Author's personal copy - Limnologylimnology.missouri.edu/pdf/north/Northetal2012.pdf · Author's personal copy Distribution of seston and nutrient concentrations in the eastern basin

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Distribution of seston and nutrient concentrations in the eastern basin of Lake Eriepre- and post-dreissenid mussel invasion

Rebecca L. North a,⁎, Ralph E.H. Smith a, Robert E. Hecky a,1, David C. Depew a,2, Luis F. León a,3,Murray N. Charlton b, Stephanie J. Guildford a,1

a Department of Biology, University of Waterloo, 200 University Ave. W., Waterloo, ON, Canada N2L 3G1b National Water Research Institute, Environment Canada, 867 Lakeshore Road, P.O. Box 5050, Burlington, ON, Canada L7R 4A6

a b s t r a c ta r t i c l e i n f o

Article history:Received 29 September 2011Accepted 20 June 2012Available online 19 July 2012

Communicated by Hunter Carrick

Index words:EutrophicationLake ErieLakesNutrient cyclingPhytoplanktonDreissenids

Increased human population growth, reduction of phosphorus (P) loading, and the invasion of dreissenidmussels may have changed the spatial pattern and relationships between the nearshore and the offshoreseston and nutrient concentrations in the eastern basin of Lake Erie over the past 30 years. We comparedseston characteristics, nutrient concentrations, and phytoplankton nutrient status between nearshore andoffshore zones in years before (1973–1985) and after (1990–2003) the dreissenid invasion. In 1973 (theonly pre-dreissenid year nearshore data was collected), chlorophyll a (chla) and nutrient concentrationswere higher nearshore than offshore. In post-dreissenid years, nearshore chla concentrations became signif-icantly lower than the offshore, while carbon (C):chla ratios became higher, which was related to musselgrazing and possibly photoacclimation. Phosphorus deficiency in the phytoplankton increased over the30-year period, and in the post-dreissenid years was less acute in the nearshore than offshore. Mean watercolumn irradiance became higher in the nearshore relative to the offshore in the post-dreissenid years. Thenutrient changes and phytoplankton physiology were consistent with the expected effects of nutrient cyclingby mussels and diminished demand by phytoplankton despite increased demand from benthic algae in thenearshore. This basin-scale study suggests that dreissenid mussel invasion can be associated with alterationsin the spatial pattern of water column properties in large lakes even on open coasts with vigorous circulationand exchange.

© 2012 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.

Introduction

Lake Erie is the shallowest, most productive, and most heavily im-pacted of the Laurentian Great Lakes of North America. Over the lastthree decades there have been many changes to Lake Erie includingincreased population growth and concentrated animal feeding opera-tions (CAFOs) in the watershed, nutrient controls on point sources,and the introduction of exotic invaders such as dreissenid mussels.All of these changes might be expected to impact the nearshore

regions of the lake differently than the offshore, and make Lake Eriean interesting model system for investigating spatially-structuredlake processes.

In response to excessive algal growth in the 1970s (Burns et al.,1976), nutrient controls were initiated with the U.S.–Canada GreatLakes Water Quality Agreements (GLWQAs) of 1972 and 1978.They specified a 30 μmol L−1 total phosphorus (TP) effluent limiton municipal sewage treatment plants discharging in excess of3, 800 m3 day−1 and bans on phosphorus (P) in detergents (Dolanand McGunagle, 2005). Reduced P load from municipal sources waspredicted to lead to a reduction in the total algal biomass in the lakevia P limitation. These P controls did not show immediate results,but by the mid-1980s, declines in lake-wide P loadings were obvious(Dolan and McGunagle, 2005), and decreased total phytoplanktonbiomass (Makarewicz, 1993) suggested that Lake Erie water qualitywas improving. By the mid 1980s, the TP target load set by theGLWQA of 11,000 metric tonnes year−1 (MTA) was achieved (Dolanand McGunagle, 2005). From 1995 to 2007 there was an increasingtrend in the absolute soluble reactive P (SRP) loads, despite the factthat TP loading has remained relatively constant (inclusive ofweather-induced variability; Ohio Lake Erie phosphorus task force,2010). The increasing SRP is of particular concern because it is more

Journal of Great Lakes Research 38 (2012) 463–476

⁎ Corresponding author at: Department of Chemistry, Trent University, 1600 WestBank Drive, Peterborough, ON, Canada K9J 7B8. Tel.: +1 705 748 1011x7243.

E-mail addresses: [email protected] (R.L. North), [email protected](R.E. Hecky), [email protected] (D.C. Depew), [email protected] (L.F. León),[email protected] (M.N. Charlton), [email protected] (S.J. Guildford).

1 Current address: Department of Biology and Large Lakes Observatory, University ofMinnesota Duluth, 2205 East 5th Street, Duluth, MN, USA, 55812. Tel.: +1 218 7267926.

2 Current address: School of Environmental Studies, Queen's University, 116 BarrieSt., Kingston, ON, Canada K7L 3N6.

3 Current address: National Water Research Institute, Environment Canada, 867Lakeshore Road, P.O. Box 5050, Burlington, ON, Canada L7R 4A6.

0380-1330/$ – see front matter © 2012 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.doi:10.1016/j.jglr.2012.06.012

Contents lists available at SciVerse ScienceDirect

Journal of Great Lakes Research

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


Fig. 1. The eastern basin of Lake Erie showing survey stations. A)Historical basin-wide survey stations sampled by the EPA (1973), Environment Canada (1979–1990) and theDFO (1993, 1994,1998). Ten stationswere removed (shown as crossed out station numbers) based on circulation patterns as identified in theMethods section. B) Vertically averaged circulation patterns for theeastern basin of Lake Erie in August 2003 showing the 6 basin-wide survey stations sampled by Environment Canada in the post-dreissenid period (1994–1999) and the additional 15basin-wide survey stations sampled from 2001 to 2003. Four stationswere removed based on circulation patterns, STP discharges, and tributary influences as identified in theMethods section.C) North-shore survey stations sampled in 2001, 2002 and 2003. Contour lines are shown in 5 m increments with the 20 m contour identified by shading.

464 R.L. North et al. / Journal of Great Lakes Research 38 (2012) 463–476


bioavailable to support algal growth than TP (Ohio Lake Eriephosphorus task force, 2010).

The zebra mussel (Dreissena polymorpha) invaded Lake Erie in1988 (Hebert et al., 1989) and first appeared in the eastern basin inAugust of 1989 (Mackie, 1991), followed closely by the quagga mus-sel (Dreissena bugensis; Mills et al., 1993). Herein, both zebra andquagga mussels will be collectively referred to as dreissenid mussels.In this study examining changes in the distribution of seston and nu-trients, we focused on the eastern basin of Lake Erie. In 2002, 90% ofdreissenid biomass in Lake Erie on both soft and hard substrateswas found in the eastern basin (Patterson et al., 2005). As well, tribu-tary influences are less than the other basins because the east basinreceives most of its nutrient load from the offshore waters of the cen-tral basin (Dolan and McGunagle, 2005). The evidence of dreissenidimpacts in Lake Erie is mixed, depending on the season and basinstudied. While some authors attribute temporal changes in nutrientsand phytoplankton to dreissenids (Barbiero et al., 2006; Conroy et al.,2005; Makarewicz et al., 2000), others report that most of the changehappened prior to mussel invasion (Charlton et al., 1999). The major-ity of these studies focused on offshore waters.

Hecky et al. (2004) hypothesized that dreissenid mussels wereimpacting nutrient cycling through their filter-feeding behavior andthe impact was most strongly affecting the shallow, nearshore re-gions of large lakes. Their “nearshore shunt” hypothesis suggeststhat by intercepting, detaining, and redirecting both energy and nu-trient flow, mussels have changed the ecological function of the near-shore and its relationship to the offshore. There have been a numberof studies examining the effects of P loading controls and mussels onthe nearshore of Lake Erie, including the eastern basin (Depew et al.,2006; Nicholls and Hopkins, 1993; Smith et al., 2007), and collective-ly, they suggest substantial decreases of chla, primary production,phytoplankton biomass, and chla:TP ratios associated with arrival ofmussels in the nearshore. None of these studies, however, made sys-tematic comparisons between nearshore and offshore on the basinscale and over extended time periods (i.e., more than 1 or 2 years).The most systematic studies of the impact of dreissenid mussels onnearshore areas published to date come from Saginaw Bay of LakeHuron, where decreases of chla and particulate nutrients and in-creases in Secchi transparency depths were documented in the near-shore areas and not in the offshore (Fahnenstiel et al., 1995; Heath etal., 1995; Johengen et al., 1995). However, because Saginaw Bay ismore isolated from offshore waters compared to open shore-lines which characterize eastern Lake Erie, it is not clear wheth-er the mussel-associated changes documented for Saginaw Baycan be generally applied to other nearshore locations on theGreat Lakes.

The seston and nutrient concentrations reported for the nearshoreand offshore regions of large lakes have historically been different,with the nearshore regions of Lakes Erie, Ontario, and Michigan gen-erally reported to be more eutrophic than the offshore waters(Beeton and Edmondson, 1972; Glooschenko et al., 1974; Heathcoteet al., 1981). The importance of understanding nearshore processesin North American Great Lakes is critical, as current water quality is-sues affecting human health (e.g., toxic algal blooms and taste andodor issues in drinking water (Watson et al., 2008) and nuisanceCladophora blooms (Auer et al., 2010)) have been linked to processesoccurring or originating in the nearshore. Evidence is accumulatingthat responses in the offshore (Barbiero et al., 2006; Charlton andMilne, 2004) may no longer be adequate predictors of change in thenearshore (Auer et al., 2010; Hecky et al., 2004).

Three research questions are addressed in this study: 1) has thereduced external P loading altered phytoplankton biomass and nutri-ent status as well as nutrient concentrations in the eastern basin? 2)has the nearshore-offshore distribution of seston and nutrients chan-ged in the eastern basin after mussel establishment? 3) are thereconsistent post-dreissenid differences in nearshore–offshore seston

concentrations and nutrient status? To answer these questions, datacollected in the eastern basin of Lake Erie by multiple agencies frompre-dreissenid years (1973–1985) were compared to data frompost-dreissenid years (1990–2003) to examine patterns in nearshoreand offshore measurements of chla, light parameters, and nutrientchemistry (P, nitrogen (N), and silica (Si)) over three decades of ob-servations. For the second question concerning changing spatial pat-terns before and after mussel invasion, only 1 year of nearshore andoffshore survey data were available.

Methods

Study site and surveys

The eastern basin is the deepest of Lake Erie's three basins (maxi-mum depth of 64 m) and is considered to be oligotrophic (Charlton etal., 1999; Fig. 1). This study combined surveys conducted in the east-ern basin by multiple agencies including the U.S. Environmental Pro-tection Agency (EPA), Environment Canada (Env. Can.), the CanadianDepartment of Fisheries and Oceans (DFO), and the University of Wa-terloo (UW), to examine long-term patterns in nearshore and off-shore distributions of seston and nutrients (Table 1). In this studywe define nearshore or coastal waters as b20 m and offshore watersas ≥20 m. This classification criterion was selected to remain consis-tent with previous studies in the eastern basin (Depew et al., 2006),and because the 20-m depth is the typical depth limit for themid-summer surface mixed layer. In offshore areas deeper than20 m, the phytoplankton in the mixed layer will be isolated from di-rect mussel influence as a result of thermal stratification during thesummer period. Water circulation patterns were also used to identifyexceptions to this simple nearshore–offshore classification based ondepth. Although most of the surveys contained data spanning April–November, we only used the data from the stratified period of June–September due to complex early- and late-season circulation dynam-ics such as thermal bars or fall storms. The summer months were alsoselected with reference to the surface water temperature (data notshown) as the season when mussels would have the most impactdue to their optimal temperature range (8–25 °C; Stanczykowska etal., 1975) for filtration and growth in the nearshore.

Station selection

While nearshore–offshore classifications were primarily based onwater depth, we refined this classification by using water circulationpatterns. Of the 21 basin-wide stations sampled in 2001–2003, fourstations (444, 447, 448 and 442; Fig. 1B) were removed from analysesdue to horizontal water circulation evidence challenging their near-shore–offshore classification or evidence of significant tributary andsewage treatment plant (STP) influences. We examined monthlyvertically-averaged water circulation patterns for May–October,2001–2003 as predicted by the hydrodynamic model ELCOM (Leónet al., 2005) in the eastern basin of Lake Erie. In Fig. 1B, we showthe pattern from August 2003 to illustrate typical circulation patternspresent when seston and nutrient concentrations were measured(June–September). The most obvious circulation features during Au-gust are the gyres present in the center of the basin (Fig. 1B), whichare one of the important mechanisms for nearshore–offshore trans-port in the Great Lakes (Rao and Schwab, 2007). Based on circulationpatterns and STP influences (identified by seasonal chloride (Cl−)patterns (data not shown) and proximity), stations 444, 447, and448 (Fig. 1B) were removed from all further analyses as being, tovarying extents, more individual cases than general representativesof their category. Station 442 (Fig. 1B) was removed based on circula-tion patterns and its location downstream of the Grand River, thelargest river system entering the north-shore of the eastern basin(He et al., 2006). The Grand River plume has been identified as one

465R.L. North et al. / Journal of Great Lakes Research 38 (2012) 463–476


of the main sources negatively affecting the water quality of the sur-rounding area (Rao and Schwab, 2007), and Nicholls et al. (1983)demonstrated an impact of the Grand River on north-shore nutrientsand phytoplankton.

Select historical EPA, Env. Can., and DFO stations were also re-moved from the analyses using the same criteria as applied to thebasin-wide 2001–2003 surveys for reasons of consistency andunder the assumption that the major circulation patterns in theeastern basin have not changed in the last 35 years. Ten historicalstations sampled by the EPA in 1973 were removed from the analy-ses to give a spatially comparable set of stations to compare withother years. This EPA survey is the only pre-dreissenid year availableto us to compare nearshore and offshore distributions of nutrientparameters. Station 11 was removed due to its proximity to theGrand River discharge, stations 8 and 16 were removed as offshorestations due to nearshore influences and stations 17–22 and 64were removed due to their proximity to and influence by centralbasin waters (Fig. 1A).

To examine the differences between nearshore and offshore loca-tions for temporal consistency at a higher spatial resolution, indepen-dent surveys were conducted along the north shore of the easternbasin of Lake Erie in 2001–2003 (Fig. 1C). In 2001, a five station tran-sect was sampled across increasing depths and distance from thecoast (stations 500 (deepest), 501, 502, 503, and 504; Fig. 1C). In2002 and 2003, a six-station grid composed of three, 5-m stations(503, 507, and 508) and three, 20-m stations (501, 505, and 506)was used to increase replication of nearshore and offshore stations(Fig. 1C).

Water quality parameters

Variables describing physical conditions, and nutrient and chlaconcentrations extracted from each government agency and surveyare referenced in Table 2. The EPA survey in 1973 was the onlypre-dreissenid survey that sampled nearshore stations, and they col-lected discrete water samples from a 1-m depth at all stations (GreatLakes Laboratory, 1974). Env. Can. collected epilimnetic whole watersamples from either discrete depths within the epilimnion (ideallytwice the Secchi depth) or with an integrating sampler from thesurface to a maximum depth of 20 m, or to the top of the thermocline,defined by a change in water temperature gradient >1 °C m−1. In theDFO surveys water was collected from 1 m below the surface to 1 mabove the epilimnetic mixing depth (Zmix) using a diaphragm pumpin 1993 (Dahl et al., 1995) and 1994 (Graham et al., 1996), and awater column integrator at deep stations in 1998 (MacDougall et al.,2001). In our basin-wide surveys (2001–2003) epilimnetic wholewater samples were collected with an integrating sampler followingthe same methods as the Env. Can. surveys. In our north-shore highspatial resolution surveys, epilimnetic whole water samples were col-lected with an integrating tube sampler from surface to a maximumdepth of 10 m or to the top of the thermocline.

For the surveys from 2001 to 2003, picoplankton size fractionationwas conducted by filtering 500 mL of lake water through 2-μmpore-size polycarbonate membrane filters. The filtrate (b2 μm) wasthen treated exactly as the sample water for the determination ofpicoplankton (0.2–2 μm) particulate C and chla concentrations. Thehistorical surveys did not phaeophytin-correct their chla concentra-tions, while the basin-wide (2001–2003) surveys did (Table 2). To re-solve these methodological differences, we compared phaeophytin-corrected chla concentrations with uncorrected concentrations forrepresentative samples during 2001–2003 for which both methodswere applied to the same samples. The relationship between UWbasin-wide phaeophytin-corrected chla concentrations and Env.Can. (NLET) phaeophytin-corrected chla concentrations (r2=0.544,pb0.001, n=61) was compared to the relationship between UWbasin-wide phaeophytin-corrected chla concentrations and NLETuncorrected chla concentrations (r2=0.587, pb0.001, n=89). Dueto the fact that the NLET uncorrected chla concentrations weremore comparable to the UW basin-wide phaeophytin-corrected chlaconcentrations, and that there was more uncorrected chla historicaldata available, all of the historical chla concentrations reportedwere uncorrected for phaeophytin concentrations.

Samples for phytoplankton analysis from the north-shore surveys(2002 and 2003) were preserved with Lugol's iodine, followed byformaldehyde, and subsequently analyzed by the inverted micro-scope method to species resolution (Algal Taxonomy and Ecology,Winnipeg, MB, CAN).

Phosphorus deficiency indicators

Phytoplankton phosphorus deficiency was characterized usingparticulate C:P and N:P stoichiometric ratios (Guildford et al., 2005;Healey, 1973) and alkaline phosphatase activity (APA; Healey andHendzel, 1979a). Phosphorus deficiency was assessed according tothe criteria developed by Healey and Hendzel (1979b). A C:P molarratiob129 indicated no P deficiency, 129–258 indicated moderate Pdeficiency and C:P>258 indicated extreme P deficiency. Alkalinephosphatase activity is an enzymatic assay used to determine P defi-ciency. Alkaline phosphatase is an enzyme localized on the cell sur-face of algal and bacterial cells that is produced when the algae areP deficient. This enzyme removes the phosphate molecules from dis-solved organic P compounds, therefore, utilizing an otherwiseunavailable source of P (Ammerman et al., 2003). Alkaline phospha-tase activity was measured fluorometrically (Healey and Hendzel,1979a), using 5 μmol L−1 of o-methyl-fluorescein-phosphate as thesubstrate. Parallel determinations were made of total and soluble ac-tivities to distinguish between APA associated with particles and APAin solution, the soluble activity being that passing through 0.2 μmpore-size filters. The difference was reported as particulate activity.Alkaline phosphatase activity (μmol Pμg chla−1 h−1)b0.003 indicat-ed no P deficiency, 0.003–0.005 indicated moderate P deficiency and>0.005 indicated extreme P deficiency.

Table 1Overview of historical, basin-wide, and north shore surveys.

Year Station name Agency Data source

1973 1–22, 62–64 EPA Great Lakes Laboratory, 1974; Fig. 1A1979, 1983–85, 1990 1–4 Env. Cana. Fig. 1A1993, 1994, 1998 E1, E2, E3 DFO Dahl et al., 1995; Graham et al., 1996; MacDougall et al., 2001; Fig. 1A1994, 1996–1999 23, 931, 934–936, 938 Env. Can.a Fig. 1B2001–2003 23, 438–452, 931, 934–936, 938 UW This study, basin-wide survey; Fig. 1B2001–2003 500–508 UW This study, north shore survey; Fig. 1C

a Not all stations were sampled in all months or all years. The Env. Can. historical data were obtained from files of the National Water Research Institute (NWRI) that were gath-ered by Environment Canada, Ontario Region, at the Canada Centre for Inland Waters (CCIW).



Statistical analyses

We assessed changes in water quality parameters in relation to Ploading, temporal (pre-dreissenid (1973–1985) versus post-dreissenid(1990–2003) years) and spatial (nearshore versus offshore) effectsusing a linearmixed effectsmodel (Bates et al., 2008). The use of this hi-erarchical model allowed us to consider the contributions of the differ-ent sources of variance (P loading rates to the lake, nearshore vs.offshore, pre- vs. post-dreissenid invasion, year, station), and did not re-quire equal sample sizes, thus allowing us to make inter-annual com-parisons. Prior to analysis, all variables were log10 transformed tomeet the conditions of normality as assessed by a Shapiro–Wilk test(p>0.05). For the basin-wide survey data, we used the followingmodel:

log Yij

� �¼ β0 þ β1Ploadþ β2Temporalþ β3Spatialþ β4Temporal

�Spatialþ STNi þ YEARj þ εij

Where log(Yij) is the log of parameter Y at station i in year j, Ploadis the 3 year preceding mean centered P load (Dolan and McGunagle,2005; Ohio Lake Erie phosphorus task force, 2010), while temporaland spatial are categorical variables representing pre- or post-dreissenid invasion and nearshore or offshore respectively. STNi is arandom effect for station i (different stations were sampled in differ-ent years) and YEARj is a random effect for year j. Random effects andthe residual error (εij) are assumed drawn from a normal distributionN (0, σ2). Pload represents the average whole lake P loads (includingLake Huron) over the previous 3 year period. A 3 year time periodwas selected based on the hydraulic residence time for Lake Erie(2.7 years; Quinn, 1992). Although non-point sources have since be-come a larger part of the total load post P control, these loads arepoorly characterized and difficult to quantify at present (Dolan and

McGunagle, 2005). The pre-dreissenid invasion data are largelyprior to P control so P control effects are likely to be minimized inour analyses. Pload was log mean centered to simplify interpretationof estimated coefficients, and Year was rescaled to Year—1973 suchthat 1973 was expressed as year “0”. For several variables (denoteda; Table 3), the dataset lacked pre-dreissenid nearshore data, there-fore, we dropped the interaction term and tested the main fixed ef-fects only. For the UW north-shore survey data we used thefollowing model structure:

log Yij

� �¼ β0 þ β1Spatialþ STNi þ YEARj þ εij:

This is a reduced form of the model since all sampling was con-ducted post-dreissenid invasion (2001–2003), but we wanted to as-sess spatial effects. Mixed models were fit using the lme4 package(Bates et al., 2008) in the statistical software ‘R’ (R DevelopmentCore Team, 2008) using restricted maximum likelihood (REML). Test-ing for significance of fixed effects in mixed models using the lme4package is not straightforward since the p values are potentiallystrongly anti-conservative due to the fact that the number of residualdegrees of freedom cannot be accurately determined (Bolker et al.,2009). Therefore, we based our assessment of the significance of afixed effect on the highest posterior density intervals (HPDI; Bolkeret al., 2009). We consider that a fixed effect was significant whenthe HPDI did not include zero. HPDI were computed from the poste-rior distribution of parameter estimates with Markov Chain MonteCarlo methods using the pvals.fnc wrapper (package ‘LanguageR’;Baayen et al., 2008) with n=5000 samples and uniform priors forfixed effects. If the interaction between spatial and temporal effectswas significant, we tested specific interaction contrasts using simulta-neous tests for general linear hypotheses using the ‘multcomp’ pack-age (Hothorn et al., 2008). A sub-section of north-shore survey data

Table 2Comparison of methods for the water quality parameters from the institutions employing the surveys.

EPA Env. Can. DFO UW(basin-wide and north shore surveys)

Physical conditionsConductivity,temperature,depth

YSI model 54 dissolvedoxygen-temperature profiler

Electronic Bathy Thermograph (EBTT) profiler Hydrolab profilingsystem

EBTT profiler (basin-wide survey)or a SeaBird™ SBE-19 profiler or aHydrolab (north shore survey)

Secchi depth NC Secchi disk Secchi disk Secchi diskKd NC Underwater quantum sensor profiles taken at

select stations from 1979–1990aUnderwater quantumsensor profiles

Underwater quantum sensor profiles taken andcalculated as in North et al. (2007)a

Mean watercolumn PAR

NC Guildford et al. (2000) Guildford et al.(2000)

Guildford et al. (2000)

Water chemistryTP USEPA (1971) NLET (1994) NLET (1994) NLET (1994)TDP USEPA (1971) NLET (1994) NLET (1994) NLET (1994)SRP NC NLET (1994) NLET (1994) NLET (1994)Particulate P Calculated by difference

(TP–TDP)Calculated by difference(TP–TDP)b

Calculated bydifference (TP–TDP)

Stainton et al. (1977)c

NO3−+NO2

− USEPA (1971) NLET (1994) NLET (1994) NLET (1994)NH4

+ USEPA(1971) NLET (1994) NLET (1994) NLET (1994)d

DRSi NC NLET (1994) NLET (1994) NLET (1994)Cl− NC NLET (1994) NLET (1994) NLET (1994)Particulate Si NC NCb NC Stainton et al. (1977)ParticulatesC and N

Great Lakes Laboratory (1974) NLET(1994)b NLET (1994) Stainton et al. (1977)

Chl a Strickland and Parsons (1968) Strickland and Parsons (1968) Strickland andParsons (1968)

North et al. (2007)

NC, not collected; Kd, attenuation coefficient for PAR.a Where underwater quantum sensor profiles were not collected, kd was estimated from the linear relationship between kd and %T (percent transmission measured by the EBTT

profiler).b Exception of 1997 samples (Guildford et al., 2005).c Exception of 2003 samples (Parsons et al., 1984).d Exception of 2002 and 2003 samples (Holmes et al., 1999).



(seven stations in August 2002 and six stations in July 2003) andthe APA data from both the basin-wide and north-shore surveyswere analyzed with a one-way ANOVA to test for spatial differ-ences (nearshore versus offshore) and a α-value of 0.05 was ap-plied. All data were pooled over the stratified summer season(June–September).

Results

Reductions in external phosphorus loading and associated changes inphytoplankton and nutrient dynamics

In the basin-wide and historical surveys, chla and TP concentra-tions were significantly higher in the eastern basin of Lake Erie inpre- compared to post-dreissenid years and there were significanteffects of P loading on the TP concentrations reflecting the success-ful P loading controls (Table 3, Figs. 2A,B). Coincident with the tem-poral decreases in chla and TP concentrations in the lake, there wasalso a difference in the proportion of chla relative to TP (Table 3,Fig. 2C).

Although there were no significant differences at the offshore sta-tions for Secchi depths between pre- and post-dreissenid years (datanot shown), the kd values indicate an improving light environment.Pre-dreissenid kds were significantly higher than post-dreissenidvalues and the linear mixed effects model showed a significant Ploading effect (Table 3, Fig. 2E).

While SRP concentrations were not significantly different betweenpre- and post-dreissenid years (Table 3, Fig. 3D), TDP concentrationswere significantly affected by P loads, spatial, and temporal effects(Table 3, Fig. 3A). In 1973, NO3

−+NO2− concentrations (mean=

6.9 μmol L−1) were lower than post-dreissenid concentrations(mean=15.5 μmol L−1). Although temporal changes were not statis-tically significant, NO3

− concentrations have more than doubled since1973 (Table 3, Fig. 3B). There were no significant temporal differencesin dissolved reactive Si (DRSi) concentrations (Table 3, Fig. 3C). Indi-cators of P deficiency, the C:P (Fig. 4A) and N:P ratios (data notshown), display trends that suggest that Lake Erie seston becamemore P deficient; however, this temporal effect was not significant(Table 3).

Nearshore–offshore distribution of seston and nutrients in pre- and post-dreissenid years

The linear mixed effects model revealed a significant interactionfor chla concentrations between the pre- and post-dreissenid years,and nearshore versus offshore (Table 3). In 1973 (the onlypre-dreissenid year nearshore data was collected), nearshore chlaconcentrations were significantly (pb0.05) higher than offshore, incontrast to the post-dreissenid years where nearshore chla concen-trations were significantly (pb0.001) lower than offshore (Table 3,Fig. 2A). Post-dreissenid (2001–2002) picoplankton (0.2–2 μm) chlaconcentrations were also significantly (pb0.01) lower in the near-shore (n=35, mean=0.50 μg L−1) than the offshore (n=48,mean=0.70 μg L−1). However, the contribution of picoplankton tototal chla was similar between the nearshore (39%) and the offshore(41%). There were no significant spatial differences between near-shore and offshore TP concentrations in either 1973 or post-dreissenid years (Table 3, Fig. 2B). The linear mixed effects model re-vealed a significant spatial–temporal interaction for chla/TP ratios be-tween 1973 and post-dreissenid years, and nearshore versus offshore(Table 3). In 1973 there was no significant difference between near-shore and offshore chla/TP ratios; however, in the post-dreissenidyears there was a substantial difference in the ratio with 39% morechla per unit of TP offshore compared to nearshore (pb0.001,Table 3, Fig. 2C).

Table 3Mixed model output for basin-wide survey (Figs. 1A,B) data showing phosphorus load-ing (preceding 3-year mean), spatial (nearshore versus offshore) and temporal(pre-dreissenid (1973, 1979, 1983–1985) versus post-dreissenid (1990, 1993–1994,1996–1999, 2001–2003) invasion) effects on water quality parameters in the easternbasin of Lake Erie. All parameters were log transformed for statistical analyses. Interac-tions (spatial and temporal effects represented only) shown below only if significant.Significant differences are indicated by ***b0.001, **b0.05, *b0.10.

Parameter Fixed effect Estimate MCMCmean

LPDI HPDI

Chl a Intercept 0.130 0.139 −0.034 0.304P loads −0.142 −0.119 −0.648 0.415Spatial** 0.425 0.422 0.308 0.528Temporal** 1.173 1.160 0.763 1.539Interaction: spatial*temporal**

−0.733 −0.735 −0.995 −0.492

TP Intercept −1.261 −1.262 −1.432 −1.107P loads* 0.532 0.528 −0.009 1.030Spatial −0.024 −0.023 −0.111 0.073Temporal* 0.333 0.353 −0.038 0.774

Chl a/TP Intercept −2.047 −2.040 −2.248 −1.839P loads** −0.747 −0.725 −1.372 −0.088Spatial 0.457 0.455 0.314 0.597Temporal 0.787 0.755 0.254 1.236Interaction: spatial*temporal**

−0.689 −0.676 −0.987 −0.358

Kda Intercept −1.252 −1.255 −1.365 −1.144

P loads* −0.237 −0.245 −0.535 0.058Spatial −0.038 −0.038 −0.155 0.079Temporal* 0.195 0.199 0.008 0.385

ParticulateC

Intercept 3.019 3.017 2.766 3.252P loads 0.230 0.225 −0.493 1.020Spatial 0.176 0.181 0.057 0.310Temporal 0.439 0.465 −0.086 0.977Interaction: spatial*temporal*

−0.309 −0.337 −0.688 0.003

C: chl a Intercept 2.805 2.803 2.559 3.022P loads 0.153 0.140 −0.538 0.839Spatial −0.209 −0.204 −0.347 −0.049Temporal −0.537 −0.516 −1.056 0.008Interaction: spatialtemporal*

0.362 0.343 −0.041 0.718

SRPa Intercept −3.763 −3.761 −4.071 −3.428P loads 0.001 −0.018 −1.210 1.239Spatial** −0.261 −0.264 −0.446 −0.086Temporal 0.059 0.059 −0.548 0.599

TDP Intercept −2.014 −2.013 −2.158 −1.868P loads** −0.570 −0.569 −1.037 −0.099Spatial** −0.079 −0.079 −0.157 −0.002Temporal** 0.397 0.384 0.048 0.933

ParticulateP

Intercept −2.059 −2.059 −2.326 −1.785P loads* 0.876 0.873 0.016 1.696Spatial 0.094 0.094 −0.059 0.234Temporal 0.427 0.449 −0.126 1.010

NO3−+NO2

− Intercept 2.617 2.613 2.445 2.776P loads** −1.058 −1.055 −1.572 −0.555Spatial* −0.113 −0.104 −0.208 −0.001Temporal 0.028 0.035 −0.368 0.448

NH4+ Intercept 0.287 0.286 −0.097 0.692

P loads* 1.509 1.500 0.253 2.828Spatial** −0.331 −0.328 −0.548 −0.128Temporal −0.665 −0.612 −1.644 0.377

ParticulateNa

Intercept 1.293 1.299 0.913 1.688P loads 0.418 0.431 −0.554 1.582Spatial* 0.121 0.119 −0.021 0.252Temporal −0.204 −0.214 −0.756 0.267

DRSia Intercept 2.478 2.475 2.153 2.774P loads 0.451 0.456 −0.858 1.782Spatial*** −0.284 −0.282 −0.426 −0.147Temporal 0.289 0.281 −0.593 1.093

C:P Intercept 4.885 4.893 4.586 5.174P loads** −1.326 −1.324 −2.180 −0.505Spatial 0.084 0.082 −0.102 0.267Temporal 0.006 −0.028 −0.684 0.656

a Parameters for which the dataset lacked a pre-dreissenid nearshore measurementso the interaction term was dropped and the separate main fixed effects were tested;MCMC, Markov Chain Monte Carlo; LPDI, lowest posterior density intervals; HPDI,highest posterior density intervals.



The linearmixed effectsmodel revealed a significant spatial–temporalinteraction for particulate C concentrations between 1973 and post-dreissenid years, and nearshore versus offshore (Table 3). As observedfor the chla/TP ratio, there were no nearshore–offshore differences in1973 particulate C concentrations; however, within the post-dreissenidyears, the particulate C was significantly (pb0.05) lower nearshore thanoffshore (Table 3). In contrast to the size fractioned chla, size fractioned(0.2–2 μm)particulate C concentrations and the percentage of particulateC represented by picoplankton, were similar between the nearshore(mean=13.01 μmol L−1; 50%) and offshore (mean=13.90 μmol L−1;48%). The linear mixed effects model also revealed a significant interac-tion for C:chl a ratios between pre- and post-dreissenid years, andnearshore versus offshore (Table 3). The difference between nearshoreand offshore 1973 ratioswas not significant, while the post-dreissenid ra-tios were significantly (pb0.05) lower in the offshore compared to thenearshore (Table 3, Fig. 2D).

There was no significant difference between nearshore and offshorekd values in the post-dreissenid years (Table 3, Fig. 2E). The parameter

mean water column PAR, is a function of kd and the mixing depth ofthe epilimnion (Zmix); and the post-dreissenid nearshore mean PAR(as percent of surface incident PAR) was significantly (pb0.001) highernearshore than the offshore (Fig. 2F).

In 1973, there were no significant differences between nearshoreand offshore TDP concentrations, while in the post-dreissenid years,nearshore TDP and SRP concentrations were significantly higherthan offshore (Table 3, Figs. 3A,D). The 1973 particulate P concentra-tions (mean=0.96 μmol L−1) were higher than the post-dreissenidconcentrations (mean=0.14 μmol L−1; data not shown), althoughthere were no significant spatial or temporal effects (Table 3). In1973, there were no significant differences between nearshore andoffshore dissolved inorganic N (NO3

−, NH4+) concentrations (Table 3).

In the post-dreissenid years, the NO3− and NH4

+ concentrations weresignificantly higher in the nearshore than the offshore (Table 3,Figs. 3B,E), similar to the pattern observed for SRP and TDP. Therewere no significant differences in particulate N concentrations betweenpre- and post-dreissenid years, but concentrations were significantly

Fig. 2. Comparison of basin-wide nearshore and offshore historical trends in mean annual water quality parameters. A) Chla concentrations (n=413, log axes), B) TP concentrations(n=470, log axes), C) chla:TP ratio (n=404, log axes), D) C:chla ratios (n=268), E) the attenuation coefficient for PAR (kd; n=392), F) the mean water column light intensity as apercent of surface irradiance (mean water column PAR; n=230). The values shown are pooled seasonally (June–September) over all stations (Figs. 1A and B) and exclude the northshore survey data (Table 4 and Fig. 1C). Error bars represent one standard error of the mean and the arrow indicates the year of dreissenid invasion to the eastern basin (1989).



lower in the nearshore than the offshore in the post-dreissenid years(Table 3, data not shown). Within the post-dreissenid years, DRSi con-centrations were significantly higher in the nearshore than the offshore(Table 3, Fig. 3C). There were no significant differences between thenearshore and offshore for post-dreissenid (2001–2003) particulate Siconcentrations (data not shown).

In both pre- and post-dreissenid years, there were no significantspatial effects on C:P ratios; an P deficiency indicator (Table 3,Fig. 4A). A second P deficiency indicator, APA, showed that thepost-dreissenid years (1997, 2001, and 2002) were all moderatelyto extremely P deficient. The APA indicated significantly (pb0.05)stronger P deficiency in the offshore than the nearshore for both2001 and 2002 (Fig. 4B).

Post-dreissenid nearshore–offshore differences in seston and nutrientdynamics

In the north shore survey of the post-dreissenid years (2001–2003),nearshore chla concentrations were significantly lower than offshore(Table 4). Picoplankton (0.2–2 μm) chla concentrations were alsosignificantly (pb0.05) lower in the nearshore (n=35, mean=0.35 μg L−1) than the offshore (n=31, mean=0.51 μg L−1; data notshown). However, the percentage of total chla that is represented bypicoplankton was similar between the nearshore (40%) and the

offshore (39%). Conversely, the TP concentrations were significantlyhigher in the nearshore relative to the offshore, and a comparison ofthe chla:TP ratio revealed a significantly lower ratio in the nearshore(Table 4).

There were no significant differences between nearshore andoffshore particulate C concentrations (Table 4). Size-fractioned(0.2–2 μm) particulate C concentrations and the percentage of partic-ulate C represented by picoplankton were similar between the near-shore (15.1 μmol L−1; 58%) and offshore (14.4 μmol L−1; 57%),respectively (data not shown). A selection of microscopic phyto-plankton counts and corresponding biovolume estimates (seven sta-tions in August 2002 and six stations in July 2003) supported thechla data (Fig. 5). Phytoplankton biovolumes were significantlylower (p=0.041, n=13) in the nearshore relative to the offshore,as were chla concentrations (p=0.019, n=13) in the north shoresurveys of 2002 and 2003 (Fig. 5). The particulate C concentrationsdid not differ statistically (p=0.320, n=13) between the two re-gions (Fig. 5).

The kd values were significantly higher in the nearshore relative tothe offshore, however, mean water column PAR indicated a signifi-cantly higher light climate in the nearshore (Table 4). There weresignificantly higher nearshore NH4

+ and particulate P and Si concen-trations and there were no differences between nearshore and off-shore TDP, SRP, NO3

−, particulate N and DRSi concentrations (Table 4).

0.1

0.2

0.3

0.40.50.60.7

TD

P(µm

olL

-1)

Dreissenids

Year

0.05

0.10

0.150.200.25

SRP(

µmol

L-1

)

10

20

304050

Nitr

ate

& N

itrite

(µm

olL

--1 )

1970 1975 1980 1985 1990 1995 2000 2005

0.10

1.00

10.00

Am

mon

ium

(µm

olL

-1)

Dreissenids

1970 1975 1980 1985 1990 1995 2000 2005

10

20

304050

Dis

solv

ed R

eact

ive

Silic

a (µ

mol

L-1

)

Env. Can. OffshoreEPA OffshoreDFO OffshoreEnv. Can. NearshoreEPA NearshoreDFO Nearshore

A

B

C

D

E

Fig. 3. Comparison of nearshore and offshore basin-wide trends in mean seasonal dissolved nutrient concentrations. A) TDP (n=471), B) NO3− (n=403), C) DRSi (n=269), D) SRP

(n=381), E) NH4+ (n=395) concentrations are shown on log axes. Note that the 1973 NO3

− concentration estimates do not include NO2−. Error bars represent one standard error of

the mean.



The C:P molar ratio indicated significantly higher seston P defi-ciency offshore (borderline extreme P deficiency) than in the near-shore (moderately P limited; Table 4). The indicator APA showed nosignificant differences between the nearshore and the offshore, al-though the nearshore phytoplankton had a lower average value, onthe border between moderate and extreme P deficiency. The averageoffshore APA was consistently well above the threshold for extreme Pdeficiency (Table 4).

Discussion

The key elements of the design of this study were to use both theobserved trends over time before and after mussel invasion in theeastern basin of Lake Erie in combination with the observed differ-ences between nearshore and offshore to test for spatial changes aswell. Independently, both basin-scale and more spatially resolvednorth-shore surveys were tested for nearshore–offshore differences.We found lower post-dreissenid nearshore chla concentrations, indi-cators of P deficiency, and higher C:chla ratios associated with highermean irradiance, relative to the offshore. The results suggest a com-bined influence of mussel grazing, nutrient excretion, and light avail-ability as key factors determining recent patterns between nearshoreand offshore in the eastern basin of Lake Erie.

Basin-scale comparisons may be limited in their power be-cause natural variability is likely to be large at this scale, makingsignificant differences difficult to detect because of inter-annualvariability. Methodological differences between years and gov-ernmental institutions also pose a challenge for the basin-scale

comparisons, as illustrated by Esterby and Bertram (1993)wherein a three-ship inter-comparison study was performed inthe central basin of Lake Erie. They report laboratory-specific dif-ferences in TP concentrations from samples collected at the sametime and location that were large enough to interfere with the as-sessment of temporal trends. Our 2001–2003 basin-wide andnorth shore survey data showed good agreement between pa-rameters collected by Environment Canada and the University ofWaterloo. Also, our nearshore–offshore comparison within thesame year and agency validated the trends seen in combinedagency data over earlier years. While caution should be exercisedin cross-agency comparisons, multiple sources of evidence appearto yield similar conclusions.

Pelagic oligotrophication due to P-loading controls

Absolute concentrations of both TP and chla have decreased overthe decades of nutrient control and mussel invasion. The applicationof mass balance modeling to the central basin has revealed that totallake loadings are capable of explaining the current offshore TP concen-trations (Rockwell et al., 2005). In 2002, Dolan andMcGunagle (2005)estimated that non-point source loading is now larger (61%) thanpoint source loading (20%) and non-point sources constitute the dom-inant external loading to the east basin, although they have not in-creased in absolute terms over time.

Phytoplankton in the eastern basin of Lake Erie becamemore P de-ficient over our observation period. There was a significant increase inC:P ratios, similar to previous reports for the western basin of LakeErie (Makarewicz et al., 2000). Offshore eastern basin phytoplanktonhave been characterized as not P deficient in 1979 (Lean et al., 1983)to moderately P deficient in 1997 (Guildford et al., 2005). The near-shore of Lake Erie also showed signs of becoming more P deficientpost-mussel invasion. Nicholls et al. (1999b) showed that TN:TP ra-tios were higher at nearshore sites in the eastern basin of Lake Eriepost-dreissenids (1990–1994) than pre-dreissenids (1984–1987), andthe present results show that APA values are in the P deficiency rangein both nearshore and offshore in post-dreissenid years, although near-shore rates are consistently lower than offshore rates. Control of P loadsto Lake Erie appears to have successfully reduced P concentrations and in-creased the intensity of P limitation for phytoplankton in the easternbasin even as mussels have colonized the lake.

Are changes in basin-scale nearshore–offshore patterns consistent withexpected dreissenid effects?

Mussels have been present throughout the eastern basin of LakeErie from 1989 to present and in 2002 densities of mussels (entirelyquaggas) averaged 9480 m−2 in the east basin (Patterson et al.,2005) around the time of our surveys. An examination of thebasin-wide temporal trends reveals discontinuities before and afterthe mussel invasion. Total phosphorus in the eastern basin wastrending downward before invasion, as P loading decreased (Dolanand McGunagle, 2005), but this trend appears to have been arrestedat about the time of invasion. The chla:TP ratio was increasing beforeinvasion but the trend ceased or even reversed in the eastern basinafter invasion. In contrast, chla decreased significantly over thewhole observation period with no change in trend between pre-and post-dreissenid periods. There was no statistically significantchange in the total P load in the post-dreissenid period of observation(1991–2004; Dolan and McGunagle, 2005) but the east basinsupported less chla, consistent with offshore observations duringthe early spring period (Barbiero et al., 2006). This change was stron-gest in the nearshore, where chla:TP ratios became lower than off-shore ratios. The north-shore survey data confirmed that thenearshore supported lower chla despite, in that study area, signif-icantly higher TP than the offshore. Diminished chla/TP ratios can

0

100

200

300

400

500C

:P(m

olar

ratio

)

Env. Can. OffshoreEPA OffshoreDFO OffshoreEnv. Can. NearshoreEPA NearshoreDFO Nearshore

A

B

1970 1977 1984 1991 1998 20050.00

0.01

0.02

0.03

0.04

0.05

APA

(µm

olPµ

gChl

a-1h-1

)

Fig. 4. Basin-wide historical trends in mean indicators of P deficiency. A) C:P ratios(n=253), and B) APA normalized to chla (n=34). The dashed lines represent criteriafor P deficiency as defined by Healey and Hendzel (1979b). Values above the upperdashed lines are indicative of severe P deficiency, while values between the upperand lower dashed lines are indicative of moderate P deficiency, and values below thedashed lines are not considered to be P deficient. Error bars represent one standarderror of the mean. The 1997 data are from Guildford et al. (2005).



be expected where mussel grazing rates are high enough to causethe combined phytoplankton loss rates to exceed replacementrates (Padilla et al., 1996). In lake-wide studies, post-dreissenidoffshore chla/TP ratios (0.11–0.33; Charlton and Milne, 2004)

were higher than post-dreissenid nearshore ratios (0.02–0.05;Nicholls et al., 1999a) from water intake samples. Nicholls et al.(1999a) found that chla and TP in the nearshore were highly corre-lated, with 75% of the chla variance explained by TP before invasion

Table 4Arithmetic means (one standard error of the mean in parentheses) for nearshore (n=37) and offshore (n=33) north shore survey (Fig. 1C) data from 2001 to 2003. Significantdifferences between nearshore and offshore water quality parameters resulting from the mixed model output are indicated by ***b0.001, **b0.05, *b0.10. All parameters werelog transformed. Bolded values indicate P deficiency according to the deficiency indicators as defined by Healey and Hendzel (1979b).

Parameter Mean (SE) Fixed effect Estimate MCMC mean LPDI HPDI

Chl a (μg L−1) ns 0.9 (0.09) Intercept −0.225 −0.221 −0.810 0.323os 1.4 (0.13) Spatial** 0.471 0.469 0.196 0.726

TP (μmol L−1) ns 0.33 (0.03) Intercept −1.214 −1.218 −1.659 −0.7963os 0.25 (0.01) Spatial* −0.226 −0.225 −0.472 0.012

Chl a/TP ns 0.11 (0.01) Intercept −2.409 −2.403 −2.931 −1.876os 0.18 (0.02) Spatial** 0.616 0.602 0.296 0.895

Kd (m−1) ns 0.47 (0.03) Intercept −0.834 −0.840 −1.263 −0.456os 0.24 (0.01) Spatial** −0.643 −0.637 −0.802 −0.468

Mean water column PAR (%) ns 42.94 (2.19) Intercept 0.663 0.664 0.555 0.797os 29.97 (2.04) Spatial*** −0.114 −0.115 −0.172 −0.052

Particulate C (μmol L−1) ns 27.98 (1.47) Intercept 3.290 3.289 3.126 3.439os 26.74 (1.29) Spatial −0.045 −0.044 −0.211 0.106

TDP (μmol L−1) ns 0.18 (0.01) Intercept −1.791 −1.796 −2.102 −1.495os 0.15 (0.01) Spatial −0.102 −0.096 −0.241 0.061

SRP (μmol L−1) ns 0.05 (0.01) Intercept −3.121 −3.123 −3.724 −2.605os 0.04 (0.01) Spatial −0.117 −0.120 −0.381 0.125

Particulate P (μmol L−1) ns 0.18 (0.01) Intercept −1.809 −1.801 −2.365 −1.294os 0.13 (0.01) Spatial** −0.376 −0.377 −0.653 −0.095

NO3−+NO2

− (μmol L−1) ns 13.8 (0.8) Intercept 2.567 2.562 2.320 2.811os 14.0 (0.8) Spatial 0.014 0.017 −0.171 0.202

NH4+ (μmol L−1) ns 0.87 (0.11) Intercept −0.093 −0.087 −0.77 0.553

os 0.60 (0.11) Spatial** −0.494 −0.493 −0.837 −0.146Particulate N (μmol L−1) ns 3.5 (0.29) Intercept 1.179 1.176 0.845 1.461

os 3.2 (0.14) Spatial −0.050 −0.049 −0.228 0.136DRSi (μmol L−1) ns 12.0 (1.4) Intercept 2.274 2.275 1.787 2.731

os 10.8 (1.1) Spatial −0.001 0.003 −0.438 0.427Particulate Si (μmol L−1) ns 3.6 (0.6) Intercept 1.031 1.027 0.554 1.458

os 2.0 (0.2) Spatial** −0.445 −0.442 −0.765 −0.134C:P (molar ratio) ns 177 (17) Intercept 5.097 5.090 4.623 5.593

os 255 (20) Spatial** 0.337 0.346 0.057 0.596APA (μmol P μg chl a−1 h−1) ns 0.009 (0.003)

os 0.015 (0.005)

ns, nearshore; os, offshore; MCMC, Markov Chain Monte Carlo; LPDI, lowest posterior density intervals; HPDI, highest posterior density intervals.

Nearshore Offshore

0.0

0.5

1.0

1.5

2.0

2.5

Chl

a (

gL-1

)

0

100

200

300

400

Phytoplankton biovolume (m

g m-3)

Part

icul

ate

C (

m

ol L

-1)

10

20

30

40

Particulate C

BiovolumeChla

Fig. 5. Comparison of nearshore and offshore north-shore survey phytoplankton biomass indicators from surveys in August 2002 and July 2003. Triangles represent the particulate Cconcentrations (n=13), circles represent the chla concentrations (n=13), and squares represent the phytoplankton biovolume estimates (n=13). Error bars represent one stan-dard error of the mean.



but only 14% after invasion and a lower chla at the same TP.“Decoupling” of the chla–TP relationship in the nearshore of LakeErie appears to be a result of lower chla concentrations and notchanges in TP.

The evidence presented here supports previous evidence thatmussels can have an impact on chla, even in large lakes. They alsostrengthen the evidence that the impact is likely to be strong inshallower, nearshore waters and produce changes in patterns be-tween nearshore and offshore. However, grazing in the nearshoreby zooplankton could also explain the lower chla concentrations.There is evidence that Daphnia grazing was more important thanmussel grazing after the invasion in the western basin of Lake Erie(Wu and Culver, 1991) where average mussel densities in 2002were relatively low (601 m−2, Patterson et al., 2005). Nicholls andHopkins (1993) concluded that most of the chla and phytoplanktonbiomass declines in Lake Erie water intake samples were due to mus-sels, not zooplankton, because observed effects were strong in sea-sons of typically low cladoceran grazing impacts. Total zooplanktondensities in the east basin did not reveal significant differences be-tween nearshore and offshore samples in 1993, 1994, or 1998, andthe cladoceran densities were lower in the nearshore than the off-shore; significantly so for Daphnia species (MacDougall et al., 2001).The relative importance of dreissenids' grazing impact to those ofzooplankton varied among years in the eastern basin (Zhang et al.,2011). Zooplankton grazing is therefore unlikely to explain the con-sistently lower chla concentrations in the nearshore of the easternbasin of Lake Erie after the mussel invasion.

If mussel grazing lowers nearshore phytoplankton biomass, thenthere should be a decline in particulate nutrients (C, P, N, Si) and anincrease in dissolved nutrients due to mussel excretion and decreasedinstantaneous nutrient demand by phytoplankton. Previous studieshave found lower particulate nutrient concentrations consequent tomussel invasion in the more enclosed waters of Oneida Lake (Idrisiet al., 2001) and Saginaw Bay (Johengen et al., 1995). The basin-wide survey particulate C and N concentrations show expected effectsof mussel grazing, however the particulates P and Si and north-shoresurvey particulate C, N, P and Si concentrations gave a varied and/ornon-significant difference between nearshore and offshore in theeastern basin of Lake Erie. The patterns in particulate nutrients willreflect altered phytoplankton composition, nutrient status, and con-tributions of non-algal material to the seston. There is potential foroffsetting effects. For example, P status indicators (C:P, APA) showlower P deficiency in the nearshore, which may reflect higher intra-cellular P concentrations than the offshore or more detrital P contri-bution to the seston in the nearshore. The lack of statisticallysignificant spatial differences and in some cases, higher particulatenearshore concentrations, can probably be explained by either shore-line eroded material, or resuspension of particles from the bottomsediments at shallow well-mixed locations. As erosion of shorelinematerial is generally highest in spring, it is likely that the summerparticulate C, N, P and Si concentrations in the nearshore reflectresuspension which can include mussel feces and pseudofeces. Thismay explain the similar concentrations of particulate C nearshoreand offshore, even when the phytoplankton biomass and chla con-centrations were lower nearshore. Detritus of benthic origin maybe more important in the nearshore seston after mussel coloniza-tion. Nearshore seston had similar δ13C and δ15N signatures to mus-sel biodeposits collected from our stations 504, 503, and 502 in2002 (Fig. 3 in Szabo, 2004; Upsdell, 2005), suggesting thatresuspension of benthic material may mask the loss of planktonicparticles from the water column. Photoacclimation in response togreater mean PAR could also contribute by allowing phytoplanktonto accumulate more C per biovolume as well as per chla (Thompsonet al., 1991).

The higher dissolved nutrient concentrations in the nearshore ofthe eastern basin relative to the offshore after mussel colonization

were consistent with spatial variation of both increased biologicallyavailable P (SRP) from the catchment (Ohio Lake Erie phosphorustask force, 2010), and the impact of mussel grazing and nutrient ex-cretion that is hypothesized by the nearshore shunt (Hecky et al.,2004). The differences in nutrient concentrations were also reflectedin the pattern of P deficiency, which was stronger offshore than near-shore. Our data may underestimate the strength of the enhancementof nearshore dissolved nutrients, as concentrations (especially for dis-solved P) were widely variable among years and limited the statisticalpower of our analyses. The variability may have been due to high bi-ological demand by bacteria, phytoplankton, and benthic algae aswell as variable advection at our open coastal sites that makes the dis-solved P quite dynamic. Mussel-regenerated SRP was rapidly re-moved by bacteria and phytoplankton during experiments inSaginaw Bay (Heath et al., 1995), for example. Model simulationsfor the eastern basin of Lake Erie in 1997–1999 demonstrated only a1% influence of mussels (including both excretion and grazing pro-cesses) on water column SRP, which they attribute to mixing process-es (Zhang et al., 2011). In addition, the higher mean water columnPAR in the nearshore has facilitated the growth of benthic algae, espe-cially Cladophora, that exhibit P limited growth (Auer et al., 2010).Davies and Hecky (2005), studying in situ benthic productivity at anopen coastal site near our study locations, observed enhanced pro-ductivity of benthic algae in mussel beds suggesting that P demandby benthic algae would also be high. Increased demand for P by ben-thic algae has likely limited the extent of increased dissolved P con-centrations in the nearshore.

Previous work has demonstrated the potential of mussels to alterphytoplankton P status. Arnott and Vanni (1996) measured musselexcretion rates in the western basin of Lake Erie and found that thedissolved inorganic forms of N and P excreted by mussels were at alower ratio (i.e. below 16:1), which could relieve phytoplankton ofP deficiency. Heath et al. (1995) reported that phytoplankton sub-jected to mussel grazing in mesocosms in Saginaw Bay had more Pavailable to them per cell and measured this release from P defi-ciency as a decline in the rate of phosphate uptake to less than1% of the controls. The present results support previous indicationsthat mussels can alter the P status of phytoplankton in the near-shore relative to the offshore even in the large, open coastal near-shore of the eastern basin of Lake Erie as evidenced by lowerindicators of P deficiency in the nearshore, relative to the offshore(North et al., 2007).

Mussels also excrete NH4+ (Arnott and Vanni, 1996), perhaps con-

tributing to the elevated concentrations we observed in the near-shore. However, the realization of the low N:P ratio in dreissenid-excreted nutrients, while relieving the inherent P limitation of thephytoplankton, could also be promoting secondary N deficiency aspostulated by Moon and Carrick (2007). In the offshore of the centralbasin in the late summer and fall of 2002–2003, Moon and Carrick(2007) reported symptoms of N deficiency in bioassay experiments.Although Lake Erie is primarily P limited, there are sometimes symp-toms of N deficiency (Guildford et al., 2005). During such an occur-rence of N deficiency in the summer stratified season, North et al.(2007) showed was less severe in the nearshore than the offshoreof the north-shore of the eastern basin after mussel colonization.Upsdell (2005) reported that in 2002, seston from the nearshorenorth-shore survey stations (Fig. 1C) was on average enriched in15N relative to the offshore. This too is consistent with effects of mus-sel excretion, which produces NH4

+ that is 15N enriched compared toNO3

− (McCusker et al., 1999).Similar to NH4

+, NO3− concentrations were higher in the nearshore

than offshore in post-mussel years, and NO3− also showed a trend to

higher concentrations in recent years. There is evidence that theratio of N:P and absolute concentrations of NO3

− in water enteringthe eastern basin from the watershed have been increasing overmuch of the period considered here. At the mouth of the Grand



River (Dunnville monitoring station), NO3− concentrations have

increased significantly (1980–1998), coinciding with a decliningtrend (1973–1997) in TP concentration and load (OntarioMinistry of Environment, 2002). Our lake data reflect these trendsand show an increase in NO3

−+NO2− and a significant decrease in

TP concentrations from 1973 to 2003 in both the nearshore andthe offshore of the eastern basin. Chloride (Cl−) concentrationscan serve as a conservative tracer for terrestrial runoff and, unlikeNO3

−, did not reveal any significant differences between nearshoreand offshore in post-dreissenid years (data not shown). The Cl−

data indicate that the higher dissolved nutrients, specifically NO3−,

found in the nearshore are not due to local runoff effects. Enhanced pro-duction of NH4

+ in the mussel-colonized nearshore, due to mussel ex-cretion activity as well as decomposition of detritus in the benthiccommunity, may be supporting enhanced rates of nitrification and con-tributing to the observed higher NO3

− concentrations.Higher DRSi concentrations in the nearshore could be due to re-

duced uptake by diatoms, combined with mussels breaking downandmetabolizing phytoplankton cells causing structural silicon to un-dergo redissolution (Dame et al., 1991). Post-dreissenid Si increases(Guildford et al., 2005) and pelagic diatom biomass decreases(Barbiero et al., 2006) have been documented for Lake Erie. In thenearshore of the eastern basin, we could be seeing a shift from diatomdominated benthos to green alga dominated (Cladophora). However,the impact of this shift on Si dynamics is unclear as Cladophora canalso increase the surface area available for colonization of epiphyticdiatoms as seen in Lake Ontario (Malkin et al., 2009). In SaginawBay, Johengen et al. (1995) also found higher Si concentrations atmussel-impacted inner bay sites relative to outer bay and controlsites.

What do trends in Chla represent?

Present and previous results give strong evidence that musselgrazing can lower chla concentrations in many natural settings, butchla can have a variable relationship to phytoplankton biomass(Conroy et al., 2005; Fitzpatrick et al., 2007; but see Carrick et al.,2005). Chla was not significantly correlated to phytoplankton wetbiomass in the offshore of Lake Erie (Conroy et al., 2005). In westernLake Erie, Fitzpatrick et al. (2007) reported declines in chla concen-trations from 1970 to 2001, however, the total phytoplankton bio-mass remained the same and there was a nearly 50% decrease inthe chla to phytoplankton biomass ratio. They attributed the differ-ences to changes in composition, size–structure, and photo-adaptivestate of the phytoplankton community (Fitzpatrick et al., 2007). How-ever, Carrick et al. (2005) showed strong correlations between phyto-plankton biomass (measured as C determined from cell counts) andchla (r2=0.91) in the central basin of Lake Erie in 2002. Their 2002central basin offshore summer chla concentrations (2.0 μg L−1;Carrick et al., 2005) were also comparable to our 2002 eastern basinoffshore concentrations (1.8 μg L−1). Our north-shore survey dataalso indicated that chla concentrations in the nearshore were reflec-tive of similar differences in phytoplankton biomass (Fig. 5). Addi-tional evidence of lower phytoplankton biomass in the nearshore(Stn. E3) relative to the offshore (Stn. E2) of the eastern basin wasalso found in the 1998 DFO survey (MacDougall et al., 2001) byMunawar et al. (2008).

While the evidence from limited samples from the eastern basindoes point to a selective decrease of phytoplankton and associatedchla in the nearshore, the changing light conditions in the lakecould also be expected to affect chla concentrations and the relation-ship between chla and phytoplankton biomass. Photoacclimation is aphysiological response that manifests as a reduction in photosynthet-ic pigment content in response to increased irradiance (MacIntyre etal., 2002) and mean irradiance has increased in the east basin (pres-ent results; Auer et al., 2010; Barbiero et al., 2006). Examination of

the differences in the basin-wide C:chla ratio between the nearshoreand offshore of pre- and post-dreissenid years allowed us to inferphotoacclimation as indicated by a large C:chla ratio (Fig. 2D;MacIntyre et al., 2002). We are cognizant of the fact that a high C:chla ratio can also reflect nutrient deficiency (Healey and Hendzel,1979b) as well as increased contribution of heterotrophic organismsand detrital C to sestonic C concentrations (Carrick et al., 2005). How-ever, our inference of photoacclimation is also supported by thehigher mean water column irradiance in the nearshore as a result ofmussel grazing reducing turbidity and also by the evidence for gener-ally higher dissolved nutrient concentrations and reduced nutrientdeficiency in the nearshore. These changes make it unlikely that theincrease in C:chla is a result of increased nutrient deficiency, butmore probable that increased ratios are a result of increased PARavailability in the nearshore post-dreissenids. However, our resultsalso indicate a post-dreissenid increase in C:chla in the offshorewhere nutrient deficiency and the nutrient effect on C:chla may bestronger. More directed research will be necessary to establish theimportance of photoacclimation effects on chla concentrations inthe eastern basin.

Benthic–pelagic interactions and future conditions

Smith et al. (2005) reported that offshore waters of post-dreissenid Lake Erie exhibited high primary production rates relativeto TP when compared with mussel-free Great Lakes locations. A com-panion study to this one in the eastern basin was conducted byDepew et al. (2006) who reported significantly lower areal rates ofprimary production on a daily and seasonal basis in the nearshore rel-ative to the offshore. They confirmed thatmussels have the ability to de-press primary production and influence the distribution of productionbetween nearshore and offshore. Their work was limited, however, tothe epilimnion of the nearshore and did not include the benthos. Daviesand Hecky (2005) measured post-dreissenid (1997–1998) benthic pho-tosynthesis rates in the nearshore of the eastern basin close to our station504. They reported some of the highest benthic photosynthetic rates evermeasured in freshwaters and determined that areal benthic photosyn-thetic rates would greatly exceed areal pelagic photosynthesis in depthsshallower than 5 m in the eastern basin, a finding recently extended tocomparable waters in Lake Ontario (Malkin et al., 2010). A key findingof Davies and Hecky (2005) was that algal–mussel mixed benthic assem-blages can have greater gross photosynthetic rates than comparable algalcommunities without mussels. With increasing light transparency in thenearshore, benthic algae may offset the loss of pelagic production interms of the flow of trophic energy to consumers (Davies and Hecky,2005; Malkin et al., 2010).

Post-dreissenid Lake Erie remains highly efficient at translating TPinto primary production, which is still higher than comparable oligo-trophic lakes (Smith et al., 2005). There is no basis for relaxation of Pcontrols as suggested by Ryan et al. (1999) who expressed concernsregarding the decreasing TP concentrations on fish production. InOneida Lake, despite the increases in mussel grazing rates and associ-ated decreases in phytoplankton biomass, production at the primary,secondary, and tertiary levels did not decline in association with mus-sels; and no significant mussel effects were found on biomass,growth, or production of young-of-the-year yellow perch (Idrisi etal., 2001).

Mussel effects may be diminishing in Lake Erie. Mussel densitiesdecreased significantly in the eastern basin in 2004 relative to2001–2002 following a large increase in round goby (Neogobiusmelanastomus) abundance (Barton et al., 2005). These changes mayportend a decreased impact of mussels in future years as seen infreshwater systems of the former Soviet Union and eastern Europe(Karatayev et al., 1997), where mussels have been established longer.However, evidence in 27 north-temperate lakes indicate thatdreissenid effects on chla are prolonged, with no indication of



diminishing within 7–10 years of invasion (Higgins et al., 2011). It isalso not clear, whether diminished mussel abundance would neces-sarily reverse the changes associated with the nearshore shunt(Hecky et al., 2004).

Acknowledgments

The authorswish to acknowledge the technical assistance provided byA. Lublin, J. Pridham, T. Nelson, A. Desellas, and S. Ronzio. The Natural Sci-ences and Engineering Research Council of Canada (NSERC) providedfunding for this research. As well, R.L. North was supported by an OntarioGraduate Scholarship and NSERC postgraduate scholarships. We ac-knowledge the following analytical support: Chemical Engineering Ana-lytical Chemistry Services in the Department of Chemical Engineering atthe University of Waterloo, the Analytical Laboratory at the FreshwaterInstitute, and the National Laboratory for Environmental Testing (NLET)at the Canada Centre for Inland Waters (CCIW). Logistical support wasprovided by the Captain and crewof the C.C.G.S. Limnos and Technical Op-erations of Environment Canada aswell as the OntarioMinistry of NaturalResources (Port Dover). The 1973 EPA data wasmade available by the ef-forts of R. Kreis, D. Rockwell, and M. Tuchman and we would also like toacknowledge H.J. Kling (Algal Taxonomy and Ecology) for the phyto-plankton analysis.

References

Ammerman, J.W., Hood, R.R., Case, D.A., Cotner, J.B., 2003. Phosphorus deficiency in theAtlantic: an emerging paradigm in oceanography. Eos 84, 165–170.

Arnott, D.L., Vanni, M.J., 1996. Nitrogen and phosphorus recycling by the zebra mussel(Dreissena polymorpha) in the western basin of Lake Erie. Can. J. Fish. Aquat. Sci. 53,646–659.

Auer, M.T., Tomlinson, L.M., Higgins, S.N., Malkin, S.Y., Howell, E.T., Bootsma, H.A., 2010.Great Lakes Cladophora in the 21st century: same algae-different ecosystem. J. Gt.Lakes Res. 36, 248–255.

Baayen, R.H., Davidson, D.J., Bates, D.M., 2008. Mixed effects modeling with crossedrandom effects for subjects and items. J. Mem. Lang. 59, 390–412.

Barbiero, R.P., Rockwell, D.C., Warren, G.J., Tuchman, M.L., 2006. Changes in spring phy-toplankton communities and nutrient dynamics in the eastern basin of Lake Eriesince the invasion of Dreissena spp. Can. J. Fish. Aquat. Sci. 63, 1549–1563.

Barton, D.R., Johnson, R.A., Campbell, L., Petruniak, J., Patterson, M., 2005. Effects ofround gobies (Neogobius melanostomus) on dreissenid mussels and other inverte-brates in eastern Lake Erie, 2002–2004. J. Gt. Lakes Res. 31, 252–261.

Bates, D.M., Maechler, M., Dai, B., 2008. lme4: linear mixed effects models using S4classesR package version 0.999375-32, http://lme4.rforge.r-project.org/2008.

Beeton, A.M., Edmondson, W.T., 1972. The eutrophication problem. J. Fish. Res. BoardCan. 29, 673–682.

Bolker, B., Brooks, M.E., Clark, C.J., Geange, S.W., Poulsen, J.R., Stevens, M.H.H., White,J.S.S., 2009. Generalized linear mixed effects models: a practical guide for ecologyand evolution. Trends Ecol. Evol. 24, 127–135.

Burns, N.M., Williams, J.D.H., Jaquet, J.M., Kemp, A.L.W., Lam, D.C.L., 1976. A phosphorusbudget for Lake Erie. J. Fish. Res. Board Can. 33, 564–573.

Carrick, H.J., Moon, J.B., Gaylord, B.F., 2005. Phytoplankton dynamics and hypoxia inLake Erie: a hypothesis concerning benthic–pelagic coupling in the central basin.J. Gt. Lakes Res. 31, 111–124.

Charlton, M.N., Milne, J.E., 2004. Review of thirty years of change in Lake Erie waterquality. National Water Research Institute, Aquatic Ecosystem Management Re-search Branch. NWRI Contribution No. 04-167.

Charlton, M.N., Le Sage, R., Milne, J.E., 1999. Lake Erie in transition: the 1990's. In:Munawar, M., Edsall, T., Munawar, I.F. (Eds.), State of Lake Erie (SOLE)—Past, Pres-ent and Future. Backhuys Publishers, Leiden, The Netherlands, pp. 97–123.

Conroy, J.D., Kane, D.D., Dolan, D.M., Edwards, W.J., Charlton, M.N., Culver, D.A., 2005.Temporal trends in Lake Erie plankton biomass: roles of external phosphorus load-ing and dreissenid mussels. J. Gt. Lakes Res. 31, 89–110.

Dahl, J.A., Graham, D.M., Dermott, R., Johannsson, O.E., Millard, E.S., Myles, D.D., 1995.Lake Erie 1993, Western, West Central and Eastern basins: change in trophic sta-tus, and assessment of the abundance, biomass and production of the lower tro-phic levels. Great Lakes Laboratory for Fisheries and Aquatic Sciences, BayfieldInstitute, Fisheries and Oceans Canada. Canadian Technical Report of Fisheriesand Aquatic Sciences No. 2070.

Dame, R., Dankers, N., Prins, T., Jongsma, H., Smaal, A., 1991. The influence of musselbeds on nutrients in the Western Wadden Sea and Eastern Scheldt Estuaries.Estuaries 14, 130–138.

Davies, J.M., Hecky, R.E., 2005. Initial measurements of benthic photosynthesis and res-piration in Lake Erie. J. Gt. Lakes Res. 31, 195–207.

Depew, D.C., Guildford, S.J., Smith, R.E.H., 2006. Nearshore–offshore comparison ofchlorophyll a and phytoplankton production in the dreissenid-colonized easternbasin of Lake Erie. Can. J. Fish. Aquat. Sci. 63, 1115–1129.

Dolan, D.M., McGunagle, K.P., 2005. Lake Erie total phosphorus loading analysis and up-date: 1996–2002. J. Gt. Lakes Res. 31, 11–22.

Esterby, S.R., Bertram, P.E., 1993. Compatibility of sampling and laboratory proceduresevaluated for the 1985 three-ship intercomparison study on Lake Erie. J. Gt. LakesRes. 19, 400–417.

Fahnenstiel, G.L., Lang, G.A., Nalepa, T.F., Johengen, T.H., 1995. Effects of zebra mussel(Dreissena polymorpha) colonization on water quality parameters in SaginawBay, Lake Huron. J. Gt. Lakes Res. 21, 435–448.

Fitzpatrick, M.A.J., Munawar, M., Leach, J.H., Haffner, G.D., 2007. Factors regulating pri-mary production and phytoplankton dynamics in western Lake Erie. Fundam. Appl.Limnol. 169, 137–152.

Glooschenko, W.A., Moore, J.E., Vollenwider, R.A., 1974. Spatial and temporal distribu-tion of chlorophyll a and pheopigments in surface waters of Lake Erie. J. Fish. Res.Board Can. 31, 265–274.

Graham, D.M., Dahl, J.A., Millard, E.S., Johannsson, O.E., White, L.L., 1996. Assessment ofthe abundance, biomass and production of the lower trophic levels in the easternbasin of Lake Erie, 1994. Great Lakes Laboratory for Fisheries and Aquatic Sciences,Bayfield Institute, Fisheries and Oceans Canada. Canadian Technical Report ofFisheries and Aquatic Sciences No. 2110.

Great Lakes Laboratory, 1974. Lake Erie nutrient control program: an assessment of itseffectiveness in controlling eutrophication–eastern basin. 1973 Annual Report tothe Environmental Protection Agency.

Guildford, S.J., Bootsma, H.A., Fee, E.J., Hecky, R.E., Patterson, G., 2000. Phytoplanktonnutrient status and mean water column light intensity in Lakes Malawi and Supe-rior. Aquat. Ecosyst. Health Manag. 3, 35–45.

Guildford, S.J., Hecky, R.E., Smith, R.E.H., Taylor,W.D., Charlton, M.N., Barlow-Busch, L., North,R.L., 2005. Phytoplankton nutrient status in Lake Erie in 1997. J. Gt. Lakes Res. 31, 72–88.

He, C., Rao, Y.R., Skafel, M.G., Howell, T., 2006. Numerical modelling of the Grand Riverplume in Lake Erie during unstratified period. Water Qual. Res. J. Can. 41, 16–23.

Healey, F.P., 1973. Inorganic nutrient uptake and deficiency in algae. CRC Crit. Rev.Microbiol. 3, 69–113.

Healey, F.P., Hendzel, L.L., 1979a. Fluorometric measurements of alkaline phosphataseactivity in algae. Freshw. Biol. 9, 429–439.

Healey, F.P., Hendzel, L.L., 1979b. Indicators of phosphorus and nitrogen deficiency infive algae in culture. Can. J. Fish. Aquat. Sci. 36, 1364–1369.

Heath, R.T., Fahnenstiel, G.L., Gardner, W.S., Cavaletto, J.F., Hwang, S.J., 1995. Ecosys-tem-level effects of zebra mussels (Dreissena polymorpha): an enclosure experi-ment in Saginaw Bay, Lake Huron. J. Gt. Lakes Res. 21, 501–516.

Heathcote, I.W., Weiler, R.R., Tanner, J.W., 1981. Lake Erie nearshore water chemistry atNanticoke, Ontario, 1969–1978. J. Gt. Lakes Res. 7, 130–135.

Hebert, P.D.N., Muncaster, B.W., Mackie, G.L., 1989. Ecological and genetic studies onDreissena polymorpha (Pallas): a new mollusc in the Great Lakes. Can. J. Fish.Aquat. Sci. 46, 1587–1591.

Hecky, R.E., Smith, R.E.H., Barton, D.R., Guildford, S.J., Taylor, W.D., Charlton, M.N., Howell,T., 2004. The nearshore phosphorus shunt: a consequence of ecosystem engineeringby dreissenids in the Laurentian Great Lakes. Can. J. Fish. Aquat. Sci. 61, 1285–1294.

Higgins, S.N., Vander Zanden, M.J., Joppa, L.N., Vadeboncoeur, Y., 2011. The effect ofdreissenid invasions on chlorophyll and the chlorophyll: total phosphorus radioin north-temperate lakes. Can. J. Fish. Aquat. Sci. 68, 319–329.

Holmes, R.M., Aminot, A., Kerouel, R., Hooker, B.A., Peterson, B.J., 1999. A simple andprecise method for measuring ammonium in marine and freshwater ecosystems.Can. J. Fish. Aquat. Sci. 56, 1801–1808.

Hothorn, T., Bretz, F., Westfall, P., 2008. Simultaneous inference in general parametricmodels. Biom. J. 50, 346–363.

Idrisi, N., Mills, E.L., Rudstam, L.G., Stewart, D.J., 2001. Impact of zebra mussels(Dreissena polymorpha) on the pelagic lower trophic levels of Oneida Lake, NewYork. Can. J. Fish. Aquat. Sci. 58, 1430–1441.

Johengen, T.H., Nalepa, T.F., Fahnenstiel, G.L., Goudy, G., 1995. Nutrient changes inSaginaw Bay, Lake Huron, after the establishment of the zebra mussel (Dreissenapolymorpha). J. Gt. Lakes Res. 21, 449–464.

Karatayev, A.Y., Burlakova, L.E., Padilla, D.K., 1997. The effects ofDreissena polymorpha (Pallas)invasion on aquatic communities in eastern Europe. J. Shellfish. Res. 16, 187–203.

Lean, D.R.S., Abbott, A.P., Charlton, M.N., Rao, S.S., 1983. Seasonal phosphate demandfor Lake Erie plankton. J. Gt. Lakes Res. 9, 83–91.

León, L.F., Imberger, J., Smith, R.E.H., Hecky, R.E., Lam, D.C.L., Schertzer, W.M., 2005.Modeling as a tool for nutrient management in Lake Erie: a hydrodynamicsstudy. J. Gt. Lakes Res. 31, 309–318.

MacDougall, T.M., Benoit, H.P., Dermott, R., Johansson, O.E., Johnson, T.B., Millard, E.S.,Munawar, M., 2001. Lake Erie 1998: assessment of abundance, biomass and produc-tion of the lower trophic levels, diets of juvenile yellow perch and trends in the fish-ery. Great Lakes Laboratory for Fisheries and Aquatic Sciences, Canada Centre forInlandWaters. Canadian Technical Report of Fisheries and Aquatic Sciences No. 2376.

MacIntyre, H.L., Kana, T.M., Anning, T., Geider, R.J., 2002. Photoacclimation of photosyn-thesis irradiance response curves and photosynthetic pigments in microalgae andcyanobacteria. J. Phycol. 38, 17–38.

Mackie, G.L., 1991. Biology of the exotic zebra mussel, Dreissena polymorpha, in relationto native bivalves and its potential impact in Lake St. Clair. Hydrobiologia 219,251–268.

Makarewicz, J.C., 1993. Phytoplankton biomass and species composition in Lake Erie,1970–1987. J. Gt. Lakes Res. 19, 258–274.

Makarewicz, J.C., Bertram, P., Lewis, T.W., 2000. Chemistry of the offshore surfacewaters of Lake Erie: pre-and post-Dreissena introduction (1983–1993). J. Gt.Lakes Res. 26, 82–93.

Malkin, S.Y., Sorichetti, R.J., Wiklund, J.A., Hecky, R.E., 2009. Seasonal abundance, com-munity composition, and silica content of diatoms epiphytic on Cladophoraglomerata. J. Gt. Lakes Res. 35, 199–205.



Malkin, S.Y., Bocaniov, S.A., Smith, R.E.H., Guildford, S.J., Hecky, R.E., 2010. In situmeasure-ments confirm the seasonal dominance of benthic algae over phytoplankton in near-shore primary production of a large lake. Freshw. Biol. 55, 2468–2483.

McCusker, E.M., Ostrom, P.H., Ostrom, N.E., Jeremiason, J.D., Baker, J.E., 1999. Seasonalvariation in the biogeochemical cycling of seston in Grand Traverse Bay, LakeMichigan. Org. Geochem. 30, 1543–1557.

Mills, E.L., Leach, J.H., Carlton, J.T., Secor, C.L., 1993. Exotic species in the Great Lakes: ahistory of biotic crises and anthropogenic introductions. J. Gt. Lakes Res. 19, 1–54.

Moon, J.B., Carrick, H.J., 2007. Seasonal variation of phytoplankton nutrient limitationin Lake Erie. Aquat. Microb. Ecol. 48, 61–71.

Munawar, M., Munawar, I.F., Fitzpatrick, M., Niblock, H., Bowen, K., Lorimer, J., 2008. Anintensive assessment of planktonic communities in the Canadian waters of LakeErie, 1998. In: Munawar, M., Heath, R. (Eds.), Checking the Pulse of Lake Erie.Aquatic Ecosystem Health and Management, New Delhi, pp. 297–346.

Nicholls, K.H., Hopkins, G.J., 1993. Recent changes in Lake Erie (north shore) phyto-plankton: cumulative impacts of phosphorus loading reductions and the zebramussel introduction. J. Gt. Lakes Res. 19, 637–647.

Nicholls, K.H., Taylor, R., Hamdy, Y., 1983. The influence of the Grand River on phytoplank-ton near thenortheastern shore of Lake Erie during1979. Arch.Hydrobiol. 98, 146–172.

Nicholls, K.H., Hopkins, G.J., Standke, S.J., 1999a. Reduced chlorophyll to phosphorus ra-tios in the nearshore Great Lakes waters coincide with the establishment ofdreissenid mussels. Can. J. Fish. Aquat. Sci. 56, 153–161.

Nicholls, K.H., Standke, S.J., Hopkins, G.J., 1999b. Effects of dreissenid mussels on nitro-gen and phosphorus in north shore waters of Lake Erie. In: Munawar, M., Edsall, T.,Munawar, I.F. (Eds.), State of Lake Erie (SOLE)—Past, Present and Future. BackhuysPublishers, Leiden, The Netherlands, pp. 323–336.

NLET, 1994. The National Laboratory for Environmental Testing. Environment Canada, Bur-lington, ON.

North, R.L., Guildford, S.J., Smith, R.E.H., Havens, S.M., Twiss, M.R., 2007. Evidence forphosphorus, nitrogen, and iron colimitation of phytoplankton communities inLake Erie. Limnol. Oceanogr. 52, 315–328.

Ohio Lake Erie phosphorus task force, 2010. Ohio Lake Erie Phosphorus Task Force FinalReport. Ohio EPA.

Ontario Ministry of Environment, 2002. Surface Water Monitoring and Assessment: 1998Lake Erie Report.

Padilla, D.K., Adolph, S.C., Cottingham, K.L., Schneider, D.W., 1996. Predicting the con-sequences of dreissenid mussels on a pelagic food web. Ecol. Model. 85, 129–144.

Parsons, T.R., Maita, Y., Lalli, C.M., 1984. A Manual of Chemical and Biological Methodsfor Seawater Analysis. Pergamon.

Patterson, M.W.R., Ciborowski, J.J.H., Barton, D.R., 2005. The distribution and abundance ofDreissena species (Dreissenidae) in Lake Erie, 2002. J. Gt. Lakes Res. 31, 223–237.

Quinn, F.H., 1992. Hydraulic residence times for the Laurentian Great Lakes. J. Gt. LakesRes. 18, 22–28.

R Development Core Team, 2008. R: a language and environment for statistical com-puting. R Foundation for Statistical Computing, Vienna, Austria3-900051-07-0.URL http://www.R-project.org.

Rao, Y.R., Schwab, D.J., 2007. Transport and mixing between the coastal and offshorewaters in the Great Lakes: a review. J. Gt. Lakes Res. 33, 202–218.

Rockwell, D.C., Warren, G.J., Bertram, P.E., Salisbury, D.K., Burns, N.M., 2005. The U.S.EPA Lake Erie indicators monitoring program 1983–2002: trends in phosphorus,silica, and chlorophyll a in the central basin. J. Gt. Lakes Res. 31, 23–34.

Ryan, P.A., Witzel, L.D., Paine, J.R., Freeman, M.J., Hardy, M., Sztramko, K.L., 1999. Recenttrends in Lake Erie fish stocks within a changing trophic state and food web(1980–1994). In: Munawar, M., Edsall, T., Munawar, I.F. (Eds.), State of Lake Erie(SOLE)—Past, Present and Future. Backhuys Publishers, Leiden, The Netherlands, pp.241–290.

Smith, R.E.H., Hiriart-Baer, V.P., Higgins, S.N., Guildford, S.J., Charlton, M.N., 2005.Planktonic primary production in the offshore waters of dreissenid-infested LakeErie in 1997. J. Gt. Lakes Res. 31, 50–62.

Smith, R.E.H., Parrish, C.C., Depew, D.C., Ghadouani, A., 2007. Spatial patterns of sestonconcentration and biochemical composition between nearshore and offshorewaters of a Great Lake. Freshw. Biol. 52, 2196–2210.

Stainton, M.P., Capel, M.J., Armstrong, F.A.J., 1977. The chemical analysis of freshwater.Can. Fish. Mar. Serv. Misc. Spec. Publ. 25.

Stanczykowska, A., Lawacz, W., Mattice, J., 1975. Use of field measurements of con-sumption and assimilation in evaluation of the role of Dreissena polymorpha Pall.in a lake ecosystem. Pol. Arch. Hydrobiol. 22, 509–520.

Strickland, J.D., Parsons, T.R., 1968. A practical handbook of seawater analysis. Bull. Fish.Res. Board Can. 169.

Szabo, R.A., 2004. The benthic food web on nearshore hard substrates at Peacock Point, easternLake Erie, M. Sc. thesis, Department of Biology, TheUniversity ofWaterloo,Waterloo, ON.

Thompson, P.A., Harrison, P.J., Parslow, J.S., 1991. Influence of irradiance on cell-volume and carbon quota for 10 species of marine-phytoplankton. J. Phycol. 27,351–360.

Upsdell, B.L., 2005. The carbon and nitrogen composition of suspended particulate matterin Lake Erie, selected tributaries, and its outflow, M. Sc. thesis, Department of Biology,The University of Waterloo, Waterloo, ON.

USEPA, 1971. Methods for Chemical Analysis of Water andWastes. E.P.A. Water QualityOffice. Analytical Quality Control Laboratory.

Watson, S.B., Ridal, J., Boyer, G.L., 2008. Taste and odour and cyanobacterial toxins: im-pairment, prediction, and management in the Great Lakes. Can. J. Fish. Aquat. Sci.65, 1779–1796.

Wu, L., Culver, D.A., 1991. Zooplankton grazing and phytoplankton abundance: an assess-ment before and after invasion of Dreissena polymorpha. J. Gt. Lakes Res. 17, 425–436.

Zhang, H., Culver, D.A., Boegman, L., 2011. Dreissenids in Lake Erie: an algal filter or afertilizer? Aquat. Invasions 6, 175–194.


Documents

Author's personal copy - Limnologylimnology.missouri.edu/pdf/north/Northetal2012.pdf · Author's personal copy Distribution of seston and nutrient concentrations in the eastern basin