia, D.L., Miller, S.E., 2014. Distribution, abundance, and verticalmigrations of Leptodora kindtii in a mainstem Missouri reservoir Montana. J.Freshwat. Ecol. 29, 171–186.

Bowling, L.C., Merrick, C., Swann, J., Green, D., Smith, G., Neilan, B.A., 2013. Effects ofhydrology and river management on the distribution, abundance and persistence ofcyanobacterial blooms in the Murray River, Australia. Harmful Algae 30, 27–36.

Bryson, R.A., Hare, F.K., 1974. Climates of North America. Elsevier.Bunting, L., Leavitt, P.R.,Wissel, B., Laird, K.R., Cumming, B.F., St. Amand, A., Engstrom, D.R.,

2011. Sudden ecosystem state change in LakeWinnipeg, Canada, caused by eutrophi-cation arising from crop and livestock production during the 20th century. FinalReport to Manitoba Conservation and Water Stewardship (67 pp. http://www.gov.mb.ca/waterstewardship/water_quality/lake_winnipeg/pdf/report_lake_wpg_paleolimnology_2011.pdf).

Cantin, A., Beisner, B.E., Gunn, J.M., Prairie, Y.T., Winter, J.G., 2011. Effects of thermoclinedeepening on lake plankton communities. Can. J. Fish. Aquat. Sci. 68, 260–276.

Chen, H., Guo, S.L., Xu, C.Y., Singh, V.P., 2007. Historical temporal trends of hydro-climaticvariables and runoff response to climate variability and their relevance in waterresource management in the Hanjiang basin. J. Hydrol. 344, 171–184.

Christiansen, E.A., 1960. Geology and ground-water resources of the Qu'Appelle area,Saskatchewan. Geology Division Report No. 1, Saskatchewan Research Council,Saskatoon, Canada.

Coakley, J.P., Hamblin, P.F., 1969. Investigation of Bank Erosion and Nearshore Sedimenta-tion in Lake Diefenbaker. Canada Center for Inland Waters, Burlington.

Crossman, J., Whitehead, P.G., Futter, M.N., Jin, L., Shahgedanova, et al., 2013. The interac-tive responses of water quality and hydrology to changes in multiple stressors, andimplications for the long-term effective management of phosphorus. Sci. TotalEnviron. 454, 230–244.

R Development Core Team, 2014. R: A Language and Environment for Statistical Comput-ing. R Foundation for Statistical Computing, Austria.

Donald, D.B., Bogard, M.J., Finlay, K., Leavitt, P.R., 2011. Comparative effects of urea,ammonium, and nitrate on phytoplankton abundance, community composition,and toxicity in hypereutrophic freshwaters. Limnol. Oceanogr. 56, 2161–2175.

Donald, D.B., Bogard, M.J., Finlay, K., Bunting, L., Leavitt, P.R., 2013. Phytoplankton-specificresponse to enrichment of phosphorus-rich surface waters with ammonium, nitrateand urea. PLoS One 8, e53277.

Dröscher, I., Finlay, K., Patoine, A., Leavitt, P.R., 2008. Daphnia control of the spring clear-water phase in six polymictic lakes of varying productivity and size. Verh. Int. Ver.Limnol. 30, 186–190.

Dröscher, I., Patoine, A., Finlay, K., Leavitt, P.R., 2009. Climate control of the spring clear-water phase through the transfer of energy and mass to lakes. Limnol. Oceanogr.54, 2469–2480.

Dubourg, P., North, R.L., Hunter, K., Vandergucht, D., Abirhire, O., et al., 2015. Light and nu-trient co-limitation of phytoplankton communities in a large reservoir: Lake Diefen-baker, Saskatchewan, Canada. J. Great Lakes Res. 41 (Supplement 2), 129–143.

Elser, J.J., Luecke, C., Brett, M.T., Goldman, C.R., 1995. Effects of food web compensationafter manipulation of rainbow trout in an oligotrophic lake. Ecology 76, 52–69.

Environment Canada, 1988. Lake Diefenbaker and Upper South Saskatchewan Riverwater quality study. Report WQ 111. Environment Canada, Burlington, Ont.

Fang, X., Pomeroy, J.W., 2007. Snowmelt runoff sensitivity analysis to drought on theCanadian prairies. Hydrol. Process. 21, 2594–2609.

Fernandez, C., Caceres, E.J., Parodi, E.R., 2014. Phytoplankton development in a highlyeutrophic man-made lake from the Pampa plain of Argentina — a functionalapproach. Int. J. Environ. Res. 8, 1–14.

90 R.J. Vogt et al. / Journal of Great Lakes Research 41 Supplement 2 (2015) 81–90

Figueredo, C.C., Giani, A., 2005. Ecological interactions between Nile tilapia (Oreochromisniloticus L.) and the phytoplankton community of the Furnas Reservoir (Brazil).Freshwat. Biol. 50, 1391–1403.

Finlay, K., Leavitt, P.R., Wissel, B., Prairie, Y.T., 2009. Regulation of spatial and temporalvariability of carbon flux in six hard-water lakes of the northern Great Plains. Limnol.Oceanogr. 54, 2553–2564.

Finlay, K., Leavitt, P.R., Patoine, A., Wissel, B., 2010. Magnitudes and controls of organicand inorganic carbon flux through a chain of hard-water lakes on the northernGreat Plains. Limnol. Oceanogr. 55, 1551–1564.

Guo, L.G., Li, Z.J., 2002. Effects of nitrogen and phosphorus from fish cage-culture on thecommunities of a shallow lake in middle Yangtze River basin of China. Aquaculture226, 201–212.

Hall, R.I., Leavitt, P.R., Dixit, A.S., Quinlan, R., Smol, J.P., 1999a. Limnological succession inreservoirs: a paleolimnological comparison of two methods of reservoir formation.Can. J. Fish. Aquat. Sci. 56, 1109–1121.

Hall, R.I., Leavitt, P.R., Quinlan, R., Dixit, A.S., Smol, J.P., 1999b. Effects of agriculture, urban-ization, and climate on water quality in the northern Great Plains. Limnol. Oceanogr.44, 739–756.

Hart, D.D., Calhoun, J.K., 2010. Rethinking the role of ecological research in the sustainablemanagement of freshwater ecosystems. Freshwat. Biol. 55 (Suppl. 1), 258–269.

Hecker, M., Khim, J.S., Giesy, J.P., Li, S., Ryu, J., 2012. Seasonal dynamics of nutrient loadingand chlorophyll a in a northern prairies reservoir, Saskatchewan Canada. J. WaterResour. Protect. 4, 180–202.

Hecky, R.E., Guildford, S.J., 1984. Primary productivity of Southern Indian Lake before,during, and after impoundment and Churchill River diversion. Can. J. Fish. Aquat.Sci. 41, 591–604.

Hewlett, C., North, R.L., Johansson, J., Vandergaucht, D., Hudson, J., 2015. Contribution ofshoreline erosion to nutrient loading of the Lake Diefenbaker reservoir, Saskatche-wan, Canada. J. Great Lakes Res. 41 (Supplement 2), 110–117.

Hudson, J., Vandergucht, D., 2015. Spatial and temporal patterns in physical propertiesand dissolved oxygen in Lake Diefenbaker, a large reservoir on the Canadian Prairies.J. Great Lakes Res. 41 (Supplement 2), 22–33.

Huisman, J., Sharples, J., Stroom, J., Visser, P., Kardinaal, W., et al., 2004. Changes in turbu-lent mixing shift competition for light between phytoplankton species. Ecology 85,2960–2970.

Hupfer, M., Lewandowski, J., 2008. Oxygen controls the phosphorus release from lakesediments — a long-lasting paradigm in limnology. Int. Rev. Hydrobiol. 93, 415–432.

Hurrell, J.W., 1995. Decadal trends in the North-Atlantic Oscillation — regional tempera-tures and precipitation. Science 269, 676–679.

Jackson, R.B., Carpenter, S.R., Dahm, C.N., McKnight, D.M., Naiman, R.J., et al., 2001. Waterin a changing world. Ecol. Appl. 11, 1027–1045.

Joeckel, R.M., Diffendal, R.F., 2004. Geomorphic and environmental change around a large,aging reservoir: Lake C.W. McConaughy, Western Nebraska USA. Environ. Eng.Geosci. 10, 69–90.

Johnk, K.D., Huisman, J., Sharples, J., Sommeijer, B., Visser, P.M., Stroom, J.M., 2008.Summer heatwaves promote blooms of harmful cyanobacteria. Glob. Chang. Biol.14, 495–512.

Lapp, S.L., St. Jacques, J.M., Barrow, E., Sauchyn, D.J., 2012. GCM projections for the PacificDecadal Oscillation under greenhouse forcing for the early 21st century. Int. J.Climatol. 32, 1423–1442.

Lapp, S.L. St, Jacques, J.M., Sauchyn, D.J., Vanstone, J.R., 2013. Forcing of hydrologicvariability in the northwestern Great Plains since AD 1406. Quat. Int. 310, 47–61.

Leavitt, P.R., Hodgson, D.A., 2001. Sedimentary pigments. In: Smol, J.P., Birks, H.J.B., Last,W.M. (Eds.), Tracking Environmental Change Using Lake SedimentsTerrestrial,Algal, and Siliceous Indicators vol. 3. Kluwer, pp. 295–325.

Leavitt, P.R., Brock, C.S., Ebel, C., Patoine, A., 2006. Landscape-scale effects of urban nitro-gen on a chain of freshwater lakes in central North America. Limnol. Oceanogr. 51,2262–2277.

Leavitt, P.R., Fritz, S.C., Anderson, N.J., Baker, P.A., Blenckner, T., et al., 2009.Paleolimnological evidence of the effects on lakes of energy and mass transferfrom climate and humans. Limnol. Oceanogr. 54, 2330–2348.

Leigh, C., Burford, M.A., Roberts, D.T., Udy, J.W., 2010. Predicting the vulnerability of reser-voirs to poor water quality and cyanobacterial blooms. Water Res. 44, 4487–4496.

Lucas, B.T., Liber, K., Doig, L.E., 2015. Reconstructing diatom and chironomid assemblagesto infer environmental spatiotemporal trends within Lake Diefenbaker, a narrowriver-valley reservoir on the Canadian Prairies. J. Great Lakes Res. 41 (Supplement2), 45–55.

Lv, H., Yang, J., Liu, L.M., Yu, X.Q., Yu, Z., Chiang, P.C., 2014. Temperature and nutrients aresignificant drivers of a seasonal shift in phytoplankton community from a drinkingwater reservoir, subtropical China. Environ. Sci. Pollut. Res. 21, 5917–5928.

Lynott, S.T., Bryan, S.D., Hill, T.D., Duffy, W.G., 1995. Monthly and size-related changesin the diet of rainbow trout in Lake Oahe, South Dakota. J. Freshwat. Ecol. 10,399–407.

Magnuson, J.J., Webster, K.E., Assel, R.A., Bowser, C.J., et al., 1997. Potential effects ofclimate changes on aquatic systems: Laurentian Great Lakes and Precambrian ShieldRegion. Hydrol. Process. 11, 825–871.

Mantua, M.J., Hare, S., Zhang, Y., Wallace, J.M., Francis, R.C., 1997. A Pacific interdecadalclimate oscillation with impacts on salmon production. Bull. Am. Meteorol. Soc. 78,1069–1080.

Marce, R., Rodriguez, M.A., Garcia, J.C., Armengol, J., 2010. El Niño–Southern Oscillationand climate trend impact reservoir water quality. Glob. Chang. Biol. 16, 2857–2865.

McCabe, G.J., Palecki, M.A., Betancourt, J.L., 2004. Pacific and Atlantic Ocean influences onmultidecadal drought frequency in the United States. Proc. Natl. Acad. Sci. U. S. A. 101,4136–4141.

McCullough, G.K., Page, S.J., Hesslein, R.H., Stainton, M.P., et al., 2012. Hydrological forcingof a recent trophic surge in Lake Winnipeg. J. Great Lakes Res. 38 (Suppl. 3), 95–105.

McGowan, S., Leavitt, P.R., Hall, R.I., 2005a. A whole-lake experiment to determine theeffects of winter droughts on shallow lakes. Ecosystems 7, 694–708.

McGowan, S., Patoine, A., Graham, M.D., Leavitt, P.R., 2005b. Intrinsic and extrinsiccontrols of lake phytoplankton synchrony. Verh. Int. Ver. Limnol. 29, 794–798.

North, R.L., Khan, N.H., Ahsan, M., Prestie, C., Korber, D.R., et al., 2014. Relationship be-tween water quality parameters and bacterial indicators in a large prairie reservoir:Lake Diefenbaker, Saskatchewan Canada. Can. J. Microbiol. 60, 243–249.

North, R.L., Johansson, J., Vandergucht, D., Doig, L.E., Liber, K., et al., 2015. Evidence for inter-nal phosphorus loading in a large prairie reservoir (Lake Diefenbaker, Saskatchewan). J.Great Lakes Res. 41 (Supplement 2), 91–99.

Nürnberg, G.K., 1998. Prediction of annual and seasonal phosphorus concentration instratified and polymictic lakes. Limnol. Oceanogr. 43, 1544–1552.

Orihel, D.M., Bird, D.F., Brylinsky, M., Chen, H., Donald, D.B., et al., 2012. High microcystinconcentrations occur only at low nitrogen-to-phosphorus ratios in nutrient-richlakes. Can. J. Fish. Aquat. Sci. 69, 1457–1462.

Paerl, H.W., Otten, T.G., 2013. Harmful cyanobacterial blooms: causes, consequences, andcontrols. Microb. Ecol. 65, 995–1010.

Patoine, A., Graham, M.D., Leavitt, P.R., 2006. Spatial variation of nitrogen fixation in lakesof the northern Great Plains. Limnol. Oceanogr. 51, 1665–1677.

Pham, S.V., Leavitt, P.R., McGowan, S., Peres-Neto, P., 2008. Spatial variability of climateand land-use effects on lakes of the northern Great Plains. Limnol. Oceanogr. 53,728–742.

Pham, S.V., Leavitt, P.R., McGowan, S., Wissel, B., Wassenaar, L., 2009. Spatial and temporalvariability of prairie lake hydrology as revealed using stable isotopes of hydrogen andoxygen. Limnol. Oceanogr. 54, 101–118.

Pomeroy, J.W., Gray, D.M., Brown, T., Hedstrom, N.R., Quinton, W.L., et al., 2007. The coldregions hydrological model: a platform for basing process representation and modelstructure on physical evidence. Hydrol. Process. 21, 2650–2667.

Post, J.R., Sullivan, M., Cox, S., Lester, N.P., Walters, C.J., et al., 2002. Canada's recreationalfisheries: the invisible collapse? Fisheries 27, 6–17.

Quiñones-Rivera, Z.J., Finlay, K., Vogt, R.J., Leavitt, P.R., Wissel, B., 2015. Hydrologic, meta-bolic and chemical regulation of water-column metabolism and atmospheric CO2 ex-change in a large continental reservoir during spring and summer. J. Great Lakes Res.41 (Supplement 2), 144–154.

Sadeghian, A., de Boer, D., Hudson, J., Wheater, H., Linderschmidt, K.-E., 2015. Lake Diefen-baker temperature and mixing model. J. Great Lakes Res. 41 (Supplement 2), 8–21.

Saskatchewan Water Security Agency (SWSA), 2012. State of Lake Diefenbaker.Government of Saskatchewan, Regina, Saskatchewan.

Schindler, D., Donahue,W.F., 2006. An impending water crisis in Canada's western prairieprovinces. Proc. Natl. Acad. Sci. U. S. A. 103, 7210–7216.

Shabbar, A., Yu, B., 2012. Intraseasonal Canadian winter temperature responses to inter-annual and interdecadal Pacific SST modulations. Atmosphere-Ocean 50, 109–121.

Shabbar, A., Bonsal, B.R., Szeto, K., 2011. Atmospheric and oceanic variability associatedwith growing season droughts and pluvials on the Canadian Prairies. Atmosphere-Ocean 49, 339–355.

Sharma, S., Magnuson, J.J., Mendoza, G., Carpenter, S.R., 2013. Influences of local weather,large-scale climatic drivers, and the ca. 11 year solar cycle on lake ice breakup dates;1905–2004. Clim. Chang. 118, 857–870.

Soumis, N., Duchemin, E., Canuel, R., Lucotte, M., 2004. Greenhouse gas emissions fromreservoirs of the western United States. Global Biogeochem. Cycles 18, GB3022.

St. Jacques, J.M., Sauchyn, D.J., Zhao, Y., 2010. Northern Rocky Mountain streamflowrecords: global warming trends, human impacts or natural variability? Geophys.Res. Lett. 37, L06407.

Starks, E., Cooper, R., Leavitt, P.R., Wissel, B., 2014. Effects of drought and pluvial periodson fish and zooplankton communities in prairie lakes: systematic and asystematicresponses. Glob. Chang. Biol. 20, 1032–1042.

Sukenik, A., Hadas, O., Kaplan, A., Quesada, A., 2012. Invasion of Nostocales (Cyanobacteria)to subtropical and temperate freshwater lakes — physiological, regional, and globaldriving forces. Front. Microbiol. 3, 86.

Thornton, K.W., Kimmel, B.L., Payne, F.E. (Eds.), 1990. Reservoir Limnology: EcologicalPerspectives. John Wiley and Sons, New York.

Trenberth, K.E., Hurrell, J.W., 1994. Decadal atmospheric–ocean variations in the Pacific.Clim. Dyn. 9, 303–319.

Tse, T.J., Doig, L.E., Leavitt, P.R., Quiñones-Rivera, Z.J., Codling, G., et al., 2015. Long-termspatial trends in sedimentary algal pigments in a narrow river-valley reservoir,Lake Diefenbaker, Canada. J. Great Lakes Res. 41 (Supplement 2), 56–66.

van der Kamp, G., Keir, D., Evans, M., 2008. Long-termwater level changes in closed-basinlakes of the Canadian Prairies. Can. Water Resour. J. 33, 23–28.

Vogt, R.J., Rusak, J.A., Patoine, A., Leavitt, P.R., 2011. Differential effects of energy and massinflux on the landscape synchrony of lake ecosystems. Ecology 92, 1104–1114.

Wang, X.L., Wan, H., Swail, V.R., 2006. Observed changes in cyclone activity in Canada andtheir relationships to major circulation regimes. J. Clim. 19, 896–915.

Wetzel, R.G., 2001. Limnology. Academic Press.Weyhenmeyer, G.A., Blenckner, T., Pettersson, K., 1999. Changes of the plankton spring

outburst related to the North Atlantic Oscillation. Limnol. Oceanogr. 44, 1788–1792.Winder, M., Schindler, D.E., 2004. Climatic effects on the phenology of lake processes.

Glob. Chang. Biol. 10, 1844–1856.Winder, M., Sommer, U., 2012. Phytoplankton response to a changing climate.

Hydrobiologia 598, 5–16.Winter, T., Rosenberry, D., 1998. Hydrology of prairie pothole wetlands during drought

and deluge: A 17-year study of the cottonwood lake wetland complex in NorthDakota in the perspective of longer term measured and proxy hydrological records.Clim. Chang. 40, 189–209.

Yip, H.D., Johansson, J., Hudson, J.J., 2015. A 29-year assessment of the water clarity andchlorophyll-a concentration of a large reservoir: Investigating spatial and temporalchanges using Landsat imagery. J. Great Lakes Res. 41 (Supplement 2), 34–44.