The Development of a Planktonic Index of Biotic Integrity

Home , Cercopagis pengoi, Lake ecosystem

THE DEVELOPMENT OF A PLANKTONIC INDEX OF BIOTIC INTEGRITY

FOR LAKE ERIE

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of the Ohio State University

Douglas D. Kane, B.S., M.S.

*****

The Ohio State University 2004

Dissertation Committee:

Dr. David A. Culver, Adviser

Dr. Paul C. Baumann Approved by

Dr. Steven I. Gordon

Dr. Michael A. Hoggarth ______

Dr. Jeffrey M. Reutter Adviser

Dr. Roy A. Stein Department of Evolution, Ecology, and Organismal Biology

ABSTRACT

Herein, I provide a water quality monitoring tool for the offshore waters of Lake

Erie. The Planktonic Index of Biotic Integrity (P-IBI) was developed using Lake Erie zooplankton and phytoplankton data. I reviewed literature related to Indices of Biotic

Integrity and plankton ecology (Chapter 1). To construct a valid IBI, I conducted temporal and spatial analyses for different parameters of Lake Erie plankton and plankton sampling regimens (Chapter 2). Using this information, I then determined appropriate temporal and spatial sampling frequency needed for the P-IBI. Further, using Lake Erie monitoring data I provide information on the invasive predatory cladoceran, Cercopagis pengoi (Chapter 3), and on the temporal and spatial distribution of Limnocalanus macrurus (Chapter 4), a pollution intolerant calanoid copepod. The P-IBI was developed using phytoplankton and zooplankton data from 1970 and 1996, validated with respect to total phosphorus and chlorophyll a concentrations, and applied to approximately 10 years of Lake Erie plankton data (Chapter 5). P-IBI candidate zooplankton metrics included an abundance ratio of calanoid copepods to cladocerans and cyclopoid copepods, percentage density of large Daphnia, rotifer composition, density of the calanoid copepod

Limnocalanus macrurus, percentage biomass of the predatory invasive zooplankters

Bythotrephes and Cercopagis, biomass of crustacean zooplankton/ biomass of phytoplankton, and biomass of crustacean zooplankton. Candidate phytoplankton metrics included a generic index of diatoms, abundance ratio of centric diatoms to

ii pennate diatoms, biomass of inedible algae taxa, percentage biomass of bluegreen algae of the total phytoplankton biomass, percentage biomass of the potentially toxic

Microcystis, Anabaena, and Aphanizomenon of total phytoplankton biomass, and biomass of edible algae taxa.

Discriminant analysis was used to determine which metrics reflected levels of degradation in Lake Erie, as based on total phosphorus and chlorophyll a concentrations.

A Kappa statistic was calculated to determine the classification accuracy of the significant metrics obtained from the discriminant analysis. Significant metrics for June included biomass of edible algae taxa, percentage Microcystis, Anabaena, and

Aphanizomenon of total phytoplankton biomass, and an abundance ratio of calanoids to cladocerans and cyclopoids. The only significant metric for July was the density of

Limnocalanus macrurus, while both the zooplankton abundance ratio and crustacean zooplankton biomass were significant metrics for August. The P-IBI was significantly correlated with a measure of eutrophication (e.g., total phytoplankton biomass) and showed an increase of water quality between 1970 and the mid-1990’s (1996 and 1997) in Lake Erie, with declining water quality in the late 1990’s and early 2000’s.

iii

Dedicated to my father, Donald E. Kane

ACKNOWLEDGMENTS

I wish to thank my adviser, Dr. David A. Culver, for teaching me to think more scientifically, broadening my academic horizons, and for providing both intellectual and financial support for this project.

I thank a number of scientists who have helped me throughout my research. First of all, I thank my committee members: Drs. Roy Stein, Jeff Reutter, Paul Baumann,

Michael Hoggarth, and Steve Gordon, for their expertise and suggestions. I especially thank Dr. Steve Gordon, Department of City and Regional Planning, the Ohio State

University, and the Ohio Supercomputer Center, for consultation on statistical methodology and Roy Stein for his careful criticism of drafts of this dissertation. I would also like to thank Drs. John Gannon, David Jude and C. Edward Herdendorf for helpful advice. I thank Roger Thoma, Ohio EPA, who believed in the importance of developing a P-IBI, as well as the whole Lake Erie Quality Index Team for helpful feedback during the development of the Planktonic IBI.

I thank both the Ohio Journal of Science for permission to use my publication

“The characteristics and potential ecological effects of the exotic crustacean zooplankter

Cercopagis pengoi (Cladocera: Cercopagidae), a recent invader of Lake Erie” (Ohio

Journal of Science 103: 79-83) and the Journal of Great Lakes Research for permission to use my publication “The status of Limnocalanus macrurus (Copepoda: Calanoida:

v Centropagidae) in Lake Erie” ( Journal of Great Lakes Research 30: 22-30) as chapters in my dissertation. Further, I thank Dr. Roy Stein for suggesting that these publications be included as part of my dissertation.

For the Cercopagis research, I would like to again thank Dr. David Jude, Center for Great Lakes and Aquatic Sciences- University of Michigan, for showing me samples of Cercopagis pengoi from Lake Michigan that allowed me to identify it in Lake Erie. I would also like to thank two anonymous reviewers who greatly improved the quality of this manuscript. For the Limnocalanus research, I thank Jeff Tyson and Carey Knight of the ODW, Donald Einhouse of New York State Department of Environmental

Conservation (NYSDEC), and Roger Kenyon of the Pennsylvania Fish and Boat

Commission for rainbow smelt trawl data. I thank Katherine Simons and Doug Hunter,

Biological Sciences, Oakland University, Rochester, MI and Bruce Davis, USGS/BRD,

Great Lakes Science Center, Ann Arbor, MI for Lake St. Clair Limnocalanus data.

Finally, I thank two anonymous reviewers, whose constructive comments allowed us to greatly improve this manuscript.

I would also like to thank all of the technicians who have analyzed plankton samples for the Lake Erie Plankton Abundance Study (LEPAS). I am especially grateful to Erin Haas and Nate Gargasz for the many years they spent analyzing zooplankton and phytoplankton samples, as well as maintaining the LEPAS database. I thank Joe Conroy for assistance in the preparation of a number of figures and helpful discussions. Finally, I

vi thank Kelsey Reider for conducting the VIP research and for data entry and Maryke

Swartz for data entry.

I thank the Ohio Lake Erie Protection Fund for supporting this work (LEQI 01-

05). Further, plankton analyses for this project were funded by the Federal Aid in Sport

Fish Restoration Program (F-69-P, Fish Management in Ohio), administered jointly by the U.S. Fish and Wildlife Service and the Ohio Division of Wildlife. A Raymond C.

Osborn Memorial Fellowship from the Graduate School at the Ohio State University and the Department of Evolution, Ecology, and Organismal Biology at OSU provided me additional summer support. Further, I thank the Ohio Division of Wildlife of the Ohio

Department of Natural Resources, Ontario Ministry of Natural Resources, and the

National Water Resources Institute of Environmental Canada for extensive Lake Erie plankton sample collection from the 1990’s until present. I am also greatly indebted to

Dr. Murray Charlton of the National Water Resources Institute for access to site, nutrient, chlorophyll a, and zooplankton data from 1970, Dr. Mohiuddin Munawar of the

Department of Fisheries and Oceans for phytoplankton data for 1970, and Dr. David

Dolan of the University of Wisconsin- Green Bay for phosphorus loading data. Finally, I thank Heather Niblock of Dr. Munawar’s lab for assistance with data acquisitio n.

Finally, I would like to thank my wife Melissa for being there for me through much of this research, and my mother, Carol L. Kane, for providing emotional and financial support.

vii

VITA

January 8, 1977………………………Born – Cleveland, Ohio

1999………………………………….B.S. Zoology, with Minors in Plant Biology and

History of Art, The Ohio State University

2002………………………………….M.S. Evolution, Ecology, and Organismal Biology,

The Ohio State University

1999 – present……………………….Graduate Teaching and Research Associate,

The Ohio State University

PUBLICATIONS

Research Publications

1. Kane, D.D. 2002. Science in the art of the Italian Renaissance I: Ghiberti’s Gates of Paradise- linear perspective and space. Ohio Journal of Science 102: 110-112.

2. Kane, D.D. 2002. Science in the art of the Italian Renaissance II: Leonardo

DaVinci’s representations of animals in his works. Ohio Journal of Science 102: 113-

115.

viii

3. Therriault, T. W., Grigorovich, I. A., Kane, D. D., Haas, E. M., Culver, D. A., and

MacIsaac, H. J. 2002. Range expansion of the exotic zooplankter Cercopagis pengoi

(Ostroumov) into western Lake Erie and Muskegon Lake. Journal of Great Lakes

Research 28: 698-701.

4. Kane, D.D., Haas, E. M., and Culver, D. A. 2003. The characteristics and potential ecological effects of the exotic crustacean zooplankter Cercopagis pengoi

(Cladocera: Cercopagidae), a recent invader of Lake Erie. Ohio Journal of Science 103:

79-83.

5. Kane, D. D., Gannon, J.E., and Culver, D.A. 2004. The status of Limnocalanus macrurus (Copepoda: Calanoida: Centropagidae) in Lake Erie. Journal of Great Lakes

Research 30: 22-30.

FIELDS OF STUDY

Major Field: Evolution, Ecology, and Organismal Biology

TABLE OF CONTENTS

Page

Abstract…………………………………………………………………………………...ii

Dedication………………………………………………………………………………..iv

Acknowledgments…………………………………………………………………….….v

Vita……………………………………………………………………………………...viii

List of Tables……………………………………………………………………………xiii

List of Figures……………………………………...………………………………..…xviii

Chapters:

1. Introduction…………………………………………………………………….…1

2. Characterizing a changing Lake Erie with phytoplankton and zooplankton…….18

Introduction………………………………………………………………………18

Materials and Methods…………………………………………………………...22

Results……………………………………………………………………………38

Discussion……………………………………………………………………..…46

Conclusion……………………………………………………………………….54

x 3. The characteristics and potential ecological effects of the exotic crustacean

zooplankton Cercopagis pengoi (Cladocera: Cercopagidae), a recent invader

of Lake Erie……………………………………………………………………...56

Introduction……………………………………………………………………...56

Materials and Methods……………………………………………………..……58

Results……………….…………………………………………………………..60

Discussion……………………………………………………………………….61

Conclusion………………………………………………………………………69

4. The status of Limnocalanus macrurus (Copepoda:Calanoida:

Centropagidae) in Lake Erie………………………………………………….…70

Introduction………………………………………………………………….…..70

Materials and Methods…………………………………………………….…….75

Results…………………………………………………………………………...78

Discussion………………………………………………………………………..80

Conclusion……………………………………………………………………….84

xi 5. Development, validation, and application of a Planktonic Index of Biotic

Integrity…………………………………………………………………………..86

Introduction………………………………………………………………….…...86

Materials and Methods…………………………………………………….……..92

Results…………………………………………………………………………..128

Discussion………………………………………………………………………132

Conclusion……………………………………………………………………...145

Appendices:

A. Tables……..…………………………………………………………………….146

B. Figures…………………………………………………………………………..207

C. LEPAS Methods…………………………………………………………..……236

List of References………………………………………………………………………252

xii

LIST OF TABLES

Table Page

1 Interactions among Beneficial Use Impairments(BUIs) (Hartig et al. 1997) and planktonic characteristics in Lake Erie. Characteristics impacting specific BUIs are found in the right column (Culver, unpublished). Plankton characteristics- A. Taste and odor production, B. Toxins from Cyanophyta, C. Floating algal mats or blooms, D. Dominant phytoplankton inedible by zooplankton, E. Predation by introduced planktonic invertebrates, F. Non-indigenous species (Dreissena planktonic larvae), G. Large planktonic predators, H. Large Daphnia spp. dominant……………..147

2 Number of phytoplankton and zooplankton samples collected from each of the three Lake Erie basins (1996-2002) and the length of season sampled for temporal and spatial biomass comparisons (J.D. Conroy, The Ohio State University, personal communication)…………………………..….148

3 Numbers of sites and site identities of full and reduced suites of sites analyzed in the western basin of Lake Erie (1996-2002). For the location of site numbers see Figure 2…………………………...………...……149

4 Comparison of sampling, enumeration, and biomass calculation methods for phytoplankton and zooplankton in the western, central, and eastern basins of Lake Erie (1970-2002)…………………….……………………………………154

xiii 5 Temporal comparisons of sampling regimens in Lake Erie for phytoplankton and zooplankton biomass for 1996-2002. Number of samples taken during the entire sampling season (full number of samples), and number of samples taken during a month (reduced number of samples) are given. Bonferroni corrected alpha value are given using α = 0.05/ number of comparisons for individual statistical test in a given data set. Results of regression analyses include r2 values and F-ratios with associated p-values. Results for paired t-tests include % mean difference ((mean biomass for entire season – mean biomass for month)/ mean biomass for entire season) * 100) and t-values with associated p-values. Bold p-values indicate significance at the appropriate Bonferroni corrected alpha level……………..158

6 Spatial comparisons of sampling regimens in Lake Erie for phytoplankton and zooplankton biomass for 1996-2002. Number of sites sampled in the western basin during the entire sampling season (full number of sites), and the diminished number of sites that served as surrogates (reduced number of sites) are given. Bonferroni corrected alpha value are given using α = 0.05/ number of comparisons for individual statistical test in a given data set. Results of regression analyses include r2 values and F-ratios with associated p-values. Results for paired t-tests include % mean difference ((mean biomass for full number of sites- mean biomass for reduced number of sites)/ mean biomass for full number of sites) * 100) and t-values with associated p-values. Bold p-values indicate significance at the appropriate Bonferroni corrected alpha level.……………………………….………………..….………………...167

7 Spatial and temporal variability of phytoplankton and zooplankton biomass in Lake Erie (1996-2002). F-ratios and p values are from ANOVA analyses. Bold p-values are significant at the Bonferroni corrected α level. Bonferroni corrected alpha value are given using α = 0.05/ number of comparisons for individual statistical test in a given data set (i.e., phytoplankton, zooplankton biomass) in a given year (i.e., phytoplankton biomass- α = 0.05/24 comparisons = 0.002, zooplankton biomass- α = 0.05/15 = 0.0033). Also given is % of the variance explained by basin, site within basin, and month, along with the % residual variance….…………..….172

8 Breakpoint analysis of (a) total phytoplankton biomass and (b) total crustacean zooplankton biomass from 1970-2002. The breakpoint year and r2 and the regression slope, intercept, r2, and p-value are given for the regression equations before and after the breakpoint year (af ter Conroy et al. 2004a).…………………………………………….….……….……………191

xiv

9 Lengths, weights, densities, and biomasses of Cercopagis pengoi sampled during August and September 2001 in the western basin of Lake Erie. Average length, average weight, density, and biomass of Leptodora kindti from the same dates are included for comparison. No individuals of Bythotrephes longimanus were found. See Figure 7 for site identification………………….……….………...…………………………….192

10 Spatial and temporal abundances of Limnocalanus macrurus throughout sampling season in each basin of Lake Erie 1995-2000. Total numbers of samples, samples containing L. macrurus and frequency of occurrence of L. macrurus in samples, expressed as a percentage. EB = eastern basin, CB= central basin, WB= western basin …….…………………………..…....……..193

11 Monthly variation in abundance of Limnocalanus macrurus in the western basin of Lake Erie (1995-2000)…………………………………...….194

12 Zooplankton candidate metrics for Lake Erie P-IBI, along with their description/ ecological relevance, what they measure, and their hypothesized response to degradation. Metrics included in the final multimetric P-IBI are in bold. …………………………………………..….....195

13 Phytoplankton candidate metrics for Lake Erie P-IBI, along with their description/ ecological relevance, what they measure, and their hypothesized response to degradation. Metrics included in the final multimetric P-IBI are in bold. ……………………………………….…..……196

14 Comparison of sampling dates, sampling frequency, number of sites, sampling method, enumeration method, and biomass calculation method for phytoplankton and zooplankton in Lake Erie (1970 and 1995-2002)....………197

15 Ranges of total phosphorus (µg P/L) and chlorophyll a (µg/L) for each trophic class (Chapra and Dobson 1981). Individual and summed (total phosporus + chlorophyll a) metric values are given for each class, as well as possible combinations that give the summed metric values………..………198

16 95th percentile of P-IBI metrics for all Lake Erie observations (3 basins, 1970 and 1996). Cutoff values were determined by trisection of the 0-95th percentile range for metric scores for each metric and assigned values of 1, 3, or 5. For metrics that have a 0.00 value in the 1 scoring class, the cutoff minimal values should be read from left to right, while those with a 0.00 value in the 5 scoring class should be read from right to left.…………………199

17 Significant metrics for each month, as determined by stepwise discriminant analysis. The metric, its partial r2, F, and p values are all given. Significance level to enter and remove was set at p = 0.20. Metric numbers are the same as those in Tables 2 and 3.………..……………………200

18 Metrics excluded/included in final P-IBI. Reasons for not including rejected candidate metrics are also given. DA= included in discriminant analysis, IBI= included in final IBI.……………………………………….….201

19 95th percentile of P-IBI metrics for all Lake Erie observations (3 basins, 1970 and 1996). Cutoff values were determined by trisection of the 0-95th percentile range for metric scores for each metric and assigned values of 1, 3, or 5. For metrics that have a 0.00 value in the 1 scoring class, the cutoff minimal values should be read from left to right, while those with a 0.00 value in the 5 scoring class should be read from right to left………..……….203

20 Sample calculation of P-IBI for mean site score. XXX refers to a month when the particular metric is not significant in the discriminant analysis and thus not used as part of the P-IBI. EB= eastern basin of Lake Erie. Although western and central basin mean site score calculations are not shown, they follow the same methodology as the eastern basin example given below. Metric 1 = zooplankton ratio (Calanoida/ (Cladocera + Cyclopoida)), metric 4 = Limnocalanus macurus density (#/L), metric 7 = crustacean biomass (µg/L), metric 12 = % Microcystis, Anabaena, and Aphanizomenon of total phytoplankton biomass, and metric 13 = Biomass of edible algae taxa (µg/L). Sum of metrics across June, July, August equals 24. Mean site score equals 24 (13 + 1 + 10) divided by 6 (the number of metrics summed)……….……………………………………..….204

xvi

21 Sample calculation of mean basin score and mean lakewide score for the P-IBI. Mean basin score equals sum across sites (21.01) divided by the number of sites (6), which is 3.50. Mean lakewide score equals the sum of the mean basin scores divided by the number of basins in Lake Erie (3). For this example, mean western basin score equals 3.69, mean central basin score equals 3.69, and mean eastern basin score equals 3.50. Therefore, the mean lakewide score equals the sum of these three values (10.88) divided by 3, which is 3.63………………………………..205

22 Basin and lakewide P-IBI scores in Lake Erie for 1970, and 1995-2002. The basin scores are the means of all the site scores in the basin. The lakewide score is the mean of the three basin scores. Trophic status classifications are based on IBI scores and are based on the scale of <3 reflecting eutrophic conditions, 3 -4 reflecting mesotrophic conditions, and >4 reflecting oligotrophic conditions. WB = western basin, CB = central basin, and EB = eastern basin; E = eutrophic, M = mesotrophic, and O = oligotrophic…………………….206

xvii

LIST OF FIGURES

Figure Page

1 Locations of Lake Erie Plankton Abundance Study (LEPAS) sampling sites in Lake Erie from 1996-2002 (every site was not sampled every year). Squares indicate sites sampled in 1996, crosses indicate sites sampled in 1997, and circles indicate sites typically sampled during 1998-2002. Sites 84 and 1279 in the central basin were used in spatial comparisons………………………………………………………..208

2 Lake Erie Plankton Abundance Study (LEPAS) sites sampled in the western basin of Lake Erie during 1996-2002. The identities of subsets used in spatial comparisons are given in Table 3………………………….…209

3 Volumetric Index of the Plankton (VIP) regression analyses. (a) Relationship between log crustacean zooplankton biomass (µg/L) and log standardized crustacean zooplankton volume in Lake Erie during 1998. (b) Relationship between log cyanophyte biomass (µg/L) and log standardized cyanophyte volume in Lake Erie during 1998………….……...210

4 Phosphorus loading (kilotonnes) in Lake Erie 1969-2001. Each year represents a water year. A water year begins on October 1 and ends on September 30 of the following year and is referred to by the year in which it ends. For example, the phosphorus loading values for the year 1995 are calculated from October 1994-September 1995 (i.e. the 1995 water year). Water years are used to calculate loading because data from which loading is calculated (USGS stream flow data) are compil ed by water year. Data are from Dolan 1993 and Dolan unpublished ([email protected], University of Wisconsin - Green Bay). The target loading level of 11 kilotonnes is based upon the Great Lakes Water Quality Agreement. ………………………………………………..…212

xviii

5 Temporal change in Lake Erie seasonal average phytoplankton and crustacean zooplankton biomass (mg/L) in the western (a,d), central (b,e), and eastern (c,f) basins. Phytoplankton data (a,b,c) from 1970 are from Munawar and Munawar (1976), 1978 data are from Devault and Rockwell (1986), 1983-87 data are from Makarewicz (1993a), and 1996-2002 data are from the Lake Erie Plankton Abundance Study (LEPAS). Zooplankton data (d,e,f) from1970 are from Bean (1980), 1974-75 data (western basin only) are from Weisgerber (2000), 1984-87 data are from Makarewicz (1993b), and 1996-2002 data are from LEPAS. Error bars are +1 standard error and are only calculated for the 1996-2002 data. Arrows indicate breakpoint years (Table 8) (after Conroy et al. 2004a)……………………….…………………………………………….….213

6 Lake Erie seasonal average phytoplankton biomass (mg/L) (a,b,c) and seasonal average crustacean zooplankton biomass (mg/L) (d,e,f) as a function of lake-wide annual estimated total phosphorus loading (ktonnes) for the western (a, d), central (b,d), and eastern (c,f) basins from 1970 to 2001 (after Conroy et al. 2004a)……………………...……………………….214

7 Sites sampled in the western basin of Lake Erie during 2001. Filled-in circles denote sites whose samples contained Cercopagis pengoi, sampled sites that did not contain C. pengoi are denoted by crosses. Site 27 is located at 41” 46.32’ N, 83” 3.17’ W; Site 29 is located at 41” 51.59’ N, 83” 8.07’ W.…………………………………………..…….………….…….216

8 Vertical distribution of Limnocalanus macrurus in the Steinsfjord, Norway demonstrating its existence between narrow environmental limits (redrawn from Gannon and Beeton 1971, after Strøm 1946)….………217

9 Lake Erie Plankton Abundance Study (LEPAS) sampling sites 1995-2000 (open circles), sites where Limnocalanus macrurus was found on one or more dates in Lake Erie 1995-2000 (crosses), and mean abundance (#/m3) of Limnocalanus macrurus (contour lines) in Lake Erie 1995-2000 (all sampling dates). Contour intervals are 10/ m3………………...….……...218

xix 10 Uncorrected chlorophyll a concentrations (µg/L) (a) and unfiltered total phosphorus concentrations (µg/L) (b) in the three basins of Lake Erie in 1970 and 1996. The boxes indicate the 25th percentile, median, and 75th percentile, whiskers extend from the 10th to the 90th percentile, while dots indicate the 5th and 95th percentile values………………….……...219

11 1970 and 1996 sites used for Planktonic IBI development in Lake Erie. Black triangles represent 1970 sites, while white circles represent 1996 sites……………………………………………………………………………221

12 Edible phytoplankton biomass versus trophic status (as determined by total phosphorus and chlorophyll a) regression analysis for June (1970 and 1996). oligotrophic =1, mesotrophic=2, eutrophic=3….……………………..222

13 Sample graph demonstrating boxplot conventions and trisection technique (based on Karr et al. 1996). The boxes in the boxplots are comprised of horizontal lines designating the 25th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentile of the cumulative frequency distribution of edible phytoplankton biomass. These conventions are followed for all subsequent boxplots. Also demonstrated is the trisection technique used and individual metric score assignment (lines are drawn for demonstration purposes and thus are approximations). In this case low values (0 - 849 µg/L) are assigned the highest water quality value, 5………………………………………………………………………...223

14 Comparison of cumulative frequency distributions for P-IBI metrics in June (a-c), July (d), and August (e-f) in 1970 and 1996 across all Lake Erie basins. The boxes indicate the 25th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentiles. These conventions are followed for all subsequent boxplots. Please note that due to different calculation techniques, the 95th percentiles in these visual izations may differ slightly from those in Table 9. However, the values in Table 9 are the values from which all classifications for the P-IBI were made……..…………..….…224

xx 15 Comparison of cumulative frequency distributions of P-IBI metrics versus trophic status in June (a-c), July (d), and August (e-f) in 1970 and 1996 across all Lake Erie basins. The boxes indicate the 25th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentiles. H and p-values are from the Kruskal-Wallis tests that were used to test the equality of medians. Significance is judged at a 0.05 α level and is indicated in bold……………………….…………….…………….……….226

16 Comparison of cumulative frequency distributions of P-IBI metrics between 1970 and 1996 in June (a-c), July (d), and August (e-f) across all Lake Erie basins. The boxes indicate the 25 th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentiles. W and p-values are from the Mann-Whitney tests. Significance is judged at a 0.05 α level and is indicated in bold that were used to test the equality of medians. Due to zero values in 1996, Mann-Whitney tests could not be performed for Limnocalanus macrurus abundance (d)………………………………………228

17 Mean lake-wide P-IBI score versus year (1970, 1996-2002) (a) and mean basin P-IBI score versus year (b) (1970, 1996-2002)..……………………….230

18 Relationship between external phosphorus loading (ktonnes) to Lake Erie and mean lakewide P-IBI score (1970, 1996-2001). For external phosphorus loading, each year represents a water year. A water year begins on October 1 and ends on September 30 of the following year and is referred to by the year in which it ends. For example, the phosphorus loading values for the year 1995 are calculated from October 1994-September 1995 (i.e., the 1995 water year). Water years are used to calculate loading because data from which loading is calculated (USGS stream flow data) are compiled by water year. Phosphorus l oading data are from Dolan 1993 and Dolan unpublished ([email protected], University of Wisconsin - Green Bay)……………………………………….231

19 Mean chlorophyll a (µg/L) versus mean P-IBI score (a) at the same site (1970, 1996-2000). For 1970 uncorrected chlorophyll a data were used, while for 1996-2000 corrected chlorophyll a data were used. Mean total phosphorus (µg/L) versus mean P-IBI score (b) at the same site (1970, 1996-2000)………………………………………….…………….…..232

xxi 20 Mean basin total phytoplankton biomass (mg/L) versus mean basin P-IBI score (a) (1970, and 1996-2002) and offshore fish community IBI score (all ages of fish, all sites in the western and central basins of Lake Erie, Karr IBI method) (Kershner and Hopkins 2003) versus mean lakewide P-IBI score (b) (1970, and 1996-1999). Offshore fish community IBI scores were estimated from Figure 16 (Kershner and Hopkins 2003)…………………………………………………….………..…234

xxii

CHAPTER 1

INTRODUCTION

Responding to problems in the 1960’s, the U.S. government took steps to protect the quality of water in the U.S. Legislation was enacted in 1972 (Water Pollution Control

Act Amendments of 1972) and in 1977 (Clean Water Act) that called for the restoration and maintenance of not only the physical and chemical integrity of U.S. waters, but also the biological integrity of these waters (Karr 1991). Because of the explicit inclusion of biological integrity into this definition, physical and chemical measures of water were no longer sufficient. Thus, a number of biotic measures of water quality were subsequently developed (e.g., Index of Biotic Integrity (IBI), Invertebrate Community Index (ICI)).

Indices of biotic integrity are tools to measure the biological water quality of ecosystems.

As Karr (1991) notes, “the solution of water resource problems will not come from better regulation of chemicals or the development of better assessment tools to detect degradation.” What is needed is a better understanding of biological water quality and what organisms are appropriate for its measurement. Although various indices exist to measure Lake Erie water quality, a majority of these focus on nearshore areas. Few biological measures of water quality have been applied to the offshore waters of Lake

Erie; however, understanding of these offshore waters is critical in any management plan

1 for Lake Erie. Therefore I sought to develop a biological water quality monitoring tool, a

Planktonic Index of Biotic Integrity (P-IBI), for the offshore waters of Lake Erie.

Plankton is ideal for offshore water quality monitoring, because it is inexpensive to collect and highly sensitive to changes in ecosystems, two requirements of useful monitoring programs (Schindler 1987). For the remainder of this introductory chapter I will 1) review indices in ecology, particularly with regards to lake classification, 2) review indices of biotic integrity, 3) justify the need for a P-IBI to measure water quality in Lake Erie, 4) review variability in plankton communities, and 5) review stressors that affect plankton communities.

A. G. Tansley (1935) who first defined the term ecosystem as a “particular category of physical systems that consist of organisms and inorganic components in a relatively stable equilibrium,” provided a basis for our subsequent examination of ecosystem health. Various ecologists expanded his definition in the following years. E.

P. Odum (1969) extended the definition of ecosystem to mean “any unit that includes all of the organisms (i.e., the “community”) in a given area interacting with the physical environment so that the flow of energy leads to clearly defined trophic structure, biotic diversity, and material cycling (i.e., exchange of materials between living and nonliving parts) within the system is an ecological system or ecosystem.” As ecologists tried to define ecosystems, they also tried to apply quantitative measures to the properties of systems. For example, Lindeman (1942) created a framework for studying trophic dynamics, while Shannon and Weaver (1949) and Wiener (1948) developed indices that

2 were used to measure species diversity as measures of the potential number of interactions between taxa.

The creation of indices in ecology reflects Elster’s (1974) three reasons that science attempts to order its findings: 1) to gain an overall view, 2) to simplify the understanding of complex systems by characterizing a few common factors, and 3) to predict properties or relationships of parts of systems from other measured properties.

Accordingly, the use of indices to characterize ecosystems led to systems to classify lakes in the 20th century. Leach and Herron (1992) reviewed the history of lake classification attempts, noting that the first attempt to characterize lakes according to trophic status was made by Weber (1907), who divided lakes into the familiar oligotrophic-mesotrophic- eutrophic continuum. A number of limnologists developed indices for placing lakes into this framework, including the use of single parameters such as: hypolimnetic oxygen loss

(Hutchinson 1957), primary production (Rodhe 1958), total phosphorus (Chapra and

Robertson 1977), total nitrogen (Vollenweider 1968), chlorophyll a (Dobson et al. 1974), and Secchi disk transparency (Vallentyne et al. 1969).

Composite indices also have been developed, including the Morphoedaphic Index for fish yield prediction in north-temperate lakes (Ryder 1965), the Trophic State Index

(TSI) using Secchi disk transparency, specific conductance, total organic nitrogen, total phosphorus, primary production, chlorophyll a, and Pearson’s cation ratio (Shannon and

Brezonik 1972). The TSI was later modified by Carlson (1977) to include only Secchi disk transparency, chlorophyll a, and total phosphorus. Finally, the Lake Condition Index

3 was developed to incorporate dissolved oxygen, transparency, fish kills, and use impairment into one index (Uttormark and Wall 1975).

As noted above, the Clean Water Act sought to assure the health of aquatic systems by restoring and maintaining biological integrity. Biological integrity can be defined as “the ability of an aquatic ecosystem, to support and maintain a balanced, integrated, adaptive community of organisms having a species composition, diversity, and functional organization comparable to that of the natural habitats of a region” (Karr and Dudley 1981). Another view, one of the associated concepts of ecosystem health comes from the philosopher J. Baird Callicott (1995) who defined ecosystem health as

“the normal occurrence of ecological processes and function.” He goes on to state that health is “an objective condition of ecosystems, although the concept of ecosystem health allows some room for personal and social determination or construction.”

Ecologists have tried to measure ecosystem health and ecosystem integrity through the development of Indices of Biotic Integrity (IBIs). Karr (1981) first devised an index to measure biological integrity in a stream, using fish as indicator species.

Karr’s (1981) Index of Biological Integrity (IBI) has been adopted by many management agencies (e.g., Ohio EPA 1988), been modified to use benthic macroinvertebrates as indicators (Fore et al. 1996), and has even been modified to evaluate the integrity of estuarine ecosystems (Weisberg et al. 1997). Although IBIs have been applied or modified for many purposes, all developed to date are limited to streams or nearshore communities, which are relatively restricted in size. Such indices are thus poorly suited to assess the overall health of very large lakes. In particular, a P-IBI or Planktonic Index

4 of Biotic Integrity developed for large lakes would be of great utility for forming or testing the results of major management decisions and in informing the public of the relative health of the aquatic ecosystem.

Arguably all components of lake function are influenced in major ways by the dynamics of the phytoplankton and zooplankton. Phytoplankters are the primary source of energy driving large lake ecosystems, and the zooplankton is the central trophic link between primary producers and fish (Schriver et al. 1995, Tatrai et al. 1997).

Zooplankton dynamics influence the recruitment of young-of-year fish. In Lake Erie, this includes economically important walleye (Sander vitreus) and yellow perch (Perca flavescens) (Gopalan et al. 1998, Culver and Wu 1997, Wu and Culver 1992, 1994).

Contaminants reaching fish, birds, and humans have been previously bioconcentrated by the phytoplankton and zooplankton. Recently, algal consumption by zebra mussels

(Dreissena spp.) has further concentrated contaminants from plankton through assimilation from ingested algae or through release of toxic feces and pseudofeces that can be eaten by other benthic organisms, such as amphipods (Fisher et al. 1993, Bruner et al. 1994). Nuisance algal blooms have been a problem in all of the lower Great Lakes, but toxic blooms of Microcystis, Anabaena, and Aphanizomenon have been particularly common in Lake Erie, where blooms of Microcystis occurred in 1995 (Budd et al. 2002) and infrequently in subsequent years. Microcystis can produce a liver toxin, microcystin, whereas the other two taxa can produce neurotoxins, including saxitoxin, the compound responsible for paralytic shellfish poisoning in the marine environment (Carmichael

1997). These taxa as well as other algae can produce geosmin and 2-methylisoborneol

(MIB) (Sugiura et al. 1998), compounds responsible for taste and odor problems in water supplies (Jones and Korth 1995) and off-flavors in fish (Persson 1980). Clearly, zooplankton and phytoplankton dynamics have a large impact on aquatic ecosystems and on humans who use these environments.

Further, humans have a number of negative effects on aquatic ecosystems and the organisms that reside in them. Input of contaminants, habitat modifications, and alteration of energy sources are just three of the negative impacts humans have had on aquatic ecosystems (Karr 1981). Often, water quality has been used as the indicator of damage to these aquatic systems. In measuring the physical and chemical properties of water, biotic properties have often been overlooked. Likewise, monitoring the concentrations of various chemicals (heavy metals, PCBs, etc.) often misses a number of human-induced problems, such as habitat degradation and flow alterations. In addition, physical and chemical attributes are usually poor surrogates for the measure of biological properties (Karr and Dudley 1981). A more informative method would be to use the biota to measure community health directly, rather than the use of abiotic measures that indirectly affect the biota. Karr (1981) originally developed the Index of Biological

Integrity for evaluating the “health” of a watershed. Herein, I develop a multimetric

Index of Biotic Integrity, for use in the evaluation of the “health” of the open waters of

Lake Erie, similar to those developed for other suites of organisms, but based on plankton. This index would be used to evaluate and communicate changes in Lake Erie to the public, resource managers, politicians, and other scientists. To develop the P-IBI,

6 information on the variability of plankton communities, and stressors that affect these communities is needed.

Variability of Plankton Communities

Zooplankton spatial patchiness has significant effects on the local rates of nutrient regeneration, phytoplankton grazing, and the feeding activity of planktivores (Pinel-

Alloul 1995). Zooplankton patchiness is a three-dimensional phenomenon, with both vertical and horizontal components. For example, both diel horizontal migration and diel vertical migration contribute to this patchiness and are hypothesized adaptations to predation pressure from zooplanktivores (Lauridsen and Buenk 1996, Lampert 1989).

However, other factors, including phytoplankton abundance and composition and metabolic factors may affect such migrations (Lampert 1989). Physical factors, such as currents and stratification patterns also can affect zooplankton (Pinel-Alloul 1995) and phytoplankton patchiness (Fogg 1991). Further, some phytoplankton can control their position in the water column by the use of flagella or gas vacuoles, allowing for vertical and horizontal movements (Fogg 1991). This patchy distribution has important implications for sampling regimes and the development of an IBI based on plankton.

Zooplankton and phytoplankton communities also vary in their composition and abundance over time, both on a seasonal (Sommer et al. 1986) and an interannual basis

(Kratz et al. 1987, Müller-Navarra et al. 1997). Thorough studies of seasonal succession of zooplankton and phytoplankton communities in north temperate lakes (Sommer et al.

1986) have provided important paradigms to general ecology (Lampert 1997). For

7 example, the Plankton Ecology Group (PEG) model of seasonal succession in the plankton describes 24 steps in the seasonal succession of an idealized lake (Sommer et al.

1986). This model incorporates both biotic and abiotic factors that influence the seasonal succession of the plankton, and has been applied to a variety of lakes. Abiotic factors, such as temperature, light, and nutrient availability, predominantly influence succession early and late in the year, whereas biotic interactions influence succession in the middle of the year. Although the sequence of events of seasonal succession generally remains the same from year-to-year, interannual variation is evident in the timing and magnitude of such events as the clear water phase (Müller-Navarra et al. 1997) and maximal abundance of zooplankton (Kratz et al. 1987). An understanding of these temporal dynamics of plankton communities is necessary for the development of an IBI based on plankton.

Spatial and temporal analyses of both multi-year (Evans and Sell 1983) and single year (Gannon 1975, Stockwell and Sprules 1995, Stockwell et al. 2002) zooplankton sampling programs have been conducted in the Great Lakes. Further, spatial and temporal analyses of plankton have been conducted at a multi-lake level (Barbiero and Tuchman 2001, Barbiero et al. 2001), multi-basin level (Stockwell and Sprules 1995,

Stockwell et al. 2002), and single basin level (Gannon 1975, Frost and Culver 2001). In

Lake Michigan, Gannon (1975) found that zooplankton distribution along a cross-lake transect was relatively uniform during fall, winter, and spring. However, in summer a number of taxa were more prevalent in the waters 0-18 km from the shore, while others were more prevalent in the open water. Further, both the zooplankton and phytoplankton

8 communities in all five of the Laurentian Great Lakes varied considerably in composition and abundance between the lakes during spring (March- early May) and summer surveys

(August- early September) in 1998 (Barbiero and Tuchman 2001, Barbiero et al. 2001).

Zooplankton and phytoplankton abundance and community composition also differered between the spring and summer.

Similarly, temporal and spatial patterns of abundance, biomass, and composition of zooplankton communities have been found to be variable in Lake Erie.

In a lakewide study during the early 1970’s, Watson (1976) found that crustacean zooplankton abundances increased from January up to June and July and declined afterward. Accompanying this abundance peak was a peak in crustacean zooplankton biomass (Watson 1976, Watson and Carpenter 1974). Taxonomically, cyclopoid copepods and cladocerans made up a larger percentage of the total crustacean zooplankton community during summer, than during winter, when more calanoid copepods were present. Frost and Culver (2001) found the same temporal pattern in zooplankton biomass to hold true in the western basin of Lake Erie during the mid-

1990’s, with highest biomasses occurring in late June and early July. Further, with respect to spatial distribution, they found zooplankton biomass in nearshore areas to be greater than offshore areas, which matches spatial patterns found by Watson (1974).

Also in the mid-1990’s, Stockwell and Sprules (1995) found considerable temporal and spatial variation in zooplankton biomasses (wet weight) based on lake-wide Optical

Plankton Counter data. Lowest lakewide mean zooplankton biomass values (463 µg/L) occurred in spring (May), highest biomass values (1366 µg/L) in summer (June), and

9 intermediate values (642 µg/L) were found in the fall (September). In spring, basin mean zooplankton biomass was greatest in the western basin (1401 µg/L), lowest in the eastern basin (157 µg/L) and intermediate in the central basin (536 µg/L). Further, summer mean basin zooplankton biomass remained lowest in the eastern basin (599 µg/L).

However, summer mean central basin zooplankton biomass (1748 µg/L) was greater than western basin mean biomass (1312 µg/L). Finally, fall mean zooplankton biomass was greater in the western basin (1172 µg/L) than in the eastern basin (557 µg/L) (central basin not sampled). Hence, great temporal and spatial variability exists in the zooplankton community of Lake Erie.

The phytoplankton community in Lake Erie has also been found to be quite variable. Munawar and Munawar (1976) found in 1970 that the western basin had the highest mean phytoplankton biomass (wet weight-5.3 g/m3), with the central basin having a lower mean phytoplankton biomass (3.2 g/m3), and the eastern basin having the lowest mean phytoplankton biomass (2.4 g/m3). Mean phytoplankton biomasses in the western and central basins were greatest in the spring (April- June), decreased in the summer

(July- September), and were least in the fall (October- December). However, central basin biomasses were greatest in summer and lowest in spring. Further, the percentage composition of major taxonomic groups of phytoplankton (i.e., Cyanophyta, Chlorophyta etc.) varied between basins and between seasons. Frost and Culver (2001) found a clear water phase with low phytoplankton biomass in mid-June throughout the western basin of

Lake Erie. Further, they found phytoplankton biomasses to be greatest in late July of

1996. Previous studies have also found great temporal and spatial variab ility in the phytoplankton community of Lake Erie.

Stressors that Affect Plankton Communities

To develop an Index of Biotic Integrity for plankton, the major anthropogenic stressors that affect zooplankton and phytoplankton must be identified. These incl ude pesticides, poly-chlorinated biphenyls (PCBs), metals, and nutrient enrichment.

Although a large body of research deals with the effects of nutrient enrichment on plankton communities, much less research has been conducted on the effects of the other three stressors on plankton communities.

Pesticides influence zooplankton and phytoplankton communities. Endosulfan

(Fernandez-Casalderrey et al. 1994), diazinon (Fernandez-Casalderrey et al. 1994), and lindane (Gliwicz and Sieniawska 1986, Hartgers et al. 1999) reduce filtering rate in

Daphnia. Endosulfan reduces zooplankton abundance, which leads to increases in phytoplankton abundance (Barry and Logan 1998). Endosulfan also shifts zooplankton community composition from a community containing cladocerans, calanoid copepods, and cyclopoid copepods to a community consisting of only cyclopoid copepods. High pyridaben concentrations (> 3.4 µg/ L, which is equivalent to 0.01% surface run off and

5.0% aerial drift from application of maximum labeled rate by aircraft) in microcosms cause a decline in zooplankton populations (rotifers and crustaceans) and an increase in

Chlorophyta, Pyrrhophyta, and Chrysophyta abundances (Rand et al. 2001).

The carbamate insecticide carbaryl has been thoroughly studied for its effects on zooplankton. Carbaryl (tradename: Sevin, among others) impairs growth and reproduction in Daphnia longicephala at concentrations > 0.32 µg/ L (Barry 1999). Low oxygen, Chaoborus kairomones, and carbaryl were found to interact synergistically reducing juvenile growth rate, size at maturity, clutch size, and neonate body size in

Daphnia (Hanazato and Dodson 1995). Carbaryl affects Daphnia behavior. Daphnia showed a spinning response to high levels of carbaryl (40 ppb), irritation response to low levels of carbaryl (1 ppb), and no change in behavior to kairomones of a zooplanktivorous insect (water from a tank containing the phantom midge Chaoborus) at

0 ppb carbaryl. Daphnia with spinning behavior were also more likely to be eaten by bluegill (Dodson et al. 1995). Carbaryl has greater effects on larger zooplankton, such as

Daphnia. In fact, large cladocerans (including Daphnia spp.) can have a 50% reduction in biomass, due to the addition of carbaryl to mesocosms (Havens 1993). Because carbaryl has a greater effect on large zooplankton, smaller zooplankton, such as copepods and rotifers, may increase in density and biomass (Hanazato 2001), directly because the smaller zooplankton are less susceptible to carbaryl and indirectly by the removal of large cladocerans, which are better competitors for phytoplankton. Carbaryl can thus affect zooplankton from the individual level to the community level.

Unfortunately, a number of problems exist with the use of data from pesticide exposures to construct a Planktonic Index of Biotic Integrity for Lake Erie. First of all, the work done has been conducted at the individual level for zooplankt on (i.e., LC50 and behavioral studies) or quantified physiological parameters (e.g., clearance rate), and

12 typically at pesticide concentrations 100-10,000 times higher than those found in large lakes. Second, most studies of pesticide effects on plankton communities have focused on zooplankton, with indirect effects on phytoplankton. Further, most of the studies of the effects of pesticides on plankton communities have been done at the mesocosm or microcosm scale, with few data on their effects on zooplankton and phytoplankton in lakes as large as Lake Erie (but see Swackhammer et al. (1998) who studied toxaphene).

Swackhammer at al. (1998) monitored toxaphene (a chlorinated pesticide mixture) in all five of the Laurentian Great Lakes and found concentrations ranging from 0.17 + 0.07 ng/L (Lake Ontario) to 1.12 + 0.18 ng/L (Lake Superior). Further, they found concentrations in the biota ranging from 51.3 + 30.4 ng/L in the phytoplankton to 2373 +

1454 ng/L in Lake Trout (Salvelinus namaycush), indicating that toxaphene bioaccumulates from lower to upper trophic levels in the Great Lakes. However, studies on this scale are few. Thus development and application of a Planktonic Index of Biotic

Integrity for Lake Erie based on pesticide effects on the plankton is limited.

The use of the response of planktonic communities to the effects of polychlorinated biphenyls (PCBs) or heavy metals to construct an Index of Biotic

Integrity for Lake Erie is equally problematic. Seasonal dynamics of levels of PCBs ha ve been measured in phytoplankton and zooplankton in the Great Lakes (Epplett et al. 2000,

Stapleton et al. 2002), and studies from other lakes have focused on bioaccumulation of

PCBs in phytoplankton and zooplankton (Berglund et al. 2000), and the effect of lake size (Paterson et al. 1998) and trophic status (Berglund et al. 2001, Larsson et al. 1998)

13 on PCB dynamics. However, little research has been conducted on the effect that PCBs have on zooplankton or phytoplankton biomass or community composition.

Similarly, a number of studies have examined heavy metal levels in lake zooplankton (Chen and Folt 2000, Chen et al. 2000), fluxes of trace metals to and from zooplankton in both the laboratory (Yu and Wang 2002) and the field (Twiss 1996), and the effects of heavy metals on molecular and population parameters of zooplankton

(Chen et al. 1999). However, few studies have examined the effect of heavy metals on zooplankton community composition (but see Stemberger and Chen 1998), particularly in the Great Lakes. Stemberger and Chen found significant relationships between zinc and mercury concentration in fish and zooplankton community structure (i.e. complexity, number of feeding links) in data from 38 northeastern U.S. lakes. However, few studies of this type have been conducted in the Great Lakes. In summary, not enough research has been conducted examining the effects of PCBs and heavy metals on phytoplankton or zooplankton community composition for their use in a planktonic IBI.

The other major stressor that impacts plankton communities is nutrient enrichment. In contrast to the stressors discussed above, effects of nutrient enrichment on both phytoplankton and zooplankton communities have been well documented.

Phosphorus limits the growth of phytoplankton in lakes (Schindler 1977). Further, total phosphorus is positively related to chlorophyll a, a surrogate for phytoplankton concentrations in lakes (Dillon and Rigler 1974), including the Laurentian Great Lakes

(Johengen et al. 1994). The eutrophication of lakes favors many cyanophyte taxa that compete better at high phosphorus concentrations (Schindler 1977, Smith 1979) or low

N:P ratios (Smith 1983). Cyanophytes have inhibitory and toxic effects on crustacean zooplankton, such as Daphnia (Lampert 1981, Lampert 1982, DeMott et al. 1991), and can affect crustacean zooplankton taxa differentially (Vaga et al. 1985, DeMott et al.

1991). Further, eutrophication can lead to low dissolved oxygen levels that can adversely affect zooplankters intolerant of low dissolved oxygen conditions. Decreases in the abundance of Limnocalanus macrurus, a large calanoid copepod, were found in Lake

Erie concomitant with eutrophication and low dissolved oxygen levels (Gannon and

Beeton 1971). Unlike pesticides, PCBs, and heavy metals, the extensive literature on the effects of nutrient enrichment on plankton communities supports the development of a P-

IBI.

Similarly, the use of zooplankton and phytoplankton as indicators of nutrient enrichment has been well studied. Gannon and Stemberger (1978) examined the utility of zooplankton communities as indicators of lake trophic condition (i.e. eutrophic, mesotrophic, oligotrophic). The ratio of calanoid copepod abundance to cladoceran plus cyclopoid copepod abundance led to distinguishing among different trophic conditions in

Lake Michigan, the Straits of Mackinac, and Lake Huron. Gannon and Stemberger also found that certain rotifer species assemblages were associated with eutrophic, mesotrophic, and oligotrophic conditions in Lake Huron and the Saginaw Bay.

Stemberger and Miller (1998) found that zooplankton assemblages could be used as indicators of the N:P ratio of lakes, due to differential requirements of zooplankters for these nutrients. Further, zooplankton composition could be used to infer changes in watershed characteristics and nutrient loading.

Phytoplankton has also been widely used as an indicator of nutrient conditions in lakes. Rawson (1956) reviewed the use of phytoplankton as indicators of lake trophic status. Measures such as the ratio of centric to pennate diatoms, and a number of other ratios of phytoplankton taxa were discussed as being indicative of trophic conditions.

Dixit et al. (1992) reviewed the many uses of diatoms in monitoring environmental change, including changes in lake trophic status, lake acidification, and climate regimes.

Because the effects of eutrophication on plankton communities and the use of plankton as indicators of trophic status have been so well studied, the responses of plankton to nutrient enrichment in Lake Erie will be the foundation for the construction of the P-IBI.

To contruct a Planktonic Index of Biotic Integrity for Lake Erie, I first reviewed relevant literature dealing with the variability of plankton communities and effects of major stressors on these communities (Chapter 1). I then examined the appropriate spatial and temporal scales to sample phytoplankton and zooplankton in Lake Erie and how these communities can measure changes in the Lake Erie ecosystem (Chapter 2). I then used data from monitoring programs to document the invasion of Cercopagis pengoi

(Chapter 3) and temporal and spatial changes in the distribution of Limnocalanus macrurus in Lake Erie (Chapter 4). Finally, I modified a general methodology for IBI development (Karr and Chu 1997) to specifically construct a planktonic IBI for Lake

Erie. Their steps, or project goals, as applied to the Planktonic IBI were 1) Classify environments to define homogenous sets, 2) Select measurable attributes that provide reliable and relevant signals, 3) Develop sampling protocols, 4) Devise the multimetric

16 index analytical procedures, and 5) Communicate results to citizens, policy makers, and other scientists (Chapter 5).

CHAPTER 2

CHARACTERIZING A CHANGING LAKE ERIE WITH PHYTOPLANKTON AND

ZOOPLANKTON

INTRODUCTION

Responding to problems in the 1960’s, the U.S. government took steps to protect the quality of water in the U.S. Legislation was enacted in 1972 (Water Pollution Control

Act Amendments of 1972) and in 1977 (Clean Water Act) that called for the restoration and maintenance of, not only the physical and chemical integrity of U.S. waters, but also the biological integrity of these waters (Karr 1991). Because of the explicit inclusion of biological integrity into this definition, physical and chemical measures of water were no longer sufficient. Thus, a number of biotic measures of water quality were subsequently developed (e.g., Index of Biotic Integrity (IBI), Invertebrate Community Index (ICI)).

Karr (1981) first devised an Index of Biological Integrity (IBI) to measure biological integrity in a stream, using fish as indicator species. In general, biotic integrity is an ecosystem property that can be defined as “the capability of supporting and maintaining a balanced, integrated, adaptive community of organisms having a species composition, diversity, and functional organization comparable to that of natural habitat of the region”

(Karr and Dudley 1981). Indices of biotic integrity provide measures of this ecosystem property.

Water is essential to human survival. However, less than 5% of the earth’s water is fresh, and most of this water is in ice caps, glaciers, and groundwater (Beeton 2002).

Thus the Laurentian Great Lakes represent the world’s largest (accessible) freshwater source (Bolsenga and Herdendorf 1993). Further, Naiman et al. (1995) listed research priorities for freshwater resources and their management to be ecological restoration and rehabilitation and ecosystem goods and services. A number of Beneficial Use

Impairments (BUIs) to ecosystem goods and services have been identified for the

Laurentian Great Lakes (Table 1), including Lake Erie (Hartig et al. 1997). BUIs describe impairments to the human utilization of Great Lakes water resources, and directly affect human health (e.g., restrictions on drinking water), ecosystem function

(e.g., degradation of benthos), or both (e.g., eutrophication or undesirable algae).

Characteristics of the plankton communities impact many Beneficial Use Impairments

(BUIs). In fact many of the BUIs deal with plankton communities explicitly (i.e., degradation of phytoplankton and zooplankton populations) or implicitly (i.e., taste and odor problems in water, often due to algae) (Table 1). Although various indices now exist to measure Lake Erie water quality and reflect the BUIs, a majority of these focus on nearshore areas (e.g., contaminated harbor sediments, beach bacterial pollution, point source pollution) (Ohio Lake Erie Commission 1998). Few biological measures of water quality have been applied to the offshore waters of Lake Erie due to logistical issues in sampling these areas; however, understanding these offshore waters, which

19 contribute a large portion of Lake Erie’s area and volume, is critical to any management plan for Lake Erie. By using the plankton to construct an IBI, this situation can be remedied. Plankton is sensitive to environmental changes (Schindler 1987), a necessary component of a useful monitoring program. Further, plankton is inexpensive to collect, another requirement of useful monitoring programs (Schindler 1987). In addition, samples can be stored for long periods and do not take up large amounts of space and historical samples can be analyzed and compared with current samples. However, plankton monitoring is not without drawbacks. Plankton taxonomic enumeration is time intensive and thus methods to expedite this process are highly advantageous. Below, I examine one possible way to speed up the estimation of cyanophyte and crustacean zooplankton biomass.

In order to recognize ecological restoration and rehabilitation targets of our freshwaters, human-induced degradation of systems must be recognized (Naiman et al.

1995). In the Laurentian Great Lakes, the United States Environmental Protection

Agency and Environment Canada have programs designed to assess the state of Great

Lakes waters (Neilson et al. 2003) with regards to degradation. These agencies monitor a variety of indicators, including pressure indicators (e.g., nutrient and toxic chemical concentrations), which describe natural and human processes that impact or threaten environmental quality, and state indicators (e.g., organismal abundance and diversity, fish and wildlife health), which directly reflect the biological, chemical, physical variables and ecological function of the environment (Neilson et al. 2003). These indicators are used to determine the U.S. and Canadian compliance with the Great Lakes Water Quality

Agreement and are reported at the State of the Lakes Ecosystem Conference (SOLEC)

(Neilson et al. 2003). Further, the Ohio Lake Erie Commission has established the Lake

Erie Quality Index (LEQI) with the goal of establishing benchmarks for monitoring and evaluating progress in the restoration of Lake Erie (Ohio Lake Erie Commission 1998).

To achieve the goals of both SOLEC and the LEQI, development of indicators of biological integrity is essential.

Lake Erie is a changing ecosystem and these changes have direct ties to the plankton and BUIs. Physically, Lake Erie water levels have fluctuated greatly over the past century (Quinn 2002, Changnon 2004). Chemically, phosphorus loading to Lake

Erie declined between 1970 and mid-1980’s (Dolan 1993), and summer total phosphorus concentrations decreased significantly in the three basins of Lake Erie between the periods 1968-1972 and 1994-1996 (Charlton et al. 1999). With regard to biota, declines in phytoplankton biomass were evident by the late 1970’s (Nicholls et al. 1977) and continued in the mid-1980’s (Makarewicz et al. 1993a). Further, a fish community indicative of more oligotrophic conditions returned to Lake Erie during the 1980’s and first half of the 1990’s (Ludsin et al. 2001). The numerous, recent biotic invasions of

Lake Erie also typifies the changing Lake Erie ecosystem (Mills et al. 1994). The crustacean zooplankter Bythotrephes longimanus invaded Lake Erie in the mid 1980’s

(Bur et al.1986), while the zebra mussel (Dreissena polymorpha) and quagga mussel (D. bugensis) invaded Lake Erie shortly thereafter (Hebert et al. 1989, May and Marsden

1992). Further, dressenids may affect nutrient cycling in Lake Erie, due both to consumptive (MacIsaac et al. 1992) and excretory (Arnott and Vanni 1996) processes.

Thus, physical, chemical, and biotic changes and their interactions reveal that Lake Erie is in a state of perpetual change, which has consequences for the plankton community and BUIs.

During this chapter, I ask two major questions relating to Lake Erie plankton that have impacts on the development of a planktonic Index of Biotic Integrity (P-IBI). 1)

What is an optimal sampling design for Lake Erie plankton (one that captures natural variability and informs management decisions, while remaining practical) and 2) How have nutrient stressors on the lake (namely phosphorus) affected Lake Erie plankton communities over the long term? Therefore, my goals are to 1) use information regarding natural temporal and spatial variability of plankton to determine appropriate spatial and temporal scales for a sampling program for Lake Erie plankton, 2) determine if a rapid assessment tool can serve as a surrogate for more time-consuming measures of plankton biomass, and 3) characterize changes regarding Lake Erie plankton biomass during the time period 1970-2002, with respect to changes in phosphorus loading. With information on both natural variability and variability due to sampling and long-term changes in Lake Erie plankton, a more informed plankton sampling strategy can be developed, which can aid in the development of the P-IBI for Lake Erie.

MATERIALS AND METHODS

First, to determine an optimal sampling design for Lake Erie plankton, I analyzed a very large phytoplankton and zooplankton biomass data set (1996-2002) with respect to temporal and spatial variability to determine the minimal number of samples required to

22 characterize the lake. Second, I analyzed the ability of a rapid method for assessing cyanophyte and crustacean zooplankton biomass to serve as a surrogate for more time- intensive enumeration and biomass calculations. Third, I analyzed the effects of phosphorus-loading changes on zooplankton and phytoplankton biomass during 1970 through 2002.

Optimal Sampling Design

Sampling Methods- 1996-2002

1996-2002 zooplankton and phytoplankton samples were collected in Lake Erie by the Ohio Department of Natural Resouces (ODNR) and Environment Canada’s

National Water Research Institute (NWRI) as part of the Ohio State University’s Lake

Erie Plankton Abundance Study (LEPAS). This ongoing study (1995-present) seeks to monitor long-term trends in lower trophic level interactions that affect the recruitment of fish species in Lake Erie. Sites in all three basins of Lake Erie were sampled during this study (Figure 1) for phytoplankton and zooplankton (Table 2). General methods for the sampling and analysis of plankton samples can be found in Frost and Culver (2001).

More specific methods follow (APPENDIX C). The LEPAS database is stored on

Microsoft Access 2000 (Microsoft Corporation) and contains phytoplankton and zooplankton abundance and biomass data, as well as nutrient and site data. Lake Erie maps (Figures 1,2) were created using Surfer Version 6.01 (Golden Software

Incorporated 1993-95)

Historical Plankton Datasets

Methods Comparison

To facilitate comparisons of methods between various sources of plankton abundance and biomass data, I made relevant comparisons among sampling dates, sampling frequency, numbers of sites, sampling methods, enumeration methods, and biomass calculation methods for both phytoplankton and zooplankton in Lake Erie

(1970-2002) (Table 4). More detailed descriptions follow.

Phytoplankton

In 1970, phytoplankton samples were collected from April to December at 4-week intervals in all three basins of Lake Erie from 25 different stations (Munawar and

Munawar 1976). Samples were collected in a VanDorn bottle from 1- and 5m depth, mixed, and then preserved with Lugol’s solution. Another subsample was used to identify living, motile phytoflagellates (Munawar and Munawar 1976). For 1978, samples were collected on 9 cruises from May to November in the three basins of Lake

Erie at 87 stations (Devault and Rockwell 1986). When the lake was thermal ly stratified, samples were collected in an 8 L opaque Niskin bottle from 1 meter, 1 meter above the metalimnion, 1 meter above the thermocline, 1 meter above the hypolimnion, and 1 meter above the bottom. Samples were collected at 1 meter, mid -depth, and 1 meter above the bottom in unstratified situations. Phytoplankton samples contained 500 mL of water and were preserved with 10 mL of modified Lugol’s solution (Devault and Rockwell 1986).

From 1983-1987, phytoplankton was collected at 21 different stations in the three basins of Lake Erie during 33 cruises during the spring, summer, and autumn. An 8 L PVC

Niskin bottle mounted on a General Oceanics Rosette sampler was used to collect phytoplankton. For deeper waters, phytoplankton samples were composed of equal aliquots of water from depths of 1, 5, 10, and 20m. In the western basin, samples were taken from 1m, mid-depth, and 1m above bottom. One-liter combined samples were preserved with Lugol’s solution, with formalin added later (Makarewicz 1993a ). From

1996-2002, phytoplankton was collected from 30 to 80 stations throughout Lake Erie’s three basins typically from late April/early May until late September/ early October.

Phytoplankton water samples were obtained with an integrated water sampler (2.5cm diameter tube) from the surface to twice the Secchi depth (volume of sample=π*1.252

* twice Secchi depth). Collected water was poured into a plastic bucket from which a

500mL sample was taken. Each sample was preserved in a canning jar with Lu gol’s solution (Frost and Culver 2001).

Phytoplankton enumeration for all years generally followed Utermöhl (1958).

For the 1970 samples, aliquots of 5-25mL, depending on phytoplankton density, were settled and examined with an inverted microscope (Wild Heerbrugg M40, phase contrast). Common netplankton species were counted in two transects under 300x magnification. Less common and rare species were enumerated in the entire chamber at

300x. Finally, nannoplankton and microalgae were counted in two transects under 300x magnification. In each sample at least 300 units or entities were enumerated, where each colony was treated as a unit. Diatoms were identified by making permanent slides with a

HYRAX mounting medium (Munawar and Munawar 1976). For 1978, 10 mL of sample were settled and phytoplankton were identified and enumerated using an inverted microscope (Leitz Ultralux) (Devault and Rockwell 1986). Organisms greater than 10

µm were enumerated and identified at 250X, while organisms less than 10 µm were enumerated and identified at 500X, with both methods enumerating 2 perpendicular strips

13.6 mm long. Phytoplankton densities were then expressed in cells/mL. The dimensions of at least 10 cells, but usually many more, were measured in surface samples in selected samples for cell volume calculations (Devault and Rockwell 1986). For 1983-1987 samples phytoplankton (other than diatoms) were identified and enumerated at a magnification of 500x. Diatoms were identified at 1,250x by making permanent slides with a HYRAX mounting medium (Makarewicz 1993a). When present, at least 10 specimens of each species were measured for cell volume calculations. For 1996-2002, sample jars were inverted 25 times to fully mix samples and 250mL from each jar was poured into a graduated cylinder, and allowed to settle for 3 days in a dark-chamber.

Each sample was then concentrated down to 30mL by siphoning off liquid from the top and transferring the remaining sample to a 36.97mL vial. Mixed subsamples of 3- to

5mL were obtained from the concentrated samples and placed into a counting chamber.

Exact volumes were determined by weighing the counting chamber. All phytoplankton genera were identified and counted using a Wild inverted microscope at 400x. Repeated transects were counted until 100 algal units (cells, filaments, or colonies) of the most common taxa were recorded. In all samples, however, all algal units in at least two transects were counted for each sample even if 100 algal units of the most common taxon

26 was enumerated before the full two transects were completed. Dimensional measurements were recorded for the first 20 algal units for each genus enumerated. For filamentous algal taxa, however, all filament lengths were measured, summed, and recorded as the total filament length for each taxon (Frost and Culver 2001).

Total biomass of phytoplankton samples was determined in all years by summing species-specific total biomass over all species present in a particular sample. Cell volume was computed for each species present using the mean algal dimensions for each species in a sample. These average dimensions were then used in volumetric equations that best described the shape of each species. For colonies, the mean number of cells per colony was calculated and multiplied by the average volume per cell to determine volume per colony. Subsequently, volumes were converted to biomass assuming the specific gravity of phytoplankton to be 1.0 (Munawar and Munawar 1976, Devault and Rockwell 1986,

Makarewicz 1993a, Frost and Culver 2001). Consequently, all reported phytoplankton biomasses are wet weights (mg/L).

Phytoplankton total seasonal average biomass data (SABp) by basin were available in the literature from 1970 (Munawar and Munawar 1976), 1978 (Devault and

Rockwell 1986), and 1983-87 (Makarewicz 1993a). Sampling, enumeration, and total biomass calculation methods were broadly similar to those used in LEPAS because the analysis of SABp was based upon the methods of the previously published studies in order to maximize direct comparability. The most consequential difference between the published and new datasets is the timing of sampling and this difference may lead to changes in the interpretation and comparability of the computed SABp among samples

(Conroy et al. 2004a, in review). Samples from 1970 were collected monthly from 25 stations distributed throughout Lake Erie’s three basins from April to December. SABp for these data comes from Munawar and Munawar’s (1976) Table 7. Samples from 1978 were collected on nine cruises from May to November in the three basins of Lake Erie at

87 stations (Devault and Rockwell 1986). Samples from 1983-87 were collected from between 11 and 20 stations sampled from 21 total stations distributed throughout the lake

(Makarewicz 1993a). Data reported and used as the SABp estimate were from April and

August. SABp comes directly as reported in Table 10 of Makarewicz (1993a).

Zooplankton

In 1970, zooplankton samples were collected at 30 stations located in all three basins of Lake Erie on 10 cruises from April through December. Samples were collected with a 64µm mesh, 0.4m diameter unmetered net, using single vertical hauls from 2m above the bottom (or 50m from surface). These samples were preserved in 4% formalin

(Watson 1976). During 1974-1975, zooplankton were collected at 19 stations in the western basin of Lake Erie using a 64µm mesh, 0.5m diameter metered net between April and December (Weisgerber 2000). Although this net was metered, meter data were unavailable, thus calibration tows indicating that the net was 70% efficient were used for calculating zooplankton densities (Weisgerber 2000). Vertical tows were made starting

1m above the bottom. Zooplankton were anesthetized with club soda and samples were preserved with 5% MgCO3 buffered formalin. From 1983-1987, zooplankton was collected at 21 stations in the three basins of Lake Erie during 33 cruises during the

28 spring, summer, and autumn. Samples were collected with a 62µm mesh, 0.5m diameter, metered (Kahl flow meter, Model OOSWA200) net, using vertical tows either from 20m to the surface, or if the water column was less than 20m deep, from 1m above the bottom to the surface and were preserved in 5% formalin (Makarewicz 1993b). From 1996-

2002, zooplankton was collected at 30-80 sites in all three basins of Lake Erie typically from late April/early May until late September/ early October from 1996-2002.

Zooplankton sampling and enumeration methodology is published in detail elsewhere

(Frost and Culver 2001). Briefly, vertical tow samples were collected using a front- weighted zooplankton net (0.5m diameter, 64µm mesh) fitted with a General Oceanics

2030R model flow meter and 500mL jar. The net was lowered with the open end pointing downward until the 2kg weight fastened to the front bridle by a 1m line hit bottom. The net was then retrieved, allowing the water column to be sampled both as the net was lowered and as it was pulled up, while avoiding collecting mud from the bottom.

Samples were then concentrated and preserved with a 4% sugar formaldehyde solution

(Haney and Hall 1973).

For 1970 samples, 1mL subsamples were analyzed under inverted microscope until 200 individuals of each taxon were enumerated (Watson 1976). Numbers/m3 were calculated from the percent of sample counted, assuming 100% net efficiency, tow depth, and net diameter (Watson and Carpenter 1974). Adult calanoid and cyclopoid copepods were identified to species and sex, while nauplii were enumerated but not identified to species. Cladocerans were identified to species and sex, when possible (Watson 1976).

Length measurements of 10-50 individuals (based on percent abundance) and egg

29 densities were obtained at a later time (Bean 1980). The number of eggs per female of a taxon was determined by counting the number of eggs and dividing by the number of females to obtain eggs per female. This number was multiplied by the number of females/m3 to give the eggs/m3 for each taxon (Bean 1980). For 1974-1975 samples, crustacean zooplankton were enumerated using a Wild M5 dissecting microscope. Sub - sampling was conducted to assure a count of approximately 200 of a single adult crustacean taxon was made for each sample. At a later time, length-frequency measurements (±0.05mm) were performed on 20 individuals of each crustacean taxon

(Weisgerber 2000). For 1983-1987 samples, enumeration followed Gannon (1971).

Adult calanoid and cyclopoid copepods were identified to species, as were cladocerans.

Copepodites were identified as cyclopoid or calanoid, while nauplii were only identified as copepods (Makarewicz 1993b). For samples taken during 1996-2002, each sample was diluted to a known volume, typically from 500mL to 3000mL unless the sample contained an extremely small or large amount of zooplankton, requiring lower or larger dilution volumes. After dilution, all zooplankton in at least two subsamples of 5- to

10mL were identified and enumerated. All enumeration was done using a Wild M5A dissecting microscope fitted with a calibrated ocular micrometer for body measurements

(±0.05mm). Cladocerans and copepods were identified to species and sex while rotifers were identified to genus. Additional subsamples were analyzed until at least 100 individuals of the most common taxon were recorded.

For 1970, 1974-1975, and 1996-2002 samples, zooplankton biomass was calculated by determining the average individual biomass for each species counted and

30 multiplying by the number of individuals per cubic meter. Average individual biomass was determined using length-weight regressions developed by Culver et al. (1985) and the lengths measured from historical samples (1970 and 1974-1975) or during sample enumeration (1996-2002). Species-specific regression equations converted from length

(in mm) to dry-weight biomass (µg) values for each crustacean zooplankton taxon

(including eggs). Species-specific total biomasses were summed over all taxa giving the total crustacean zooplankton biomass at a given sampling site, for a given date.

Similarly, total zooplankton biomass for samples from 1983-1987 was also determined using average length of individuals for each crustacean species found. At least 20 individuals were measured (Makarewicz 1993b); adult copepod and cladoceran lengths were converted to dry weight using length-weight regressions (Downing and Rigler 1984,

Makarewicz and Likens 1979), while the dry weight of copepod nauplii followed

Hawkins and Evans (1979). These individual taxon biomasses were summed over all taxa giving a total crustacean biomass for each site, on each date.

Zooplankton total crustacean seasonal average biomass (SABz) measurements were available from the literature from 1970 (abundance data from Watson 1976, Watson and Carpenter 1974; biomass calculations by Bean 1980), 1974-75 (abundance data from

Center for Lake Erie Research; biomass calculations by Weisgerber 2000), and 1984-87

(Makarewicz 1993b) (Conroy et al. 2004a, in review). Field sampling, en umeration, and biomass calculations in all historical studies were similar to those in LEPAS because the current study used these historical studies as examples for methods. Sampling in all studies used a 62-64µm mesh net, while abundance (numbers/m3) of taxa was determined

31 using net efficiency, tow depth, and net diameter. Abundance calculations were more precise in later studies (1984-87 and 1996-2002) due to the use of a flow meter on the mouth of the net. Enumeration of samples, measurement of lengths for biomass calculations, and conversion of lengths to weights using published regression equations

(Culver et al. (1985) for samples from 1970, 1974-75, and 1996-2002; Downing and

Rigler (1984) and Makarewicz and Likens (1979) for 1984 -87 samples) for individual taxa were similar across all years. Sampling dates for 1970 (Watson 1976) and 1984-87

(Makarewicz 1993b) samples are identical to those given for phytoplankton (see above).

Samples from 1974-75 were collected only in the western basin at 19 stations from April to December (Weisgerber 2000). SABz for the 1970 samples was calculated by averaging the data by basin from Bean’s (1980) Table 16. SABz for the 1974-75 samples was calculated using the data that contributed to Weisgerber’s (2000) Table 1. SABz for the 1984-87 samples was calculated from Makarewicz’s (1993b) Table 2 by subtracting the “Rotifera” biomass column from the “Mean Abundance” biomass column.

Comparison of Sampling Regimens: Temporal and Spatial Variability

To test more extensive (i.e., more sites sampled, more frequent sampling) temporal and spatial plankton sampling regimens versus less intensive regimens (1996-

2002), regression analyses and paired t-tests were performed on phytoplankton and zooplankton mean biomasses using MINITAB 13.0 (Minitab Incorporated 2000).

Bonferroni corrections were made on alpha levels for the temporal, spatial, and variability analyses (below). Following Cabin and Mitchell (2000), the explicit rationale

32 and process by which I made Bonferroni corrections follow. Because hypotheses were tested using statistical tests on individual data sets within each year, Bonferroni corrections were made separately for each statistical test in each independent data set for the number of comparisons made. For example, because six temporal comparisons (May,

June, July, August, September, October) made with regression analyses for the 1996 phytoplankton biomass data set, my Bonferroni corrected α value was 0.05/ 6 comparisons or 0.0083.

For temporal analyses, mean biomass was calculated by taxon for phytoplankton

(i.e., Cyanophyta, Pyrrophyta, Cryptophyta, Chlorophyta, and Chrysophyta) and zooplankton (i.e., Cladocera, Cyclopoida, and Calanoida). Further, the mean biomass (by taxon) for an individual month was compared with the mean biomass for the entire sampling season (typically May-October) across all three basins of Lake Erie.

Percentage mean differences were calculated from paired t-tests using the mean biomass from the entire season minus the mean biomass from a given month and then standardized by dividing by the mean biomass for the entire season. This measure gave the average percentage that a month overestimated or underestimated mean phytoplankton or zooplankton biomass.

For the spatial analyses, mean biomass of phytoplankton and zooplankton biomass were compared between a full suite of sites and reduced suites in the western basin of Lake Erie (Table 3, Figure 2). Reduced suites of sites included at least 5 sites, but usually 10, 15, and 20 sites, with the addition of 30 and 40 sites in 1996 (Table 3).

Reduced sites were chosen to spatially reflect the full suite of sites. When possible, the

33 same sites were used in a reduced suite across years. However, when sites sampled differed between years, spatially close sites (<10 km apart) were utilized (e.g., in 1997 site 4 was substituted for site 3 (which was sampled during 1996 and 1998-2002)).

Percentage mean differences were calculated in the same manner as the temporal analyses.

To determine whether a more inshore site (1279, Figure 1) could serve as a surrogate for a mid-basin offshore site (84, Figure 1) of the same depth (approximately

20 m) in the central basin of Lake Erie, phytoplankton and zooplankton biomasses were compared at the taxonomic level (same resolution as above) across 1997-2000 using t- tests. I conducted these analyses because mid-basin sites were not sampled in 1996 and thus validation of the use of more inshore sites as surrogates for mid -basin sites was necessary.

Sources of Variability

Lake Erie phytoplankton and zooplankton biomass data were analyzed across all samples taken during each year (1996-2002), by month, and by basin. To determine the sources of variability in the sampling of phytoplankton and zooplankton biomass,

ANOVAs were conducted using a General Linear Model of Yijk = µ + Bi + Sj:i + Mk + eijk, where Yijk = biomass, µ = grand mean, Bi = effect of basin, Sj:I = effect of site within basin, and Mk = effect of month. Phytoplankton biomass data were analyzed as total phytoplankton, at the functional level of edible and inedible phytoplankton, and at the taxonomic level of class (e.g., Chlorophyta). Zooplankton biomass data were analyzed as

34 total zooplankton (rotifers + crustaceans + dreissenid ve liger larvae), crustacean zooplankton, and at the taxonomic level (Cladocera, Cyclopoida, and Calanoida).

Percentage of variance explained by the three effects and the unexplained (residual) variance were calculated using the effect (basin, site (basin), and month) or error

(unexplained) sequential sum of squares divided by the total sum of squares and multiplying by 100. Evans and Sell (1983) and Rusak et al. (2002) previously used this technique in zooplankton studies to determine the relative variance contributed by each of the ANOVA components.

Volumetric Index of the Plankton (VIP) Analyses

Plankton samples were collected in 1998 in all three basins of Lake Erie with a

64-µm plankton net fitted with a General Oceanics 2030R flowmeter and preserved in sugar formalin (Haney and Hall 1973). Samples were taken to the lab for analyses.

Zooplankton and phytoplankton biomasses were estimated using standard enumeration techniques (APPENDIX C). These results were compared with the quicker VIP technique. All VIP lab analyses were performed by an undergraduate student that had only 1 week of previous experience with aquatic biology. Samples were poured into a

100-mL graduated cylinder, which was then capped with plastic film, inverted 3-5 times, and then allowed to settle for at least 24 h. After 24 h the two different layers that represented phytoplankton and zooplankton were measured to the nearest mL. Due to density and sinking rate differences, the zooplankton settled to the bottom of the cylinder first, with much of the phytoplankton then settling on top of the zooplankton. The

35 zooplankton layer was lighter in color (light brown) than the phytoplankton layer, which typically was a blue-green color when cyanophytes were present in high densities. The measured volumes of phytoplankton and zooplankton were then standardized by dividing them by the amount of lakewater originally sampled, which could be obtained from flowmeter readings. Regression analyses then were conducted to determine whether these volumetric estimates were correlated with biomass estimates that had previously been calculated from enumeration of zooplankton and phytoplankton. All standardized volumes and biomass estimates were log 10 transformed to adjust for high variability in the data. Further, the regressions between cyanophyte biomass and standardized phytoplankton biomass were calculated only for cases when measured phytoplankton volumes in VIP samples and cyanophyte biomass by enumeration were greater than zero.

Long Term Trend Analyses of Phosphorus Loading and Plankton Biomass

External Phosphorus Loading Estimation

Phosphorus loading calculations differed depending on the source of the loading.

To estimate loads from point sources, the following calculations can be used to report annual average point source phosphorus loads (Dolan 1993):

∑C Q Loading = i i n

th where: C i is the average total phosphorus effluent concentration for the i

month,

th Qi is the mean effluent flow for the i month,

n is the number of months of monitoring.

Calculations for point sources are performed on a “per pipe” basis and the estimates are summed (for multi-pipe facilities) to provide loads on a “per facility” basis. For monitored tributaries, the Stratified Beale’s Ratio Estimator (Beale 1962, Tin 1965,

Dolan et al. 1981) was used. Daily tributary loads were calculated on a annual basis for each tributary and then these data were stratified into one or more stratum(a) depending on the nature of the flow and concentration relationship within each stratum. For unmonitored tributaries, a unit area load (UAL) was estimated from nearby monitored tributaries and applied to the unmonitored basin area (Rathke and McCrae 1989). For atmospheric loadings, the flux of phosphorus in units of mass per area was estimated from precipitation collectors and applied to the lake area that the collector represents

(Rathke and McCrae 1989).

Statistical Analyses

Piecewise linear or “breakpoint” regression has been identified as a valuable tool for detecting thresholds in ecology (Toms and Lesperance 2003). In this study, I wished to determine in which year trend changes (i.e., when change in directionality of trend occurred) in plankton biomass occurred. Therefore, I used breakpoint regression, a nonlinear regression technique, to identify years during which trend changes occurred in zooplankton and phytoplankton biomass. Individuals performing previous plankton

(Winkler et al. 2003) and algal studies (Dodds et al. 2002) also have used this technique to determine thresholds. Breakpoint analyses were conducted using a quasi-Newton

37 search method with Statistica 6.0 software (StatSoft 1998). All years for which there were plankton biomass data were tested as possible breakpoint years. Following Dodds et al. (2002), the year that explained the greatest proportion of the variance in the plankton biomass data (i.e., the highest r2 value in the breakpoint analysis) was identified as the breakpoint year. After identifying the breakpoint year, two linear regressions were constructed in Minitab Version 13.1 (Minitab 2000) following the methodology of

Bascompte and Rodriguez (2001). The pre-breakpoint regression included all the years up to and including the breakpoint year, while the post-breakpoint regression included all years after the breakpoint year (not including the breakpoint year).

Plankton Biomass versus External Phosphorus Loading

To test the hypothesis that increased phosphorus loading to Lake Erie may cause increases in phytoplankton and zooplankton biomass, regression analyses in Minitab

Version 13.1 (Minitab 2000) were used to determine trends in phosphorus loading vs. plankton biomass (phytoplankton and zooplankton) for each of Lake Erie’s three basins.

RESULTS

Optimal Sampling Design

Comparison of Sampling Regimens: Temporal Variability

Monthly mean phytoplankton biomass (by taxon) was quite temporally variable and was significantly correlated with the entire season mean in only 6 out of 40 comparisons (Table 5). Further, the significant months (i.e., months that predicted entire

38 season biomass) differed between years (May in 1996 and 1997, June in 2000, July in

1997, August in 1998, and September in 1998) (Table 5). The results of the paired t-tests indicate that entire season and monthly phytoplankton biomass differed significantly only once (total phytoplankton in October, p = 0.006, Table 5). However, the absolute percentage mean differences were frequently >50% (11 out of 40 comparisons), and differences were as high as 251.15% (Table 5).

Monthly mean zooplankton biomass (by taxon) was also quite temporally variable and was significantly correlated in even fewer comparisons than phytoplankton biomass

(2 out of 40, June of 1996 and August of 2002) (Table 5). Further, results from the paired t-tests indicate that entire season and monthly zooplankton biomass did not significantly differ in any of the comparisons; however, almost half of the comparisons

(19 out of 40) had absolute percentage mean differences >50%, with differences as high as 122.39% (Table 5).

Comparison of Sampling Regimens: Spatial Variability

For phytoplankton biomass, reduced suites of sites were successful at describing the full suite of sites in many cases. Mean phytoplankton biomass at a full suite of sites was significantly linearly related to mean phytoplankton biomass at a reduced suite of 40 sites (1996), 30 sites (1996), 20 sites (1996-1998), 15 sites (1996-1998), 10 sites (1996-

2002) and 5 sites (1996-1999 and 2001-2002) (Table 6). For paired t-tests, phytoplankton mean biomass at the full suite of sites and at reduced suites of sites did not differ significantly in any of the comparisons (Table 6). Further, almost all comparisons

(19 out of 21) had absolute percentage mean differences <25%, with differences as low as 0.32% (Table 6).

For zooplankton biomass, reduced suites of sites were also successful at describing the full suite of sites in many cases. Mean zooplankton biomass at a full suite of sites was significantly linearly related to mean zooplankton biomass at a reduced suite of 40 sites (1996), 20 sites (1997 and 1998), 15 sites (1996-1998), 10 sites (1996, 1999, and 2002), and 5 sites (1998, 1999, 2001, and 2002) (Table 6). Like phytoplankton mean biomass, zooplankton mean biomass at the full suite of sites and at reduced suites of sites did not differ significantly in any of the paired t-tests (Table 6). Further, nearly all comparisons (19 out of 21) had absolute mean differences <25%, with differences as low as 0.10% (Table 6).

Neither phytoplankton nor zooplankton biomasses differed (p > 0.10 for all comparisons) between an inshore site (1279) and a mid-basin offshore site (84) of the same depth (20 m) in the central basin of Lake Erie (1997-2000). Thus a more inshore site could serve as an effective surrogate for a mid-basin site, if they were both at the same depth.

Sources of variability

From phytoplankton biomass ANOVAs, significant effects varied among groups and years, indicating that processes affecting phytoplankton biomass differened at spatial, temporal, and taxonomic levels. The effect of basin was significant for Chlorophyta

(2000), Cryptophyta (1997-2001), edible phytoplankton (2000), and total phytoplankton

(2000) (Table 7). The site within basin effect was significant for Chrysophyta (2001),

Cryptophyta (1997), edible phytoplankton in (2001), and total phytoplankton (2001)

(Table 7). The effect of month was significant for Chrysophyta (1996 and 1997),

Cryptophyta (1996 and 2000), Cyanophyta (1996 and 1998), edible phytoplankton (1996 and 1997), inedible phytoplankton (1996 and 1998), and total phytoplankton (1996 and

1997) (Table 7). Of these three effects, site within basin typically explained most of the variance in the phytoplankton biomass comparisons (n = 56) (mean + standard error % =

22.92 + 1.19%), followed by month (6.94 + 0.69%), and basin (n = 4.00 + 0.54%). Thus the total residual or unexplained variance was nearly twice (66.19 + 1.22%) the variance explained by these three effects.

Like phytoplankton analyses, significant effects on zooplankton biomass varied among groups and years, again indicating that processes affecting zoolankton biomass differed at spatial, temporal, and taxonomic levels. The effect of basin was significant for Cyclopoida (1998, 2000, and 2001), Calanoida (2001 and 2002), Cladocera (1996,

1998, 2000, and 2001), crustacean zooplankton in (1998, 2000, and 2001), and total zooplankton in (1996, 1998-2002) (Table 7). The site within basin effect on zooplankton biomass was significant for Cyclopoida (1996, 1998, and 2002), Calanoida (1996, 1999,

2001, and 2002), Cladocera (2002), crustacean zooplankton (1996 and 2002), and total zooplankton (2002) (Table 7). The effect of month on zooplankton biomass was significant for Cyclopoida (1996-2002), Calanoida (1996 and 1998), and Cladocera

(1996-2002), and crustacean zooplankton (1996-2002), and total zooplankton (1996,

1998-2001) (Table 7). These three effects followed the same pattern in the

41 phytoplankton analyses, with site within basin typically explaining the greatest percentage of variance in the zooplankton biomass comp arisons (n = 35) (30.10 +

3.07%), followed by month (14.31 + 1.57%), and basin (6.76 + 1.08%). The total residual or unexplained variance was about equal (48.83 + 2.50%) to the variance explained by these three effects, and thus these three effects explained more of the variance in zooplankton biomass (approximately 50%) than phytoplankton biomass

(approximately 33%).

Volumetric Index of the Plankton (VIP) Analyses

The volumetric estimates for zooplankton and phytoplankton varied in their strength of correlation with biomass estimates. Log10 standardized zooplankton volume was significantly (p < 0.001) correlated with log crustacean biomass in 1998 (Figure 3a), but only explained approximately one-third of the variability in the data during 1998 (r2 =

0.312). Further, log standardized cyanophyte volume was significantly correlated with cyanophyte biomasss in 1998 (p = 0.001) (Figure 3b), with the regression equation explaining only one-fifth of the variability in the data (r 2 = 0.206).

Long Term Trend Analyses of Phosphorus Loading and Plankton Biomass

Historic Plankton Datasets

Although the methods used in each of the datasets (Table 4) were similar (e.g., the

Utermöhl technique is used for phytoplankton enumeration, length-weight regressions are used for zooplankton biomass estimates), differences among years may affect the results

42 and conclusions drawn across studies. First, the sampling dates for Makarewicz (1993 a,b) were restricted to only 2 months, whereas other studies sampled across at least 4 consecutive months. Average seasonal biomass would be affected if biomass during either of those months not sampled differed from the mean biomass that would have been obtained across a longer time span (e.g., April-September). Another important difference among the studies is the lack of meter data for the 1970 and 1974 -1975 zooplankton hauls. For the 1970 zooplankton hauls, numbers/m3 were calculated from the percent of sample counted, assuming 100% net efficiency, tow depth, and net diameter (Wat son and

Carpenter 1974). However, the efficiency of plankton nets is seldom 100%, due to net shape, mesh size, mesh area, netting porosity, filtering area, and the mesh area to mouth opening ratio (Sameoto et al. 2000) and thus abundance estimates from 1970 are likely underestimates. For 1974-1975, zooplankton abundance data was corrected to 70% net efficiency (Weisgerber 2000). Although the two problems mentioned above could lead to over- or underestimates of biomass, the fact that many of the sampling and enumeration techniques used to obtain biomass estimates are similar among these studies, combined with the paucity of historical data regarding phytoplankton and zooplankton biomass data in Lake Erie, make the use of these studies allowable.

External Phosphorus Loading

External phosphorus loading decreased from greater than 25 kilotonnes in the early 1970’s to less than the target loading limit of 11 kilotonnes (set by the Great Lakes

Water Quality Agreement) in the 1980’s (Figure 4). However, loading fluctuated around

43 this 11 kilotonne level in the 1990’s, with three consecutive years of loading above the 11 kilotonne target in 1996, 1997, and 1998.

Phytoplankton

Seasonal mean total phytoplankton biomass in all three basins declined between

1970 and the mid-1990’s (western, central, and eastern basins, Figures 5a,b,c). Further, the results of the breakpoint regression analysis reveal that the significant declining trends in phytoplankton biomass ended in 1984 for the central basin (r2 = 0.997, p =

0.002, Table 8a) and in 1997 for the western (r2 = 0.85, p < 0.001) and eastern basins (r 2

= 0.67, p = 0.007) (Table 8a). After these “breakpoint” years the trends in phytoplankton biomass were not significant (i.e., slopes were not different from zero) (0.05 < r2 < 0.35,

0.291< p < 0.719, Table 8a). However, visual inspection of the western basin (Figure 5a) and the eastern basin (Figure 5c) phytoplankton biomass in 1997 and after indicates that increases in phytoplankton biomass may be occurring, even if not statistically significant

(Table 8a).

Zooplankton

Seasonal mean crustacean zooplankton biomass in the western and the eastern basins (Figure 5d,f) showed declines between 1970 and the mid-1980’s, while zooplankton biomass in the central basin declined from 1970 to 1999 (Figure 5e).

Further, the results of the breakpoint regression analysis indicate that the significant declining trends in zooplankton biomass occurred between 1970 and 1986 (r2 = 0.85, p =

0.009, Table 8b), but declines were not significant in the central (r2 = 0.41, p = 0.063) or eastern basins (r2 = 0.42, p = 0.553) (Table 8b). Further, after “breakpoint” years trends in zooplankton biomass over time were not significantly from zero in any of the three basins of Lake Erie (0.07 < r2 < 0.90, 0.208 < p < 0.543, Table 8b). When 1970 data were removed from the phytoplankton and zooplankton biomass regressions, slopes were not significantly different from zero in any basin.

Plankton Biomass versus External Phosphorus Loading

Seasonal average total phytoplankton biomass (mg/L) was significantly positively correlated with lakewide external phosphorus loading (kilotonnes) from 1970 to 2001 in the western (slope = 0.1460, r2 = 0.31, p = 0.048, Figure 6a) and central basins (slope =

0.0972, r2 = 0.49, p = 0.007, Figure 6b), but not in the eastern basin (slope = 0.0185, r2 =

0.01, p = 0.735, Figure 6c). Further, seasonal average crustacean zooplankton biomass

(mg/L) was positively correlated with lakewide external phosphorus loading (kilotonnes) from 1970 to 2001 in the western basin (slope = 0.0100, r2 = 0.43, p = 0.015, Figure 6d) and eastern basin (slope = 0.0043, r2 = 0.44, p = 0.026, Figure 6f), but not in the central basin (slope = 0.0031, r2 = 0.13, p = 0.273, Figure 6e). However, when the 1970 data are removed, I found no relationship between total phytoplankton biomass and lake-wide phosphorus load for the western (r2 = 0.006, p = 0.811), central (r2 = 0.02, p = 0.703), or eastern (r2 = 0.13, p = 0.249) basins nor between total crustacean zooplankton biomass and lake-wide total phosphorus loading for the western (r2 = 0.16, p = 0.204), central (r2

= 0.06, p = 0.511), or eastern (r2 = 0.04, p = 0.588) basins.

DISCUSSION

Zooplankton and phytoplankton communities vary with respect to time and space across different scales in the Laurentian Great Lakes. Studies have been conducted on long temporal scales (decades), (Davis 1964, Nicholls 1997), large spatial scales

(Barbiero et al. 2001) (Barbiero and Tuchman 2001), short temporal scales (hours)

(Smith et al. 1998), and small spatial scales (meters) (Hubschman 1960). The variety of scales used in the Lake Erie plankton research indicates that choosing the correct scale is an important aspect of the study of planktonic communities, due to the spatial and temporal variability of these communities. Further, designing and implementing a sampling program that coincides with the goal and scale of a study is critical.

Lake Erie’s phytoplankton and zooplankton communities vary temporally.

Temporal variability in phytoplankton was evident in 1970 with total phytoplankton biomass declining from spring to fall in the western basin (spring = 7.1 mg/L > summer =

5.4 mg/L > fall = 3.3 mg/L) and eastern basin (spring = 2.9 mg/L > summer = 2.7 mg/L > fall = 1.5 mg/L) and declines occurring from summer (4.9 mg/L) to fall (2.7 mg/L) in the central basin of Lake Erie (Munawar and Munawar 1976). Similarly, my results indicate that phytoplankton biomass varies over the time scale of months, as evident from both the different strengths of correlations (regression analyses- 0.005 < r2 < 0.985) and percentage mean differences (paired t-tests- 1.64 < absolute % mean difference < 251.15) between individual months and the entire sampling season and the significance of the effect of month on a number of phytoplankton groups (n = 6) and years (across groups) (n

= 12) (ANOVAs). Zooplankton biomass in the western basin of Lake Erie varies temporally, tripling between early and late June (Frost and Culver 2001). Similarly,

Watson and Carpenter (1974) found large increases in crustacean zooplankton biomass between early June and early July in 1970. Similar to phytoplankton, zooplankton biomass varied over the time scale of months. This is evident from both the different strengths of correlations (regression analyses- 0.033 < r2 < 1.00) and percentage mean differences (paired t-tests- 11.90 < absolute % mean difference < 122.39) between individual months and the entire sampling season and the significa nce of the effect of month on a number of zooplankton groups (n = 5) and years (across groups) (n = 28)

(ANOVAs). Particularly convincing is the repeated significance of the effect of month on cyclopoid, cladoceran, crustacean zooplankton, and total zooplankton biomass during the time period of this study (1996-2002).

Lake Erie’s phytoplankton and zooplankton communities also vary spatially.

Munawar and Munawar (1976) found in 1970 that the western basin had the highest mean total phytoplankton biomass (wet weight-5.3 mg/L), with the central basin having a lower mean phytoplankton biomass (3.2 mg/L), and the eastern basin having the lowest mean phytoplankton biomass (2.4 mg/L). Similarly, my long-term trend analyses (1970-

1996) indicate that seasonal average total phytoplankton biomass is greatest in the western basin, less in the central basin, and least in the eastern basin during most years.

However, during the late 1990’s eastern basin seasonal average total phytoplankton biomass approached values observed in the western basin. Stockwell and Sprules (1995) found that mean zooplankton biomass (wet weight) decreased from west to east in May

47 of 1994 (western basin = 1401.3 µg/L > central basin = 539.5 µg/L > eastern basin =

157.2 µg/L). Likewise, in my long-term trend analyses, seasonal average crustacean zooplankton biomass typically declined from west to east in Lake Erie. I also found that the basin had a significant effect on total zooplankton biomass, as well as a number of its components.

The high residual or unexplained variance in both the phytoplankton (~66%) and zooplankton (~50%) analyses (ANOVAs) indicate that more than basin, site, and short- term temporal effects drive the variability of these communities. Rusak et al. (2002) found that in a large-scale study of north temperate lakes in North America the effect of region (as a surrogate for geology, climate, and glacial history), at times, c ould explain greater than 40% of the variability in crustacean zooplankton abundance data. Further, both nutrients and zooplankton grazing combined explained greater than 80% of the spring and summer variability in phytoplankton in the offshore waters of Lake Michigan during 1983-1992 (Makarewicz et al. 1998). Also in Lake Michigan, Makarewicz et al.

(1995) found evidence that planktivorous fish could partially explain zooplankton biomass patterns. Thus the analyses I conducted did not take into account regional spatial scales, bottom-up (nutrients) or top-down processes (predation and grazing) that could explain variability in phytoplankton and zooplankton biomass. However, my goal was not to characterize all components of plankton variability. Rather, my goal was to determine variability that was associated with sampling regimens and determine ways to minimize sampling effort, while maximizing information gained regarding plankton variability.

Given the quantitative data for both zooplankton and phytoplankton biomass in

Lake Erie across space and time, the utility of using diminished sampling schemes (i.e., sampling only during certain months, sampling at only a few sites) may be determined.

First, monthly mean biomass was rarely significantly correlated to entire season mean biomass for either phytoplankton and zooplankton. Further, the absolute percentage mean differences between monthly and entire season mean absolute biomass were frequently >50% (11 out of 40 comparisons), and differences were as high as 250% for mean phytoplankton biomass and almost half of the crustacean zooplankton biomass comparisons (19 out of 40) had absolute percentage mean differences >50% between monthly and entire season biomass, with differences exceeding 120%. The lack of a single month that consistently predicts entire season zooplankton and phytoplankton biomass, combined with large percentage mean differences, indicate that phytoplankton and zooplankton sampling cannot be reduced to 1 or 2 months in Lake Erie. Thus sampling May through October (or late September) is necessary to capture the variability of the plankton. The late spring, summer, and early fall are time periods of great variability in the plankton of north temperate lakes (Sommer et al. 1986), including Lake

Erie.

Although temporal reduction of plankton sampling in Lake Erie is not justifiable, using a reduced suite of sites as a surrogate for a larger suite of sites may be warranted.

As few as 5 or 10 sites could serve as surrogates for a fu ll suite of sites in Lake Erie’s western basin, even one containing 40 sites. Further, almost all phytoplankton spatial comparisons (19 out of 21) had absolute percentage mean differences <25% and nearly

49 all crustacean zooplankton spatial comparisons (20 out of 21) had absolute mean differences <25%. Further, the percentage mean differences were <1% in some cases. In the central basin, mean phytoplankton and zooplankton biomass at two sites of the same depth (20m) did not differ, even though these sites differed in distance from shore by approximately 20km. Further, these results were obtained across 4 years (1997-2000).

Hall et al. (2003) used a depth of 20m to distinguish nearshore and offshore sites in Lake

Ontario. Using this depth to define the offshore zone of the central basin of Lake Erie, both sites 83 and 1279 would be considered offshore sites. This, combined with the lack in differences in phytoplankton and zooplankton biomass between the sites indicates that more inshore (but still offshore) sites can be used as surrogates for mid-basin offshore sites.

Evans and Sell (1983) recommended that zooplankton sampling in Lake Michigan should be conducted “at least monthly and preferably weekly or biweekly. However,…it is not necessary to intensively sample the entire study area.” I concur with this recommendation for both zooplankton and phytoplankton in Lake Erie. Greater frequency of sampling is more important than sampling a greater number of sites in an area, even as large as a basin. Further, differences in the patterns of variability of phytoplankton and zooplankton during 1996-2002 demonstrate that inter-annual differences also must be taken into account for plankton monitoring programs. Large inter-annual differences found in plankton communities (Kratz et al. 1987) reinforce the need for long-term plankton monitoring programs, a conclusion viewed as a priority in ecological research (Magnuson 1990). In conclusion, for each year I would recommend

50 sampling 5-10 sites in each basin at a temporal frequency of at least monthly (May-

October). I also agree that the spatial variability in plankton biomass, especially between years and basins in Lake Erie, is consistent with the call that for zooplankton “long-term monitoring programs, stratified sampling by basin should be considered an absolute minimum to characterize lakewide trends” (Stockwell et al. 2002).

Because the temporal variability of plankton communities is great during the summer in north temperate lakes (Sommer et al. 1986), sampling should also occur throughout the months of May-October in Lake Erie. Sampling in only one month could lead to highly overestimating or underestimating the actual seasonal average biomass.

Further, by not sampling during a number of months, a m onitoring program may miss plankton characteristics that reflect Beneficial Use Impairments For example blue -green algal blooms are strongly tied to weather and physical conditions of the water column

(Sorrano 1997) and can be unpredictable as to the exact location and timing of their occurrence. Although not addressing the spatial aspect of blue-green blooms, a monthly temporal sampling scheme would be necessary to detect such stochastic events. For example, in 1996 cyanophyte biomass in the offshore waters of Lake Erie was less than

50 µg/L for most of June and July, but increased to greater than 1000 µg/L in late July and August (Culver, unpublished data). Monitoring in June and July would not have allowed the observation of this great increase in the Cyanophyta, a taxon that can cause taste and odor problems in water, as well as negatively affect human health.

The goal of rapid bioassessment is to minimize the effort needed to obtain information, but retain the ability to detect changes in community structure (Van Dam et

51 al. 1998). Although a number of studies have developed and utilized methods of rapid bioassesment to characterize invertebrate stream communities (Chessman 1995, Resh et al. 1995), few studies have attempted to rapidly assess plankton communities. Further, studies that have developed techniques for rapid assessment of plankton have met limited success. Mischke and Zimba (2001) used screens of different mesh size to size- fractionate zooplankton communities, but found that zoopla nkton 50% larger or smaller than the mesh size could be retained on a given screen. The Volumetric Index of the

Plankton (VIP) has numerous advantages over traditional plankton enumeration. First, to employ the technique little knowledge of plankton dynamics, little training, and limited materials are necessary. Second, general trends in cyanophyte and crustacean zooplankton can be determined in a relatively short time (i.e., 24 h sample processing).

Further, since the samples are not destroyed in the process, careful microscopic enumeration and identification can be conducted at a later time, if funding, time, and the questions asked warrant. Because the VIP explains very little of the variability (<33%) in cyanophyte or crustacean zooplankton biomass, even after log transformations, it cannot be used as a quantitative estimator of plankton biomass. However, the VIP may still be valuable as a rapid, qualitative method to assess plankton communities (e.g., the presence of cyanophyte blooms, low crustacean zooplankton densities during the clear-water phase in Lake Erie (Wu and Culver 1991)). Further, the VIP should be applied to other systems

(e.g., small lakes, reservoirs, hatchery ponds) to determine if it has more predictive power in these systems than it does in Lake Erie.

Phosphorus loading to Lake Erie declined between 1970 and the mid-1980’s, and summer total phosphorus concentrations decreased significantly in the three basins of

Lake Erie between the periods 1968-1972 and 1994-1996 (Charlton et al. 1999).

Phosphorus limits the growth of phytoplankton in lakes (Schindler 1977). Therefore, declining phosphorus loads into Lake Erie (Dolan 1993) should lead to a decline in phytoplankton biomass. A decline of phytoplankton biomass with oligotrophication has been found in other lakes (Edmondson and Lehman 1981, Sommer et al. 1993). Further, this relationship has been found in Lake Erie (Nicholls et al. 1977, Nicholls 1997). Less evidence exists for a relationship between phosphorus loading and z ooplankton biomass.

However, Beeton (1965) reviewed studies that indicated zooplankton abundance increased with the increased eutrophication of Lake Erie. Further, Johannsson et al.

(1999b) found a decline in Lake Erie zooplankton biomass between the 197 0’s and

1980’s associated with decreased phosphorus loading. Finally, a fish community indicative of more oligotrophic conditions returned to Lake Erie during the 1980’s and first half of the 1990’s (Ludsin et al. 2001), after declines in external phosphorus loading.

In general, long-term phytoplankton and zooplankton biomass changes track with changes in phosphorus loading to Lake Erie (1970-2001). For example, as phosphorus loading declined from 1970 through the 1980’s and early 1990’s, so too did phytoplankton, and to a lesser extent zooplankton biomass. However, these results also indicate that significant declining trends in phytoplankton and zooplankton ended by the mid-1990’s. Since this time, phytoplankton biomass appears to have increased in both the western and eastern basins of Lake Erie. In fact, average seasonal phytoplankton

53 biomasses have approached or exceeded those found in 1970. Further, greatest increases in phytoplankton biomass in these basins occurred after three consecutive years of higher phosphorus loading (>11 kilotonne target) to the lake (1996, 1997, and 1998). Increased phytoplankton biomass has been accompanied by increases in blooms of cyanophytes, such as Microcystis in the western basin of Lake Erie during this same time period (late-

1990’s) (Budd et al. 2002). Microcystis blooms have the potential to contain microcystins, liver toxins, which can affect human health (Brittain et al. 2002). Further, in recent years increased areas of hypoxia/anoxia are evident in the central basin of Lake

Erie which may be attributed to the increased phytoplankton biomass, the subsequent settling of dead phytoplankton to the bottom, and consumption of oxygen by bacteria breaking down this material (USEPA 2004). Further, dreissenid mussel s may contribute to this problem by remineralizing nutrients and stimulating phytoplankton growth

(Conroy et al. 2004b, in review). Blue-green algal blooms, anoxic waters, and invasive species, such as dreissenid mussels, can influence a number of the Beneficial Use

Impairments in Lake Erie. In the future, by monitoring phytoplankton and zooplankton biomass changes in Lake Erie, managers can gather data that is relevant to BUIs, as well as relevant to the changing structure and function of the Lake Erie ecosystem.

CONCLUSION

In this chapter, I presented two questions that are imperative for understanding the changing Lake Erie ecosystem. First, can we sample plankton in an effective way, so as to maximize information, while minimizing effort? Second, how have changes in

54 external phosphorus loading correlated with changes in phytoplankton and zooplankton biomass? To minimize effort without losing information, sampling can be restricted to a few stations per basin in Lake Erie, preferably stations that are spatially distant from one another. Sampling should be conducted at least monthly during May -October. Finally, long-term trends in phytoplankton and zooplankton biomass tracked well with changes in phosphorus loading to Lake Erie. However, other stressors both biotic (e.g., dreissenid mussels), chemical (e.g., toxic metals), and physical (e.g., water level changes) may be affecting plankton communities in Lake Erie and the impacts of these additional stressors must be explored. In conclusion, the understanding of both plankton variability and changes associated with phosphorus loading (both external and internal) are important for understanding the changing Lake Erie ecosystem. With an understanding of the changing

Lake Erie ecosystem, more effective monitoring and management are possible.

CHAPTER 3

THE CHARACTERISTICS AND POTENTIAL ECOLOGICAL EFFECTS OF THE

EXOTIC CRUSTACEAN ZOOPLANKTER Cercopagis pengoi (CLADOCERA:

CERCOPAGIDAE), A RECENT INVADER OF LAKE ERIE

INTRODUCTION

Responding to problems in the 1960’s, the U.S. government took steps to protect the quality of water in the U.S. Legislation was enacted in 1972 (Water Pollution Control

Karr (1991) notes, “the solution of water resource problems will not come from better regulation of chemicals or the development of better assessment tools to detect degradation.” What is needed is a better understanding of biological water quality and what organisms are appropriate for its measurement.

Nonindigenous species have invaded the Great Lakes at an accelerating pace since the 1800s and this biological pollution causes declines in biotic integrity. The

Great Lakes is home to at least 139 nonindigenous fishes, invertebrates, fish disease pathogens, plants, and algae (Mills et al. 1994). Recently researchers have been paying closer attention to the biogeography of these invaders. Seventy percent of the invading species that have been discovered since 1985 are native to the waters of the Ponto -

Caspian region (Black, Caspian, and Azov Seas) (Ricciardi and MacIsaac 2000).

Organisms from this region have had negative effects on the Great Lakes, causing extinctions or extirpations of native species, alterations to habitat, and disruptions of foodwebs (Ricciardi and MacIsaac 2000). This suite of Ponto-Caspian invaders has succeeded due to its broad range of environmental tolerances, its taxonomic diversity, and its increased colonization of European ports (Ricciardi and MacIsaac 2000).

Crustaceans previously composed only 4% of the invaders to the Great Lakes

(Mills et al. 1994). Recently, however, the Great Lakes have been invaded by a number of crustacean species, including an amphipod, Echinogammarus ischnus (Witt et al.

1997), and the cladocerans Bythotrephes longimanus (Bur et al. 1986, Lange and Cap

1986), Cercopagis pengoi (MacIsaac et al.1999), and Daphnia lumholtzi (Muzinic 2000).

Three of these species are of Ponto-Caspian origin (E. ischnus, B. longimanus, and C. pengoi). Ricciardi and Rasmussen (1998) predicted that other amphipods (Corophium spp.) and mysids may be the next crustacean invaders from the Ponto-Caspian area to invade the Great Lakes. The increase of Ponto-Caspian invaders, especially since the introduction of Dreissena polymorpha and Dreissena bugensis into the Great Lakes, may

57 be a case of what Simberloff and Von Holle (1999) call “invasional meltdown.” In this process the presence of one or more non-native species facilitates the invasion of other species that share a common geographic and evolutionary history with initial invader(s).

Samples from August and September 2001 from the western basin of Lake Erie contained Cercopagis pengoi (Therrialt et al. 2002), and C. pengoi has been previously found in Lake Ontario (MacIsaac et al. 1999) and Lake Michigan (Charlebois et al.

2001). Cercopagis pengoi biofouls fishermens’ lines (MacIsaac et al. 1999) and may have a number of negative impacts on the biota of Lake Erie. The characteristics of

Cercopagis and its possible ecological effects are discussed in this chapter.

MATERIALS AND METHODS

Zooplankton samples were collected by the Ohio Department of Natural Resouces

(ODNR) as part of the Ohio State University’s Lake Erie Plankton Abundance Study

(LEPAS). This ongoing study (1995-present) seeks to monitor long-term trends in lower trophic level interactions that affect the recruitment of fish species in Lake Erie. The discovery of Cercopagis pengoi was not an intended goal for LEPAS, rather it was a fortuitous byproduct of a rigorous spatial and temporal sampling regimen that allowed us to detect this new invader. Sampling was conducted approximately once every two weeks from April 30, 2001 to October 2, 2001 at eight stations in the western basin of

Lake Erie (Figure 7). A total of 79 zooplankton samples was collected from western

Lake Erie and analyzed during 2001. An additional 28 samples from central Lake Erie and 61 samples from eastern Lake Erie were collected and analyzed the same way, but

58 contained no Cercopagis. Zooplankton sampling and enumeration methodology is outlined below and followed the methodology used by previous researchers (Frost and

Culver 2001).

Vertical tow samples were collected using a front-weighted zooplankton net (0.5- m diameter, 64-µm mesh) fitted with a General Oceanics 2030R model flow meter and

500-ml jar. The net was lowered with the open end pointing downward until the 2 kg weight fastened to the front bridle by a 1-m line hit bottom. The net was then retrieved, allowing the water column to be sampled both as the net was lowered and as it was pulled up, while avoiding collecting mud from the bottom. Samples were then concentrated and preserved with a 4% sugar formaldehyde solution (Haney and Hall 1973).

In the laboratory, each sample was diluted to a known volume, typically from 500 ml to 3000 ml unless the sample contained an extremely small or large amount of zooplankton, requiring lower or larger dilution volumes. After dilution, all zooplankton in at least two subsamples of 5-10 ml were identified and counted. Samples were processed using a Wild M5A dissecting microscope fitted with a calibrated ocular micrometer so body measurements could be obtained to the nearest 0.05 mm.

Cladocerans and copepods were identified to species while rotifers were identified to genus. Additional subsamples were analyzed until at least 100 individuals of the most common taxon were recorded. Cercopagis was initially discovered in one subsample from station 27 (Figure 7). After C. pengoi was discovered in this subsample, all 2001 samples (n = 21) from station 27 and 29 (Figure 7) were examined in their entirety for the

59 presence of C. pengoi and another cercopagid, Bythotrephes longimanus. Zooplankton data on taxa other than C. pengoi will be reported elsewhere.

Cercopagis pengoi individuals were measured from the top of the head to the base of the caudal process, which can be over five times the length of the body (Makarewicz et al. 2001). Biomass was calculated from length measurements using Grigorovich et al.’s

(2000) equation: logW= 0.375 + 2.442 log L, where W is dry body mass (µg) and L is length (mm). The volume of water (L) sampled was calculated using the number of flowmeter revolutions multiplied by the net constant (5.2765 L/ revolut ion). Densities and biomass of individuals, as well as average biomass, were then calculated.

RESULTS

Samples containing C. pengoi were collected at two sites in the western basin of

Lake Erie, south of the Detroit River during 2001 (Figure 7). This area is near where other researchers reported sampling Cercopagis pengoi in late August 2001 (H. J.

MacIsaac, University of Windsor, pers. comm.). Average length of the body (less the tail) at the two sites was 1.0 mm + 0.07 (mean + standard error), while the average weight of individuals at the sites was 2.4 + 0.36 µg. Only seven individuals were collected and both densities and biomasses were extremely low (Table 9). For both sites combined, densities were 1.0 + 0.7 /m3. Average biomass was also very low at 3.0 + 0.2

µg/m3.

Average body length of individuals found in this study was somewhat less than values for parthenogenetic females found in Lake Ontario (0.99 mm vs. 1.36 mm)

(MacIsaac et al. 1999). Median densities of C. pengoi in Lake Ontario (295/ m3)

(Ojaveer et al. 2001) greatly exceeded our mean of 1/ m3. Likewise, median dry-weight biomass in Lake Ontario (13,400 µg DW/ m3) (Ojaveer et al. 2001) exceeded our average biomass value of 3 µg/ m3. Because no Bythotrephes individuals were found on these sampling dates at these sites, data on Leptodora kindti from these samples are included for comparison of Cercopagis with another predatory cladoceran (Table 9). Average

Leptodora total body length is almost twice that of Cercopagis excluding the tail.

Further, average Leptodora biomass is almost three times that of Cercopagis. Finally,

Leptodora average density is more than five hundred times the density obtained for

Cercopagis, while the average biomass of Leptodora is greater than one thousand times the biomass obtained for C. pengoi. Cercopagis from the western basin of Lake Erie appear to be of small size, low density, and low biomass when compared to C. pengoi from Lake Ontario and Leptodora from the western basin of Lake Erie.

DISCUSSION

Cercopagis pengoi has successfully invaded lakes Ontario and Michigan

(MacIsaac et al. 1999, Charlebois et al. 2001), and now appears to be established in Lake

Erie (Therriault et al. 2002). Individuals found in the western basin of Lake Erie likely entered the lake via the Detroit River from the upper lakes. Therefore currents could have carried individual colonizers into this area from the Detroit River.

Now that C. pengoi is present in Lake Erie, it is important to study the extent to which it will become established and the impact it will have on the lake ecosystem.

Characteristics of C. pengoi, as they relate to ten characteristics of successful invaders outlined below (Ricciardi and Rasmussen 1998), may provide some answers.

1. Abundant and Widely Distributed in Original Range

Cercopagis pengoi is native to a large area (> 3.5 million km2) containing the

Caspian and Aral Seas, the Don Estuary and the Sea of Azov, and to coastal lakes and the

Dneiper-Bug and Dniester estuaries of the Black Sea (MacIsaac et al.1999). Further, densities of Cercopagis between 0.1 and 8/L, and as high as 26/L have been recorded in its native range (Makarewicz et al. 2001).

2. Wide Environmental Tolerance

Cercopagis pengoi exhibits a wide environmental tolerance to a number of abiotic factors. It is euryhaline, occurring in waters with salinities ranging from 0.1 to 14%

(MacIsaac et al. 1999), including numerous sites in the Baltic Sea, such as the gulfs of

Riga (Ojaveer et al. 1998) and Finland (Krylov et al. 1999). Cercopagis pengoi is also very eurythermic, persisting in waters as cold as 8ºC and, at lower abundances, at temperatures as high as 30ºC (MacIsaac et al. 1999).

3. Short Generation Time

Although no current studies evaluating generation time in C. pengoi have been conducted, generation times in other cladocerans (e. g., Daphnia) are short (days), especially at higher temperatures (Hall 1964, Edmondson and Litt 1982). Total

62 developmental time at 21ºC for the cercopagid Bythotrephes, from embryonic stage through instar III, has been measured as approximately 10 days (Lehman and

Branstrator 1995). Cercopagis pengoi most likely has short generation times, based on those of other cladocerans and its rapid increases in abundance during the year. More research needs to be conducted with regards to Cercopagis pengoi population dynamics.

4. Rapid Growth

Although no current studies evaluate growth rates of Cercopagis pengoi, growth rates in Daphnia (Edmondson and Litt 1982) and Bythotrephes are rapid (Sprules et al.

1990), especially at higher temperatures. Studies of the cercopagid Bythotrephes have found both high individual growth efficiencies (conversion of 25% of prey biomass into predator biomass) (Lehman and Branstrator 1995) and population growth rates (0.075/ day at 8ºC to 0.146/day at 15ºC, assuming no juvenile or adult daily mortality) (Sprules et al. 1990). Rapid growth would also be consistent with the large increases in C. pengoi observed at various times during the year.

5. High Reproductive Capacity

C. pengoi can reproduce both parthenogenetically and gametogenetically, depending on a number of environmental cues (Grigorovich et al. 2000). Since parthenogenetic females are present at most times of the year and asexual reproduction is rapid relative to sexual reproduction C. pengoi’s reproductive capacity should be high.

The closely related Bythotrephes has an early age-at-first-reproduction (13 days), short

63 egg development time (8 days), and a high intrinsic rate of increase (as high as 0.146

/day, assuming no predation mortality) at 15ºC (Sprules et al. 1990). Decreased ages at first reproduction and egg development times were found for Bythotrephes as temperatures increased. At 8ºC, age at first reproduction was 26 days, and e gg development time was 16 days, as compared to 13 and 8 days, respectively, at 15ºC (13 days, 8 days) (Sprules et al. 1990). Increased temperature leads to shorter developmental times and higher reproductive rates in other cladocerans, such as Daphnia spp.

(Edmondson and Litt 1982). Therefore, C. pengoi’s reproduction rate may be high during high water temperatures in the summer, due both to these factors and its tolerance of high temperatures.

6. Early Sexual Maturity

Although sexual maturity may only occur during times of abiotic or biotic stresses, C. pengoi can reproduce parthenogenetically at instars I, II, and III (Grigorovich et al. 2000). Further, instar II sexual females have, in rare cases, been found to be fecund as well as instar III females (Grigorovich et al. 2000); early sexual maturity is consistent with C. pengoi being a successful invader.

7. Broad Diet (Opportunistic Feeding)

Uitto et al. (1999) have shown that Cercopagis pengoi in European lakes eats a broad range of prey foods including 60% copepods (nauplii and copepodites of the genera Acartia, Eurytemora, and Temora), 20% rotifers, and 20% podonids (Evadne

64 nordmanni). These three groups vary considerably in size and trophic position, thus providing evidence of a broad diet.

8. Gregariousness

Cercopagis pengoi can achieve densities up to 26,000/ m3 (Makarewicz et al.

2001). Studies have shown that abundances are usually highest in the epilimnion in both

Lake Ontario (Ojaveer et al. 2001) and the Gulf of Finland (Krylov et al. 1999). Further, in order for sexual reproduction to occur, males and females must overlap spatially and temporally. All of these reasons indicate that C. pengoi is, to some extent, gregarious.

This gregariousness may increase invasion potential by providing for greater abundances of initial colonizers.

9. Possessing Natural Mechanisms of Rapid Dispersal

Cercopagis pengoi resting eggs have been found to pass undamaged through fish digestive systems (Antsulevich and Valipakka 2000). Thus migratory fish, and even waterfowl could provide rapid dispersal of C. pengoi. Further, since Cercopagis pengoi is a planktonic organism, currents within lakes, between lakes, and in rivers may cause it to disperse rapidly.

10. Commensal with Human Activity (e.g. Ship Ballast Transport)

Cercopagis pengoi was most likely introduced into the Great Lakes and Baltic

Sea through ballast water (MacIsaac et al. 1999, Cristescu et al. 2001). Since C. pengoi

65 is planktonic and produces resting eggs, either of the se life stages could have been transported via shipping. Furthermore, it is possible that even if ships carrying the organisms flushed their ballast tanks with saline water, the euryhaline C. pengoi would have survived (MacIsaac et al. 1999). Given the f requent passage of ships from both lakes Ontario and Michigan through lakes Huron, St. Clair, and Erie, it is surprising that

Cercopagis did not become established in these lakes earlier than it did.

Potential Ecological Effects

Now that C. pengoi is widespread in the Great Lakes, it is important to determine whether it will become abundant in the western basin of Lake Erie and to study its possible ecological consequences for this ecosystem.

In order to determine whether C. pengoi will become abundant in western Lake

Erie, a comparison with the unsuccessful establishment of the cercopagid Bythotrephes longimanus in western Lake Erie is necessary. There are a number of hypotheses as to why Bythotrephes has been an unsuccessful invader to western Lake Erie. Previous studies have found that Bythotrephes’ abundance in western Lake Erie was negatively correlated with water temperature (Berg and Garton 1988) and it thus was absent from the western basin for much of the summer. Further, Bythotrephes abundance was also negatively correlated with Leptodora abundance. This negative correlation between

Bythotrephes and Leptodora has also been found in Lake Michigan, and has been ascribed to either competition between the two species or predation by Bythotrephes on

Leptodora (Branstrator 1995). Further, Garton et al. (1990) found that Bythotrephes

66 could not seasonally acclimate to temperature (measured by the interaction of season, temperature, and mortality) and that reduced fecundity, presence of sexual eggs, and greater presence of males were evidence that Bythotrephes had reached an environmental limit with respect to the high summer temperatures (23-25 ºC) of western Lake Erie.

However, in a number of bodies of water in Europe Cercopagis pengoi has been shown to have its highest abundance (> 100,000 individuals/m3, in one case) in water temperatures between 20-30º C (MacIsaac et al. 1999), temperatures that are similar or even higher than summer water temperatures found in western Lake Erie. Although

Bythotrephes has failed to become abundant in western Lake Erie, the differing thermal tolerances of Cercopagis make its success more likely.

If Cercopagis does become established in the western basin, it may have a number of ecological effects. C. pengoi eats the calanoid copepod Eurytemora (Uitto et al. 1999). Eurytemora affinis is present in Lake Erie and has been shown to be an important part of the diet of the rainbow smelt (Osmerus mordax) (Faber and Jermolajev

1966, Gopalan et al. 1998), which is a commercially important introduced species in the

Great Lakes. Because both Cercopagis and rainbow smelt consume Eurytemora, competition may occur between them. Young-of-year rainbow smelt are often caught in high numbers in the western portion of the central basin of Lake Erie (i.e. 184 individuals caught per hour trawling (CPHT) in October monitoring by the Ohio Department of

Natural Resources (1999), as compared to 1.1 and 2.3 CPHT for the zooplanktivorous alewife (Alosa pseudoharengus) and gizzard shad (Dorosoma cepedianum)). Further,

Cercopagis is eaten by a broad range of fish families in Europe. In the Baltic Sea, for

67 example, C. pengoi is important in the diets of herring (Clupeidae), sticklebacks

(Gasterosteidae), bleak (Cyprinidae), and smelt (Osmeridae) (Ojaveer et al. 1998). The authors found that C. pengoi made up 50% of herring food at some sites and almost

100% at others. Clupeids such as the alewife and gizzard shad occur in Lake Erie

(ODNR 1999), and alewives have been shown to prey upon Bythotrephes in the Great

Lakes (Branstrator and Lehman 1996), so clupeids should most likely consume

Cercopagis, which has a similar morphology and ecology. Therefore, C. pengoi may provide a new source of food for these planktivores, as well as for rainbow smelt. Future research will determine to what extent competition with, and predation upon, Cercopagis will affect the fish of the Lake Erie ecosystem.

Native predaceous cladocerans (Leptodora kindti), and the herbivorous Daphnia spp. decreased contemporaneously with the invasion of Bythotrephes into Lake Michigan

(Lehman and Caceres 1993). Others (e. g., Dumitru et al. 2001) have suggested that predation by Bythotrephes has influenced plankton dynamics, including the possibility that Daphnia abundance is controlled by Bythotrephes predation, rather than by the native yellow perch (Perca flavescens) (Hoffman et al. 2001). Further, Bythotrephes has quickly become abundant in the diets of planktivorous fish, such as the lake herring

(Coregonus artedii) (Coulas et al. 1998). Cercopagis may have similar trophic consequences.

CONCLUSION

The environment and suite of organisms found in the western basin of Lake Erie appear to be amenable to the establishment of Cercopagis pengoi. However, further monitoring, and lab and field experiments will be needed in order to determine whether

C. pengoi will become established in the western basin of Lake Erie. Further, the discovery of Cercopagis pengoi through monitoring efforts demonstrates the need to continue such programs to track future invaders to the Lake Erie ecosystem.

CHAPTER 4

THE STATUS OF Limnocalanus macrurus (COPEPODA: CALANOIDA:

CENTROPAGIDAE) IN LAKE ERIE

INTRODUCTION

Responding to problems in the 1960’s, the U.S. government took steps to protect the quality of water in the U.S. Legislation was enacted in 1972 (Water Pollution Control

A number of recent studies have documented the recovery of the Lake Erie ecosystem. These studies range in scale from the recovery of individual populations

(Cornelius et al. 1995, Krieger et al. 1996) to ecosystem-level recovery (Makarewicz and

Bertram 1991, Ludsin et al. 2001). Resurgence in abundance and/or distribution of a number of taxa has been documented for organisms as diverse as aquatic macrophytes

(Stuckey and Moore 1995), mayflies (Hexagenia spp.) (Krieger et al. 1996), and lake trout (Salvelinus namaycush) (Cornelius et al. 1995). Ecosystem stability is one of the fundamental properties of an ecosystem (Odum 1969), and can be divided into both resistance and resilience. Resistance is defined as an ecosystem’s ability to resist external perturbations (Odum 1969), whereas resilience refers to its ability to return to previous abiotic and biotic conditions after a perturbation (Gunderson 2000). This chapter investigates to what extent populations of the copepod Limnocalanus macrurus, a large, centropagid, calanoid copepod, have recovered in the three basins of Lake Erie, following improvements in water quality subsequent to the cultural eutrophication of the mid 20th century, all within the context of the resilience of Lake Erie.

This chapter will analyze the status of Limnocalanus macrurus in Lake Erie using new data collected in 1995-2000 in comparison with earlier 20th century zooplankton studies. In order to assist in interpreting temporal and spatial changes in

Limnocalanus abundance and distribution in Lake Erie, a brief review is included of its distribution, life history, ecology, narrow physiological tolerances, and its responses to abiotic and biotic perturbations. Hence I discuss the status of Limnocalanus macrurus in

Lake Erie relative to changing characteristics of the Lake Erie ecosystem. I use an

71 historical approach to: 1) document changes in Limnocalanus densities during the 20 th century, 2) document changes in possible stressors on Limnocalanus, namely rainbow smelt (Osmerus mordax) and dissolved oxygen, 3) measure correlations between

Limnocalanus densities and these possible stressors, and 4) interpret the above changes within the context of the physiological ecology ofL. macrurus and Lake Erie ecosystem change.

Distribution and Abundance

Limnocalanus macrurus is circumpolar and occurs in both fresh and salt waters from Scandinavia, the Baltic and Caspian seas and in all the Laurentian Great Lakes, as well as lakes in Canada, Poland, Russia, and the Arctic (Balcer et al. 1984). In the St.

Marys River, the outflow of Lake Superior, L. macrurus was found to reach an average density of 33 /m3 in a year-round study (Selgeby 1975) and generally reaches moderate densities (600-4000 /m3) in lakes Ontario, Huron, and portions of Lake Michigan (Balcer et al. 1984), with lower densities occurring in Lake Erie, and usually fewer than 100 /m3 in the western basin of Lake Erie (Chandler 1940). Historically (pre-1960’s),

Limnocalanus macrurus was found in all three basins of Lake Erie, but in the western basin only during the winter and spring (Gannon and Beeton 1971). Gannon and Beeton

(1971) posited that populations of Limnocalanus originate in Lake Huron and move through the St. Clair River, Lake St. Clair, and the Detroit River before entering the western basin. By the late 1960’s, however, Limnocalanus was essentially absent in Lake

Erie (Gannon and Beeton 1971). Further, mean lake-wide averages of <10 adults/m3 and

<25 juveniles/m3 were found during all 10 months of a study in 1970 (Watson and

Carpenter 1974).

Life History and Ecology

The reproductive cycle of L. macrurus has been studied in all of the Laurentian

Great Lakes. Mating occurs after fall overturn (Torke 1975). Females do not brood their eggs, but deposit them directly into the water after fertilization. The eggs fall to the bottom and stick to detritus particles (Balcer et al. 1984). The nauplii subsequently hatch, and grow slowly during the winter. During the spring phytoplankton bloom, nauplii molt into the copepodid form and grow rapidly with increased phytoplankton food sources

(Gannon 1972). Adult copepodites can often reach lengths of 2.4-2.9 mm or greater

(Balcer et al. 1984).

Limnocalanus macrurus is omnivorous and has been shown to consume diatoms and other chrysophytes, as well as cladocerans, copepods, and its own nauplii (Balcer et al. 1984). Its large size and nutritious body composition also make L. macrurus an excellent food source for planktivorous fish (Balcer et al. 1984). Fish predation, in fact, may have contributed to the decline of this species in areas of the Great Lakes (Gannon and Beeton 1971, Wells 1970).

Physiological Tolerances

L. macrurus is a cold-water stenotherm and is intolerant of low dissolved oxygen levels. Strøm (1946) found that during summer stratification it was restricted in

Steinsfjord, Norway, to below 10 m where water temperature was less than 14ºC. In addition, it was only found above 18 m, where dissolved oxygen was more than 5.6 mg/L. Limnocalanus was most abundant at 13 m at 12ºC and 6.4 mg O2/ L (Strøm 1946)

(Figure 8). The narrow thermal tolerance of Limnocalanus macrurus has also been documented in Lake Michigan as < 11ºC (McNaught and Hasler 1966) and in Lake Erie as < 14ºC (Gannon and Beeton 1971). Studies of temperature and dissolved oxygen tolerances of Limnocalanus eggs and nauplii also document narrow physiological tolerances. Roff (1972) found that L. macrurus eggs that were incubated at 16ºC showed

78% mortality and the nauplii that did hatch from these eggs lived less than a day. He also found that eggs collected from a lake with a 0.8 mg/ L near-bottom dissolved oxygen concentration were not viable. Thus in both European and North American field and lab studies, immature and mature stages of Limnocalanus macrurus have been shown to have a narrow range of physiological tolerances to dissolved oxygen and temperature.

Response to Abiotic and Biotic Perturbations

Gannon and Beeton (1971) suggested that the combination of low dissolved oxygen and increase in planktivores, especially rainbow smelt (Osmerus mordax), in

Lake Erie during the 1960’s most likely led to a drastic reduction of L. macrurus densities there. Low dissolved oxygen or anoxia was recorded throughout the 1950s -

1970s in the western basin (Britt 1955, Beeton 1961) and central basin (Charlton 1980).

In the deep waters of the eastern basin, dissolved oxygen levels were found to be as low as 5.5 mg/L in September 1961 (Carr 1962). However, since dissolved oxygen

74 concentrations were usually greater than this in the eastern basin, Gannon and Beeton

(1971) posited it was the additional stress of increasing planktivorous fish populations that decreased Limnocalanus population size.

Although a number of zooplanktivorous fish species increased in numbers during the 1960s, rainbow smelt in Lake Erie arguably had the largest increase. In 1962, a record total commercial catch of 8.6 million kg of smelt was achieved in Lake Erie, with smelt accounting for 30% of the entire catch of fish that year (Christie 1974). L. macrurus was found to be one of the most common components of smelt diets in the spring of 1958 in the western basin of Lake Erie (Price 1963). Furthermore,

Limnocalanus abundances were low in the environment during this period, ranging from

2-34 individuals /m3 in stations where it was present, and it was absent at many stations, suggesting that rainbow smelt may have been preying selectively upon it.

MATERIALS AND METHODS

Zooplankton samples were collected by the Ohio Department of Natural Resouces

(ODNR) and the National Water Research Institute (NWRI) of Environment Canada as part of the Ohio State University’s Lake Erie Plankton Abundance Study (LEPAS)

(1995-present). LEPAS seeks to monitor long-term trends in lower trophic level interactions that affect the recruitment of fish species in Lake Erie, so sampling sites for this study are found throughout the three basins of Lake Erie (Figure 9). Vertical tow samples were collected using a front-weighted zooplankton net (0.5-m diameter, 64-µm mesh) fitted with a General Oceanics 2030R model flow meter and 500-ml jar. The net

75 was lowered with the open end pointing downward until the 2 kg weight fastened to the front bridle by a 1-m line hit bottom. The net was then retrieved, allowing the water column to be sampled both as the net was lowered and as it was pulled up, while avoiding collecting mud from the bottom. Samples were then concentrated and preserved with a

4% sugar formaldehyde solution (Haney and Hall 1972).

All zooplankton samples were transported to The Ohio State University for analysis. Each sample was diluted to a known volume, usually ranging from 500 ml to

3000 ml unless the sample contained an extremely small or large number of zooplankton.

After dilution, all of the zooplankters in at least two subsamples of 5-10 ml were identified and counted. Additional complete subsamples were analyzed until at least 100 individuals of the most common taxon were recorded. Actual counts of Limnocalanus in the subsamples ranged from 1-18 individuals, when present. However, 55% of the time

Limnocalanus was present, it was present with > 2 individuals counted. Using a Poisson distribution for count data, as has previously been applied to plankton enumeration (Lund et al. 1958), the 95% confidence interval for 2 organisms counted would range from 0-8, for 10 organisms would range from 5-19, and for 18 organisms would range from 11-29.

Thus overestimates of Limnocalanus counts could be at most 4 (8/2) times greater than the count for 2 organisms counted and 2 (29/18) times greater than the count for 18 organisms counted. Underestimates of Limnocalanus counts could be at most 2 (18/11) times less than the count for 18 organisms counted. Although we acknowledge that more accurate counts of Limnocalanus could have been obtained by scanning complete samples, the purpose of LEPAS is broader than documenting the abundance of

Limnocalanus and thus precluded such attention to one taxon. When Limnocalanus was detected, from 5 to1650 individuals were present in the total sample, calculated from the number of individuals in a subsample, its volume, and the volume of the diluted sample.

For example, if we detected 1 individual in a 10 ml subsample, there would be 50 total individuals in a sample with a 500 ml dilution volume, 100 individuals with a 1000 ml dilution volume, and 300 individuals with a 3000 ml dilution volume. Length measurements to 0.05 mm were made using a Wild M5A dissecting microscope fitted with a calibrated ocular micrometer.

Temperature was measured infrequently, using a YSI probe at the surface and within one 1 m of the bottom, at the same sampling stations as the zooplankton. Average bottom temperatures at sites with Limnocalanus present were compared with those at sites where Limnocalanus was absent using the nonparametric Wilcoxon two-sample rank sum test (Minitab Version 13.1, 2000).

Young-of-year (Y-O-Y) and one year-and-older (Y-A-O) rainbow smelt trawl data for the period of 1996-2000 were obtained from the Ohio Department of Natural

Resources (western and central basin) (Jeff Tyson and Carey Knight, personal communication) and the New York State Department of Environmental Conservation

(eastern basin) (Donald Einhouse, personal communication). All fish samples were taken as part of summer (western and central basins) and fall (western, central, and eastern basins) index trawl programs, the data from which are summarized annually (Great Lakes

Fishery Commission 2001). For many stations, the zooplankton samples were taken at the same time and location as the trawl samples. In order to examine relationships

77 between rainbow smelt abundances and Limnocalanus abundances, simple regression analyses were performed on rainbow smelt data and Limnocalanus data (Minitab Version

13.1, 2000). Limnocalanus abundances were converted from number per meter cubed to number per meter squared by multiplying Limnocalanus density for each sample by the depth of the water column at its sampling site. Finally, the rainbow smelt abundances were transformed (log10) to remove problems of larger variances at greater abundances and to reduce positive skewness in the smelt data.

RESULTS

Limnocalanus macrurus was present in all three basins of Lake Erie during the study period of 1995-2000 (Figure 9). However, in the eastern basin and central basin it was collected very infrequently, (1 out of 101 (1%) and 8 out of 338 samples (2%) respectively). Limnocalanus was more frequently collected in the western basin, being collected in 99 out of 1107 samples (9%) (Table 10). Limnocalanus also reached a much higher average abundance (11.8 individuals /m 3) in the western basin (Table 10, Figure

9). For comparison with two other large ( > 1 mm) calanoid copepods, Epischura lacustris densities averaged 845.0, 900.0, and 79.8 indivi duals/m3 and Eurytemora affinis densities averaged 114.6, 125.3, and 407.9 individuals/m 3 for the same time periods in the eastern, central, and western basins, respectively.

Limnocalanus was typically present in May and June in the western basin of Lake

Erie, but was present in other months in some years (Table 11). It was most abundant in

May in the western basin of Lake Erie, except in 1998 when it was most abundant in July

(Table 11).

Median bottom temperatures from 1995-2000 in the western basin of Lake Erie were significantly lower at sites where Limnocalanus was present (14.5ºC, N=78), compared to those without Limnocalanus (21.7ºC, N=964) (nonparametric Wilcoxon two-sample rank sum test, W= 10879.5, p<0.0001). In addition, in 2000 the mean bottom temperature in May (13.8ºC) was nearly 9ºC lower than in June-August (22-23ºC) and approximately 5ºC less than the September mean temperature (19.1ºC). Monthly temperature averages were similar in other years to those reported above but are not reported here.

Mean abundances of young-of-year (Y-O-Y) smelt from 1996-2000 were greatest in the eastern basin of Lake Erie (0.11/m2), less in the central basin (0.05/m2), and least in the western basin (0.02/m2). During the same time period, year-and-older (Y-A-O) smelt were more abundant in the central basin (0.16/m2) than in the eastern (0.05/m2) or western basin (<0.01/m2).

Across the three basins of Lake Erie, the relationship between mean abundances of Y-O-Y smelt and Limnocalanus was not significant (r2 = 0.001, p = 0.907), while the relationship between mean abundances of Y-A-O smelt and Limnocalanus was significant (r2 = 0.368, p = 0.028) and negative.

DISCUSSION

Abundances of Limnocalanus macrurus in Lake Erie have varied tremendously.

The most complete review of Limnocalanus macrurus abundance and distribution in

Lake Erie (Gannon and Beeton 1971) notes that L. macrurus was present in the western basin of Lake Erie in 1929-1930 in low numbers (100-200 individuals /m3) (Wright

1955). It was also present in nearly every site sampled in the central and eastern basins during June 1929, comprising 74% of the total microcrustacean community in the bottom waters of the deepest (62 m) station in the eastern basin (Fish 1960). In February and

May of 1939, fewer than 100 individuals /m3 were found in the Bass Island region of the western basin (Chandler 1940). Limnocalanus was collected in the central basin at densities of 300 and 700 individuals /m 3 in January and May of 1951 (Davis 1954). By

1957, however, Limnocalanus was found at abundances of only 1-20 individuals /m3 in the western basin, and less than 1 individual /m3 in the central basin and eastern basins.

By the late 1960’s Limnocalanus was collected infrequently in Lake Erie, even with intensive sampling. Similarly, in 1970 mean cruise abundances of 30 or more individuals

(juveniles + adults) per cubic meter were only encountered twice (early May and early

July) during a 10-month lake-wide study (Watson and Carpenter 1974). By the mid-

1980’s, Limnocalanus was either in extremely low numbers or absent in lake-wide offshore zooplankton studies for 1983-87 (Makarewicz 1993b).

Data from this study demonstrates that Limnocalanus macrurus densities in the western basin during the late 1990’s have returned to nearly the same magnitude of average spring densities (10-100 individuals /m3) seen in 1939. However, Limnocalanus

80 has not increased in the central and eastern basins, where both low densities (<10 individuals /m3) and infrequent presence (<2% of samples containing Limnocalanus) continues. Although our sampling regimen is not as rigorous in these two basins (Figure

9), we still have over 100 samples from both the eastern and central basins and 5 years sampled (1996-2000). Furthermore, Barbiero et al. (2001) found similar abundances

(35.3/m3) in the western basin of Lake Erie during a spring cruise in 1998. They found much lower abundances in the central basin (fewer than 2 individuals /m 3) during both spring and summer cruises, similar to the densities we obtained that year (0.9 individuals

/m3). Finally, Barbiero et al. did not find Limnocalanus in the eastern basin, whereas we found Limnocalanus in only 1 of 101 samples taken in the eastern basin from 1996-2000.

While Limnocalanus macrurus has returned to historical abundances in the western basin of Lake Erie, it has not recovered in the central and eastern basins.

Gannon and Beeton (1971) posited that populations of Limnocalanus do not reproduce in the western basin of Lake Erie, but originate in Lake Huron and move through the St. Clair River, Lake St. Clair, and the Detroit River before entering the western basin, consistent with its current spatial distribution. Furthermore, Limnocalanus was found in a western basin sample with a bottom temperature of 24.1ºC, a temperature that is much higher than Limnocalanus is reported to tolerate (Strøm 1946), suggesting that populations continuously enter Lake Erie from the upper lakes. Perhaps these populations can persist for a short time in high temperatures but then die or migrate to cooler waters. Further, during a spring-summer zooplankton study in 2000,

Limnocalanus was present in Lake St. Clair only during the colder months of May and

September, with densities of < 80/m3 and was found at a number of sites, including sites in and near the shipping channel in Lake St. Clair that runs from the St.Clair River to the

Detroit River (Doug Hunter, Oakland University, Rochester, MI, unpublished data). The timing and magnitude of Limnocalanus densities in Lake St. Clair in 2000 were similar to those in Lake Erie in 2000 and other years. These similarities, along with the spatial distribution of Limnocalanus in Lake St. Clair, are consistent with the suggestion that

Lake Huron may contain a source population. More laboratory comparisons of thermal tolerances and genetic composition of Lake Huron and Lake Erie populations would help answer questions regarding source populations of Limnocalanus.

Gannon and Beeton (1971) attributed the decline in the 1960’s to both hypolimnetic dissolved oxygen depletion and predation by planktivorous fish, most notably the rainbow smelt (Osmerus mordax), which was first documented in Lake Erie in the late 1930’s (Van Oosten 1937). Changes in these two factors in the last thirty years could provide for the resurgence of Limnocalanus.

The implementation of the Great Lakes Water Quality Agreement of 1972 led to the reduction of phosphorus loading into Lake Erie (Dolan 1993). Phosphorus levels have decreased greatly and subsequently dissolved oxygen conditions have improved

(DePinto et al. 1986). Although portions of the central basin still have low dissolved oxygen or even anoxic conditions in the summer, the dissolved oxygen depletion rates in the basin have decreased (Bertram 1993). Further, dissolved oxygen depletion may be a natural event based on basin morphology (Charlton 1980), and may have occurred before

European settlement of the basin (Delorme 1982). The western basin becomes anoxic

82 infrequently, although in the 1960s periods of low dissolved oxygen were observed (Carr et al.1965). Low dissolved oxygen events were most likely responsible for the decline of

Hexagenia populations in western Lake Erie (Britt 1955). Recently improved water quality and more infrequent low dissolved oxygen events have allowed for the resurgence of Hexagenia populations in the western basin of Lake Erie (Krieger et al. 1996), beginning in the mid-1990’s and continuing to the present. Like Hexagenia,

Limnocalanus is intolerant of low dissolved oxygen conditions, so the resurgence of

Limnocalanus populations at the same time as the Hexagenia recovery (ca. early 1990’s) suggests dissolved oxygen levels are involved.

Although water quality has improved throughout Lake Erie, Limnocalanus abundances have only increased significantly in the western basin. As Gannon and

Beeton (1971) noted, planktivorous fish, especially rainbow smelt, may have combined with low dissolved oxygen to reduce Limnocalanus populations. Lake Erie rainbow smelt of all lengths feed heavily on Limnocalanus, with up to 20% of smelt containing

Limnocalanus in their stomachs in 1958 (Price 1963). Further, smelt were most abundant in the eastern basin during the 1960’s and undergo seasonal and diurnal vertical migrations that are similar to those of Limnocalanus, making smelt a likely factor in the decline of Limnocalanus (Gannon and Beeton 1971). Smelt are still abundant in the eastern basin of Lake Erie. Average (mean of annual Geometric Mean Catch Per Hour

Trawling) (CPHT) smelt abundances from fall trawl surveys done from 1996-2000 record

Y-O-Y (young-of-year) smelt abundances as 694, and 75 for Y-A-O (one year-and-older) smelt (Ohio Department of Natural Resources 2003). Smelt were less abundant in the

83 central basin from 1996-2000 with 116 CPHT for Y-O-Y rainbow smelt and 16 for Y-A-

O smelt (Ohio Department of Natural Resources 2003). In the western basin, smelt abundances were lower still. From 1996-2000, CPHT for Y-O-Y smelt was 18, and <1 for Y-A-O fish (Ohio Department of Natural Resources 2003), similar to our analyses for

1996-2000. For this period, Y-O-Y and Y-A-O smelt abundances (CPHT) were consistently greatest in the eastern basin and consistently least abundant in the western basin of Lake Erie, with intermediate abundances occurring in the central basin. Further, there is a significant negative relationship between mean Y-A-O smelt abundance and

Limnocalanus in Lake Erie from 1996-2000. These results indicate that rainbow smelt predation upon Limnocalanus could be preventing its resurgence in the central and eastern basins.

CONCLUSION

Munawar et al. (2002) argue that Lake Erie is a resilient ecosystem and that phosphorus abatement (and possibly dreissenid mussels) have allowed for its oligotrophication. They note that reduction of phytoplankton biomass and blooms, the relative decrease of eutrophic algal species, and the increase of mesotrophic-oligotrophic species are signs that Lake Erie is returning to a mesotrophic-oligotrophic trophic status.

Further, increases in the relative proportion of calanoid copepods and a decrease in the relative proportion of cladocerans and the resurgence of Hexagenia all are signs of Lake

Erie’s resilience. Although Limnocalanus populations may not be reproducing in the western basin of Lake Erie, its presence in the 1990’s alone indicates that conditions

84 favoring its survival in Lake Erie have improved since the 1960’s-1980’s, when it was virtually absent from the ecosystem. Further, factors (e.g., improved water quality) that increase Limnocalanus survival during transit through the connecting channels corridor

(St. Clair River, Lake St. Clair, and Detroit River) between Lake Huron and Lake Erie, will likely result in an increased population in Lake Erie’s western basin. Water quality has indeed improved in these connecting channels (Panek et al. 2003) and may contribute to the increase in the abundance of Limnocalanus in western Lake Erie. Our present study documents the increase in the abundance of Limnocalanus macrurus in the western basin of Lake Erie during the late 1990’s and provides a further example of Lake Erie’s return to a more mesotrophic state, typical before the cultural eutrophication of the mid

20th century.

CHAPTER 5

DEVELOPMENT, VALIDATION, AND APPLICATION OF A PLANKTONIC

INDEX OF BIOTIC INTEGRITY

INTRODUCTION

Responding to problems in the 1960’s, the U.S. government took steps to protect the quality of water in the U.S. Legislation was enacted in 1972 (Water Pollution Control

(Karr and Dudley 1981). Indices of biotic integrity provide measures of this ecosystem property.

Water is essential to human survival. However, less than 5% of the earth’s water is fresh, and most of this water is in ice caps, glaciers, and groundwater (Beeton 2002).

Impairments (BUIs) to ecosystem goods and services have been identified for the

(e.g., degradation of benthos), or both (e.g., eutrophication or undesirable algae).

Characteristics of the plankton communities impact many Beneficial Use Impairments

87 contribute a large portion of Lake Erie’s area and volume, is critical to any management plan for Lake Erie. By using the plankton to construct an IBI, this situation can be remedied. Plankton is sensitive to environmental changes (Schindler 1987), a necessary component of a useful monitoring program. Further, plankton is inexpensive to collect, another requirement of useful monitoring programs (Schindler 1987). In addition, samples can be stored for long periods and do not take up large amounts of space and historical samples can be analyzed and compared with current samples. However, plankton monitoring is not without drawbacks. Plankton taxonomic enumeration is time intensive and thus methods to expedite this process are highly advantageous. Below, I examine one possible way to speed up the estimation of cyanophyte and crustacean zooplankton biomass.

In order to recognize ecological restoration and rehabilitation targets of our freshwaters, human-induced degradation of systems must be recognized (Naiman et al.

1995). In the Laurentian Great Lakes, the United States Environmental Protection

Agency and Environment Canada have programs designed to assess the state of Great

Agreement and are reported at the State of the Lakes Ecosystem Conference (SOLEC)

(Neilson et al. 2003). Further, the Ohio Lake Erie Commission has established the Lake

Erie Quality Index (LEQI) with the goal of establishing benchmarks for monitoring and evaluating progress in the restoration of Lake Erie (Ohio Lake Erie Commission 1998).

To achieve the goals of both SOLEC and the LEQI development of indicators of biological integrity is essential.

Toward this end, Karr (1981) first devised an Index of Biological Integrity (IBI) to measure biological integrity in a stream, using fish as indicator species. Indices of biotic integrity are tools to measure the biological water quality of ecosystems. Karr’s index has been adopted by many management agencies (e.g., Ohio EPA 1988), been modified to use benthic macroinvertebrates as indicators (Fore et al. 1996), and has even been modified to evaluate the integrity of estuarine ecosystems (Weisberg et al. 1997).

Although IBIs have been applied or modified for many purposes, all developed to date are limited to streams or nearshore lake habitats, which are relatively restricted in size.

Such indices are thus poorly suited to assess the overall health of large lakes, which have large open water zones. In particular, a Planktonic Index of Biotic Integrity (P-IBI) developed for large lakes would be of great utility for forming or testing the results of major management decisions and in informing the public of the relative health of t he aquatic ecosystem.

Furthermore, arguably all components of lake function are influenced in major ways by the dynamics of the phytoplankton and zooplankton. Phytoplankters are the primary source of energy driving large lake ecosystems; and the zooplankton is the

89 central trophic link between primary producers and fish (Schriver et al. 1995, Tatrai et al.

1997). Zooplankton dynamics influence fish recruitment. In Lake Erie, this includes economically important walleye (Sander vitreus) and yellow perch (Perca flavescens)

(Wu and Culver 1992, 1994, Culver and Wu 1997, Gopalan et al. 1998). Contaminants reaching fish, birds, and humans have been previously bioconcentrated by the phytoplankton and zooplankton. Recently, algal consumption by zebra mussels

(Dreissena spp.) has further concentrated contaminants from plankton through assimilation from ingested algae or through release of toxic feces and pseudofeces that can be eaten by other benthic organisms, such as amphipods (Fisher et al. 1993, Bruner et al. 1994). Nuisance algal blooms have been a problem in all of the lower Great Lakes, but blooms of the cyanophytes Microcystis, Anabaena, and Aphanizomenon have been particularly common in Lake Erie, where blooms of Microcystis occurred in 1995 (Budd et al. 2002) and infrequently in subsequent years. Microcystis can produce a liver toxin, microcystin, whereas the other two taxa can produce neurotoxins, including saxitoxin, the compound responsible for paralytic shellfish poisoning in the marine environment

(Carmichael 1997). These taxa as well as other algae can produce geosmin and 2- methylisoborneol (MIB) (Sugiura et al. 1998), compounds responsible for taste and odor problems in water supplies (Jones and Korth 1995) and off-flavors in fish (Persson

1980). These examples given are not exhaustive; however, they exemplify the types of information regarding ecosystem processes and effects on humans to be gained by analysis of plankton communities.

Plankton communities have been successfully used to characterize lakes, with respect to nutrient conditions. For example, Gannon and Stemberge r (1978) examined zooplankton communities as indicators of lake trophic status (i.e., eutrophic, mesotrophic, oligotrophic) in Lake Michigan, the Straits of Mackinac, and Lake Huron and found the ratio of calanoid copepod abundance to cladoceran plus cyclopoid copepod abundance distinguished among different trophic conditions. They also found that certain rotifer species assemblages were associated with eutrophic, mesotrophic, and oligotrophic conditions in Lake Huron and the Saginaw Bay. Stemberger and Mi ller

(1998) found that zooplankton assemblages could be used as indicators of the N:P ratio of lakes, due to differential requirements of zooplankters for these nutrients. Further, zooplankton composition could be used to infer changes in watershed characteristics and nutrient loading. Phytoplankton also has been widely used as an indicator of nutrient conditions in lakes. Rawson (1956) discussed measures such as the ratio of centric to pennate diatoms, and a number of other ratios of phytoplankton taxa as being indicative of trophic conditions. Similarly, Dixit et al. (1992) reviewed the many uses of diatoms in monitoring environmental change, including changes in lake trophic status, lake acidification, and climate regimes.

The overall aim of this chapter is to document changes in the offshore water quality of Lake Erie during the second half of the 20th century using a newly developed

P-IBI that includes both zooplankton and phytoplankton metrics. With the P -IBI, lake managers, scientists, and the public can track the biological water quality of Lake Erie based on offshore areas in the lake, areas for which there previously have been few

91 standardized measurements of biological water quality. Further, the P-IBI can be modified and validated for other large lakes, including the other Laurentian Great Lakes.

MATERIALS AND METHODS

The methodology that was used to develop the P-IBI followed Karr and Chu

(1997) and included five goals: 1) Classify environments to define homogenous sets, 2)

Select measurable attributes that provide reliable and relevant signals, 3) Develop sampling protocols, 4) Devise the multimetric index analytical procedures, 5)

Communicate results to citizens, policy makers, and other scientists.

Goal 1. Classify environments to define homogenous sets

The data used to develop the P-IBI were defined in both a spatially and temporally explicit fashion. Spatially, Lake Erie’s three basins vary widely from one another in morphometry (Bolsenga and Herdendorf 1993), temperature and seasonal stratification patterns (Lam and Schertzer 1987, Schertzer et al. 1987), and nutrient

(Yaksich et al.1985, Fraser 1987) and contaminant (Richards et al.1996) loadings. The ecology of the plankton in the three basins reflects these differences (Davis 1 964, Gannon and Beeton 1971).

The three basins of Lake Erie differ in a number of important physical characteristics. With respect to morphometry, Bolsenga and Herdendorf (1993) note that the western basin is the shallowest basin, with a mean depth of 7.4 m and a maximum depth of 18.9 m. Further, they found that only 12.8% of Lake Erie’s area and 5.1% of the

92 volume of Lake Erie are contained in the western basin. The central basin has a mean depth of 18.5 m and a maximum depth of 25.6 m, contains 62.9% of the lake’s area and

63.0% of its volume. The eastern basin is the deepest of the three basins with a mean depth of 24.4 m and a maximum depth of 64.0 m (24.3% of the lake’s area and 31.9% of

Lake Erie’s volume). Further, the lake basins differ in their water storage capacity, the amount of land they drain, and the mean tributary inflow into each basin. The western basin has the shortest water storage capacity (51 days), the largest drainage basin (37,000 km2), and the highest mean tributary inflow (5,300 m3/s). The central basin has the longest water storage capacity (635 days), drains 15,000 km 2, and has a mean tributary inflow of 200 m3/s. Finally, the eastern basin has a water storage capacity of 322 days, drains 6,800 km2, and has a mean tributary inflow of 200 m3/s (Bolsenga and Herdendorf

1993).

The three basins of Lake Erie also differ in temperature and seasonal stratification patterns (Schertzer et al. 1987). During a majority of the year, the western basin has the highest surface temperature (approx. 4-24 ºC) and vertically integrated water temperature

(approx. 4-24 ºC), and the lowest monthly average difference (approx. 2-3 ºC difference) between surface and bottom temperatures. The eastern basin has the lowest surface temperature (approx. 0-22 ºC) and vertically integrated water temperature (approx. 1-16

ºC), and the greatest monthly average difference (approx. 0-16 ºC difference) in surface and bottom temperatures. The central basin lies between the other two basins for these three quantities. The characteristics of each basin also differ with respect to thermocline formation and stratification. Seasonal thermoclines (Lam and Schertzer 1987) and

93 stratification (Bolsenga and Herdendorf 1993) occur yearly in both the central and eastern basins, while both are infrequent in the western basin. However, short periods of seasonal stratification of the western basin occurred during 1928-30, 1938-42, 1953, and

1963 (Carr et al. 1965). Thus, the western basin can stratify seasonally, though it does not do so in every year.

Contaminant and nutrient loadings also differ among the three basins of Lake

Erie. The western basin typically has higher loadings of the herbicides atrazine, alachlor, metolachlor, metribuzin, and cyanazine than does the central basin (Richards et al.1996).

Contaminant loadings of metals and PCBs in Lake Erie are reflected in decreasing concentrations of these contaminants in surficial sediments from west to east (Marvin et al. 2002). Further, phosphorus loading is typically greatest in the western basin

(including the Maumee, Sandusky, and Detroit River inflows), less in the central basin, and least in the eastern basin (Yaksich et al. 1985, Fraser 1987).

The interaction of basin morphometry, temperature and seasonal stratification patterns, and the history of nutrient and contaminant loadings in Lake Erie has had profound effects on the plankton communities of Lake Erie. First, the eutrophication of

Lake Erie in the 1950’s led to an alteration in the phytoplankton composition of the lake, with increased phytoplankton abundance and increases in taxa favored by high nutrient concentrations (Davis 1964). Further, the combination of increased nutrient loads, temperature and stratification patterns, and basin morphometry led to reoccurring anoxia/hypoxia in the central basin (Carr 1962, Charlton 1980). Further, the infrequent thermal stratification in the western basin has led to periodic low oxygen events at those

94 times (Britt 1955, Carr et al.1965). These low oxygen events in the western and central basin were hypothesized to have contributed to the decline of Limnocalanus macrurus, a calanoid copepod intolerant of high nutrient/ low oxygen conditions (Gannon and Beeton

1971). The interaction of a number of factors has thus affected Lake Erie plankton in each of its basins.

In addition to spatially defining Lake Erie by basin, calibration of the P-IBI is based on sets of temporal data from two distinct periods in the lake’s history (1970 and

1996). Within each of these years, data collected from May-September were analyzed, including zooplankton, phytoplankton, and nutrient/ trophic status data (i.e., total phosphorus and chlorophyll a concentrations). The 1970 data reflect Lake Erie during the height of its eutrophication (early 1970’s), while the 1996 data reflect Lake Erie at a more mesotrophic state (late 1990’s), as measured by chlorophyll a and total phosphorus concentrations. In Lake Erie, average (mean + standard error) uncorrected chlorophyll a concentrations (µg/L) were greater in 1970 than in 1996 in the western basin (12.58 +

1.82 > 5.40 + 0.22), the central basin (5.90 + 0.36 > 3.17 + 0.54), and the eastern basin

(5.17 + 0.38 > 1.67 + 0.18). Spatially, median uncorrected chlorophyll a concentrations generally declined from west to east in Lake Erie during each year (Figure 10a). Further, average (mean + standard error) total phosphorus concentrations (µg/L) were greater in

1970 than in 1996 in the western basin (41.53 + 2.68 > 29.75 + 1.39), and eastern basin

(14.84 + 0.82 > 7.74 + 0.28), but similar in the central basin (18.67 + 1.00 vs. 19.72 +

6.19). Spatially, median total phosphorus concentrations declined from west to east in

Lake Erie during both 1970 and 1996 (Figure 10b). Charlton et al. (1999) found similar

95 patterns in total phosphorus in Lake Erie. For example, summer total phosphorus concentrations decreased significantly in the three basins of Lake Erie between the periods 1968-1972 and 1994-1996 (Charlton et al. 1999). In the eastern basin, summer total phosphorus concentrations declined from 14.7 µg/L during 1968-1972 to 8.0 µg/L during 1994-1996. During the same time period, summer total phosphorus concentrations in the central basin declined from 14.1 µg/L to 8.4 µg/L. Declines in the western basin were the greatest, dropping from 41.2 µg/L during 1968-1972 to 17.7 µg/L during 1994-1996. Thus by 1994-1996, the eastern and central basins could be classified as oligotrophic, while the western basin would be considered mesotrophic, according to the classification system of Chapra and Dobson (1981). Thus the data used for the development of the P-IBI both were classified in a spatial (by basin) and temporal (by year and date) fashion.

Goal 2. Select measurable attributes that provide reliable and relevant signals

A number of plankton characteristics that can cause Beneficial Use Impairments

(BUIs) to the waters of Lake Erie (Table 1) were used in developing the P-IBI. To identify candidate plankton metrics, a USEPA (1998) technical document was first consulted to determine general types of plankton metrics (e.g., % abundance, % biomass of certain taxa, etc.) that could be used in a P-IBI. An extensive literature search then identified planktonic metrics that reflect BUIs and produced 7 candidate zooplankton metrics (Table 12) and 6 candidate phytoplankton metrics (Table 13). Candidate metrics were defined as individual metrics that would be considered further for inclusion in the P -

IBI. The 13 candidate metrics, their descriptions, ecological relevance, what they measure, hypothesized response to degradation, and literature references were used to select the metrics included in the final P -IBI (Tables 12, 13). More detailed descriptions of the candidate metrics follow below. Included in these descriptions are whether they were analyzed statistically for inclusion in the P -IBI.

1. Zooplankton Ratio

Gannon and Stemberger (1978) examined the utility of zooplankton communities as indicators of lake trophic condition. In general, calanoid copepods are indicative of more oligotrophic conditions, while cyclopoid copepods and cladocerans are indicative of more eutrophic conditions. Gannon and Stemberger used this information to develop an index based on the ratio of calanoid copepod abundance to cladoceran plus cyclopoid abundance to distinguish between lake trophic conditions. With this ratio, trophic conditions in Lake Michigan, the Straits of Mackinac, and Lake Huron could be distinguised. Greater values for the ratio (> 2) occurred in the more oligotrophic offshore waters, whereas lower values (< 2) were calculated for more eutrophic nearshore waters.

Similar trends in zooplankton community composition were evident on a cross-lake transect of Lake Michigan from Milwaukee, Wisconsin to Ludington, Michigan, with lower ratio values found in more eutrophic waters (Gannon 1975).

This crustacean zooplankton ratio also has been applied to Lake Erie

(Makarewicz 1993b, Johannsson et al. 1999b, Barbiero et al. 2001, Frost and Culver

2001). Johannsson et al. (1999b) found that 1970 crustacean ratio values (using April

97 and August abundance data from Watson 1976), averaged 0.13, 0.20, and 0.33 for the western, central, and eastern basins, respectively. These values were lower (and thus more eutrophic) than the averages of 0.27, 0.58, and 0.69 for the same three basins found by Makarewicz (1993) for the period 1983-1987, also calculated using April and August abundances. Further, Johannsson et al. (1999b) found that average ratio values were not significantly different in the western and eastern basin between the 1983-1987 and 1993-

1994 periods. However, with ratio values determined from combining spring (April) and summer (August) 1998 zooplankton abundances, Barbiero et al. (20 01) found a general increase in ratio values (0.35, and 0.86 for the western and central basins, respectively), compared to the 1983-1987 and 1993-1994 periods. These changes in the abundance ratio reflect the changes in nutrient concentrations in Lake Erie. Because spring total phosphorus concentrations have declined in the central basin of the lake from 15.1 µg/L in 1970 to approximately 10 µg/L in 1991 (Bertram 1993), we know water quality improved. Over the same period, the crustacean ratio increased , suggesting it is an individual metric whose possible use was analyzed statistically.

2. Mean Zooplankton Length

In freshwater ecological research, the study of the factors influencing zooplankton size has contributed greatly to general ecological theory (Lampert 1997). For example, the size-efficiency hypothesis, which seeks to explain the relative influence of size- selective fish predation and competition among herbivorous zooplankton for algae, on zooplankton size distribution (Brooks and Dodson 1965), has been an important

98 paradigm in freshwater zooplankton ecology (Lampert 1997). This hypothesis states that during times of low predation, the zooplankton community will be dominated by larger zooplankters, that are more efficient feeders and thus can outcompete smaller zooplankters for algae. Conversely, during times of high fish predation, smaller zooplankters are more abundant because zooplanktivorous fish feed selectively on larger zooplankters. Finally, under moderate predation, large and small zooplankters will co- occur. Accordingly, mean zooplankton length has been suggested as a metric for evaluating the relative health of fish communities, e.g., the balance of predators and prey.

Mills and Schiavone Jr. (1982) compared mean crustacean zooplankton length with piscivorous and planktivorous fish abundances. When planktivory by fish was high, the mean zooplankton size was hypothesized to be lower than when piscivory removed planktivorous fish from these lakes. They found that in lakes where piscivorous predators controlled planktivore density, mean body lengths of crustacean zooplankton were greater than 1.0 mm, while mean body lengths of crustacean zooplankton (copepods and cladocerans) were lower when piscivores did not control planktivores. Mills et al.

(1987) expanded and refined the use of mean crustacean length and found that lakes with

< 0.8 mm crustacean mean length had piscivore to planktivore abundance ratios of approximately 0.2 or less. Thus in lakes with low piscivory, planktivores can select larger zooplankton and reduce the mean length of zooplankton. Mills et al.’s method has been modified for a variety of situations, such as the use of the mean size of Daphnia ephippia to reconstruct planktivorous fish abundance in paleolimnological studies (Jeppesen et al.

2002), and the correction of the method for plankton nets of different mesh size

(Johannsson et al. 1999a).

Crustacean zooplankton mean size and Daphnia mean size measures are less useful in a planktonic IBI. First of all, the abundance of planktivorous fish may not be linked to degraded conditions (as measured by lake trophic status). Canfield and Jones

(1996) found no significant correlation between zooplankton size-distribution and trophic status, as measured by algal chlorophyll concentrations, of midwestern lakes. Further, the abundance of planktivorous fish is not the only factor that is important in influencing mean size of crustacean zooplankton. Blooms of inedible algae may cause a shift towards smaller zooplankton, because large algae cause greater interference with the feeding of large cladocerans than small algae (Gliwicz 1980, Vaga et al. 1985). In addition, the abundance of zooplankton in north temperate lakes varies tremendously seasonally, and large herbivorous zooplankton can overgraze the phytoplankton (the clear-water phase) (Sommer et al. 1986). Subsequently, the abundance of large, herbivorous zooplankton declines, while smaller taxa may increase in abundance. Thus, both top-down (e.g., planktivory by fish) and bottom-up factors (e.g., nutrient concentrations and phytoplankton abundance and composition) play a role in determining the size of zooplankton. Because it is impossible to disentangle these confounding factors, the mean size of zooplankton is not a reliable signal of degradation or even the piscivore to planktivore ratio and will not be used as a metric for the P -IBI.

100

3. Rotifer Composition

Rotifer taxonomic composition has been a widely applied indicator of lake trophic status. Gannon and Stemberger (1978) found that certain rotifer species assemblages were associated with eutrophic, mesotrophic, and oligotrophic conditions in Lake Huron and Saginaw Bay. Taxa such as Brachionus spp., Filinia longiseta, Keratella cochlearis, and Trichocerca multicrinis were associated with eutrophic waters, while other rotifer taxa were associated with more mesotrophic or oligotrophic conditions (i.e., Kellicottia longispina, Keratella hiemalis, Polyarthra dolichoptera) (Gannon and Stemberger 1978).

Makarewicz (1993b) used rotifers as indicators of eutrophic lake conditions in Lake Erie and found that during 1983-1987, Brachionus spp., Filinia longiseta, and Trichocerca spp. were significantly more abundant in the western basin, or that thes e taxa only occurred there. Rotifers have also been used to measure trophic status in Swedish lakes

(Berzin and Pejler 1989) and even estuaries (Attayde and Bozelli 1998). Rotifer genera also were used as a part of a Wetland Zooplankton Index, developed using data from throughout Great Lakes wetlands (Lougheed and Chow-Fraser 2002).

Nevertheless, the inclusion of rotifers in a Planktonic Index of Biotic Integrity for

Lake Erie is problematic. First of all, the identification of rotifers is a specialize d task that requires a great amount of taxonomic understanding and effort. The second and more severe problem is that the method of sampling that is used to collect typical zooplankton samples does not allow for accurate sampling of rotifers. Plankton nets with a mesh size of 60+ µm do not sample rotifers quantitatively or even qualitatively, allowing many of the smaller and softer rotifer taxa to pass through the net. In a

101 comparison of different rotifer sampling techniques in Lake Ontario, Mazumder et al.

(1992) found that rotifer abundance was two to three times higher when collected on a

30-µm screen than when collected with a 64-µm plankton net. Although rotifers reflect trophic status, sampling methods used to collect zooplankton in this study do not allow for the use of rotifers as part of the P-IBI for Lake Erie.

4. Density of Limnocalanus macrurus

The calanoid copepod Limnocalanus macrurus is intolerant of low dissolved oxygen levels. Strøm (1946) found that during summer stratification L. macrurus was restricted to areas where dissolved oxygen was more than 5.6 mg/L and was most abundant where the water contained 6.4 mg O2/L (Strøm 1946). Gannon and Beeton

(1971) suggested that low dissolved oxygen brought about by eutrophication in Lake Erie during the 1960’s most likely led to a drastic reduction of L. macrurus densities. Low oxygen or anoxia was recorded throughout the 1950s-1970s in the western basin (Britt

1955, Beeton 1961) and central basin (Charlton 1980). In the deep waters of the eastern basin, oxygen levels were as low as 5.5 mg/L in September 1961 (Carr 1962).

Limnocalanus abundances declined as Lake Erie became more eutrophic and oxygen levels declined. Gannon and Beeton (1971) conducted an extensive review of

Limnocalanus abundance and distribution in Lake Erie. Before the height of cultural eutrophication, Limnocalanus was quite abundant and widespread. It was present in nearly every site sampled in the central and eastern basins during June 1929, comprising

74% of the total microcrustacean community in the eastern basin, and was particularly

102 abundant in bottom sites (Fish 1960). In February and May of 1939, however, fewer than

100 individuals/m3 were found in the Bass Island region of the western basin (Chandler

1940). Limnocalanus was collected in the central basin at densities of 300 and 700 individuals/m3 in January and May of 1951. However, once eutrophication increased,

Limnocalanus abundance decreased. In fact, by 1957 Limnocalanus was found at abundances of only 1-20 individuals/m3 in the western basin, and less than 1 individual/m3 in the central and eastern basins. By the late 1960’s Limnocalanus was collected infrequently in Lake Erie, even with intensive sampling. Similarly, mean cruise abundances of 30 or more individuals (juveniles + adults) per cubic meter were only encountered twice in 1970 (early May and early July) during a 10-month lake-wide study

(Watson and Carpenter 1974). By the mid- 1980’s, Limnocalanus was either in extremely low numbers or absent in lake-wide offshore zooplankton studies for 1983-87

(Makarewicz 1993b). However, with the implementation of nutrient abatement programs, Limnocalanus abundances in Lake Erie have increased. Limnocalanus macrurus densities in the western basin during th e late 1990’s have returned to nearly the same magnitude of average spring densities seen in 1939 (10 -100 individuals /m3). At times, densities reached several hundred individuals/m3 and on one occasion in 1996, reached over 1000 individuals/m3 (Kane et al. 2004, Chapter 4). Because of its known intolerance to low oxygen conditions and the historic record of Limnocalanus decline concomitant with increased eutrophication in Lake Erie, Limnocalanus density is a good candidate metric and its possible use in a multimetric index was analyzed statistically.

103

5. Percentage Biomass of Invasive Predatory Cladocerans (Bythotrephes and

Cercopagis)

Two invasive, predatory cladocerans may have the ability to alter food web structure in Lake Erie. Bythotrephes longimanus, which invaded Lake Erie in the mid

1980’s (Bur et al. 1986), and Cercopagis pengoi, which invaded Lake Erie in 2001

(Therriault et al. 2002, Kane et al. 2003b, Chapter 3), are large, predatory cladocerans.

These confamiliar genera can have profound effects on the dynamics of aquatic food webs. Native predaceous cladocerans (Leptodora kindti), and the herbivorous Daphnia spp. decreased contemporaneously with the invasion of Bythotrephes into Lake Michigan

(Lehman and Caceres 1993). Others have suggested that predation by Bythotrephes has influenced plankton dynamics, including the possibility that Daphnia abundance can be controlled by Bythotrephes predation (Hoffman et al. 2001). Further, Cercopagis also has been implicated in causing decreases in crustacean zooplankton abundance in Finland

(Uitto et al. 1999). Decreased zooplankton abundance due to Bythotrephes and

Cercopagis predation may reduce zooplanktivorous fish recruitment by reducing food available to fish such as yellow perch (Perca flavescens) (Hoffman et al. 2001).

Further, Cercopagis may also compete with zooplanktivorous fish and increase food chain length (Vanderploeg et al. 2002).

The percentage biomass of Bythotrephes and Cercopagis that these two taxa contribute to the total cladoceran biomass will be used as a candidate metric. Biomass will be used because these cladocerans are large, and even when they are low in abundance they may contribute substantially to the overall cladoceran biomass. Further,

104 dividing by the total cladoceran biomass standardizes between samples, allowing the ability to distinguish what these taxa contribute to the total cladoceran biomass.

Presumably, when biomass of these predaceous cladocerans makes up a high percentage of the total cladoceran biomass, they may have a greater potential or realized predatory impact on other zooplankton taxa. Since these two cercopagid cladocerans may both cause a decrease in zooplankton and cause the elongation of food chains, their inclusion in a Planktonic Index of Biotic Integrity should be considered. Thus the percentage biomass of these taxa was a metric whose use was analyzed statistically.

6. Ratio of Zooplankton Biomass to Phytoplankton Biomass

The ratio of zooplankton biomass to phytoplankton biomass was proposed as an indicator of lake ecosystem health for the Ecological Modeling Method (EMM) developed by Xu et al. (2001). This ratio is very low during algal blooms and more eutrophic conditions, due both to an increase in phytoplankton biomass and a decrease in zooplanktion biomass due to interference to zooplankton grazing caused both by chemicals produced by the algae and large filament size. Further, this ratio decreases under other types of chemical pollution. For example, certain pesticides that kill zooplankton (e. g. endosulfan (Barry and Logan 1998), and pyridaben (Rand et al. 2001)) indirectly cause algal blooms by releasing algae from grazing pressure. Also, the ratio of zooplankton biomass to phytoplankton biomass is also a measure of energy transfer efficiency between lower and upper trophic levels (Havens 1998). When the ratio is low, less energy is available to trophic levels above the zooplankton. Due to the lack of rotifer

105 biomass data from 1970, the current high abundance of zebra mussel veligers, and the inaccurate sampling of rotifers, only crustacean zooplankton will be used as an estimate for zooplankton biomass. The ratio of crustacean zooplankton biomass to total phytoplankton biomass is thus a candidate metric that reflects lake trophic status, available energy to upper trophic levels, and may reflect other stressors on lake ecosystems. Therefore, the possible use of this candidate metric was analyzed statistically.

7. Biomass of Crustacean Zooplankton

Crustacean zooplankters are important components of the diet of a number of larval fishes, as well as some adult fishes. Fish typically select for large, crustacean zooplankton in their diet (Brooks and Dodson 1965). For example, the composition, abundance, and biomass of the crustacean zooplankton community can affect growth and survival of larval gizzard shad (Dorosoma cepedianum) (Bremigan and Stein 1999).

Further, crustacean zooplankton biomass comprises a large proportion of the diet of Lake

Erie young-of-year trout-perch (Percopsis omiscomaycus), freshwater drum (Aplodinotus grunniens), white perch (Morone americana), yellow perch (Perca flavescens), gizzard shad (Dorosoma cepedianum), alewife (Alosa pseudoharengus), emerald shiner (Notropis atherinoides), spottail shiner (Notropis hudsonicus), and rainbow smelt (Osmerus mordax) (Gopalan et al. 1998) and in Lake Erie adult freshwater drum (Aplodinotus grunniens) (Bur 1982).

106

Crustacean biomass is not only important to fish as food, but has changed in Lake

Erie with changes in nutrient loading. In 1970, at the height of eutrophication, crustacean dry weight biomass, averaged across May-October, in the western Lake Erie was over 30

µg/L, while by the mid-1980’s zooplankton biomass had dropped below 10 µg/L

(Chapter 2). This decrease in zooplankton biomass reflected decreases in loading of total phosphorus to the lake (Dolan 1993). Crustacean zooplankton biomass can be used to track changes in the trophic status of Lake Erie and is important as fish food, thus the possible use of crustacean zooplankton biomass as a metric was analyzed statistically.

8. Generic Index of Diatoms

Wu (1999) developed a Generic Index of diatom assemblages to characterize organic pollution in the Keelung River of Taiwan. The Generic Index (GI) consists of the ratio of abundance of Achnanthes, Cocconeis, and Cymbella to Cyclotella, Melosira, and

Nitzschia. The genera in the numerator were found to be more abundant in unpolluted environments, whereas the genera in the denominator were more abundant in polluted environments. Generic Index was negatively correlated with such measures as total phosphorus and turbidity and positively correlated with dissolved oxygen concentrations.

Thus, higher GI values indicated better water quality conditions. The GI also was calculated separately using both planktonic and epilithic diatoms, with little difference in the final GI value between the sampling techniques.

107

Although a generic assessment of diatoms may be a good metric for measuring water quality in rivers (see also Chessman et al. 1999), it has not been tested in lakes.

Further, some of the taxa used to calculate the ratio are benthic (i.e., Achnanthes and

Cocconeis) (Prescott 1978). This may not be a problem in a shallow system where one may be able to accurately sample tychoplanktonic taxa. However, in a deeper lake system, such taxa may be underrepresented in the plankton. Because the Generic Index of diatoms is designed for river systems and Lake Erie has deep areas it will not be a par t of the multimetric P-IBI.

9. Centrales/ Pennales Ratio

Nygaard (1949) presented a diatom quotient that could be used to distinguish eutrophy and oligotrophy in lakes. This quotient consists of the ratio of centric diatom abundance to pennate diatom abundance (Centrales/ Pennales). Typically centric diatoms

(e.g., Melosira) are more abundant in eutrophic situations, and thus higher values of this ratio indicate more eutrophic conditions. Rawson (1956) investigated the use of diatom and other phytoplankton quotients in the Great Lakes and western Canadian lakes and found that the presence of certain indicator taxa could be used to judge lake trophic status. Further, Stockner and Benson (1967) used the relative percentages of the

Araphidineae and Centrales in the sediments of Lake Washington to measure changes in lake trophic status. From a practical standpoint, the use of the ratio of centric to pennate diatoms may be advantageous because identification of diatoms to tribe, genus, or species

108 is not necessary to calculate this ratio. Thus the use of this ratio will be analyzed statistically for consideration for the P-IBI.

10. Biomass of Inedible Algae

In general, large, colonial, filamentous, spiny, or toxic algae are poorly grazed by crustacean zooplankton (DeMott 1989, Sterner 1989). Cladocerans are unable to ingest large algal cells. Indeed, the presence of large particles can interfere with feeding on all algal taxa (reviewed in Sterner 1989). Large colonial and filamentous algae were found to decrease filtering rates in large cladocerans, as compared to smaller cladocerans

(Gliwicz 1980). Further, growth, survivorship, ingestion, and filtering rate of Daphnia were lower and rejection rates were higher when fed a toxic strain of Anabaena flos- aquae, as compared to Daphnia fed Chlamydomonas reinhardi (Porter and Orcutt 1980).

The authors concluded that A. flos-aquae had nutritional, handling, and toxic properties that made it a poor food source for Daphnia.

The edibility of algal taxa clearly has important effects on crustacean zooplankton. Taking into account size, morphology, colonial organization, and potential toxic properties, Lake Erie algal taxa can be classified as edible or non -edible. The biomass of inedible phytoplankton taxa will be analyzed statistically for possible inclusion in the P-IBI.

109

11. Percentage Biomass of Blue-Green Algae

Because of their ability to reduce filtering rate, growth, and survival in crustacean zooplankton, cyanophytes affect aquatic ecosytems (see above). Lampert (1981) found that blue-green algae could have direct, toxic effects on Daphnia. Further, Microcystis aeruginosa decreased filtering rates in a variety of cladocerans (Lampert 1982). DeMott and Moxter (1991) found that copepods rejected toxic cyanobacteria on the basis of their toxicity and not their filamentous nature. Although the response of crustacean zooplankton to blue-green algae is variable and depends on both the properties of the inedible alga taxon and the zooplankton taxa, th e response is usually negative. Further, large blooms of cyanophytes may inhibit transfer of energy to herbivorous zooplankton and subsequently to zooplanktivorous fish (Havens 1998). The eutrophication of lakes favors many cyanophyte taxa that are superior competitors at high phosphorus concentrations (Schindler 1977, Smith 1979) or low N:P ratios (Smith 1983). During the eutrophication of Lake Erie, cyanophyte taxa became more abundant (Davis 1964).

Havens (1999) suggested the percentage biomass of the Cyanophyta compared to the total biomass of phytoplankton as a potential metric of energy transfer in plankton communities. Higher percentages of cyanophytes negatively affect grazers, and cause a reduction in energy transfer to higher trophic levels. However, because this metric is redundant to the percentage biomass of Microcystis, Anabaena, and Aphanizomenon compared to total phytoplankton biomass (see analysis below), it will not be used in the

P-IBI.

110

12. Percentage Biomass of Microcystis, Anabaena, Aphanizomenon

The cyanophyte genera Microcystis, Anabaena, and Aphanizomenon are bloom- forming phytoplankters (Paerl 1988) that have the ability to produce toxins (Carmichael

1986). These taxa increased in abundance as Lake Erie became more eutrophic i n the middle of the 20th century (Davis 1964). As mentioned above, these taxa have profound effects on zooplankton survival and reproduction. Further, toxins produced from these taxa have the ability to reach upper trophic levels (Babcock-Jackson et al. 2002). Large blooms of Microcystis have occurred in the western basin of Lake Erie in recent years

(Budd et al. 2002). The recent Microcystis blooms have contained hepatotoxins that may have effects on fish and human health (Brittain et al. 2000).

An analysis of the composition of the cyanophyte community in Lake Erie during

1996 showed that when blue-green algae are present in Lake Erie, Microcystis,

Anabaena, and Aphanizomenon compose a large proportion of the cyanophyte biomass

(mean = 70%, median = 100%). Further, the relationship between cyanophyte biomass and biomass of these three taxa is highly significant (p < 0.001) and strongly correlated

(r2 = 1.00). Therefore, the biomass of these three taxa will be used in the P -IBI, rather than the biomass of all blue-green taxa. Because of their documented response to eutrophication and negative impacts on zooplankton, fish, macroinvertebrates, and humans, the cyanophyte genera Microcystis, Anabaena, and Aphanizomenon will be analyzed statistically for possible inclusion in the P -IBI.

111

13. Biomass of Edible Algae

Algal particles that are approximately 3-20 µm in length and that lack protective covering or spines are highly edible by cladocerans (Sterner 1989). Further, bead experiments reveal that particles >40 µm can be grazed by large cladocerans (Sterner

1989). Thus smaller taxa of phytoplankton, such as cryptophytes and small chlorophytes, as well as some diatoms, are also edible to cladocerans. Havens (1998) suggests the percentage biomass of nanoflagellates compared to total phytoplankton biomass as a potential metric of energy transfer in plankton communities. For the P-IBI for Lake Erie, the biomass of all edible taxa will be used as a candidate metric. Further, the absolute biomass of the edible algae will be used, because this is the biomass that is a vailable to grazers. Thus this candidate metric will be analyzed statistically for possible inclusion in the P-IBI.

Goal 3. Develop sampling protocols

Plankton, total phosphorus, and chlorophyll a data used to develop the P-IBI came from a diversity of sources. 1970 monitoring data came from M.N. Charlton, National

Water Research Institute (NWRI), Environment Canada (nutrient concentration and zooplankton abundance), D. Bean (1980) (zooplankton biomass), and M. Munawar,

Department of Fisheries and Oceans Canada (phytoplankton). 1970 Lake Erie field and lab protocols are described in Burns (1976a, 1976b) and the “STAR” methods dictionary

(NWRI, Environment Canada) for nutrients, in Glooschenko et al. (1974) for chlorophyll

112 a, in Munawar and Munawar (1976) for phytoplankton, and in Watson and Carpenter

(1974) and Watson (1976) for zooplankton. 1996 plankton data were based on samples collected by the Ohio Division of Wildlife and the NWRI and were analyzed by the Ohio

State University as part of the Lake Erie Plankton Abundance Study (LEPAS) (Frost and

Culver 2001), while 1996 nutrient data were obtained from M.N. Charlton, NWRI. 1995 and 1997-2002 plankton sampling and analyses also followed procedures outlined in

Frost and Culver (2001). 1970 and 1996 Lake Erie sampling sites used in Planktonic

Index of Biotic Integrity development were located in all three basins (Figure 11) to reflect spatial gradients in total phosphorus and chlorophyll a (Figure 10).

To facilitate comparisons of plankton data from this multitude of sampling dates, sampling frequency, number of sites, sampling method, enumeration method, and biomass calculation method for phytoplankton and zooplankton in Lake Erie (1970 and

1996), careful comparisons of methods were made (Table 14). More detailed descriptions follow.

Phytoplankton

In 1970, samples were collected from April to December at four-week intervals in all three basins of Lake Erie from 25 different sites (Munawar and Munawar 1976, Figure

11). Samples were collected in a VanDorn bottle from 1- and 5-m depth, mixed, and then preserved with Lugol’s solution. Another subsample was used to identify motile phytoflagellates (Munawar and Munawar 1976). Phytoplankton samples were collected weekly to monthly in all three basins of Lake Erie during 1996 from May –October

113

(Figure 11, Table 14). The western basin was sampled the most intensively, with the greatest number of sites (n = 51) and samples (n = 344) taken. The central basin had fewer sites (n = 39) and samples (n = 71) taken than the western basin, while the eastern basin had the least number of sites (n = 5) and samples (n = 11) taken.

Phytoplankton enumeration for both 1970 and 1996 generally followed Utermöhl

(1958). For the 1970 samples, aliquots of 5-25mL, depending on phytoplankton density, were settled and examined with an inverted microscope (Wild Heerbrugg M40, phase contrast). Common netplankton species were counted in two transects under 300x magnification. Less common and rare species were enumerated in the entire chamber at

HYRAX mounting medium (Munawar and Munawar 1976). For 1996, sample jars were inverted 25 times to fully mix samples and 250mL from each jar was poured into a graduated cylinder, and allowed to settle for 3 days in a dark-chamber. Each sample was then concentrated down to 30mL by siphoning off liquid from the top and transferring the remaining sample to a 36.97mL vial. Mixed subsamples of approximately 3- to 5mL were obtained from the concentrated samples and placed into a countin g chamber, and then weighed to determine the volume precisely. All phytoplankton genera were identified and counted using a Wild inverted microscope at 400x. Repeated transects were counted until 100 algal units (cells, filaments, or colonies) of the most common taxa were recorded. In all samples, however, all algal units in at least two transects were

114 counted for each sample even if 100 algal units of the most common taxa were enumerated before the full two transects were completed. Dimensional measurements were recorded for the first 20 algal units for each taxon enumerated. For filamentous algal taxa, however, all filament lengths were measured, summed, and recorded as the total filament length for each taxon (Frost and Culver 2001).

Total biomass of phytoplankton samples was determined in both 1970 and 1996 by summing species-specific total biomass over all species present in a particular sample.

Cell volume was computed for each species present using the mean algal dimensions for each species in a sample. These average dimensions were then used in volumetric equations that best described the shape of each species. For colonies, the mean number of cells per colony was calculated and multiplied by the average volume per cell to determine volume per colony. Subsequently, volumes were converted to biomass assuming the specific gravity of phytoplankton to be 1.0 (Munawar and Munawar 1976,

Frost and Culver 2001). Consequently, all reported biomasses are wet weights (µg/L) of phytoplankton.

Zooplankton

In 1970, zooplankton samples were collected at 33 sites located in all three basins of Lake Erie on 10 cruises during April through December. These sites generally overlapped the phytoplankton sampling sites (Figure 11), but not always. Samples were collected with a 64µm mesh, 0.4m diameter unmetered net, using single vertical hauls from 2m above the bottom (or 50m from surface). These samples were preserved in 4%

115 formalin (Watson 1976). Zooplankton samples were collected weekly to monthly in all three basins of Lake Erie during 1996 from May –October (Figure 11, Table 14). The western basin was sampled the most intensively, with the greatest number of sites (n =

51) and samples (n = 486) taken. The central basin had fewer sites (n = 46) and samples taken (n = 87), while the eastern basin had the fewest sites (n = 5) and samples (n = 14) taken.

For 1996, vertical tow samples were collected using a front-weighted zooplankton net (0.5m diameter, 64µm mesh) fitted with a General Oceanics 2030R model flow meter and 500mL jar. The net was lowered with the open end pointing downward until the 2kg weight fastened to the front bridle by a 1m line hit bottom. The net was then retrieved, allowing the water column to be sampled both as the net was lowered and a s it was pulled up, while avoiding collecting mud from the bottom (Frost and Culver 2001). Samples were then concentrated and preserved with a 4% sugar formaldehyde solution (Haney and

Hall 1973).

For 1970 samples, 1mL subsamples were analyzed under inverted microscope until 200 individuals of each taxon were enumerated (Watson 1976). Numbers/m3 were calculated from the percent of sample counted, assuming 100% net efficiency, tow depth, and net diameter (Watson and Carpenter 1974). However, the net efficiency of plankton nets is seldom 100%, due to net shape, mesh size, mesh area, netting porosity, filtering area, and the mesh area to mouth opening ratio (Sameoto et al. 2000) and thus abundance estimates from 1970 are likely underestimates. Adult calanoid and cyclopoid copepods were identified to species and sex, while nauplii were enumerated but not identified to

116 species. Cladocerans were identified to species and sex, when possible (Watson 1976).

Length measurements of 10-50 individuals (based on percent abundance) and egg densities were obtained at a later time (Bean 1980). The number of eggs per female of a taxon was determined by counting the number of eggs and dividing by the number of females to obtain eggs per female. This number was multiplie d by the number of females/m3 to give the eggs/m3 for each taxon (Bean 1980). For samples taken during

1996, each sample was diluted to a known volume, typically from 500mL to 3000mL unless the sample contained an extremely small or large amount of zoopla nkton, requiring smaller or larger dilution volumes. After dilution, all zooplankton in at least two subsamples of 5- to 10mL were identified and enumerated. All enumeration was done using a Wild M5A dissecting microscope fitted with a calibrated ocular micrometer for body measurements (±0.05mm). Cladocerans and copepods were identified to species and sex while rotifers were identified to genus. Additional subsamples were analyzed until at least 100 individuals of the most common taxon were recorded (Frost and Culver 2001).

For 1970 and 1996 samples, zooplankton biomass was calculated by determining the average individual biomass for each species counted and multiplying by the number of individuals per cubic meter. Average individual biomass was determined using length-weight regressions developed by Culver et al. (1985) and the lengths measured from historical samples (1970) or during sample enumeration (1996). Species-specific regression equations converted from length (in mm) to dry-weight biomass (µg) values for each crustacean zooplankton taxon (including eggs). Species-specific total biomasses

117 were summed over all taxa giving the total crustacean zooplankton biomass at a given sampling site, for a given date.

The sampling and analysis of phytoplankton and zooplankton in 1970 and 1996 are, in general, similar. Both studies sampled during an overlapping time frame (May -

October) and in all three basins of Lake Erie (Figure 11). Further, most sites were sampled at least monthly in both 1970 and 1996. In addition, a number of 1970 sites were located close to 1996 sites, if not overlapping (Figure 11). The 1996 sampling regimen lacked sites in the middle of the central basin; however, a more inshore site and an offshore mid-basin site (both sites approximately 20 m deep, a depth previously used to delineate offshore/nearshore in Lake Ontario (Hall et al. 2003)) in the central basin did not differ (t-test, p > 0.10) for zooplankton and phytoplankton biomass in 1997 (Chapter

2). Thus more inshore sites served as effective surrogates for offshore sites of the same depth. Although there are some differences, similar sampling regimens (multiple sites in all three basins over a number of months), sampling techniques (zooplankton net hauls to near bottom (1 m above bottom- 1996, 2 m above bottom- 1970), sampling upper meters of water column for phytoplankton), and analytical procedures (Utermöhl (1958) technique for phytoplankton enumeration, Culver et al. (1985) length-weight regressions for zooplankton biomass calculations) allow the use of phytoplankton and zooplankton data from 1970 and 1996 in the construction of a P -IBI for Lake Erie.

118

Goal 4. Devise the multimetric index analytical procedures

Metric Selection

I used data from 1970 and 1996, representing a broad range of conditions, trophic status, and candidate metrics. After examining the literature and identifying data availability, four of the candidate metrics were dropped from subsequent analysis. Thus,

9 of the original 13 candidate metrics (Tables 12,13) were included in the discriminant analysis used to form the multimetric index: zooplankton ratio, abundance of

Limnocalanus macrurus (#/L), % biomass of invasive zooplankters of the total crustacean biomass, biomass of crustacean zooplankton/ biomass of phytoplankton, biomass of crustacean zooplankton, abundance of centrales/abundance of pennales, biomass of inedible algae taxa, % biomass of Microcystis, Anabaena, and Aphanizomenon of the total phytoplankton biomass, and biomass of edible algae taxa. The data for the nine candidate metrics came from individual sites on individual sampling dates, except for the

1970 Limnocalanus macrurus abundance data, which were determined by using monthly, lakewide (all three basins) cruise averages of juveniles + adults from Watson and

Carpenter’s (1974) Table 2. This approximation was necessary due to the lack of site and date specific Limnocalanus macrurus density data for 1970.

Unfiltered total phosphorus concentrations (µg P/L), and chlorophyll a, uncorrected for pheophytin, concentrations (µg/L) from 1970 (STAR, NWRI) and 1996

(STAR, NWRI and LEPAS, OSU) were used to classify sites in Lake Erie with respect to lake trophic status (i.e., oligotrophic-mesotrophic-eutrophic continuum) (Leach and

Herron 1992) and thus reflect levels of degradation. Another measure of lake trophic

119 status, Secchi depth, was not used both because it is influenced by sediment particles

(Leach and Herron 1992) and because not all of the 1970 sites sampled had Secchi transparency data. For both 1970 and 1996, total phosphorus and chlorophyll a data only were used when taken in conjunction with zooplankton and phytoplankton samples from the same sampling site on the same date. Phytoplankton abundance and biomass data from 1996 and for most of 1970 were taken simultaneously with the nutrient, chlorophyll a, and zooplankton abundance and biomass data obtained from the same site and date.

On a few occasions in 1970, phytoplankton was not collected at a site where nutrient, chlorophyll a, and zooplankton data were collected. In such cases, phytoplankton data from an adjacent site (based on Burns 1976b) within the same basin and on the same cruise were used in conjunction with the nutrient, chlorophyll, and zooplankton data.

This extrapolation was performed on less than 30% of the individual samples (49 out of

169), at less than 30% of the sites (9 out of 32) sampled in 1970.

The trophic status classes (i.e., oligotrophic, mesotrophic, or eutrophic) assigned to total phosphorus and chlorophyll a levels were based on values determined by Chapra and Dobson (1981) for the offshore waters of the Great Lakes (Table 15). Total phosphorus and chlorophyll a concentrations that were classified as eutrophic received a metric value of 1, those classified as mesotrophic received a 3, and those classified as oligotrophic received a 5. These two metrics were then summed to form a trophic status metric that could have values of 2, 4, 6, 8, or 10. Trophic status metric values of 8 or 10 were classified as oligotrophic, values of 6 were classified as mesotrophic, and values of

2 or 4 were classified as eutrophic (Table 15). I created discrete values for trophic status

120 in order to classify a sample into a predetermined class for discriminant analysis (see below).

A multivariate statistical technique (discriminant function analysis or discriminant analysis (DA)) was used to evaluate the ability of plankton metrics to distinguish among levels of degradation. DA discriminates among pre-specified groups of samples based on a suite of variables to find gradients among groups of samples, such that variation among groups is maximized, while within group variation is minimized (McGarigal et al. 2000).

Discriminant analysis has been identified as an acceptable statistical method for the development of Indices of Biotic Integrity and can be used to identify variables that can discriminate between levels of degradation (USEPA 1998). Further, discriminant analysis has been used to develop IBIs for benthic (Engle and Summers 1999, Christman and Dauer 2003) and fish (Meng et al. 2002) communities and has been used to classify phytoplankton (Sugiura et al. 2002) and zooplankton (Shaw and Kelso 1992) communities with respect to water quality parameters.

Discriminant analyses (SAS Version 8e, SAS Institute, 1999-2001) were performed using PROC DISCRIM with prior probabilities set proportional to the number of observations in each class with all the metrics included (full model). PROC

STEPDISC with the stepwise command and significance level-to-enter/ significance level-to-remove set to 0.20 was then used to determine which variables were significant.

PROC DISCRIM was then run with only the significant metrics found from the PROC

STEPDISC procedure (reduced model). A separate discriminant analysis was performed for each month, from May-September due to the temporal variability of both zooplankton

121 and phytoplankton communities in Lake Erie (Table 16). The jackknife cross-validation procedure in SAS was used and a weighted Cohen’s Kappa statistic was calculated from these cross-validations in order to judge the accuracy of classification. Cohen’s Kappa is a statistical measure of the agreement of two raters or two rating methods. A weighted

Cohen’s Kappa statistic was calculated from the jackknife cross-validation results in order to provide an unbiased result, as suggested by McGarigal et al. (2000).

Significance for the weighted Cohen’s Kappa was judged at α= 0.05 for hypothesis testing. A 95% confidence interval was calculated using PROC FREQ (SAS Institute

1999-2001). The weighted Cohen’s Kappa was used to weight close misses in classification more heavily than misclassifications that are further apart (e.g., classifying something that is oligotrophic as mesotrophic (close misclassification) vs. classifying something as oligotrophic when it is eutrophic (further misclassification)) (Table 16).

Further, the specific metrics that were found to be significant in each month were also determined (Table 17).

Correlation between Significant Metrics and Total Phosphorus and Chlorophyll a

Regression analyses (Minitab Version 13.1, 2000) for each of the significant metrics against the phosphorus and chlorophyll a concentrations measured at the same time and place were used to determine whether the values for each of the metrics increased or decreased with increasing trophic status (i.e., more eutrophic or degraded conditions) (e.g., Figure 12). This technique of determining the relationship between a continuous individual metric and a discrete variable has been applied before to IBIs

122

(Fausch et al. 1984, Fore et al. 1996). Similarly, Miltner and Rankin (1998) assigned discrete codes to two measures of degradation. They used total inorganic nitrogen levels and total phosphorus levels to create a discrete scale (1, 2, 3, 4, 5, 6) upon which they plotted general trends of individual fish and invertebrate metric values. Finally, the five statistically significant metrics were totaled for each month in which they were found significant. Further, reasons for inclusion/ rejection of all candidate metrics are given

(Table 18).

Individual Metric Cumulative Frequency Distribution

To calculate individual metric scores, “boxplots” of the significant individual plankton metrics frequency distributions were constructed. Using the 95 th percentile as the upper boundary and zero as the lower boundary (Karr et al. 1996), each of the final individual metrics included in the multimetric P-IBI (based on significance in the discriminant analyses) was trisected into ranges that were assigned a score of 1, 3, or 5

(Table 19). The trisection technique is used across the whole variation of a metric to standardize or normalize the distribution of a metric. Thus, metrics with formerly different scales can be analyzed on a consistent scale (Barbour et al. 1995). Metrics that have a 0.00 value in the 1 scoring class (e.g., zooplankton ratio, Table 19) the cu toff minimal values should be read from left to right, while those with a 0.00 value in the 5 scoring class should be read from right to left. A 5 was always given to the least degraded (most oligotrophic) range of this trisection, while a 1 was always given to the most degraded (most eutrophic) range of the trisection. As in Figure 13, the horizontal

123 lines on the boxplots indicate the metric values associated with the 25 th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentiles (Figure 14). Boxplots also were constructed of the metric data versus trophic status (as defined in Table 15) (Figure 15) and versus year

(1970 and 1996) (Figure 16).

To compare the medians of the individual metric distributions across trophic status, a non-parametric Kruskal-Wallis test was used, while comparison of the medians of the individual metric distributions between 1970 and 1996 was performed using a non - parametric Mann-Whitney test (Minitab Version 13.1, 2000).

Temporal and Spatial Variation in the Multimetric P-IBI

In order to calculate the P-IBI, 4 steps are needed.

1) Collect phytoplankton and zooplankton samples.

2) Enumerate samples to genus/species and calculate abundances and biomasses.

3) Calculate metric scores using Table 19.

4) Estimate basin or lakewide mean scores.

To demonstrate explicitly the calculation of the P-IBI, both the formula (see below) for the P-IBI calculation and a sample calculation from the level of individual metric score assignments to the lakewide mean score (P-IBI) are given (Tables 20,21).

124

B S 1 1 1 + + + + + P-IBI = ∑ ∑ (EA jk CB jk RJ jk LM jk RA jk ZB jk ) B k =1 S j=1 M

Where,

EAjk = June biomass of edible algae taxa metric score,

CBjk = June % Microcystis, Anabaena, and Aphanizomenon of total phytoplankton biomass metric score,

RJjk = June zooplankton ratio (Calanoida/ (Cladocera + Cyclopoida) metric score,

LMjk = July Limnocalanus macrurus density metric score,

RAjk = August zooplankton ratio (Calanoida/ (Cladocera + Cyclopoida) metric score,

ZBjk = August crustacean zooplankton biomass metric score,

M = number of metrics,

S = number of sites (within a basin), and

B = number of basins

The focus in the example (Tables 20,21) is a mean site score calculation within the eastern basin and the mean basin score for this basin; however, the same calculation procedures were used in the western and central basins. The mean site score for each year was determined by summation of the individual metric scores across June, July, and

August and then dividing by the number of metrics (Table 20). A mean score across all months was warranted because of the temporal variability of the plankton and the fact

125 that not all sites were sampled with the same frequency. The basin mean for each year was then calculated by summation of the mean site scores across each basin and then dividing by the number of sites (Table 21). Finally, the lakewide Planktonic IBI scores for each year were calculated by summation of the basin mean scores and then dividing by the numbers of basins (n = 3) (Table 21 caption). Basin and lakewide P -IBI scores for

Lake Erie were calculated for those years during which data were available for all three basins (1970 and 1996-2002) (Figure 17, Table 22). 1995 western basin P-IBI scores were also calculated (Table 22).

To determine the relative importance of temporal and spatial variability in the mean site P-IBI score (1970, 1996-2002) an ANOVA was conducted for basin and year using a General Linear Model of Yijk = µ + Bi +Tj + eij, where Yij = mean site IBI score, µ

= grand mean, Bi = effect of basin, and Tk = effect of year. Significance for each of the analyses was set at an alpha level of 0.05. Further, the relationship between phosphorus loading into Lake Erie (values from Dolan 1993, Dolan unpublished) and mean lakewide

P-IBI score was investigated. In addition to a linear regression, a quadratic polynomial was fit to this relationship due to the low P-IBI values at both high and low phosphorus loading values and high P-IBI values occurring at intermediate loading values (Figure

18). The relationship between mean chlorophyll a (Figure 19a) and mean unfiltered total phosphorus (µg/L) (Figure 19b) versus mean P-IBI score at a site was determined using a quadratic polynomial (Minitab Version 13.1, 2000) due to the greatest chlorophyll a and total phosphorus concentrations occurring at intermediate P-IBI values. Due to data availability, chlorophyll a uncorrected for pheophytin (chlorophyll a + pheophytin) was

126 used in 1970, while corrected chlorophyll a (chlorophyll a only) was used durring 1996-

2000. The pheophytin correction is used to distinguish viable chlorophyll pigments fro m degraded pheophytin (Lorenzen 1967). In 1996, uncorrected chlorophyll a consisted of

16.6 + 1.1 % (mean + standard error, n = 232) pheophytin and thus uncorrected and corrected chlorophyll a did not differ substantially. Similarly, in 1970 in Lake Erie, uncorrected chlorophyll a consisted of approximately 19.0 + 3.0% (n = 26) (estimated from Glooschenko et al.’s (1974) Figure 4a,b,d). Pheophytin contributions to uncorrected chlorophyll a concentrations were typically greater than 20% during time periods not included in the P-IBI (October, November, December). For comparison with planktonic data, pheophytin was typically the dominant (>50%) pigment in benthic algal samples in the eastern basin of Lake Erie (Carrick 2004). Finally, the relationships between the P-IBI and phytoplankton biomass (Figure 20a), zooplankton biomass, and an offshore fish community IBI (Figure 20b) were determined using linear regressions

(Minitab Version 13.1 2000).

Goal 5. Communicate results to other scientists, managers, citizens, and policy makers

A major goal of IBI development is to communicate the inferred biotic integrity results to a variety of different groups of stakeholders (Karr and Chu 1997). This group ranges from scientists and managers, who may be very well acquainted with planktonic organisms and techniques involved in constructing and calculating the P-IBI, to citizens and policy makers who may only have a vague notion of the organisms involved and may

127 not be familiar with IBI’s or the techniques used in their development, but who nonetheless may have a large impact on management of water quality in the lake. The future inclusion of the P-IBI in the Lake Erie Quality Index (LEQI) (disseminated by the

Ohio Lake Erie Commission), the presentation of the final P-IBI to Great Lakes scientists at conferences like those of the International Association of Great Lakes Research (Kane et al. 2003a) and publication of this chapter allow the P-IBI to be communicated broadly to a range of Lake Erie stakeholders. Further, simple bar graphs (lakewide and basin) of

P-IBI score for each year (Figure 17 a,b) to make results accessible to the public and to show qualitative values of P-IBI-based offshore water quality that reflect the LEQI scoring system (i.e., poor, fair, good, excellent) (Ohio Lake Erie Commission 1998).

RESULTS

Metric Selection

Of the nine metrics analyzed using discriminant analysis, five were significant and included in the final P-IBI. Of the full models (models with all nine metrics included) tested for each month (May-September), only the June full model was found to be statistically significant (weighted Kappa, p = 0.0098) (Table 16). However, the stepwise discriminant analysis gave at least one significant metric for each month (Table

17). These significant metrics were then analyzed as the reduced model (model with significant metrics only) for each month. The reduced models for June, July, and August were each found to be significant (weighted Kappa, p < 0.05) (Table 16), but not May or

September (p > 0.05). The June reduced model contained metrics based on the

128 zooplankton ratio, percentage biomass of Microcystis, Anabaena, and Aphanizomenon of the total phytoplankton biomass, and the biomass of edible algae (Table 16). The July reduced model contained only the Limnocalanus macrurus density metric (Table 16).

Finally, the August reduced model contained the zooplankton ratio and crustacean zooplankton biomass metrics (Table 16). Redundancy, lack of adequate testing, sampling limitation, and insignificance in the discriminant analysis were all reasons for candidate metric rejection (Table 18), while significance in the discrimnant analyses was the only criterion for inclusion of a candidate metric in the final index.

Correlation between Significant Metrics and Total Phosphorus and Chlorophyll a

Regression analysis for each of the significant metrics against the phosphorus and chlorophyll a concentrations indicated that the values of some metrics decreased with increasing trophic status, while others increased (e.g., Figure 12). Lower values were indicative of more eutrophic conditions, as determined by total phosphorus and chlorophyll a, for the zooplankton ratio (slope = -0.055, p = 0.007 for June, slope =

-0.131, p = 0.001 for August) and Limnocalanus macrurus density (slope = -0.005, p <

0.001) metrics. For the biomass of edible algae taxa (slope = 526.808, p < 0.001) percentage biomass of Microcystis, Anabaena, and Aphanizomenon of the total phytoplankton biomass (slope = 0.201, p = 0.188) and the crustacean zooplankton biomass (slope = 14.859, p = 0.217) metrics, greater values were indicative of more eutrophic conditions, as determined by total phosphorus and chlorophyll a.

129

Individual Metric Cumulative Frequency Distribution

Each boxplot used to calculate individual metric scores (Figure 14) for June (a-c),

July (d), and August (e-f) was trisected between zero and the 95th percentile in order to give scores of 1, 3, or 5 (Table 19). Further, June median edible phytoplankton (µg/L) differed significantly among eutrophic, mesotrophic, and oligotropic classes (Figure 15a).

August median Limnocalanus macrurus density (#/L) (Figure 15d) and August abundance of Calanoida/ (Cyclopoida + Cladocera) were statistically different among trophic status classes and the median ratio decreased with increasing trophic status

(Figure 15e). August median crustacean zooplankton biomass (µg/L) was significantly different among trophic status classes and increased was greatest for the eutrophic class, and less for the mesotrophic and oligotrophic classes (Figure 15f). With respect to temporal comparisons (1970 vs. 1996), June median abundance of Calanoida/

(Cyclopoida + Cladocera) was significantly greater in 1996 versus 1970 (Figure 16c).

August median abundance of Calanoida/ (Cyclopoida + Cladocera) was statistically different between years, with 1970 having the greater median than 1996 (Figure 16e).

Finally, August median crustacean zooplankton biomass (µg/L) was significantly greater in 1970 than in 1996 (Figure 16f).

130

Temporal and Spatial Variation in the Multimetric P-IBI

The mean lakewide P-IBI score was below 3 (eutrophic) in 1970, increased to >3

(mesotrophic) during 1996-1998, declined to <3 in 1999, increased to >3 in 2000 and

2001 and then declined to <3 in 2002 (Figure 17a, Table 22). The mean basin scores generally followed this lakewide trend, however, not all basin scores varied in exactly the same manner (Figure 17b, Table 22). In 1970, the western and eastern basins were eutrophic, while the central basin was slightly mesotrophic. The western basin was characterized as eutrophic during 1996, mesotrophic between 1997 and 1999, and eutrophic between 2000 and 2002. The central basin was mesotrophic between 1996 and

1998, declined to eutrophic (1999 and 2000), became mesotrophic in 2001, and then declined again to eutrophic (2002). The eastern basin was characterized as oligotrophic

(1996), declined to mesotrophic (1997), declined further to eutrophic (1998 and 1999), became mesotrophic (2000 and 2001), and declined again to eutrophic (2002). Overall, the trends in each basin and the lakewide P-IBI score indicate an improvement in water quality between 1970 and 1996 and declines from mesotrophic (or even oligotrophic) conditions in the mid-1990’s to eutrophic conditions in 2002.

There were changes over time in mean site P -IBI, but not a significant variation in scores among basins, based on the ANOVA which indicated that there were significant differences in mean site P-IBI score among years (F2,7 = 5.00, p < 0.001), but not spatially among basins (F2,7 = 2.37, p = 0.096).

The relationship between mean lakeside P-IBI score and phosphorus loading in

1970, and 1996-2001 (Figure 18) was insignificant for both linear (p =0.862) and

131 quadratic (p = 0.221) regressions. Mean chlorophyll a (µg/L) was significantly (p <

0.001) correlated with mean P-IBI score at the same sites (1970, 1996 -2000) (Figure

19a); further, only one-eighth of the variability in chlorophyll a was explained by mean

P-IBI score (r2 = 0.127). Mean unfiltered total phosphorus (µg/L) was significantly (p =

0.024) correlated with mean P-IBI score at the same sites (1970, 1996 -2000) (Figure

19b); however, the mean P-IBI score explained even less of the variability in the mean unfiltered total phosphorus (r2 = 0.062). For both of these relationships, there were sharp declines in total phosphorus and chlorophyll a concentrations at greater P-IBI scores

(between 3-5) (Figures 19a,b). Further, total phytoplankton biomass (mg/L) was significantly (p = 0.012, r2 = 0.253) and negatively (slope = -1.322) related to P-IBI score

(Figure 20a), while mean crustacean zooplankton biomass was not significantly (p =

0.169) related to P-IBI score on a basin scale. Finally, there may be a positive trend between offshore fish IBI score (Kershner and Hopkins 2003) and mean lakewide P-IBI

(Figure 20b); however, only 5 years of data were available and the relationship was not statistically significant (p = 0.319).

DISCUSSION

There are more stressors on the plankton of large lakes than just external loading of nutrients. Pesticides, such as carbaryl can affect zooplankton from the individual

(Hanazato and Dodson 1995) to the community level (Hanazato 2001). However, little information on how pesticides affect the ecological structure and function of large lake foodwebs is now available. In the future, an ecological perspective on the effect of

132 pesticides on the plankton is needed (Hanazato 2001). Further, global warming could have a profound effect on plankton communities, causing increases in both phytoplankton and zooplankton (Magnuson et al. 1997). However, only a few studies (e.g., Chen and

Folt 1996) have evaluated the consequences of global warming on plankton communities in lakes. Finally, a careful examination of the role of dreissenid mussels and other members of the benthos is needed to understand internal phosphorus and nitrogen loading in large lakes (Culver et al. 2003). More research on the effect of a number of stressors on large lake plankton is needed to understand these stressors’ combined effects on the integrity of the plankton community.

In addition to determining the effects of other stressors on plankton communities, the P-IBI could be improved in future applications. Long-term datasets are important in ecological studies to understand temporal trends (Magnuson 1990). However, few continuous long-term plankton data sets are available for Lake Erie (an exception being the USEPA’s Spring (April) and Summer (August) sampling program (1970’s-present)).

The P-IBI could greatly be improved for application to other lakes if such datasets were available. For example, three North American lake regions (Experimental Lakes Area- northwestern Ontario, Dorset Environmental Science Centre- south-central Ontario, and

North Temperate Lakes Long-Term Ecological Research sites- Wisconsin northern highlands lakes) have long-term plankton and water quality data (Rusak et al. 2002) and thus may be candidates for application of the P-IBI. Further, monitoring programs that implement the P-IBI should consider taking samples at regular temporal and spatial intervals. For example, phytoplankton and zooplankton samples should be taken together

133 at least monthly from spatially distinct areas of a lake basin. The sites sampled should be the same each year to facilitate inter-annual comparisons in P-IBI scores. Finally, morphometric data on lakes should be used in conjunction with P-IBI application to determine the impact of morphometry on the plankton (e.g., areal or volumetric measures of abundance/ biomass). Thus modifications to this P-IBI are welcomed. Further, application of the methodology used to construct the P-IBI is encouraged for other lake ecosystems.

Although the methodologies used in each of the datasets (Table 14) are similar

(e.g., Utermöhl technique is used for phytoplankton enumeration, Culver et al. 1985 length-weight regressions are used for zooplankton biomass estimates), important differences exist among the years that may effect the results and conclusions drawn across studies. An important difference between the studies is the lack of water -volume sampled (flow meter data) for the 1970 zooplankton hauls. For these hauls, numbers/m3 were calculated from the percent of sample counted and volume sampled from tow depth, and net diameter (assuming 100% net efficiency) (Watson and Carpenter 1974).

However, the net efficiency of plankton nets is seldom 100%, due to net shape, mesh size, mesh area, netting porosity, filtering area, and the mesh area to mouth opening ratio

(Sameoto et al. 2000) and thus abundance, and consequently biomass, estimates from

1970 are likely underestimates. Likewise, Limnocalanus macrurus abundances from

1970 would also be underestimates. These underestimates would likely decrease our ability to use crustacean zooplankton abundance to discriminate between different trophic status classes. However, the significance of crustacean zooplankton biomass and

134

Limnocalanus macrurus in the dicriminant analyses indicate that although underestimates, the 1970 data can be used in conjun ction with the 1996 data to distinguish between different trophic status classes. Unlike crustacean biomass and

Limnocalanus abundance, the zooplankton ratio should not vary with different net filtering efficiencies because relative, rather than absolute, densities are used in the ratio calculations. For phytoplankton data, the samples were taken from the upper few meters of the water column, mixed, and enumerated using the same technique (Utermöhl technique) in both 1970 and 1996 and thus should be directly comparable. The use of similar sampling and enumeration techniques, combined with the paucity of historical data regarding phytoplankton and zooplankton data in Lake Erie make the use of these studies unavoidable.

Although the multimetric P-IBI contains only five unique individual metrics (one metric is used twice), these metrics reflect a number of the Beneficial Use Impairments

(BUIs) to Lake Erie (Hartig et al. 1997, Table 1) and cover a wide range of plankton taxonomic and functional groups. For example, the percentage Anabaena,

Aphanizomenon, and Microcystis individual metric directly addresses undesirable algae, restrictions on water consumption, and degradation of phytoplankton communities.

Further, the Limnocalanus macrurus individual metric directly addresses degradation of zooplankton communities. The (Calanoida/ (Cladocera + Cylopoida)) individual metric distinguishes between different trophic classes and thus addresses eutrophication. Also within the five unique metrics, all five taxa of Lake Erie phytoplankton are represented by two of the metrics, the edible phytoplankton biomass metric (Chlorophyta,

135

Chrysophyta, Cryptophyta, and Pyrrophyta) and the percentage Anabaena,

Aphanizomenon, and Microcystis metric (Cyanophyta). Further, the three taxa of crustacean zooplankton in Lake Erie (Calanoida, Cyclopoida, and Cladocera) are represented in the zooplankton ratio metric. Thus the use of a variety of plankton taxonomic groups and these groups’ varied ecologies allow the multimetric P-IBI to provide information on BUIs to Lake Erie.

Further, the P-IBI is an important tool because it is one of the few biological measures of offshore water quality available. Other than the offshore fish community IBI

(discussed below), the P-IBI is the only offshore component of the Lake Erie Quality

Index (LEQI). Thus the P-IBI can give managers information that few of the other LEQI components can. For example, phosphorus loading and the walleye LEQI components were reported as excellent in 1996 (Ohio Lake Erie Commission 1998), while the P-IBI was good (but near excellent). Thus, by providing information on lower trophic levels, the P-IBI can provide a link between nutrient input into the lake and its effect on important sportfish.

Testing the validity of the final Planktonic IBI can be achieved in part by determining whether individual metric values agree with more traditional measures of lake trophic status. Other than the biomass of edible algae taxa metric, the trends in metric values versus trophic status (as measured by regression line slopes) were in agreement with the hypothesized response to degradation (eutrophication). For example,

August abundance of Calanoida/ (Cyclopoida+Cladocera) decreased with increasing trophic status (oligotrophic = 0.216 > mesotrophic = 0.180 > eutrophic = 0.079), similar

136 to the trends found by Gannon and Stemberger (1978) in Lake Michigan. Further, June and August abundance of Calanoida/ (Cyclopoida+Cladocera) increased between 1970

(June = 0.006, August = 0.0999) and 1996 (June = 0.131, August = 0.263) concomitant with oligotrophication (based on total phosphorus and chlorophyll a concentrations,

Figure 10a,b). In the case of the biomass of edible algae metric, median algal biomass increased with increasing trophic status (oligotrophic = 1025.0 µg/L > mesotrophic =

431.4 µg/L > oligotrophic = 404.3 µg/L), which was opposite to the hypothesized response of this metric to increasing trophic status. However, at high nutrient levels in

Lake Erie, both edible and nonedible taxa benefit (Davis 1964). Thus, the increase in edible phytoplankton with increased nutrient availability actually shows agreement between the metric and traditional measures of trophic status, such as total phosphorus concentration.

The validity of the multimetric mean basin and mean lakewide P-IBI scores should also be judged within the context of changes that have occurred in Lake Erie. If the P-IBI is valid, then it should reflect trends in Lake Erie trophic status, since it was based on such measures. The changes in P-IBI values over time, specifically the trend of increasing lakewide P-IBI values from 1970 to 1996-1997 is consistent with oligotrophication of Lake Erie that has been documented for both nutrients (Charlton et al. 1999) and the plankton communities (Munawar et al. 2002, Conroy et al. 2004a, in review). Further, the decline in lakewide P-IBI values during 1998-1999 reflects increasing nutrient concentrations, reoccurrence of extensive anoxic/ hypoxic areas in the

137 central basin (Banicki 2003, USEPA 2004), and the increasing prevalence of Microcystis blooms in the western basin (Budd et al. 2002) that have recently occurred in Lake Erie.

The P-IBI should also reflect changes in Lake Erie over space and time, as well as differences between each of the three basins with respect to more traditional measures of

Lake Erie trophic status, such as phosphorus and chlorophyll a concentrations. Using basin-wide mean values of chlorophyll a observed on monthly cruises of Lake Erie from

1970, Dobson et al. (1974) characterized the western basin as eutrophic, the central basin as mesotrophic, and the eastern basin as mesotrophic-oligotrophic; using particulate phosphorus the same authors characterized the western basin as eutrophic, and the central and eastern basins as mesotrophic. In general, P-IBI values match these more traditional measures of lake trophic status. Using a scale of <3 reflecting eutrophic conditions, 3 -4 reflecting mesotrophic conditions, and >4 reflecting oligotrophic conditions, the mean P-

IBI score for Lake Erie in 1970 is less than 3 and thus eutrophic. The separate basin scores for 1970 also reflect more traditional measures of trophic status. The mean value for the western basin is in the eutrophic range (< 3), while the mean value for the central basin is at the boundary of the eutrophic and mesotrophic ranges. The only basin-wide mean P-IBI value for 1970 that disagrees with the traditional trophic measures is the eastern basin mean P-IBI value, which reflects a eutrophic condition, rather than a mesotrophic-oligotrophic condition. For the 1996 P-IBI values, the biotic integrity is greater both lakewide and in each respective basin, than it was in 1970. This increase in

P-IBI reflects declines in both chlorophyll a levels, and total phosphorus levels during the mid-1990’s, compared to the early 1970’s (Charlton et al. 1999, Figure 10). Further,

138

Charlton et al.’s (1999) data reveal that the central basin (CB) and the eastern basin (EB) had similar mean summer total phosphorus concentrations (µg/L)(CB = 8.4, EB = 8.0) and mean summer chlorophyll a (µg/L) (CB = 1.8, EB = 2.1) for 1994-1996, which may explain equivalent or even lower P-IBI scores in the eastern basin, the basin traditionally considered the most oligotrophic in Lake Erie (Munawar et al. 2002).

The sensitivity of the P-IBI to changes in Lake Erie also should be examined. On a lakewide scale, the P-IBI varied 18.75% between 1970 and 1996. Because chlorophyll a and total phosphorus concentrations declined between 1970 (more eutrophic) and 1996

(more oligotrophic), the responsiveness of the P-IBI to these traditional measures of trophic status, as well as the comparability of the P-IBI with other biotic measures of water quality must be investigated. Although there are no other Planktonic IBIs for offshore ecosystems in large lakes, a wetland zooplankton index (WZI) using zooplankton community characteristics has been developed for assessing the effects of stressors on Great Lakes wetlands (Lougheed and Chow-Fraser 2002). The authors used the WZI to evaluate the effect of the removal of a stressor (i.e., carp) on wetland water quality. The WZI differs approximately 20% (estimated from Lougheed and Chow-

Fraser’s (2002) Figure 5) in vegetated zones before and after carp exclusion. Thus the responsiveness of the P-IBI to changes in nutrient stressors in offshore waters (18.75%) is similar to the responsiveness of the WZI to a biotic stressor in the wetlands (20%) of the

Laurentian Great Lakes. Further, during 1970 and 1996, Lake Erie approaches neither the most oligotrophic end nor the most eutrophic end of the trophic continuum, as measured by total phosphorus and chlorophyll a concentrations. Wetzel (2001) reports

139 values as low as 3 µg/L total phosphorus for oligotrophic lakes (mean = 8.0, n = 21) and as high 386 µg/L total phosphorus for eutrophic lakes (mean = 84.4, n = 71). Total phosphorus concentrations in Lake Erie during 1970 and 1996 were alm ost never as low as 3 µg/L and never higher than 228 µg/L, with basinwide means ranging 7.74-41.53

µg/L. Further, Wetzel (2001) reports chlorophyll a values as low as 0.3 µg/L for oligotrophic lakes (mean = 1.7, n =22) and as high as 78 µg/L for eutrophic lakes (mean

= 14.3, n = 70). Uncorrected chlorophyll a concentrations in Lake Erie during 1970 and

1996 were never as low as 0.3 µg/L and never as high as 78 µg/L, with basinwide means ranging 1.67-12.58 µg/L. Thus the lakewide P-IBI’s variance of 18.75% is similar to the

14-15% variance between mean upper and mean lower basinwide chlorophyll a and total phosphorus means (difference between maximum and minimum basinwide means divided by the total range given by Wetzel (2001)). Thus the P-IBI has been as (or more) responsive to changes in Lake Erie as other measures of trophic status and water quality.

The ultimate test of the P-IBI is whether it correlates with the end product of eutrophication, increased algal biomass. Increases in phytoplankton biomass were evident in the mid-20th century (Davis 1964, Munawar and Munawar 1976) and were hypothesized to be in response to increased phosphorus loading. My data show a significant, negative relationship between total phytoplankton biomass (mg/L) and P-IBI score at the basin level. Recent (1998-2002) analyses indicate that phytoplankton biomass in Lake Erie (Chapter 2, Figure 5) is approaching 1970 levels in the western basin and may be exceeding 1970 levels in the eastern basin. Thus P -IBI scores equal to or less than those calculated for 1970 are not unexpected for the period of 1998-2002.

140

The results of the P-IBI also appear to agree with a recently developed method of assessing Lake Erie, an IBI based on the offshore fish assemblage of Lake Erie (Kershner and Hopkins 2003). Metrics used in this index included # intolerant fish species, # benthic fish species, # phytophilic species, # native species, # non-indigenous species, # native cyprinid species, % carnivores, % omnivores, % tolerant individuals, % nonindigenous individuals, and total and native Catch Per Unit Effort (CPUE). The authors found that Lake Erie between 1996 and 1999 could be qualitatively characterized as

“good,” for both the western and central basins, and both basins combined. This “good” classification corresponds with my mesotrophic classification. Similar to the fish IBI, I found mesotrophic scores for the western basin in 1995 and 1997-1999, for the central basin in 1996-1998, and lakewide in 1996-1998. Quantitatively, there may be a relationship between the P-IBI and the offshore fish IBI. However, there are currently only 5 years of overlapping data and more years of offshore fish IBI calculations are needed. Agreement between a fish offshore IBI and the P-IBI is important, because they are two of the few biological measures of offshore water quality currently available and all of the fish in Lake Erie are dependent on plankton at some time during their development (Gopalan et al.1998). However, the data available at this time preclude a definitive determination of the relationship between the P-IBI and the offshore fish IBI in

Lake Erie.

Increases in P-IBI values between 1970 and the mid-1990’s reflect the impact of the decreased phosphorus loading (Dolan 1993) that led to the oligotrophication of Lake

Erie in the 1980’s and 1990’s (Ludsin et al. 2001). In fact, Lake Erie in the early to mid -

141

1990’s could best be described as a mesotrophic-oligotrophic system (Munawar et al.

2002), based on a number of phytoplankton community characteristics (i.e., reduction of phytoplankton biomass and productivity, high species diversity, decrease of eutrophic species, and increase of oligotrophic species). The 1995 and 1996 Planktonic IBI values reflect the oligotrophication of the lake. The mean P-IBI score for the western basin is in the mesotrophic range (between 3 and 4) for both 1995 and 1997, as are the mean P -IBI scores for the central basin for 1996-1998. Further, the mean P-IBI scores in the eastern basin are in the oligotrophic range for 1996 and the mesotrophic range for 1997. Thus, for the mid 1990’s, the P-IBI supports the assertion that Lake Erie was a mesotrophic- oligotrophic system. In fact, the lakewide P-IBI scores for 1996 and 1997 were both within the mesotrophic range (3-4), compared to the 1970 score (<3), which is in the eutrophic range.

By the late 1990’s (1998 and 1999) and 2002, the P-IBI scores reflect a more eutrophic state within each basin and at a lakewide scale. In fact, the lakewide scores for

1999 and 2002 are similar to the value calculated for 1970. Is this apparent eutrophication a real event, or a problem with the P-IBI? In each of the years 1996-1998, phosphorus loading into Lake Erie (Dolan, unpublished data) was in excess of the 11 kilotonne target loading limit set in the 1970’s. Further, the return of large areas of anoxia in the central basin of Lake Erie has been seen as an indicator of the re- eutrophication of Lake Erie (Banicki 2003, USEPA 2004). Data from this study indicate that mean total phytoplankton biomass and P-IBI at the basin level are significantly and negatively correlated and that seasonal mean total phytoplankton biomasses are

142 approaching 1970 values in Lake Erie. Thus several lines of evidence support the assertion that Planktonic IBI scores in the late 1990’s and early 2000’s may be a real indication of degraded water quality in Lake Erie.

The P-IBI has a number of limitations. First, as an index, it distills the highly variable plankton community of Lake Erie into a number ranging from 1-5. Clearly, the

P-IBI cannot be used for a complete understanding of Lake Erie plankton dynamics.

Further, the paucity of historical samples for Lake Erie hinders comparison of 1970,

1990’s, and 2000’s P-IBI scores with earlier time periods (e.g., the early 20th century).

Another limitation of the P-IBI is the low variance (18.75%) of the P-IBI across a range of trophic conditions in Lake Erie. This low variance may hinder the ability of policymakers to detect changes in the P-IBI that represent alterations in the ecosystem health of Lake Erie. Further, from a logistical standpoint the necessity of monthly sampling Lake Erie at 5-10 sites in each basin may prevent groups, other than large government agencies, from collecting the samples needed for the P-IBI. Finally, because phytoplankton and zooplankton enumeration are time-intensive activities, P-IBI scores may not be available until months after the actual samples are taken. This would prevent rapid implementation of management strategies, in response to declines in P-IBI scores.

Along with the limitations mentioned above, the P-IBI has a number of strengths, as well as weaknesses. A weakness of the P-IBI is that it does not include rotifer composition, due to sampling limitations. Rotifer species composition could be an informative metric, because different suites of rotifers are prevalent in eutrophic, mesotrophic, and oligotrophic conditions (Gannon and Stemberger 1978). Another

143 weakness of the P-IBI may be the general public’s lack of knowledge regarding plankton.

This may hinder the interpretation of P-IBI results by the public. The P-IBI also has a number of strengths. The P-IBI allows one to compare the three basins of Lake Erie on the same scale. Further, the techniques employed constructing the P-IBI could be used to construct P-IBI for other lakes, including the other Laurentian Great Lakes. Finally, the

P-IBI integrates the physics, chemistry, and biology of Lake Erie by directly measuring the biota, rather than using a measure that indirectly relates to the biota and whose effects on the biota may not be well known.

The goal of the development of the P-IBI was to create an offshore water quality monitoring tool that reflected Beneficial Use Impairments (BUIs) to Lake Erie. The P-

IBI does reflect BUIs by addressing degradation of phytoplankton and zooplankton communities and the impacts of eutrophication, such as increases in nuisance or toxic phytoplankton taxa. Further, the P-IBI measures the biological water quality of Lake

Erie, which has previously been neglected. The P-IBI also provides an important biological link between nutrients and upper trophic level measurements. Because the P-

IBI provides this link and reflects the trophic status of Lake Erie, its applicatio n to other lakes is encouraged. Finally, the P-IBI fufills Elster’s (1974) three reasons that science attempts to order its findings: 1) to gain an overall view, 2) to simplify the understanding of complex systems by characterizing a few common factors, and 3) to predict properties or relationships of parts of systems from other measured properties.

144

CONCLUSION

The multimetric P-IBI developed for Lake Erie reflects Beneficial Use

Impairments to Lake Erie and can provide a useful, broad-scale way to monitor changes in the offshore water quality of the lake. The increase in lakewide P-IBI score from 1970 to the mid-1990’s and its subsequent decline reflect the changing trophic status of the lake. Further, the P-IBI can provide managers, policy makers, scientists and the general public information at a level appropriate for the needs of each group. Finally, the methodology used in the development of the P-IBI can be applied to other large lakes, including the other Laurentian Great Lakes. In any application of the P-IBI, availability of historical datasets and sampling frequency should be important considerations. The P-

IBI is a new tool for measuring offshore water quality of large lakes and can be used to monitor changes in these lakes stemming from anthropogenic stressors, such as nutrient addition.

145

APPENDIX A

TABLES

146

Beneficial Use Impairment Plankton impacts

Restrictions on fish and wildlife consumption none Tainting of fish and wildlife flavor A, B Degradation of fish and wildlife populations B, D, E, F, G, H Fish tumors or other deformities none

Bird or animal deformities or reproduction problems none

Degradation of benthos B, C, F Restrictions on dredging activities none Eutrophication or undesirable algae B, C, D Restrictions on drinking water consumption or taste or odor A, B

147 problems

Beach closings none Degradation of aesthetics A, C Added costs to agriculture or industry A, B, C, F

Degradation of phytoplankton and zooplankton populations B, C, D, E, F, H Loss of fish and wildlife habitat B, C, F

Table 1- Interactions among Beneficial Use Impairments (BUIs) (Hartig et al. 1997) and planktonic characteristics in Lake Erie. Characteristics impacting specific BUIs are found in the right column (Culver, unpublished). Plankton characteristics- A. Taste and odor production, B. Toxins from Cyanophyta, C. Floating algal mats or blooms, D. Dominant phytoplankton inedible by zooplankton, E. Predation by introduced planktonic invertebrates, F. Non-indigenous species (Dreissena planktonic larvae), G. Large planktonic predators, H. Large Daphnia spp. dominant.

Phytoplankton Zooplankton Begin Date End Date # of Samples Begin Date End Date # of Samples Year WB CB EB WB CB EB WB CB EB WB CB EB WB CB EB WB CB EB 1996 5/22 6/12 6/24 10/7 9/20 9/18 365 71 11 5/22 6/12 6/24 10/7 9/20 9/18 486 87 14 1997 5/9 5/7 5/7 10/7 8/29 8/28 99 49 19 5/9 5/7 6/2 10/7 8/29 8/26 147 58 20 1998 5/26 6/9 6/8 9/28 9/4 8/25 134 73 30 6/10 6/9 6/8 9/29 9/4 8/25 144 78 21

148 1999 5/17 5/18 5/17 9/28 9/10 9/8 85 55 25 5/17 5/18 5/17 9/28 9/9 9/8 77 53 24

2000 5/2 5/9 6/1 9/18 9/8 8/29 52 44 28 5/2 5/9 6/1 9/18 9/8 8/29 88 87 26 2001 4/30 5/22 4/30 10/2 8/2 9/14 71 27 80 4/30 5/22 5/1 10/2 8/2 9/14 79 28 62 2002 5/7 6/25 4/30 9/30 10/24 10/29 66 109 53 5/7 6/25 7/20 9/25 10/24 9/17 95 136 13

Table 2- Number of phytoplankton and zooplankton samples collected for the Lake Erie Plankton Abundance Study (LEPAS) from each of the three Lake Erie basins (1996-2002) and the length of season sampled for temporal and spatial biomass comparisons (J.D. Conroy, The Ohio State University, personal communication).

149 Table 3- Numbers of sites and site identities of full and reduced suites of sites analyzed in the western basin of Lake Erie (1996-2002).

For the location of site numbers see Figure 2.

Table 3

Full number Reduced number Year of sites Full sites of sites Reduced sites

1996 51 1-44, 343, 344, 357, 40 1, 2, 3, 4, 5, 7, 8, 9, 10, 12, 966, 967, 971, 972 13, 14, 16, 17, 18, 19, 21, 22, 24, 27, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 357, 966, 967, 971, 972

30 3, 4, 5, 7, 8, 9, 10, 12, 14, 16, 150 18, 19, 24, 27, 29, 30, 31, 32, 34, 36,

37, 38, 39, 40, 42, 357, 966, 967, 971, 972

20 3, 7, 8, 12, 14, 16, 24, 27, 29, 30, 34, 36, 37, 40, 42, 357, 966, 967, 971, 972

15 3, 8, 14, 16, 27, 29, 30, 36, 37, 42, 357, 966, 967, 971, 972

10 3, 8, 14, 16, 27, 29, 36, 37, 971, 972

5 8, 16, 27, 29, 37

Continued

Table 3 continued

1997 21 4, 7, 8, 9, 13, 14, 16, 27, 20 4, 7, 8, 9, 13, 14, 16, 27, 29, 30, 29, 30, 36, 37, 39, 42, 36, 37, 39, 42, 43, 357, 966, 967, 971, 972 43, 44, 357, 966, 967, 971, 972

151 15 4, 8, 14, 16, 27, 29, 30, 36, 37, 42, 357, 966, 967, 971, 972

10 4, 8, 14, 16, 27, 29, 36, 37, 971, 972

5 8, 16, 27, 29, 37

1998 23 3, 7, 8, 12, 14, 16, 19, 20 3, 7, 8, 12, 14, 16, 24, 27, 29, 30, 24, 26, 27, 29, 30, 34, 34, 36, 37, 40, 42, 357, 966, 967, 971, 972 36, 37, 40, 42, 357, 966, 967, 971, 972, 974

15 3, 8, 14, 16, 27, 29, 30, 36, 37, 42, 357, 966, 967, 971, 972

Continued

10 3, 8, 14, 16, 27, 29, 36, 37, 971, 972

5 8, 16, 27, 29, 37

152 1999 13 3, 8, 14, 16, 27, 29, 36, 10 3, 8, 14, 16, 27, 29, 36, 37, 971, 972 37, 357, 966, 967, 971, 972

5 8, 16, 27, 29, 37

2000 13 3, 8, 14, 16, 27, 29, 36, 10 3, 8, 14, 16, 27, 29, 36, 37, 971, 972 37, 357, 966, 967, 971, 972 8, 16, 27, 29, 37 5

Continued

Table 3 continued

2001 8 3, 8, 14, 16, 27, 29, 36, 5 8, 16, 27, 29, 37 37

2002 11 3, 8, 14, 16, 27, 29, 36, 10 3, 8, 14, 16, 27, 29, 36, 37, 971, 972 37, 971, 972, 973

153 5 8, 16, 27, 29, 37

Table 4- Comparison of sampling, enumeration, and biomass calculation methods for phytoplankton and zooplankton in the western, 154 central, and eastern basins of Lake Erie (1970-2002).

Table 4

PHYTOPLANKTON Munawar and Devault and Makarewicz (1993a) Frost and Culver (2001), this Munawar (1976) Rockwell (1986) paper Year(s) 1970 1978 1983-1987 1996-2002 Sampling Dates April-December May-November April and August Late April/ early May- late September/ early October, typically Sampling Frequency Monthly 9 cruises 33 cruises Weekly-Monthly Number of Sites 25 87 21 ~30-80

155 Sampling Method VanDorn bottle Niskin bottle Niskin bottle Integrated Water Sampler Size of sample ? 8 L 8 L π * 1.252 * twice Secchi depth Sampling Depth and 1- and 5-m. Samples During stratification- Deeper waters- 1, 5, Twice the Secchi depth. Mixing mixed. 1 m below surface, 1 10, and 20 m. Shallow Sample mixed. m above waters- 1 m, mid - metalimnion, at the depth, 1 m above the thermocline, bottom. Samples 1 m above the mixed. hypolimnion, 1 m above the bottom; Unstratified situations- 1 m below surface, mid-depth, and 1 m above the Continued

Table 4 continued

bottom. Sample mixing not reported. Enumeration Method Utermöhl (1958) Utermöhl (1958) Utermöhl (1958) Utermöhl (1958) Taxonomic Genus and species Genus and species Genus and species Genus resolution Biomass Calculation Volumes determined, Volumes determined, Volumes determined, Volumes determined, Method converted to wet converted to wet converted to wet converted to wet weight weight weight weight ZOOPLANKTON Watson and (Weisgerber 2000) Makarewicz (1993b) Frost and Culver (2001), this Carpenter (1974), paper Watson (1976), 156 Bean (1980) Year (s) 1970 1974-1975 1984-1987 1996-2002 Sampling Dates April-December April-December April and August Late April/ early May- late September/ early October, typically Sampling Frequency Monthly ? 33 cruises Weekly-Monthly Number of Sites 25 19, western basin 21 ~30-80 only Sampling Method Unmetered 64 µm Metered 64 µm net Metered 62 µm net Metered 64 µm net net (meter data unavailable) Sampling Depth 2m above the bottom 1m above the bottom 1m above the bottom 1m above the bottom to to surface (or 50m to to surface to surface (or 20m to surface surface) surface) Continued

Table 4 continued

Enumeration Method 200 individuals of 200 individuals of a Gannon (1971) 100 individuals of most each taxon single taxon common taxon Taxonomic Genus and species Genus and species Genus and species Genus and species Resolution Biomass Calculation Culver et al. (1985) Culver et al. (1985) Downing and Rigler Culver et al. (1985) Method (1984), Makarewicz and Likens (1979), Hawkins and Evans (1979)

157

Table 5- Temporal comparisons of sampling regimens in Lake Erie for phytoplankton and zooplankton biomass for 1996-2002. Number of samples taken during the entire sampling season (full number of samples), and number of samples taken during a month (reduced number of samples) are given. Bonferroni corrected alpha value are given using α = 0.05/ number of comparisons for individual statistical test in a given data set. Results of regression analyses include r2 values and F-ratios with associated p-values. 158 Results for paired t-tests include % mean difference ((mean biomass for entire season – mean biomass for month)/ mean biomass for entire season) * 100) and t-values with associated p-values. Bold p-values indicate significance at the appropriate Bonferroni corrected alpha level.

Full Reduced Bonferroni Temporal number of number of corrected % mean Category Year Comparison samples samples alpha value r2 F-ratio p-value difference t p-value Phytoplankton 0.05/6 biomass (by May vs. Entire comparisons taxon) 1996 Season 441 35 = 0.0083 0.951 58.83 0.005 43.74 4.58 0.010 June vs. Entire Season 441 121 0.883 22.74 0.018 -4.24 -0.18 0.864 July vs. Entire

159 Season 441 106 0.867 19.54 0.022 -72.88 -2.62 0.059

August vs. Entire Season 441 102 0.913 31.51 0.011 41.58 1.61 0.182 September vs. Entire Season 441 66 0.766 9.80 0.052 43.24 1.95 0.123 October vs. Entire Season 441 11 0.730 8.11 0.065 -35.25 -0.45 0.676

0.05/6 May vs. Entire comparisons 1997 Season 158 15 = 0.0083 0.960 71.81 0.003 -16.87 -0.47 0.663 June vs. Entire Season 158 47 0.910 30.44 0.012 34.00 1.78 0.149 July vs. Entire 26.72 Season 158 45 0.943 50.01 0.006 1.60 0.184

Continued

Table 5 continued

August vs. Entire Season 158 35 0.700 7.02 0.077 -29.17 -0.10 0.368 September vs. Entire Season 158 7 0.879 21.71 0.019 -251.15 -1.11 0.331 October vs. Entire Season 158 9 0.472 2.68 0.200 26.65 0.69 0.528

0.05/4 June vs. Entire comparisons 1998 Season 206 69 = 0.0125 0.091 0.30 0.622 60.19 0.99 0.378

160 July vs. Entire

Season 206 66 0.464 2.60 0.205 34.53 0.81 0.465 August vs. Entire Season 206 60 0.955 63.76 0.004 -80.72 -0.84 0.447 September vs. Entire Season 206 11 0.985 195.59 0.001 -143.55 -1.23 0.286

0.05/5 May vs. Entire comparisons 1999 Season 130 25 = 0.01 0.168 0.61 0.493 47.71 1.83 0.142 June vs. Entire Season 130 24 0.235 0.92 0.408 50.68 2.10 0.104 July vs. Entire Season 130 25 0.799 11.89 0.041 11.99 1.07 0.346 August vs. 130 33 0.714 7.48 0.072 -100.54 -2.14 0.111

Continued

Table 5 continued

Entire Season September vs. Entire Season 130 23 0.543 3.57 0.155 26.56 1.56 0.194

0.05/5 May vs. Entire comparisons 2000 Season 183 32 = 0.01 0.579 5.12 0.135 36.00 1.88 0.134 June vs. Entire Season 183 53 0.528 3.35 0.165 12.46 0.34 0.754 July vs. Entire

161 Season 183 36 0.626 5.02 0.111 -7.60 -0.39 0.718

August vs. Entire Season 183 44 0.207 0.78 0.441 -12.77 -0.39 0.715 September vs. Entire Season 183 18 0.295 1.26 0.344 -54.83 -1.00 0.374

0.05/7 April vs. Entire comparisons 2001 Season 154 8 = 0.0071 0.621 4.92 0.113 53.82 2.01 0.114 May vs. Entire Season 154 37 0.837 15.38 0.029 -69.27 -0.76 0.490 June vs. Entire Season 154 51 0.097 0.32 0.609 23.27 0.49 0.651 July vs. Entire 1.93 Season 154 14 0.845 16.31 0.027 38.91 0.126

Continued

Table 5 continued

August vs. Entire Season 154 14 0.287 1.21 0.352 19.82 0.58 0.593 September vs. Entire Season 154 28 0.670 6.09 0.090 1.64 0.07 0.945 October vs. Entire Season 154 2 0.227 0.88 0.418 33.82 0.86 0.439

0.05/7 April vs. Entire comparisons 2002 Season 141 4 = 0.0071 0.005 0.01 0.911 44.15 0.85 0.442

162 May vs. Entire

Season 141 18 0.789 11.19 0.044 6.34 0.16 0.881 June vs. Entire Season 141 10 0.938 45.73 0.007 -140.98 -2.57 0.062 July vs. Entire Season 141 34 0.203 0.77 0.446 36.83 1.32 0.256 August vs. Entire Season 141 23 0.711 7.37 0.073 8.53 0.24 0.825 September vs. Entire Season 141 42 0.678 6.31 0.087 -23.66 -1.01 0.371

October vs. Entire Season 141 10 0.774 10.25 0.049 67.00 5.23 0.006

Continued

Zooplankton 0.05/ 6 biomass (by May vs. Entire comparisons taxon) 1996 Season 441 35 = 0.0083 0.558 1.26 0.463 38.70 0.57 0.625 June vs. Entire Season 441 121 1.000 6653.52 0.008 -63.22 -2.45 0.134 July vs. Entire Season 441 106 0.999 784.19 0.023 -68.73 -0.97 0.436 August vs. Entire Season 441 102 0.997 343.58 0.034 70.28 1.62 0.248

163 September vs. Entire Season 441 66 0.974 38.10 0.102 81.11 1.63 0.245

October vs. Entire Season 441 11 0.994 156.27 0.051 96.59 1.79 0.215

0.05/ 6 May vs. Entire comparisons 1997 Season 158 15 = 0.0083 0.365 0.57 0.587 82.96 1.33 0.316 June vs. Entire Season 158 47 0.983 57.56 0.083 29.10 1.20 0.352 July vs. Entire Season 158 45 0.999 1287.92 0.018 -100.96 -1.22 0.348 August vs. Entire Season 158 35 0.962 25.41 0.125 29.32 0.93 0.450

Continued

Table 5 continued

September vs. Entire Season 158 7 0.914 10.64 0.189 64.31 1.05 0.403 October vs. Entire Season 158 9 0.990 103.20 0.062 50.58 2.66 0.117

0.05/4 June vs. Entire comparisons 1998 Season 206 69 = 0.0125 0.999 1866.13 0.015 -12.35 -2.89 0.102 July vs. Entire Season 206 66 0.979 47.22 0.092 -27.31 -1.18 0.360

164 August vs.

Entire Season 206 60 0.984 61.78 0.081 52.61 1.98 0.186 September vs. Entire Season 206 11 0.494 0.98 0.504 -45.45 -1.18 0.359

0.05/5 May vs. Entire comparisons 1999 Season 130 25 = 0.01 0.033 0.03 0.884 58.98 1.36 0.307 June vs. Entire Season 130 24 0.986 69.71 0.076 -115.24 -1.32 0.318 July vs. Entire Season 130 25 0.919 11.41 0.183 -11.90 -0.77 0.522 August vs. Entire Season 130 33 0.957 22.25 0.133 23.77 1.28 0.328

Continued

September vs. Entire Season 130 23 0.732 2.73 0.346 35.11 1.27 0.333

0.05/5 May vs. Entire comparisons 2000 Season 183 32 = 0.01 0.252 0.34 0.666 18.46 0.62 0.601 June vs. Entire Season 183 53 0.986 72.75 0.024 -43.46 -1.71 0.228 July vs. Entire Season 183 36 0.996 280.50 0.038 -27.38 -1.67 0.237

165 August vs.

Entire Season 183 44 0.647 1.83 0.405 27.15 1.32 0.316 September vs. Entire Season 183 18 0.317 0.47 0.619 82.94 2.67 0.116

0.05/7 April vs. Entire comparisons 2001 Season 154 8 = 0.0071 0.413 0.70 0.5555 80.78 2.06 0.175 May vs. Entire Season 154 37 0.065 0.07 0.836 13.63 0.43 0.706 June vs. Entire Season 154 51 0.999 1095.51 0.019 -36.74 -2.02 0.180

Continued

July vs. Entire Season 154 14 0.989 93.29 0.066 -31.80 -0.70 0.558 August vs. Entire Season 154 14 0.997 336.05 0.035 21.03 2.24 0.155 September vs. Entire Season 154 28 0.908 9.89 0.196 25.00 2.34 0.144 October vs. Entire Season 154 2 0.799 3.98 0.296 86.10 3.13 0.089

0.05/7

166 comparisons

April vs. Entire = 0.0071 2002 Season 141 4 0.759 3.15 0.327 97.84 2.82 0.106 May vs. Entire Season 141 18 0.955 205.28 0.044 92.84 2.60 0.121 June vs. Entire Season 141 10 0.719 2.56 0.356 14.10 0.58 0.618 July vs. Entire Season 141 34 0.915 10.78 0.188 -97.39 -1.90 0.198 August vs. Entire Season 141 23 1.000 11830000.00 <0.001 16.78 7.54 0.017 September vs. Entire Season 141 42 0.560 1.27 0.462 46.31 2.03 0.179 October vs. Entire Season 141 10 0.208 0.26 0.698 -122.39 -2.23 0.146

Table 6- Spatial comparisons of sampling regimens in Lake Erie for phytoplankton and zooplankton biomass for 1996-2002. Number of sites sampled in the western basin during the entire sampling season (full number of sites), and the diminished number of sites that served as surrogates (reduced number of sites) are given. Bonferroni corrected alpha value are given using α = 0.05/ number of 167 comparisons for individual statistical test in a given data set. Results of regression analyses include r 2 values and F-ratios with associated p-values. Results for paired t-tests include % mean difference ((mean biomass for full number of sites- mean biomass for reduced number of sites)/ mean biomass for full number of sites) * 100) and t-values with associated p-values. Bold p-values indicate significance at the appropriate Bonferroni corrected alpha level.

Full Reduced number number Bonferroni corrected alpha % mean 2 Category Year of sites of sites value r F-ratio p-value difference t p-value Phytoplankton 0.05/ 6 comparisons = biomass (by taxon) 1996 51 40 0.0083 0.996 776.19 <0.001 6.06 1.52 0.202 30 0.982 159.44 0.001 14.52 2.56 0.062 20 0.988 240.57 0.001 13.35 2.58 0.061 15 0.989 270.90 <0.001 15.93 3.98 0.016 10 0.947 71.82 0.003 18.78 2.45 0.070 5 0.941 48.25 0.006 5.68 0.44 0.680

168

0.05/ 4 comparisons = 1997 21 20 0.0125 1.000 108819.50 <0.001 2.54 1.32 0.259 15 0.986 210.27 0.001 24.80 2.43 0.072 10 0.987 234.84 0.001 24.30 2.43 0.072 5 0.988 242.25 0.001 25.85 2.25 0.088

0.05/ 4 comparisons = 1998 23 20 0.0125 0.999 5966.03 <0.001 6.37 1.12 0.325 15 1.000 9218.54 <0.001 0.79 0.75 0.497 10 0.999 2661.55 <0.001 -12.61 -1.10 0.333 5 0.986 210.42 0.001 -60.80 -1.03 0.360

Continued

0.05/ 2 comparisons = 1999 13 10 0.025 0.986 205.02 0.001 6.36 1.58 0.189

5 0.924 36.70 0.009 13.44 1.45 0.220

0.05/ 2 comparisons = 2000 13 10 0.025 0.914 31.89 0.011 1.68 0.23 0.826 5 0.438 2.33 0.224 1.49 0.06 0.959

2001 8 5 0.05/ 1 comparison = 0.05 0.999 305.29 <0.001 3.00 0.42 0.693

169

0.05/ 2 comparisons = 2002 11 10 0.025 0.999 3345.23 <0.001 0.32 0.36 0.375 5 0.864 19.03 0.022 4.90 0.43 0.689

Zooplankton biomass 0.05/ 6 comparisons = (by taxon) 1996 51 40 0.0083 1.000 19078.71 0.005 -2.67 -0.94 0.445 30 1.000 3586.14 0.011 -2.43 -0.80 0.508 20 1.000 3028.75 0.012 4.88 4.60 0.044 15 1.000 7156.34 0.008 21.21 1.63 0.244 10 1.000 11225.94 0.006 10.29 1.61 0.248 5 1.000 4101.63 0.010 1.36 0.81 0.504

Continued

Table 6 continued

0.05/ 4 comparisons = 1997 21 20 0.0125 1.000 14885.43 0.005 0.48 0.40 0.727 15 1.000 2142000 <0.001 2.72 0.72 0.547 10 0.999 1634.87 0.016 1.06 0.35 0.759 5 1.000 2173.02 0.014 -34.20 -1.12 0.380

0.05/ 4 comparisons = 1998 23 20 0.0125 1.000 13944.76 0.005 -0.10 -0.11 0.923 15 1.000 61166.87 0.003 5.67 1.83 0.209 10 0.993 151.15 0.052 2.22 0.38 0.743

170 5 1.000 6955.02 0.008 20.44 1.74 0.225

0.05/ 2 comparisons = 1999 13 10 0.025 0.995 204.29 0.044 -10.47 -0.78 0.519 5 0.999 808.76 0.022 5.16 2.50 0.130

0.05/ 2 comparisons = 2000 13 10 0.025 0.995 945.81 0.021 -4.33 -1.35 0.309 5 0.999 254.92 0.040 -13.85 -1.37 0.304

2001 8 5 0.05/ 1 comparison = 0.05 0.998 564.89 0.027 8.97 0.86 0.479

0.05/ 2 comparisons = 2002 11 10 0.025 1.000 57254.31 0.003 1.57 0.87 0.478

Table 6 continued

5 0.999 1419.65 0.017 27.96 1.70 0.231

171

Table 7- Spatial and temporal variability of phytoplankton and zooplankton biomass in Lake Erie (1996-2002). F-ratios and p values are from ANOVA analyses. Bold p -values are significant at the Bonferroni corrected α level. Bonferroni corrected alpha value are given using α = 0.05/ number of comparisons for individual statistical test in a given data set (i.e., phytoplankton, zooplankton biomass) in a given year (i.e., phytoplankton biomass- α = 0.05/24 comparisons = 0.002, zooplankton biomass- α = 0.05/15 = 0.0033). Also given is % of the variance explained by basin, site within basin, and month, along with the % residual variance.

172

Group Source F-ratio p-value % of the variance

Chlorophyta 1996 (n=441) Basin 0.81 0.446 0.18 Site (Basin) 1.32 0.038 27.04 Month 1.13 0.344 1.20 Residual 71.58

1997 (n=158) Basin 3.33 0.040 6.04 Site (Basin) 1.00 0.483 23.21 Month 3.13 0.011 9.10 Residual 61.65

1998 (n=206)

Basin 1.02 0.364 0.06 Site (Basin) 1.53 0.027 33.17 Month 1.01 0.390 1.31 Residual 65.46

1999 (n=130) Basin 1.15 0.320 2.14 Site (Basin) 1.31 0.158 30.52 Month 2.48 0.049 6.76 Residual 60.58

2000 (n=183) Basin 9.77 <0.001 6.05 Site (Basin) 1.24 0.195 19.67 Month 1.54 0.195 3.08 Residual 71.20

Continued

173

Table 7 continued

2001 (n=154) Basin 2.32 0.103 2.16 Site (Basin) 1.50 0.060 30.51 Month 0.19 0.979 7.00 Residual 60.33

2002 (n=141) Basin 0.96 0.386 4.77 Site (Basin) 0.38 1.000 12.69 Month 1.60 0.157 7.86 Residual 74.68

Chrysophyta 1996 (n=441) Basin 0.67 0.513 0.38 Site (Basin) 1.20 0.125 21.91 Month 8.58 <0.001 8.80 Residual 68.91

1997 (n=158) Basin 0.93 0.397 3.37 Site (Basin) 0.99 0.501 23.22 Month 9.05 <0.001 22.00 Residual 54.41

1998 (n=206) Basin 2.65 0.074 4.53 Site (Basin) 0.77 0.850 20.03 Month 2.48 0.063 3.54 Residual 71.90

1999 (n=130) Basin 5.60 0.005 11.10 Site (Basin) 1.32 0.153 29.55

Continued

174

Table 7 continued

Month 0.80 0.528 2.06 Residual 57.29

2000 (n=183) Basin 4.60 0.012 6.75 Site (Basin) 1.73 0.014 26.56 Month 1.84 0.125 3.28 Residual 63.41

2001 (n=154) Basin 1.00 0.370 0.74 Site (Basin) 3.19 <0.001 48.25 Month 0.74 0.618 1.96 Residual 49.05

2002 (n=141) Basin 0.37 0.691 5.68 Site (Basin) 0.46 0.996 12.61 Month 1.63 0.148 7.93 Residual 73.78

Cryptophyta 1996 (n=441) Basin 0.79 0.455 0.37 Site (Basin) 0.67 0.990 15.10 Month 6.42 <0.001 7.37 Residual 77.16

1997 (n=158) Basin 23.82 <0.001 19.32 Site (Basin) 2.68 <0.001 42.82 Month 0.11 0.990 0.20 Residual 37.66

Continued

175

Table 7 continued

1998 (n=206) Basin 10.27 <0.001 9.80 Site (Basin) 1.00 0.492 19.92 Month 5.22 0.002 6.61 Residual 63.67

1999 (n=130) Basin 8.73 <0.001 15.32 Site (Basin) 1.17 0.274 25.39 Month 1.58 0.186 3.94 Residual 55.35

2000 (n=183) Basin 9.08 <0.001 6.14 Site (Basin) 1.10 0.338 14.58 Month 9.60 <0.001 16.88 Residual 62.40

2001 (n=154) Basin 16.26 <0.001 12.99 Site (Basin) 1.74 0.017 29.01 Month 1.60 0.155 4.61 Residual 53.39

2002 (n=141) Basin 0.64 0.527 2.62 Site (Basin) 0.30 1.000 11.50 Month 0.29 0.940 1.61 Residual 84.27

Continued

176

Table 7 continued

Cyanophyta 1996 (n=441) Basin 1.37 0.255 0.13 Site (Basin) 1.22 0.105 23.76 Month 4.15 0.001 4.43 Residual 71.68

1997 (n=158) Basin 0.30 0.744 1.02 Site (Basin) 0.68 0.923 20.00 Month 1.83 0.114 6.29 Residual 72.69

1998 (n=206) Basin 1.33 0.269 2.43 Site (Basin) 0.64 0.965 16.25 Month 5.88 0.001 8.51 Residual 72.81

1999 (n=130) Basin 1.35 0.264 2.96 Site (Basin) 0.72 0.855 20.34 Month 1.11 0.356 3.65 Residual 73.05

2000 (n=183) Basin 1.35 0.262 1.19 Site (Basin) 0.89 0.649 16.92 Month 1.96 0.104 4.29 Residual 77.60

Continued

177

Table 7 continued

2001 (n=154) Basin 3.62 0.030 4.86 Site (Basin) 0.38 0.999 9.53 Month 3.79 0.002 14.57 Residual 71.04

2002 (n=141) Basin 0.01 0.985 1.03 Site (Basin) 0.50 0.993 14.23 Month 2.21 0.049 10.76 Residual 73.98

Pyrrophyta 1996 (n=441) Basin 2.22 0.110 0.98 Site (Basin) 1.53 0.003 29.83 Month 2.34 0.042 2.33 Residual 66.86

1997 (n=158) Basin 0.62 0.540 0.17 Site (Basin) 0.81 0.781 20.32 Month 3.14 0.011 10.25 Residual 69.26

1998 (n=206) Basin 2.80 0.064 3.65 Site (Basin) 1.21 0.195 25.01 Month 4.27 0.006 5.57 Residual 65.77

Continued

178

Table 7 continued

1999 (n=130) Basin 2.32 0.104 3.65 Site (Basin) 0.67 0.905 19.23 Month 0.98 0.421 3.26 Residual 73.86

2000 (n=183)

Basin 2.24 0.110 3.01 Site (Basin) 1.09 0.348 19.33 Month 1.45 0.222 3.04 Residual 74.62

2001 (n=154) Basin 0.61 0.547 1.25 Site (Basin) 0.66 0.918 14.30 Month 2.98 0.010 11.72 Residual 72.73

2002 (n=141) Basin 0.04 0.959 1.21 Site (Basin) 0.68 0.918 18.84 Month 2.34 0.038 10.68 Residual 69.27

Edible Phytoplankton 1996 (n=441) Basin 0.84 0.434 0.42 Site (Basin) 1.14 0.207 20.82 Month 10.17 <0.001 10.36 Residual 68.40

Continued

179

Table 7 continued

1997 (n=158) Basin 3.01 0.054 6.58 Site (Basin) 0.96 0.546 20.97 Month 12.76 <0.001 27.22 Residual 45.23

1998 (n=206) Basin 4.85 0.009 4.67 Site (Basin) 1.38 0.074 29.24 Month 3.14 0.027 3.89 Residual 62.20

1999 (n=130) Basin 3.08 0.051 7.34 Site (Basin) 1.21 0.238 27.45 Month 2.51 0.047 6.62 Residual 58.59

2000 (n=183) Basin 12.34 <0.001 10.03 Site (Basin) 1.46 0.065 19.80 Month 4.33 0.002 7.64 Residual 62.53

2001 (n=154) Basin 0.25 0.781 0.16 Site (Basin) 2.28 0.001 40.89 Month 0.17 0.985 0.53 Residual 58.42

Continued

180

Table 7 continued

2002 (n=141) Basin 0.58 0.560 6.03 Site (Basin) 0.63 0.947 19.80 Month 0.32 0.927 1.51 Residual 72.66

Inedible Phytoplankton 1996 (n=441) Basin 1.41 0.245 0.16 Site (Basin) 1.21 0.112 22.21 Month 8.19 <0.001 8.43 Residual 69.20

1997 (n=158) Basin 0.00 0.999 0.51 Site (Basin) 0.77 0.831 23.06 Month 2.25 0.054 7.34 Residual 69.09

1998 (n=206) Basin 1.63 0.200 3.03 Site (Basin) 0.62 0.971 15.78 Month 6.07 0.001 8.74 Residual 72.45

1999 (n=130) Basin 1.01 0.370 2.39 Site (Basin) 0.71 0.866 20.16 Month 1.30 0.277 4.27 Residual 73.18

Continued

181

Table 7 continued

2000 (n=183) Basin 3.25 0.042 3.39 Site (Basin) 0.89 0.650 16.50 Month 2.53 0.043 5.33 Residual 74.78

2001 (n=154) Basin 0.85 0.430 0.87 Site (Basin) 3.07 <0.001 46.90 Month 0.85 0.532 2.30 Residual 49.93

2002 (n=141) Basin 0.15 0.860 3.49 Site (Basin) 0.45 0.997 10.35 Month 2.63 0.021 12.72 Residual 73.44

Total Phytoplankton 1996 (n=441) Basin 1.57 0.210 0.33 Site (Basin) 1.26 0.073 21.52 Month 10.98 <0.001 10.98 Residual 67.17

1997 (n=158) Basin 1.34 0.266 4.89 Site (Basin) 1.34 0.113 30.00 Month 6.13 <0.001 14.60 Residual 50.51

Continued

182

Table 7 continued

1998 (n=206) Basin 1.72 0.182 3.54 Site (Basin) 0.67 0.950 16.58 Month 5.03 0.002 7.26 Residual 72.62

1999 (n=130) Basin 0.94 0.393 2.49 Site (Basin) 0.73 0.850 19.87 Month 2.57 0.044 8.03 Residual 69.61

2000 (n=183)

Basin 8.54 <0.001 8.25 Site (Basin) 0.94 0.573 16.03 Month 1.71 0.150 3.49 Residual 72.23

2001 (n=154) Basin 0.12 0.884 0.04 Site (Basin) 2.79 <0.001 45.53 Month 0.40 0.875 1.17 Residual 53.26

2002 (n=141) Basin 0.41 0.662 7.58 Site (Basin) 0.62 0.953 14.88 Month 2.11 0.059 9.49 Residual 68.05

Continued

183

Table 7 continued

Cyclopoida 1996 (n=441) Basin 3.77 0.024 0.67 Site (Basin) 3.17 <0.001 39.83 Month 33.57 <0.001 19.82 Residual 39.68

1997 (n=158) Basin 2.83 0.064 3.31 Site (Basin) 1.27 0.164 29.41 Month 3.84 0.003 10.32 Residual 56.96

1998 (n=206) Basin 7.25 0.001 3.76 Site (Basin) 2.14 <0.001 35.77 Month 8.43 <0.001 8.68 Residual 51.79

1999 (n=130) Basin 3.50 0.034 7.59 Site (Basin) 1.46 0.081 28.65 Month 4.30 0.003 10.32 Residual 53.44

2000 (n=183) Basin 8.24 <0.001 7.52 Site (Basin) 1.19 0.237 14.84 Month 8.28 <0.001 14.68 Residual 62.96

Continued

184

Table 7 continued

2001 (n=154) Basin 13.11 <0.001 7.79 Site (Basin) 1.58 0.040 24.45 Month 6.82 <0.001 18.25 Residual 49.51

2002 (n=141) Basin 1.07 0.346 4.29 Site (Basin) 8.39 <0.001 69.82 Month 8.54 <0.001 9.32 Residual 16.57

Calanoida 1996 (n=441) Basin 0.28 0.754 <0.01 Site (Basin) 2.16 <0.001 29.36 Month 31.44 <0.001 48.10 Residual 22.53

1997 (n=158) Basin 0.19 0.826 0.27 Site (Basin) 1.21 0.212 31.85 Month 1.02 0.411 3.11 Residual 64.77

1998 (n=206) Basin 3.79 0.025 2.99 Site (Basin) 1.36 0.081 30.66 Month 8.82 <0.001 9.90 Residual 56.45

Continued

185

Table 7 continued

1999 (n=130) Basin 1.76 0.177 3.22 Site (Basin) 2.09 0.003 41.81 Month 2.81 0.030 6.17 Residual 48.80

2000 (n=183) Basin 4.45 0.013 2.39 Site (Basin) 1.47 0.061 26.14 Month 1.67 0.161 3.20 Residual 68.27

2001 (n=154) Basin 20.86 <0.001 12.77 Site (Basin) 2.18 0.001 31.47 Month 3.20 0.006 8.22 Residual 47.54

2002 (n=141) Basin 20.26 <0.001 5.76 Site (Basin) 9.49 <0.001 77.92 Month 0.47 0.826 0.49 Residual 15.83

Cladocera 1996 (n=441) Basin 6.33 0.002 1.97 Site (Basin) 1.49 0.005 20.09 Month 33.74 <0.001 26.05 Residual 51.89

Continued

186

Table 7 continued

1997 (n=158) Basin 2.20 0.116 3.24 Site (Basin) 1.07 0.383 24.12 Month 8.11 <0.001 20.09 Residual 52.55

1998 (n=206) Basin 11.71 <0.001 9.47 Site (Basin) 0.78 0.842 17.52 Month 6.39 <0.001 8.22 Residual 64.79

1999 (n=130) Basin 3.80 0.026 5.92 Site (Basin) 1.10 0.355 20.33 Month 7.00 <0.001 17.65 Residual 56.10

2000 (n=183) Basin 8.31 <0.001 5.40 Site (Basin) 1.33 0.129 18.26 Month 9.99 <0.001 16.76 Residual 59.58

2001 (n=154) Basin 25.10 <0.001 21.69 Site (Basin) 0.33 1.000 5.75 Month 5.43 <0.001 16.46 Residual 56.10

Continued

187

Table 7 continued

2002 (n=141) Basin 1.52 0.224 8.28 Site (Basin) 2.35 <0.001 41.07 Month 5.90 <0.001 14.18 Residual 36.47

Crustacean Zooplankton 1996 (n=441) Basin 5.26 0.006 1.57 Site (Basin) 1.94 <0.001 24.23 Month 39.11 <0.001 27.30 Residual 46.90

1997 (n=158) Basin 2.56 0.082 3.41 Site (Basin) 1.32 0.123 27.29 Month 8.00 <0.001 19.00 Residual 50.30

1998 (n=206) Basin 9.69 <0.001 7.16 Site (Basin) 1.11 0.309 23.66 Month 8.53 <0.001 10.02 Residual 59.16

1999 (n=130) Basin 4.14 0.019 6.84 Site (Basin) 1.50 0.066 25.19 Month 8.36 <0.001 18.56 Residual 49.41

Continued

188

Table 7 continued

2000 (n=183) Basin 9.50 <0.001 6.46 Site (Basin) 1.62 0.028 19.97 Month 10.56 <0.001 16.87 Residual 56.70

2001 (n=154) Basin 38.42 <0.001 25.11 Site (Basin) 1.03 0.436 14.87 Month 6.73 <0.001 16.02 Residual 44.00

2002 (n=141) Basin 2.72 0.071 4.88 Site (Basin) 6.67 <0.001 67.77 Month 5.29 <0.001 7.08 Residual 20.27

Total Zooplankton 1996 (n=441) Basin 9.92 <0.001 3.14 Site (Basin) 1.07 0.337 16.82 Month 28.95 <0.001 24.10 Residual 55.94

1997 (n=158) Basin 0.78 0.460 1.15 Site (Basin) 1.11 0.323 29.56 Month 1.47 0.207 4.48 Residual 64.81

Continued

189

Table 7 continued

1998 (n=206) Basin 12.30 <0.001 5.97 Site (Basin) 1.24 0.161 24.98 Month 10.73 <0.001 12.14 Residual 56.91

1999 (n=130) Basin 16.56 <0.001 19.56 Site (Basin) 0.64 0.929 6.52 Month 13.13 <0.001 27.43 Residual 46.49

2000 (n=183) Basin 8.71 <0.001 5.53 Site (Basin) 1.47 0.064 20.88 Month 4.30 0.003 7.95 Residual 65.64

2001 (n=154) Basin 32.05 <0.001 23.80 Site (Basin) 0.75 0.831 12.31 Month 7.09 <0.001 17.70 Residual 46.19

2002 (n=141) Basin 12.53 <0.001 3.61 Site (Basin) 11.00 <0.001 80.32 Month 2.35 0.037 2.16 Residual 13.91

190

(a)

Phytoplankton Basin Breakpoint Analysis Regression Analysis Biomass- Year r2 Slope Axis r2 p Intercept Before -0.1675 334.76 0.85 <0.001 Western 1997 0.74 After -0.1211 245.00 0.05 0.719

Before -0.1590 316.36 0.997 0.002 Central 1984 0.81 After -0.0125 25.96 0.05 0.520

Before -0.0653 130.43 0.67 0.007 Eastern 1997 0.51 After -0.3625 726.80 0.35 0.291

(b)

Zooplankton Basin Breakpoint Analysis Regression Analysis Biomass- Year r2 Slope Axis r2 p Intercept Before -0.0163 32.344 0.85 0.009 Western 1986 0.83 After -0.0010 2.043 0.07 0.543

Before -0.0029 5.756 0.41 0.063 Central 1999 0.53 After -0.0290 58.160 0.90 0.208

Before -0.0240 4.848 0.42 0.553 Eastern 1985 0.82 After 0.0008 -1.513 0.15 0.307

Table 8- Breakpoint analysis of (a) total phytoplankton biomass and (b) total crustacean zooplankton biomass from 1970-2002. The breakpoint year and r2 and the regression slope, intercept, r2, and p-value are given for the regression equations before and after the breakpoint year (after Conroy et al. 2004a).

191

Average Average Site Date Number of Individuals Length (mm) Weight (µg) Density (#/m3) Biomass (µg/m3)

Cercopagis 27 8 Aug 1 1.1 2.8 0.5 1.0 29 27 Aug 1 1.1 3.2 0.3 1.0 27 6 Sep 5 1.0 2.2 2.6 6.0

192

Average 2.3 1.1 2.7 1.1 2.7

Leptodora 27 8 Aug 12 1.9 5.9 720 4,300 29 27 Aug 20 1.9 7.0 630 4,400 27 6 Sep 0 0.0 0.0 0 0

Average 16 1.9 6.5 675 4,350

Table 9- Lengths, weights, densities, and biomasses of Cercopagis pengoi sampled during August and September 2001 in the western basin of Lake Erie. Average length. average weight, density, and biomass of Leptodora kindti from the same dates are included for comparison. No individuals of Bythotrephes longimanus were found. See Figure 7 for site identification. 192

Total Number of Frequency Mean Number Samples of Abundance Standard of Containing Occurrence Site (#/m3) Error Samples Limnocalanus (%)

EB 1996-2000 1.6 1.6 101 1 1.0

CB 1996 5.2 4.5 82 2 2.4 CB 1997 5.4 3.3 51 3 5.9 CB 1998 0.9 0.9 73 1 1.4 CB 1999 0.0 0.0 50 0 0.0 CB 2000 4.5 4.5 82 2 2.4

CB 1996- 2000 3.4 1.6 338 8 2.4

WB 1995 14.6 4.1 194 33 17.0 WB 1996 11.0 3.3 458 25 5.5 WB 1997 16.1 4.4 138 20 14.5 WB 1998 8.7 3.9 144 6 4.2 WB 1999 1.3 0.9 80 2 2.5 WB 2000 17.5 6.6 93 13 14.0

WB 1995-2000 11.8 1.8 1107 99 8.9

All Basins 1995-2000 9.3 1.4 1546 108 7.0

Table 10- Spatial and temporal abundances of Limnocalanus macrurus throughout sampling season in each basin of Lake Erie 1995-2000. Total numbers of samples, samples containing L. macrurus and frequency of occurrence of L. macrurus in samples, expressed as a percentage. EB = eastern basin, CB= central basin, WB= western basin

193

Month Mean Abundance (#/m3) Standard Error

1995 May 28.2 9.5 June 19.4 7.0 July 0.0 0.0

1996 May 84.7 35.7 June 19.7 8.0 July 0.0 0.0 August 0.0 0.0 September 2.9 2.5 October 0.0 0.0

1997 May 30.7 15.8 June 29.3 9.4 July 11.5 8.9 August 0.0 0.0 September 0.0 0.0 October 0.0 0.0

1998 May 0.0 0.0 June 4.4 3.1 July 18.7 11.3 August 6.1 6.1 September 0.0 0.0

1999 May 8.1 5.6 June 0.0 0.0 July 0.0 0.0 August 0.0 0.0 September 0.0 0.0

2000 May 73.4 24.8 June 0.5 0.5 July 0.0 0.0 August 0.0 0.0 September 0.0 0.0

Table 11- Monthly variation in abundance of Limnocalanus macrurus in the western basin of Lake Erie (1995-2000).

194

Candidate Metrics Description/ Measure of Hypothesized Reference Ecological Relevance Response to Degradation 1. Zooplankton Ratio Biomass of Trophic Status Decrease Gannon and Stemberger (1978) (Calanoida/ (Cladocera + Cyclopoida) low values indicate eutrophy 2. Mean Zooplankton Larger taxa are Quality of Fish Food Decrease or Mills and Schiavone Jr. (1982), Size preferred food for Increase Mills et al. (1987) many fish

195 3. Rotifer Composition Taxa are indicative of Trophic Status Change in Gannon and Stemberger (1978) trophic conditions Taxonomic Composition 4. Density of Taxon is intolerant to Trophic Status/ Oxygen Decrease Gannon and Beeton (1971), Limnocalanus eutrophic/ anoxic Conditions Kane et al. (2004) macrurus conditions 5. % Biomass of Shifts energy away Fish Food Quality, Increase Lehman and Caceres (1993), Predatory Invasive from fish, reduces Access Food Web Hoffman et al. (2001), Zooplankters zooplankton numbers Uitto et al. (1999) 6. Biomass of Lower during blooms Trophic Status Decrease Havens 1998, Zooplankton/ Biomass of nuisance/ inedible/ Xu et al. 2001 of Phytoplankton toxic algae 7. Biomass of Preferred food for Quality of Fish Food Increase Gopalan et al.1998, Crustacean many fish Culver, Unpublished Zooplankton Table 12- Zooplankton candidate metrics for Lake Erie P-IBI, along with their description/ ecological relevance, what they measure, and their hypothesized response to degradation. Metrics included in the final multimetric P-IBI are in bold.

Candidate Metrics Description/ Ecological Measure of Hypothesized Reference Relevance Response to Degradation 8. Generic Index of % Abundance of (Achnanthes, Organic Decrease Wu (1999) Diatoms Cocconeis, and Cymbella)/ Pollution (Cyclotella, Melosira, and Nitszchia)

9. Centrales Low values indicate oligotrophy Trophic Status Increase Nygaard (1949), Abundance/Pennales Rawson (1956) Abundance

10. Biomass of Inedible Large, gelatinous, and colonial Quality of Increase DeMott and Moxter

196 Algae Taxa algae are of poor food quality Zooplankton/ (1991)

Fish Food

11. % Bluegreen Algae Bluegreen algae cause Presence of Increase Gliwicz and Siedlar (Biomass) mechanical/ chemical Inedible/ Toxic (1980) Gliwicz and interference to zooplankton Taxa Lampert (1990) feeding Carmichael (1986), (1997) 12. % Abundance of Affect human health and health Presence of Increase Carmichael (1986), Microcystis, Anabaena, of aquatic organisms Toxins (1997) Aphanizomenon

13. Biomass of Edible Provide quality nutrition for Quality of Decrease Kerfoot et al. (1988) Algae Taxa growth and reproduction of Zooplankton/ animals Fish Food Table 13- Phytoplankton candidate metrics for Lake Erie P-IBI, along with their description/ ecological relevance, what they measure, and their hypothesized response to degradation. Metrics included in the final multimetric P-IBI are in bold.

PHYTOPLANKTON Munawar and Frost and Culver (2001) Munawar (1976) Year(s) 1970 1996 Sampling Dates April-December May- October Sampling Frequency Monthly Weekly-Monthly Number of Sites 25 ~30-80 Sampling Method VanDorn bottle, 1- and Integrated Water Sampler 5m Size of Sample ? π * 1.252 * twice Secchi depth Sampling Depth and Mixing 1- and 5-m. Samples Twice the Secchi depth. mixed. Sample mixed. Enumeration Method Utermöhl (1958) Utermöhl (1958) Taxonomic Resolution Genus and species Genus Biomass Calculation Method Volumes determined, Volumes determined, converted to wet converted to wet weight weight ZOOPLANKTON Watson and Carpenter Frost and Culver (2001) (1974), Watson (1976),

197 Bean (1980)

Year (s) 1970 1995-2002 Sampling Dates April-December Late April/ early May- late September/ early October, typically Sampling Frequency Monthly Weekly-Monthly Number of Sites 25 ~30-80 Sampling Method Unmetered 64 µm net Metered 64 µm net Sampling Depth 2m above the bottom 1m above the bottom to surface to surface (or 50m to surface) Enumeration Method 200 individuals of 100 individuals of most each taxon common taxon Taxonomic Resolution Genus and species Genus and species Biomass Calculation Method Culver et al. (1985) Culver et al. (1985)

Table 14- Comparison of sampling dates, sampling frequency, number of sites, sampling method, enumeration method, and biomass calculation method for phytoplankton and zooplankton in Lake Erie (1970 and 1995-2002).

197

Oligotrophic Mesotrophic Eutrophic (least degraded) (most degraded) Total phosphorus (µg P/L) <11.0 11.0-21.7 >21.7 Chlorophyll a (µg/L) <2.9 2.9-5.6 >5.6 Individual metric value 5 3 1 Possible summed trophic status metric values 10, 8 6 2, 4 Possible combinations to give sum 5,5 3,5 5,3 3,3 1,5 5,1 1,1 1,3 3,1 198

Table 15- Ranges of total phosphorus (µg P/L) and chlorophyll a (µg/L) for each trophic class (Chapra and Dobson 1981). Individual and summed (total phosporus + chlorophyll a) metric values are given for each class, as well as possible combinations that give the summed metric values.

198

Simple Kappa Exact Test Weighted Kappa Exact Test Month/ Model Metrics in Model (95% Confidence Bounds) (P > |K|) (95% Confidence Bounds) (P > |K|) May/ Full Model 1,4,5,6,7,9,10,12,13 0.1390 (-0.0779, 0.3558) 0.1735 0.1497 (-0.0811, 0.3804) 0.2149 May/ Reduced Model 1,7,9 0.1002 (-0.1033, 0.1038) 0.3457 0.0497 (-0.1790, 0.2785) 0.7951

June/ Full Model 1,4,5,6,7,9,10,12,13 0.1462 (-0.0182, 0.3107) 0.0715 0.2167 (0.0518, 0.3816) 0.0098 June/ Reduced Model 1, 12, 13 0.1411 (-0.0010, 0.2831) 0.0296 0.2420 (0.1053, 0.3788) 0.0006

July/ Full Model 1,4,5,6,7,9,10,12,13 -0.0245 (-0.1258, 0.0768) 0.7570 0.0298 (-0.0851, 0.1448) 0.8214 July/ Reduced Model 4 0.2054 (0.0939, 0.3170) <0.0001 0.2664 (0.1274, 0.4054) <0.0001

199 August/ Full Model 1,4,5,6,7,9,10,12,13 0.0396 (-0.0869, 0.1660) 0.6378 0.0765 (-0.0805, 0.2335) 0.3952 August/ Reduced Model 1, 7 0.0951 (0.0091, 0.1810) 0.0320 0.1250 (0.0132, 0.2369) 0.0320

September/ Full Model 1,4,5,6,7,9,10,12,13 -0.0358 (-0.1479,0.0762) 0.7517 -0.191 (-0.1832, 0.1451) 1.0000 September/ Reduced Model 12 -0.0197 (-0.0572, 0.0179) 0.5472 -0.0180 (-0.0526, 0.0166) 0.5280

Table 16- 95th percentile of P-IBI metrics for all Lake Erie observations (3 basins, 1970 and 1996). Cutoff values were determined by trisection of the 0-95th percentile range for metric scores for each metric and assigned values of 1, 3, or 5. For metrics that have a 0.00 value in the 1 scoring class, the cutoff minimal values should be read from left to right, while those with a 0.00 value in the 5 scoring class should be read from right to left.

199

Month Step Metric Partial r2 F P > F May 1 7. Biomass of crustacean zooplankton/ biomass of phytoplankton 0.1958 5.97 0.0048 2 1. Zooplankton ratio 0.1385 3.86 0.0279 3 9. Centrales/ Pennales 0.1321 3.58 0.0359

June 1 13. Biomass of edible algae taxa 0.1666 8.00 0.0007 2 12. % Microcystis, Anabaena, and Aphanizomenon of total phytoplankton biomass 0.0660 2.79 0.0673 3 1. Zooplankton ratio 0.0547 2.26 0.1116

July 1 4. Limnocalanus macrurus density 0.0398 1.93 0.1511 200

August 1 1. Zooplankton ratio 0.1305 6.30 0.0028 2 7. Crustacean zooplankton biomass 0.0828 3.75 0.0277

September 1 12. % Microcystis, Anabaena, and Aphanizomenon of total phytoplankton biomass 0.0968 3.11 0.0522

Table 17- Significant metrics for each month, as determined by stepwise discriminant analysis. The metric, its partial r 2, F, and p values are all given. Significance level to enter and remove was set at p = 0.20. Metric numbers are the same as those in Tables 2 and 3.

200

Table 18- Metrics excluded/included in final P-IBI. Reasons for not including rejected candidate metrics are also given. DA=

201 included in discriminant analysis, IBI= included in final IBI.

201

Candidate Metrics Reference DA IBI Reason for Rejection 1. Zooplankton Ratio Gannon and Stemberger (1978) Yes Yes 2. % Density of Large Daphnia Mills et al. (1982), (1987) No No Confounding factors

3. Rotifer Composition Gannon and Stemberger (1978) No No Impractical, not sampled accurately 4. Density of Limnocalanus macrurus Gannon and Beeton (1971) Yes Yes

5. % Biomass of Predatory Invasive Zooplankters Lehman and Caceres (1993), Yes No Not significant in (Bythotrephes and Cercopagis) Hoffman et al. (2001), discriminant analysis Uitto et al. (1999)

202 6. Biomass of Crustacean Zooplankton/ Biomass of Havens (1998), Yes No Not significant in

Phytoplankton Xu et al. (2001) discriminant analysis 7. Biomass of Crustacean Zooplankton Xu et al. (2001) Yes Yes

8. Generic Index of Diatoms Wu (1999) No No Not tested for use in lakes

9. Centrales/Pennales Nygaard (1941) Yes No Not significant in discriminant analysis 10. Biomass of Inedible Algae Taxa Demott and Moxter (1991) Yes No Not significant in discriminant analysis 11. % Bluegreen Algae (Biomass) Gliwicz and Siedlar (1980), No No Redundant with Metric 12 Gliwicz and Lampert (1990), Carmichael (1996) 12. % Biomass of Microcystis, Anabaena, Carmichael (1996) Yes Yes Aphanizomenon 13. Biomass of Edible Algae Taxa Kerfoot et al. (1988) Yes Yes 202

Month Metric 95th Percentile Cutoff Values 1 3 5 June Biomass of edible algae taxa (µg/L) 2545.67 1697.12 848.56 0.00 % Microcystis, Anabaena, and Aphanizomenon of total phytoplankton biomass 2.00 1.33 0.67 0.00

203 Zooplankton ratio (Calanoida/ (Cladocera + Cyclopoida)) 0.43 0.00 0.14 0.29 July Limnocalanus macrurus density (#/L) 0.03 0.00 0.01 0.02

August Zooplankton ratio (Calanoida/ (Cladocera + Cyclopoida)) 0.56 0.00 0.19 0.37 Crustacean zooplankton biomass (µg/L) 220.00 146.66 73.33 0.00

Table 19- 95th percentile of P-IBI metrics for all Lake Erie observations (3 basins, 1970 and 1996). Cutoff values were determined by trisection of the 0-95th percentile range for metric scores for each metric and assigned values of 1, 3, or 5. For metrics that have a 0.00 value in the 1 scoring class, the cutoff minimal values should be re ad from left to right, while those with a 0.00 value in the 5 scoring class should be read from right to left.

203

Sum of metric scores for each Metric Metric Metric Metric Metric month Year Basin Site Month 1 Score 4 Score 7 Score 12 Score 13 Score 1997 EB 933 June 0.17 3 XXX XXX XXX XXX 0 5 70.28 5 13 1997 EB 933 July XXX XXX 0 1 XXX XXX XXX XXX XXX XXX 1 1997 EB 933 August 1.57 5 XXX XXX 22.42 5 XXX XXX XXX XXX 10

204

Table 20- Sample calculation of P-IBI for mean site score. XXX refers to a month when the particular metric is not significant in the discriminant analysis and thus not used as part of the P-IBI. EB= eastern basin of Lake Erie. Although western and central basin mean site score calculations are not shown, they follow the same methodology as the eastern basin example given below. Metric 1 = zooplankton ratio (Calanoida/ (Cladocera + Cyclopoida)), metric 4 = Limnocalanus macurus density (#/L), metric 7 = crustacean biomass (µg/L), metric 12 = % Microcystis, Anabaena, and Aphanizomenon of total phytoplankton biomass, and metric 13 = Biomass of edible algae taxa (µg/L). Sum of metrics across June, July, August equals 24. Mean site score equals 24 (13+1+10)divided by 6 (the number of metrics summed).

204

Year Basin Site Mean Site Score 1997 EB 933 4.00 EB 935 3.67 EB 936 3.67 EB 937 1.00 EB 942 5.00 EB 23 3.67

205

Table 21- Sample calculation of mean basin score and mean lakewide score for the P-IBI. Mean basin score equals sum across sites (21.01) divided by the number of sites (6), which is 3.50. Mean lakewide score equals the sum of the mean basin scores divided by the number of basins in Lake Erie (3). For this example, mean western basin score equals 3.69, mean central basin score equals 3.69, and mean eastern basin score equals 3.50. Therefore, the mean lakewide score equals the sum of these three values (10.88) divided by 3, which is 3.63.

205

WB CB EB Lakewide Year WB Trophic Status CB Trophic Status EB Trophic Status Lakewide Trophic Status 1970 2.76 E 3.07 M 2.81 E 2.88 E 1995 3.67 M No Data - No Data - Not Calculated - 1996 2.89 E 3.35 M 4.13 O 3.46 M 1997 3.69 M 3.69 M 3.50 M 3.63 M 206 1998 3.00 M 3.22 M 2.92 E 3.04 M

1999 3.01 M 2.78 E 2.56 E 2.78 E 2000 2.54 E 2.95 E 3.92 M 3.14 M 2001 2.61 E 3.89 M 3.33 M 3.28 M 2002 2.51 E 2.64 E 2.33 E 2.50 E

Table 22- Basin and lakewide P-IBI scores in Lake Erie for 1970, and 1995-2002. The basin scores are the means of all the site scores in the basin. The lakewide score is the mean of the three basin scores. Trophic status classifications are based on IBI scores and are based on the scale of <3 reflecting eutrophic conditions, 3 -4 reflecting mesotrophic conditions, and >4 reflecting oligotrophic conditions. WB = western basin, CB = central basin, and EB = eastern basin; E = eutrophic, M = mesotrophic, and O = oligotrophic.

206

APPENDIX B

207 FIGURES

207

Michigan Ontario

New York

Pennsylvania 84 N

208

Ohio 0 50 km

Figure 1- Locations of Lake Erie Plankton Abundance Study (LEPAS) sampling sites in Lake Erie from 1996-2002 (every site was not sampled every year). Squares indicate sites sampled in 1996, crosses indicate sites sampled in 1997, and circles indicate sites typically sampled during 1998-2002. Sites 84 and 1279 in the central basin were used in spatial comparisons.

208

Detroit River

Pelee Passage

209

Maumee N River

0 5 km Sandusky River

Figure 2- Lake Erie Plankton Abundance Study (LEPAS) sites sampled in the western basin of Lake Erie during 1996-2002. The identities of subsets used in spatial comparisons are given in Table 3.

209

Figure 3- Volumetric Index of the Plankton (VIP) regression analyses. (a) Relationship between log crustacean zooplankton biomass (µg/L) and log standardized crustacean zooplankton volume in Lake Erie during 1998. (b) Relationship between log cyanophyte biomass (µg/L) and log standardized 210 cyanophyte volume in Lake Erie during 1998.

210

Figure 3 (a)

y = 4.93 + 0.62 x r2 = 0.312 p < 0.001 n = 192

0 log Crustacean Biomass (ug/L) Biomass Crustacean log

-7 -6 -5 -4 log Standardized Crustacean Volume (mL/L) 211

(b)

y = 2.20 + 0.82 x r2 = 0.206 p = 0.001 5 n = 50

0 log Cyanophyte Biomass (ug/L) Biomass Cyanophyte log

0 1 2 log Standardized Cyanophyte Volume (mL/L)

211

30 28 26 24 22 20 18 16 14

12 10 Target Loading (11 kilotonnes) 8 212 6

External Phosphorus Loading (kilotonnes) Loading Phosphorus External

0 1970 1975 1980 1985 1990 1995 2000 Year

Figure 4- Phosphorus loading (kilotonnes) in Lake Erie 1969-2001. Each year represents a water year. A water year begins on October 1 and ends on September 30 of the following year and is referred to by the year in which it ends. For example, the phosphorus loading values for the year 1995 are calculated from October 1994- September 1995 (i.e. the 1995 water year). Water years are used to calculate loading because data from which loading is calculated (USGS stream flow data) are compiled by water year. Data are from Dolan 1993 and Dolan unpublished ( [email protected], University of Wisconsin - Green Bay). The target loading level of 11 kilotonnes is based upon the Great Lakes Water Quality Agreement.

212

6 0.35 (a) (d) 5 0.30

0.25 4 0.20 3 0.15 2 0.10

1 0.05

0 0.00 6 0.35 (b) (e) 5 0.30

0.25 4 0.20 3 0.15 2 0.10

1 0.05

0 0.00 Total Phytoplankton Biomass (mg/L) Biomass Phytoplankton Total 213 6 0.35 (c) (f)

5 0.30 Total CrustaceanZooplankton Biomass (mg/L)

0.25 4 0.20 3 0.15 2 0.10

1 0.05

0 0.00 1965 1970 1975 1980 1985 1990 1995 2000 2005 1965 1970 1975 1980 1985 1990 1995 2000 2005 Year Year Figure 5- Temporal change in Lake Erie seasonal average phytoplankton and crustacean zooplankton biomass (mg/L) in the western (a,d), central (b,e), and eastern (c,f) basins. Phytoplankton data (a,b,c) from 1970 are from Munawar and Munawar (1976), 1978 data are from Devault and Rockwell (1986), 1983-87 data are from Makarewicz (1993a), and 1996-2002 data are from the Lake Erie Plankton Abundance Study (LEPAS). Zooplankton data (d,e,f) from1970 are from Bean (1980), 1974-75 data (western basin only) are from Weisgerber (2000), 1984-87 data are from Makarewicz (1993b), and 1996-2002 data are from LEPAS. Error bars are +1 standard error and are only calculated for the 1996-2002 data. Arrows indicate breakpoint years (Table 8) (after Conroy et al. 2004a).

213

Figure 6- Lake Erie seasonal average phytoplankton biomass (mg/L) (a,b,c) and seasonal

214 average crustacean zooplankton biomass (mg/L) (d,e,f) as a function of lake-wide annual

estimated total phosphorus loading (ktonnes) for the western (a,d), central (b,e), and eastern (c,f) basins from 1970 to 2001 (after Conroy et al. 2004a).

214 Figure 6

6 0.35 slope=0.1460 (a) slope=0.0100 (d) 2 2 5 r =0.31, p=0.048 0.30 r =0.43, p=0.015

n=13 0.25 n=13 4 Western 0.20 Western 3 0.15 2 0.10

1 0.05

0 0.00 6 0.35 (b) slope=0.0972 slope=0.0031 (e) 0.30 5 r2=0.49, p=0.007 r2=0.13, p=0.273 n=13 0.25 n=11 4

0.20 3 Central Central 0.15 2 0.10

1 0.05

0 0.00

Total Phytoplankton Biomass (mg/L) Biomass Total Phytoplankton 6 0.35 slope=0.0185 (c) slope=0.0043 (f) 2 2 5 0.30

r =0.01, p=0.735 Total Crustacean ZooplanktonBiomass (mg/L) r =0.44, p=0.026 n=13 0.25 n=11 4 Eastern 0.20 Eastern 3 0.15 2 0.10

1 0.05

0 0.00 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Total Phosphorus Load (ktonnes) Total Phosphorus Load (ktonnes)

215

Detroit River

29 Pelee Passage

N 0 5 km Maumee River

Sandusky River

Figure 7- Sites sampled in the western basin of Lake Erie during 2001. Filled-in circles denote sites whose samples contained Cercopagis pengoi, sampled sites that did not contain C. pengoi are denoted by crosses. Site 27 is located at 41” 46.32’ N, 83” 3.17’ W; Site 29 is located at 41” 51.59’ N, 83” 8.07’ W.

216

Temperature (°C)

10 11 12 13 14 15 16 17 18 19 20 0

6 Temperature

O2 8 Temperature Limit- 14°C

10 Depth (m) 12

Limnocalanus Density 14

18 Oxygen Limit- 5.6 mg/L

22 0 1 2 3 4 5 6 7 8 9 10

3 3 Dissolved Oxygen (mg/L) and Limnocalanus Density (10 /m )

Figure 8- Vertical distribution of Limnocalanus macrurus in the Steinsfjord, Norway demonstrating its existence between narrow environmental limits (redrawn from Gannon and Beeton 1971, after Strøm 1946).

217

0 50 km

Figure 9- Lake Erie Plankton Abundance Study (LEPAS) sampling sites 1995-2000 (open circles), sites where Limnocalanus macrurus was found on one or more dates in Lake Erie 1995-2000 (crosses), and mean abundance (#/m3) of Limnocalanus macrurus (contour lines) in Lake Erie 1995-2000 (all sampling dates). Contour intervals are 10/ m3.

218

Figure 10- Uncorrected chlorophyll a concentrations (µg/L) (a) and unfiltered total phosphorus concentrations (µg/L) (b) in the three basins of Lake Erie in 1970 and 1996. The boxes indicate the 25th percentile, median, and 75th percentile, whiskers extend from the 10th to the 90th percentile, while dots indicate the 5th and 95th percentile values.

219

Figure 10 (a) 35

20 (ug/L) a

Chlorophyll 5

0 n=17 n=109 n=43 n=314 n=12 n=7

WB CB EB WB CB EB 1970 1996

(b) 100

Total phosphorus (ug/L) 0

WB CB EB WB CB EB 1970 1996 Year 220

Figure 11- 1970 and 1996 sites used for Planktonic IBI development in Lake Erie. Black triangles represent 1970 sites, while white circles represent 1996 sites.

221

y = -184.046 + 526.808 x r2 = 0.151 p < 0.001

8000

7000

6000

5000

4000

3000

2000

1000

0 Edible phytoplankton biomass (ug/L) biomass phytoplankton Edible Oligotrophic Mesotrophic Eutrophic Trophic Status

Figure 12- Edible phytoplankton biomass versus trophic status (as determined by total phosphorus and chlorophyll a) regression analysis for June (1970 and 1996). oligotrophic =1, mesotrophic=2, eutrophic=3.

222

8000

6000

4000

95 1

2000 90 75 3 median 25 5 0 10 5 Edible phytoplanktonbiomass (ug/L)

Figure 13- Sample graph demonstrating boxplot conventions and trisection technique (based on Karr et al. 1996). The boxes in the boxplots are comprised of horizontal lines designating the 25th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentile of the cumulative frequency distribution of edible phytoplankton biomass. These conventions are followed for all subsequent boxplots. Also demonstrated is the trisection technique used and individual metric score assignment (lines are drawn for demonstration purposes and thus are approximations). In this case low values (0 - 849 µg/L) are assigned the highest water quality value, 5.

223

Figure 14- Comparison of cumulative frequency distributions for P-IBI metrics in June (a-c), July (d), and August (e-f) in 1970 and 1996 across all Lake Erie basins. The boxes indicate the 25th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentiles. These conventions are followed for all subsequent boxplots. Please note that due to different calculation techniques, the 95th percentiles in these visualizations may differ slightly from those in Table 9. However, the values in Table 9 are the values from which all classifications for the P-IBI were made.

224

Figure 14

(a) (d)

4000 0.035

0.030 3000 0.025

abundance (#/L) abundance 0.020 2000 0.015

0.010 1000

0.005 Edible phytoplankton (ug/L) biomass 0 0.000 Limnocalanus macrurus macrurus Limnocalanus

(b) July (e)

1.0

0.8 of total phytoplankton biomass total of phytoplankton

0.6 2

0.4 Aphanizomenon

, and and , 0.2

0 0.0 Anabaena , Microcystis

% Abundance ofCalanoida/ (Cyclopoida + Cladocera)

0.6

0.5 300

0.4 250

0.3 200

0.2 150

0.1 100

0.0 50

0 Abundance ofCalanoida/ (Cyclopida + Cladocera) (ug/L) biomass zooplankton Crustacean June August 225

Figure 15- Comparison of cumulative frequency distributions of P-IBI metrics versus trophic status in June (a-c), July (d), and August (e-f) in 1970 and 1996 across all Lake Erie basins. The boxes indicate the 25th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentiles. H and p-values are from the Kruskal-Wallis tests that were used to test the equality of medians. Significance is judged at a 0.05 α level and is indicated in bold.

226

Figure 15

H = 24.86 (a) H = 7.70 (d) p < 0.001 p = 0.021 8000 0.035

0.030 6000 0.025

4000 abundance (#/L) 0.020

0.015 2000 0.010

0 0.005

0.000 Edible phytoplankton biomass (ug/L) Oligotrophic Mesotrophic Eutrophic Limnocalanus macrurus Oligotrophic Mesotrophic Eutrophic

(b) July (e) H = 18.91 p < 0.001

H = 0.18 2.5 p = 0.914 12 2.0

10 1.5 Aphanizomenon 8

1.0 , and and , 6

4 0.5

Anabaena 2 , 0.0

0 of total phytoplankton biomass phytoplankton total of

Oligotrophic Mesotrophic Eutrophic Microcystis Oligotrophic Mesotrophic Eutrophic Abundance of Calanoida/ (Cyclopoida+ Cladocera) % of % of

H = 5.95 (c) H = 18.75 (f) p = 0.051 p < 0.001 700 1.0

600 0.8 500

0.6 400

0.4 300

200 0.2

100 0.0

Oligotrophic Mesotrophic Eutrophic

Crustacean zooplanktonbiomass (ug/L) Oligotrophic Mesotrophic Eutrophic

Abundance of Calanoida/ (Cyclopoida/ Cladocera) (Cyclopoida/ Calanoida/ of Abundance

June August

Trophic Status Trophic Status 227

Figure 16- Comparison of cumulative frequency distributions of P-IBI metrics between 1970 and 1996 in June (a-c), July (d), and August (e-f) across all Lake Erie basins. The boxes indicate the 25th percentile, median, and 75th percentile, whiskers indicate the 10th and 90th percentiles, while dots indicate the 5th and 95th percentiles. W and p-values are from the Mann-Whitney tests that were used to test the equality of medians. Significance is judged at a 0.05 α level and is indicated in bold. Due to zero values in 1996, Mann- Whitney tests could not be performed for Limnocalanus macrurus abundance (d).

228

Figure 16

W = 108.0 (a) (d) p = 0.082

0.035 8000

0.030

6000 0.025

0.020 4000 0.015

2000 0.010

0.005 0 0.000 Limnocalanus macrurus density (#/L) density macrurus Limnocalanus Edible phytoplankton biomass(ug/L) 1970 1996 1970 1996

July (b) (e) W = 2313.0 W = 1644.5 p < 0.001 p = 0.537

12 2.5

Aphanizomenon 10

2.0 8 , and

6 1.5

4 Anabaena 1.0 ,

2 0.5

Microcystis 0.0 of totalphytoplankton biomass (ug/L) 1970 1996 Calanoida/ (CyclopoidaCladocera) + 1970 1996

Percentage

1.6 700

1.4 600

1.2 500

1.0 400 0.8 300 0.6 200 0.4

0.2 100

0.0 0

1970 1996 (ug/L) biomass zooplankton Crustacean 1970 1996

ofAbundance Cladocera) Calanoida/ (Cyclopoida + June August 229

5 (a) Excellent 4 Lake Erie Good 3 Fair P-IBI Score P-IBI 2 Poor 1 5 WB (b) CB Excellent EB 4 Good 3 Fair

P-IBI Score P-IBI 2 Poor 1 1970 1995 1996 1997 1998 1999 2000 2001 2002 2003

Year

Figure 17- Mean lake-wide P-IBI score versus year (1970, 1996-2002) (a) and mean basin P-IBI score versus year (b) (1970, 1996-2002).

230

3.65 1997 3.55

3.45 1996

3.35

3.25

3.15

3.05 P-IBI Score

2.95

2.85

2.75 1999

51525 Phosphorus Loading (ktonnes)

Figure 18- Relationship between external phosphorus loading (ktonnes) to Lake Erie and mean lakewide P-IBI score (1970, 1996-2001). For external phosphorus loading, each year represents a water year. A water year begins on October 1 and ends on September 30 of the following year and is referred to by the year in which it ends. For example, the phosphorus loading values for the year 1995 are calculated from October 1994- September 1995 (i.e., the 1995 water year). Water years are used to calculate loading because data from which loading is calculated (USGS stream flow data) are compiled by water year. Phosphorus loading data are from Dolan 1993 and D olan unpublished ([email protected], University of Wisconsin - Green Bay).

231

Figure 19- Mean chlorophyll a (µg/L) versus mean P-IBI score (a) at the same site (1970, 1996-2000). For 1970 uncorrected chlorophyll a data were used, while for 1996-2000 corrected chlorophyll a data were used. Mean total phosphorus (µg/L) versus mean P-IBI score (b) at the same site (1970, 1996-2000).

232

Figure 19 (a)

y = 3.179 + 2.133 x – 0.529 x 2 r2 = 0.127 p < 0.001 n = 119 15

5 Chlorophyll a (ug/L) a Chlorophyll

12345 P-IBI Score

(b) y = 23.556 + 5.431 x – 1.720 x 2 r2 = 0.062 p = 0.024

90 n = 119

Total Phosphorus (ug/L) Phosphorus Total 10

12345 P-IBI Score

233

Figure 20- Mean basin total phytoplankton biomass (mg/L) versus mean basin P-IBI score (a) (1970, and 1996-2002) and offshore fish community IBI score (all ages of fish, all sites in the western and central basins of Lake Erie, Karr IBI method) (Kershner and Hopkins 2003) versus mean lakewide P-IBI score (b) (1970, and 1996-1999). Offshore fish community IBI scores were estimated from Figure 16 (Kershner and Hopkins 2003).

234

Figure 20 (a) y = 5.837 – 1.322 x r2 = 0.253 p = 0.012 n = 24 5

0 Total Phytoplankton Biomass (mg/L) 2.2 3.2 4.2 P-IBI Score

(b) y = 29.723 + 1.292 x r2 = 0.321 p = 0.319 35 n = 5 1996

34 1997 Fish IBI Score IBI Fish

2.75 2.85 2.95 3.05 3.15 3.25 3.35 3.45 3.55 3.65 P-IBI Score

235

APPENDIX C

LEPAS METHODS

236

LEPAS methods for sampling and laboratory analysis of ODW-Sandusky, ODW- Fairport, and CCIW samples. After E. M. Haas (modified by D.D. Kane)

Phytoplankton

Phytoplankton water samples were obtained with an integrated water sampler (2.5 cm diameter tube) from surface to twice the Secchi depth. Collected water was poured into a clean plastic bucket from which a 250 ml sample was taken. Each sample was preserved in a Mason jar with Lugols solution, using approximately 1 ml of Lugols for every 100 ml of sample. All samples were then transported to The Ohio State University, transferred into a graduated cylinder, and allowed to settle for 3 days. Each sample was then concentrated down to 30 ml by siphoning off liquid from the top and transferring the remaining sample to a 10 dram vial. Subsamples of approximately 3-5 ml were obtained from the concentrated samples and placed into a Utermöhl counting chamber. The general procedure for phytoplankton enumeration followed the inverted microscope technique (Utermöhl 1958, Lund et al. 1958). All phytoplankton genera were identified and counted using a Wild inverted microscope at 400 X. Transects were counted until 100 algal units (cells, filaments, or colonies) of the most common taxa were recorded. However, at least two transects were counted for each sample. Measurements were recorded for the first 20 algal units for each taxon. For filamentous algal taxa, however, all filament lengths were measured and then summed and recorded as the total filament length for each taxon. All of the recorded data were entered into a spreadsheet program that calculated the density and biomass of each algal taxon, as well as the critical dimensions that were measured for each algal species or category (Table 1). Densities for solitary algal taxa were calculated using the following formula.

D = N * CA * CV T * TA * V * SV

Where, D = density (# cells/ml) N = the number of cells counted in all transects CA = chamber area in µm2 CV = concentrated sample volume in ml (usually 30 ml) T = number of transects counted TA = transect area in µm2 V = volume of the sample placed into the chamber in ml SV = original sample volume in ml (usually 250 ml)

237

For most colonial algal taxa, all cells in each colony were counted and measured so the above formula was used to calculate the density. However, for Microcystis the cells were too small to count, so the number of colonies were counted and measured. The density was then calculated using the above formula, except 'N' was the number of colonies and the units of 'D' were colonies/ml. Densities for the filamentous algal taxa were calculated using the above formula as well, except for the 1995 to 1997 samples. For these samples, the number of cells in each filament was not recorded, so the number of filaments was used for ‘N’ and the units of ‘D’ were # filaments/ml. In order to estimate the density of filamentous algal taxa in cells/ml, a conversion factor was determined. Representative samples that contained each filamentous algal taxon were examined and total filament lengths and number of cells in the filament were recorded. The average length (µm) per cell was used as the conversion factor. The following formula was then used to calculate the densities in # cells/ml.

Cell D = Tot Fil L Conv

Where, Cell D = phytoplankton density in # cells/ml Tot Fil L = total length of all filaments found in the transects Conv = average cell length for the algal taxon (µm/cell)

The conversion factors for each filamentous algal taxon, except Oscillatoria and Spirulina, were used to calculate cell densities (Table 2). For each sample that contained Oscillatoria or Spirulina, the average filament width was used as average cell length in the above formula. In order to calculate the biomass of each algal taxon, the average individual volume (for cells, filaments, or colonies) was first calculated. The measurements of each critical dimension for each taxon were averaged and then used to calculate the average individual volume. Further, volumetric equations were used for each taxon (Table 3). The total volume of each taxon was then calculated with the following formula.

TV = AIV * D

Where, TV = phytoplankton volume (µm3/ml) AIV = average individual volume ( µm3/cell, µm3/filament, or µm3/colony) D = phytoplankton density (# cells/ml, # filaments/ml, or # colonies/ml)

The phytoplankton volume was then converted to biomass using the following formula.

238

BM = TV * 103 ml * 1 cm3 * 103 mg 1 L 1012 µm3 1 cm3

Where, BM = phytoplankton biomass (µg/l) TV = phytoplankton volume (µm3/ml)

After the density and biomass of each algal taxon was calculated, they were summed together in the following categories: Chlorophyta, Chrysophyta, Cryptophyta, Cyanophyta, Pyrrophyta, Edible Phytoplankton, Inedible Phytoplankton, and Total Phytoplankton. Table 1 indicates which species are edible and which are inedible.

Beginning in 1998, phytoplankton biomass was calculated using a macro in Microsoft Excel (Table 4).

Zooplankton Zooplankton samples were collected by vertical tows, using a weighted zooplankton net (0.5 m diameter, 64 µm mesh) with a flow meter and pint jar for collection. The net was lowered to near the lake bottom (< 1m) with the open end pointing downward. This allowed the water column to be sampled as the net was lowered and as it was pulled up. After each sample was taken, the net was washed off from the outside so that all of the zooplankton was in the collection jar. The samples were then filtered, placed in plastic collection bottles, and preserved with a 4% sugar formalin solution. Flow meter numbers were recorded before and after the tow. All samples were transported to The Ohio State University for analysis. Each sample was diluted to a known volume, usually ranging from 500 ml to 3000 ml. The diluted volume may have been outside this range if the sample contained an extremely small or large amount of zooplankton. After dilution, all of the zooplankton in at least two subsamples of 5-10 ml were identified and counted. Cladocerans and copepods were identified to species while rotifers were identified to genus. Complete subsamples were analyzed until 100 individuals of the most common taxon were recorded. The most common taxon used for this rule could not be nauplii, rotifers, or veligers, because they commonly have a much higher abundance than the larger cladoceran and copepod taxa. Individuals in each taxon were categorized as either having eggs or not having eggs. Measurements were then recorded for the first 20 individuals in each these categories. If fewer than 20 individuals were found, then measurements were recorded only for the individuals found (e.g. if 10 individuals were found then 10 measurements were recorded). In addition, the number of eggs in each female and the number of loose eggs were also recorded. When analyzing the samples, some species of cladocerans and copepods were considered common and some were considered rare (Table 5). The only difference between them is that the eggs within the females for the common species were recorded

239 separately and grouped together by species. However, for the rare species any eggs found within a female were grouped together as either rare cladoceran eggs or rare copepod eggs. After each sample was analyzed, all recorded data was entered into a spreadsheet program that calculated the total volume sampled and the zooplankton densities, biomass, phosphorus and nitrogen excretion rates (Wen and Peters 1994), filtering rates, and secondary productivities. The total volume was calculated using the flow meter numbers and a net constant. The net constant is determined by using the following formulas.

Distance in meters = Difference in Counts * Rotor Constant 999999

Volume in cubic meters = π *(Net Diameter)2 * Distance 4

The rotor constant of the flow meters that were used is 26, 873, and the net diameters were 0.5 m. This gives us the following equation.

Volume (m3) = π * (0.5 m)2 * Difference in Counts * 26, 873 4 999999

Converting the volume to liters results in the following formula.

Volume (l) = Difference in Counts * 5.2765 l/revolution

Based on the above equation, the following formula was used in the spreadsheet to calculate the total volume of water sampled.

TV = (M2-M1) * NC Where, TV = Total Volume sampled in ml M1 = Flow meter number before vertical tow M2 = Flow meter number after vertical tow NC = Net Constant (5.2765 l/revolution)

The density of each zooplankton taxon was calculated according to the following formula.

D = N * DV SV * TV Where, D = zooplankton density in #/l N = number of individuals counted in the subsamples DV = diluted volume in ml

240

SV = total subsample volume in ml TV = total volume sampled in ml

The biomass of each zooplankton taxon was calculated using the following formula.

B = D * AIB

Where, B = zooplankton biomass in µg/l (dry weight) D = zooplankton density in #/l AIB = average individual biomass in µg/individual The average individual biomass for each taxon of cladocerans and copepods, except for Bythotrephes cederstroemi and Cercopagis pengoi, was determined by calculating the biomass of each individual found in the subsamples and then averaging these biomass values. The biomass of each individual was calculated by using length-weight regression equations from Culver et al. (1985). Three of the equations (Epischura lacustris, Eurtymora affinis, and Limnocalanus macrurus) were developed by Julie Richey for her Undergraduate Honors Thesis in Winter of 2003. Each equation is in the following format.

IB = a * Lb

Where, IB = dry weight of the individual in µg a = coefficient b = coefficient L = length of the individual in mm

Table 6 contains a list of the 'a' and 'b' coefficients used for each species of cladocerans and copepods. The same coefficients are used for individuals with and without eggs. Equations were not developed for all species that are found in Lake Erie, so for those species without established length-weight relationships, equations from similar species were used. The equation for Daphnia retrocurva was used for Daphnia ambigua, Daphnia longiremis, Daphnia parvula, Daphnia pulex, Daphnia schodleri, and chydorids. The equation for Diacyclops bicuspidatus thomasi was used for Cyclops scutifer, Ergasilus sp., Eucyclops agilis, Eucyclops speratus, Paracyclops fimbriatus poppei, Tropocyclops prasinus mexicanus, unidentified cyclopoids, and Canthocamptus robert cokeri. In addition, the equation for Skistodiaptomus oregonensis was used for Leptodiaptomus ashlandi, Leptodiaptomus sicilis, and Skistodiaptomus reighardi. The equation for cyclopoid nauplii was also used for harpacticoid nauplii. The average individual biomass for Bythotrephes cederstroemi was calculated using a regression equation from Yan and Pawson, 1998:

Log W = 0.939 + 3.066 log L

241

Where, L = length of the individual from the top of the head to the first tail spine (mm) W = dry weight of the individual in µg

For Cercopagis pengoi, the individual biomass was calculated using a regression equation from Grigorovich et al. 2000:

Log W = 0.375 + 2.442 log L

Where, L = length of the individual from the top of the head to the first tail spine (mm) W = dry weight of the individual in µg

For calculating the biomass of eggs, rotifers, and veligers, the average individual biomass has been predetermined (Table 7). These values were used as the average individual biomass in the calculations for all samples. After the biomass of each individual taxon was calculated, they were summed together in the following categories.

Rotifer Biomass = ∑ all rotifers Cladoceran Biomass = ∑ all cladoceran species Cyclopoid Copepod Biomass = ∑ all cyclopoid copepod species Calanoid Copepod Biomass = ∑ all calanoid copepod species Copepod Biomass = Cyclopoids + Calanoids Nauplii Biomass = Cyclopoid nauplii + calanoid nauplii + harpacticoid nauplii Cladoceran Egg Biomass = ∑ all cladoceran eggs Copepod Egg Biomass = ∑ all copepod eggs Crustacean Biomass = cladocerans + copepods + nauplii + cladoceran eggs + copepod eggs Crust-Naup Biomass = cladocerans + copepods Total Zooplankton Biomass = crustaceans + rotifers + veligers

242

Table 1: All algal genera and general categories, separated by Phylum, that were counted when analyzing phytoplankton samples. Also listed for each genus/category is the type of algae (solitary, filamentous, or colonial), whether it is edible or inedible, and what critical dimensions were used for calculating cell volumes.

Phylum Genus/Category Algal Type Edible/Inedible Critical Dimensions Chlorophyta Actinastrum Colonial Edible Length, Width of cell Ankistrodesmus Solitary Edible Length, Width Carteria Solitary Edible Diameter Characium Solitary Edible Length, Width Chlamydomonas Solitary Edible Diameter Chlorophyte filament Filament Inedible Length, Width Closteriopsis Solitary Edible Length, Width Closterium Solitary Inedible Length, Width Coelastrum Colonial Edible Diameter of colony Colonial chlorophyte with Colonial Edible Diameter of cell sheath Colonial chlorophyte Colonial Edible Diameter of cell without sheath Cosmarium Solitary Edible Diameter Crucigenia Colonial Edible Diameter of cell Dictyosphaerium Colonial Edible Diameter of cell Dimorphococcus Solitary Edible Diameter of cell Eutetramorus Colonial Edible Diameter of cell Franceia Colonial Edible Diameter of cell Golenkinia Solitary Edible Diameter Gonium Colonial Edible Diameter of cell Kirchneriella Colonial Edible Length, Width of cell Lagerheimia Solitary Edible Diameter Micractinium Solitary Inedible Diameter Oocystis Colonial Edible Diameter of cell Pandorina Colonial Edible Diameter of cell Pediastrum Colonial Edible Diameter of colony Scenedesmus Colonial Edible Length, Width of cell Schroederia Solitary Edible Length, Width Solitary Green Solitary Edible Diameter of cell Sphaerocystis Colonial Edible Diameter of cell Spiny Green Solitary Inedible Diameter of cell Spirogyra Filament Edible Length, Width Staurastrum Solitary Edible Diameter of cell Tetraedron Solitary Edible Length, Width Treubaria Solitary Edible Diameter Chrysophyta Dinobryan Colonial Edible Diameter of cell (Chrysophyceae) Mallomonas Solitary Edible Length, Width Synura Colonial Inedible Diameter of cell Chrysophyta Asterionella Colonial Edible Length, Width of cell (diatoms) Centric Diatom Solitary Edible Diameter of cell Cocconeis Solitary Edible Diameter of cell Coscinodiscus Solitary Edible Diameter of cell Cyclotella Solitary Edible Diameter of cell Cymbella Solitary Edible Length, Width Fragilaria Colonial Edible Length, Width of cell Gomphonema Solitary Edible Length, Width 1, Width 2 243

Gyrosigma Solitary Edible Length, Width Melosira Filament Inedible Length, Width Navicula Solitary Edible Length, Width Nitzschia Solitary Edible Length, Width Opephora Solitary Edible Length, Width Pennate Diatom Solitary Edible Length, Width Rhizosolenia Solitary Edible Length, Width Rhoicosphenia Solitary Edible Length, Width 1, Width 2 Stephanodiscus Solitary Edible Diameter of cell Surirella Solitary Edible Length, Width Synedra Solitary Edible Length, Width Tabellaria Colonial Edible Length, Width of cell Cryptophyta Chroomonas Solitary Edible Diameter Cryptomonas Solitary Edible Length, Width Rhodomonas Solitary Edible Length, Width Cyanophyta Anabaena Filament Inedible Length, Width Aphanizomenon Filament Inedible Length, Width Aphanocapsa Colonial Inedible Diameter of cell Aphanothece Colonial Inedible Diameter of cell Chroococcus Colonial Inedible Diameter of cell Merismopedia Colonial Inedible Diameter of cell Microcystis Colonial Inedible Diameter of colony Lyngbya Filament Inedible Length, Width Narrow Filament Filament Inedible Length, Width Oscillatoria Filament Inedible Length, Width Spirulina Filament Inedible Length, Width Pyrrophyta Ceratium Solitary Inedible Length, Spine height, Spine Width Gymnodinium Solitary Edible Length, Width Peridinium Solitary Edible Length, Width

Table 2: Conversion factors (cell lengths) that were used to convert densities of filamentous algae from filaments length/ml to # cells/ml.

Genus/Category Conversion Factor (µm/cell) Anabaena 5.61437 Aphanizomenon 2.50452 Chlorophyte Filament 20.93179 Melosira 10.78529 Narrow Filament 2.450452

244

Table 3: Volumetric equations that were used to calculate the cell volume of all phytoplankton genera or categories.

Genus/Category Volumetric Equation Actinastrum, Closterium, Cryptomonas, Gymnodinium, Mallomonas, Peridinium, (π * L * W2)/6 Rhodomonas, Scenedesmus Anabaena, Aphanizomenon, Chlorophyte Filament, Melosira, Lyngbya, Narrow (π * W2 * (TotL/# fil))/4 Filament, Oscillatoria, Spirogyra Ankistrodesmus, Characium, Closteriopsis, Kirchneriella, Schroederia, Tetraedron (π * L * W2)/12 Asterionella π * L * W2/4 Aphanocapsa, Aphanothece, Carteria, Chlamydomonas, Chroococcus, (π * D3)/6 Chroomonas, Coelastrum, Colonial Chlorophyte with sheath, Colonial Chlorophyte without sheath, Crucigenia, Dimorphococcus, Dinobryan, Dictyosphaerium, Eutetramorus, Franceia, Golenkinia, Gonium, Lagerheimia, Merismopedia, Microcystis, Micractinium, Oocystis, Pandorina, Solitary Green, Sphaerocystis, Spiny Green, Staurastrum, Synura, Treubaria Centric Diatom, Cocconeis, Coscinodiscus, Cyclotella, Stephanodiscus (D2 * π * D/2)/4 Ceratium (π/12) * (L3 + (4 * SH * SW 2)) Cosmarium (π * D2 * D/6)/6 Cymbella W * W/2 * (L – W + (π/4) * W) Fragilaria, Gyrosigma, Navicula, Nitzschia, Pennate Diatom, Opephora, L * W * W/2 Rhizosolenia, Synedra, Tabellaria Gomphonema, Rhoicosphenia W2/2 * L * ((W1 + W2)/2) Pediastrum (D2 * π * 0.15)/4 Spirulina (π * L 2 * W)/4 Surirella L * L/2 * (W – L + π/4 * L)

Table 4: Calculation of cell biovolumes. All biovolume equations represent the average biovolume and are calculated using the average measurements of all individuals that were measured in the transects (Bcell = average cell biovolume, Bcol = average colony biovolume, Bfil = average filament biovolume). Total densities (d cell) were calculated using the number of cells counted in the transects (# cell), except for Microcystis for which #col was used. For species that were measured according to colony size or filament size, the density was also calculated using the number of colonies (#col) and filaments (#fil) that were found, giving the density of colonies (dcol) or the density of filaments (dfil) in the sample. Total sample biovolumes were calculated by multiplying the average biovolume by the appropriate density (Bcell * dcell, Bcol * dcol, or Bfil * dfil).

Genus/Category Items needed for Counting Sheets Biovolume Equations Chlorophyta 2 Actinastrum # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/6 (Colonial) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 2 Ankistrodesmus # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/12 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 3 Carteria # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) 2 Characium # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/12 245

(solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 3 Chlamydomonas # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) 2 Chlorophyte # of filaments in all transects (#fill) Bfil = (π * Wfil * (LFilTot/#fil))/4 filament Total length of all filaments (LFilTot) # of cells counted in all transects (#cell) 20 individual filament widths (Wfil) 2 Closteriopsis # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/12 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 2 Closterium # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/6 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 3 Coelastrum # of cells counted in all transects (#cell) Bcol = (π * D col )/6 (colonial) # of colonies in all transects (#col) 20 diameters of colonies (D col) 3 Colonial # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 chlorophyte with 20 individual cell diameters (D cell) sheath 3 Colonial # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 chlorophyte without 20 individual cell diameters (D cell) sheath 2 Cosmarium # of cells counted in all transects (#cell) Bcell = (π * Dcell * Dcell/6)/6 (solitary) 20 individual cell diameters (D cell) 3 Crucigenia # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 3 Dictyosphaerium # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 3 Dimorphococcus # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (Dcell) 3 Eutetramorus # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 3 Franceia # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 3 Golenkinia # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) 3 Gonium # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 2 Kirchneriella # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/12 (colonial) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 3 Lagerheimia # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) 3 Micractinium # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 3 Oocystis # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 3 Pandorina # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 2 Pediastrum # of cells counted in all transects (#cell) Bcol = (Dcol * π * 0.15)/4 (colonial) # of colonies in all transects (#col) 20 diameters of colonies (D col) 2 Scenedesmus # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/6 (colonial) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 2 Schroederia # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/12 (solitary) 20 individual cell lengths (Lcell) 246

20 individual cell widths (Wcell) 3 Solitary Green # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) 3 Sphaerocystis # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 3 Spiny Green # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) 2 Spirogyra # of filaments in all transects (#fill) Bfil = (π * Wfil * (LFilTot/#fil))/4 (filamentous) Total length of all filaments (LFilTot) # of cells counted in all transects (#cell) 20 individual filament widths (Wfil) 3 Staurastrum # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) 2 Tetraedron # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/12 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 3 Treubaria # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (solitary) 20 individual cell diameters (D cell) Chrysophyta (Chrysophyceae) 3 Dinobryon # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) 2 Mallomonas # of cells counted in all transects (#cell) Bcell = (π * L cell * W cell )/6 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 3 Synura # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 individual cell diameters (D cell) Chrysophyta (diatoms) 2 Asterionella # of cells counted in all transects (#cell) Bcell = π * L cell *Wcell /4 (colonial) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) 2 Centric Diatom # of cells counted in all transects (#cell) Bcell = (Dcell * π * Dcell/2)/4 (solitary) 20 individual cell diameters (D cell) 2 Cocconeis # of cells counted in all transects (#cell) Bcell = (Dcell * π * Dcell/2)/4 (solitary) 20 individual cell diameters (D cell) 2 Coscinodiscus # of cells counted in all transects (#cell) Bcell = (Dcell * π * Dcell/2)/4 (solitary) 20 individual cell diameters (D cell) 2 Cyclotella # of cells counted in all transects (#cell) Bcell = (Dcell * π * Dcell/2)/4 (solitary) 20 individual cell diameters (D cell) Cymbella # of cells counted in all transects (#cell) Bcell = Wcell * Wcell/2 * (Lcell – (solitary) 20 individual cell lengths (Lcell) W cell + (π/4) * Wcell) 20 individual cell widths (Wcell) Fragilaria # of cells counted in all transects (#cell) Bcell = Lcell * Wcell * Wcell/2 (colonial) 20 Length of the cell in each colony (L cell

(solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) Nitzschia # of cells counted in all transects (#cell) Bcell = Lcell * Wcell * Wcell/2 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) Opephora # of cells counted in all transects (#cell) Bcell = Lcell * Wcell * Wcell/2 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell)

Pennate Diatom # of cells counted in all transects (#cell) Bcell = Lcell * Wcell * Wcell/2 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) Rhizosolenia # of cells counted in all transects (#cell) Bcell = Lcell * Wcell * Wcell/2 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) Rhoicosphenia # of cells counted in all transects (#cell) Bcell = W2cell/2 * Lcell * ((W1cell (solitary) 20 individual cell lengths (Lcell) + W2 cell)/2) 20 individual cell width1’s (W1cell) 20 individual cells width2’s (W2cell) 2 Stephanodiscus # of cells counted in all transects (#cell) Bcell = (Dcell * π * Dcell/2)/4 (solitary) 20 individual cell diameters (D cell) Surirella # of cells counted in all transects (#cell) Bcell = Lcell * L cell/2 * (Wcell – (solitary) 20 individual cell lengths (Lcell) L cell + π/4 * Lcell) 20 individual cell widths (Wcell) Synedra # of cells counted in all transects (#cell) Bcell = Lcell * Wcell * Wcell/2 (solitary) 20 individual cell lengths (Lcell) 20 individual cell widths (Wcell) Tabellaria # of cells counted in all transects (#cell) Bcell = Lcell * Wcell * Wcell/2 (colonial) 20 Length of the cell in each colony (L cell

248

2 Lyngbya # of filaments in all transects (#fill) Bfil = (π * Wfil * (LFilTot/#fil))/4 (filamentous) Total length of all filaments (LFilTot) # of cells counted in all transects (#cell) 20 individual filament widths (Wfil) 3 Merismopedia # of cells counted in all transects (#cell) Bcell = (π * D cell )/6 (colonial) 20 Diameter of a cell in each colony (D cell

Table 5: All zooplankton species or general categories that were counted during the analysis of zooplankton samples. Also listed is whether the species or category was considered common or rare. Species Commonality Cladocerans Bosmina longirostris Common Bythotrephes cederstroemi Rare Cercopagis pengoi Rare Ceriodaphnia sp. Common Chydorids Rare Daphnia ambigua Rare Daphnia galeata mendotae Common Daphnia longiremis Rare Daphnia parvula Rare Daphnia pulex Rare Daphnia retrocurva Common Daphnia schodleri Rare Diaphanosoma sp. Common Eubosmina coregoni Common Leptodora kindti Common

249

Cyclopoid Copepods Acanthocyclops vernalis Common Cyclops scutifer Rare Diacyclops bicuspidatus thomasi Common Ergasilus sp. Rare Eucyclops agilis Rare Eucyclops speratus Rare Mesocyclops edax Common Paracyclops fimbriatus poppei Rare Tropocyclops prasinus mexicanus Rare Unidentified Cyclopoids Cyclopoid Nauplii Calanoid Copepods Epischura lacustris Common Eurytemora affinis Common Leptodiaptomus ashlandi Common Leptodiaptomus minutus Common Leptodiaptomus sicilis Common Leptodiaptomus siciloides Common Limnocalanus macrurus Rare Skistodiaptomus oregonensis Common Skistodiaptomus reighardi Rare Unidentified Calanoids Calanoid Nauplii Rotifers/Veligers Asplanchna sp. Common Brachionus sp. Common Dreissena veligers Common Kellicottia sp. Common Keratella sp. Common Pleosoma sp. Common Polyarthra sp. Common Unidentified Rotifers

Table 6: 'a' and 'b' coefficients that were used in the generic length -weight regression equation, W = a * Lb, for the calculation of zooplankton biomass.

Species 'a' coefficient 'b' coefficient Cladocerans Bosmina longirostris 17.737 2.2291 Ceriodaphnia sp. 4.0216 1.9763 Daphnia ambigua, D. longiremis, D. parvula, D. 3.7847 2.6807 pulex, D. retrocurva, D. schodleri, chydorids Daphnia galeata mendota 7.4997 1.5644 Diaphanosoma sp. 5.0713 1.0456 Eubosmina coregoni 21.913 2.3371 Leptodora kindti 1.5605 1.873 Cyclopoid Copepods Acanthocyclops vernalis 7.0729 2.5563 Cyclops scutifer, Diacyclops bicuspidatus 5.6713 1.9347 thomasi, Ergasilus sp., Eucyclops agilis, E. speratus, Paracyclops fimbriatus poppei, Tropocyclops prasinus mexicanus, Unidentified cyclopoids Mesocyclops edax 6.6586 2.8945 Cyclopoid Nauplii 2.5968 1.6349 Calanoid Copepods Epischura lacustris 6.4115 1.7713 Eurytemora affinis 5.8538 1.8165 Leptodiaptomus ashlandi, Leptodiaptomus sicilis, 6.1927 1.9604 250

Skistodiaptomus oregonensis, Skistodiaptomus reighardi Leptodiaptomus minutus 7.3614 3.8564 Leptodiaptomus siciloides 5.8853 3.8498 Limnocalanus macrurus 7.9677 1.3691 Unidentified Calanoids 4.5921 1.7034 Calanoid Nauplii 3.0093 1.7064 Harpacticoid Copepods Canthocamptus robert cokeri 5.6713 1.9347 Harpacticoid Nauplii 2.5968 1.6349

Table 7: Average individual biomasses used in the calculation of egg, rotifer, and veligers biomass values.

Type of Egg/Rotifer/Veliger Average Individual Biomass (µg) Cladoceran eggs Bosmina longirostris eggs 0.51 Ceriodaphnia sp., Daphnia retrocurva, D. galeata 0.5887 mendotae, Loose Daphnia eggs, Rare cladoceran eggs Diaphanosoma sp. eggs 1.6996 Eubosmina coregoni eggs 1.0953 Leptodora kindti eggs 1.829 Loose cladoceran eggs 1.0153 Loose Bosmina eggs 0.8027 Copepod eggs Acanthocyclops vernalis eggs 0.1143 Diacyclops bicuspidatus thomasi eggs 0.168 Eurytemora affinis, Leptodiaptomus ashlandi, L. minutus, 0.1155 L. sicilis, L. siciloides, Skistodiaptomus oregonensis eggs Mesocyclops edax eggs 0.175 Loose copepod eggs, Rare copepod eggs 0.1149 Rotifers Asplanchna sp. 0.16 Brachionus sp. 0.07 Kellicottia sp. 0.062 Keratella sp. 0.5 Pleosoma sp. 0.612 Polyarthra sp. 0.053 Unidentified rotifers 0.439 Veligers Veligers 1

251

LIST OF REFERENCES

Antsulevich, A., and P. Valipakka. 2000. New important food object of the Baltic herring in the Gulf of Finland. International Review of Hydrobiology 85: 609-619.

Arnott, D. L., and M. J. Vanni. 1996. Nitrogen and phosphorus recycling by the zebra mussel (Dreissena polymorpha) in the western basin of Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 53: 646-659.

Attayde, J. L., and R. L. Bozelli. 1998. Assessing the indicator properties of zooplankton assemblages to disturbance gradients by canonical correspondence analysis. Canadian Journal of Fisheries and Aquatic Sciences 55: 1789-1797.

Babcock-Jackson, L., W. W. Carmichael, and D. A. Culver. 2002. Biodepostion by dreissenid mussels increases exposure of benthic and pelagic organisms to toxic microcystins. Verhandlungen Internationale Vereinigung für Theoretische und Angewandte Limnologie 28: 1082-1085.

Balcer, M. D., N. L. Korda, and S. I. Dodson. 1984. Zooplankton of the Great Lakes. The University of Wisconsin Press, Madison, WI.

Banicki, J. J. 2003. Dead zone update: Lake Erie again battling areas of hypoxia. Twineline 25: 5.

Barbiero, R. P., and M. L. Tuchman. 2001. Results of the U. S. EPA’s biological open water surveillance program of the Laurentian Great Lakes: I. Introduction and Phytoplankton Results. Journal of Great Lakes Research 27: 134-154.

Barbiero, R. P., R. E. Little, and M. L. Tuchman. 2001. Results from the U. S. E.P. A.’s biological open water surveillance program of the Laurentian Great Lakes: III. crustacean zooplankton. Journal of Great Lakes Research 27: 167- 184.

Barbour, M. T., J. B. Stribling, and J. R. Karr. 1995. Multimetric approach for establishing biocriteria and measuring biological condition. pp. 63-80 in Biological assessment and criteria: tools for water resource planning and decision making. W. S. Davis and T. P. Simon (eds.). CRC Press. United States of America.

252

Barry, M. J. 1999. The effects of a pesticide on inducible phenotypic plasticity in Daphnia. Environmental Pollution 104: 217-224.

Barry, M. J., and D. C. Logan. 1998. The use of temporary pond microcosms for aquatic toxicity testing: direct and indirect effects of endosulfan on community structure. Aquatic Toxicology 41: 101-124.

Bascompte, J., and M. A. Rodriguez. 2001. Habitat richness and plant species richness. Ecology Letters 4: 417-420.

Beale, E. M. L. 1962. Some uses of computers in operational research. Industrielle Organisation 31: 51-52.

Bean, D. J. 1980. Crustacean zooplankton production in Lake Erie, 1970, M.S. Thesis, The Ohio State University, Columbus, Ohio.

Beeton, A. M. 1961. Environmental changes in Lake Erie. Transactions of the American Fisheries Society 90: 153-159.

Beeton, A. M. 1965. Eutrophication of the St. Lawrence Great Lakes. Limnology and Oceanography 10: 240-254.

Beeton, A. M. 2002. Large freshwater lakes: present state, trends, and future. Environmental Conservation 29: 21-38.

Berg, D. J., and D.W. Garton. 1988. Seasonal abundance of the exotic predatory cladoceran, Bythotrephes cederstroemi, in western Lake Erie. Journal of Great Lakes Research 14: 479-488.

Berglund, O., P. Larsson, G. Ewald, and L. Okla. 2000. Bioaccumulation and differential partitioning of polychlorinated biphenyls in freshwater planktonic food webs. Canadian Journal of Fisheries and Aquatic Sciences 57: 1160-1168.

Berglund, O., P. Larsson, G. Ewald, and L. Okla. 2001. Influence of trophic status on PCB distribution in lake sediments and biota. Environmental Pollution 113: 199- 210.

Bertram, P. E. 1993. Total phophorus and dissolved oxygen trends in the central basin of Lake Erie, 1970-1991. Journal of Great Lakes Research 19: 224-236.

Berzin, B., and B. Pejler. 1989. Rotifer occurrence and trophic degree. Hydrobiologia 182: 171-180.

253

Bolsenga, S. J., and C. E. Herdendorf, (eds.). 1993. Lake Erie and Lake St. Clair Handbook. Wayne State University Press. Detroit, Michigan.

Branstrator, D. K. 1995. Ecological interactions between Bythotrephes cederstroemi and Leptodora kindtii and the implications for species replacement in Lake Michigan. Journal of Great Lakes Research 21: 670-679.

Branstrator, D. K., and J. T. Lehman. 1996. Evidence for predation by young-of-year alewife and bloater chub on Bythotrephes cederstroemi in Lake Michigan. Journal of Great Lakes Research 22: 917-924.

Bremigan, M. T., and R. A. Stein. 1999. Larval gizzard shad success, juvenile effects, and reservoir productivity: toward a framework for multi-system management. Transactions of the American Fisheries Society 128: 1106-1124.

Britt, N. W. 1955. Stratification in western Lake Erie in summer of 1953: effects on the Hexagenia (Ephemeroptera) population. Ecology 36: 239-244.

Brittain, S. M., J. Wang, L. Babcock-Jackson, W. W. Carmichael, K. L. Rinehart, and D. A. Culver. 2000. Isolation and characterization of microcystins, cyclic heptapeptide hepatotoxins from a Lake Erie strain of Microcystis aeruginosa. Journal of Great Lakes Research 26: 241-249.

Brooks, J.L., and S. I. Dodson. 1965. Predation, body size, and composition of plankton. Science 150: 28-35.

Bruner, K. A., S. W. Fisher, and P. F. Landrum. 1994. The role of the zebra mussel, Dreissena polymorpha, in contaminant cycling. II. Zebra mussel contaminant accumulation from algae and suspended particles, and transfer to the benthic invertebrate, Gammarus fasciatus. Journal of Great Lakes Research 20: 735-750.

Budd, J.W., A. M. Beeton, R. P. Stumpf, D. A. Culver, and W. C. Kerfoot. 2002. Satellite observations of Microcystis blooms in western Lake Erie. Verhandlungen Internationale Vereinigung für Theoretische und Angewandte Limnologie 27: 3787-3793.

Bur, M.T. 1982. Food of freshwater drum in western Lake Erie. Journal of Great Lakes Research 8: 672-675.

Bur, M. T., D. M. Klarer, and K. A. Krieger. 1986. 1st records of a European cladoceran, Bythotrephes cederstroemi, in lakes Erie and Huron. Journal of Great Lakes Research 12: 144-146.

254

Burns, N. M. 1976a. Temperature, oxygen, and nutrient distribution patterns in Lake Erie, 1970. Journal of the Fisheries Research Board of Canada 33: 485-511.

Burns, N. M. 1976b. Nutrient budgets for Lake Erie, 1970. Journal of the Fisheries Research Board of Canada 33: 520-536.

Cabin, R. J., and R. J. Mitchell. 2000. To Bonferroni or not to Bonferroni: when and how are the questions. ESA Bulletin 81: 246-248.

Callicott, J. B. 1995. The value of ecosystem health. Environmental Values 4: 345-361.

Canfield, T.J., and J. R. Jones. 1996. Zooplankton abundance, biomass, size-distribution in selected midwestern waterbodies and relation with trophic state. Journal of Freshwater Ecology 11: 171-181.

Carlson, R. E. 1977. A trophic state index for lakes. Limnology and Oceanography 22:361-369.

Carmichael, W.W. 1986. Algal toxins. Advances in Botanical Research 12: 47-101.

Carmichael, W. W. 1997. The cyanotoxins. pp. 211-256 in Advances in Botanical Research 27. J. A. Callow (ed.).

Carr, J. F. 1962. Dissolved oxygen in Lake Erie, past and present. Proceedings of the 5th Conference on Great Lakes Research. Great Lakes Res. Div., Univ. Michigan, Pub. No. 9: 1-14.

Carr, J. F., V. C. Applegate, and M. Keller. 1965. A recent occurrence of thermal stratification and low dissolved oxygen in western Lake Erie. Ohio Journal of Science 65: 319-327.

Carrick, H. J. 2004. Algal distribution patterns in Lake Erie: implications for oxygen balances in the eastern basin. Journal of Great Lakes Research 30: 133-147.

Chagnon, S. A. 2004. Temporal behavior of levels of the Great Lakes and climate variability. Journal of Great Lakes Research 30: 184-200.

Chandler, D. C. 1940. Limnological studies of western Lake Erie. I. Plankton and certain physical-chemical data of the Bass Islands region, from September, 1938, to November, 1939. Ohio Journal of Science 40: 291-336.

Chapra, S. C., and H. F. H. Dobson. 1981. Quantification of the lake trophic typologies of Naumann (surface quality) and Thienemann (oxygen) with special reference to the Great Lakes. Journal of Great Lakes Research 7: 182-193.

255

Chapra, S. C., and A. Robertson. 1977. Great Lakes eutrophication: the effects of point source control of total phosphorus. Science 196: 1448-1450.

Charlebois, P. M., M. J. Raffenberg, and J.M. Dettmers. 2001. First occurrence of Cercopagis pengoi in Lake Michigan. Journal of Great Lakes Research 27: 258- 261.

Charlton, M. N. 1980. Oxygen depletion in Lake Erie: Has there been any change? Canadian Journal of Fisheries and Aquatic Sciences 37: 72-81.

Charlton, M. N., R. LeSage, and J. E. Milne. 1999. Lake Erie in transition: the 1990’s. pp. 97-123. In: State of Lake Erie (SOLE)- Past, Present, and Future. M. Munawar, T. Edsall, and I. F. Munawar (eds.). Backhuys Publishers. Leiden, The Netherlands.

Chen, C. Y., and C. L. Folt. 1996. Consequences of fall warming on zooplankton overwintering success. Limnology and Oceanography 41: 1077-1086.

Chen, C. Y., and C. L. Folt. 2000. Bioaccumulation and diminution of arsenic and lead in a freshwater food web. Enviornmental Science and Technology 34: 3878- 3884.

Chen, C. Y., K. B. Sillett, C. L. Folt, S. L. Whittemore, and A. Barchowsky. 1999. Molecular and demographic measures of arsenic stress in Daphnia pulex. Hydrobiologia 401: 229-238.

Chen, C.Y., R.S. Stemberger, B. Klaue, J. D. Blum, P. C. Pickhardt, and C. L. Folt. 2000. Accumulation of heavy metals in food web components across a gradient of lakes. Limnology and Oceanography 45: 1525-1536.

Chessman, B. C. 1995. Rapid assessment of rivers using macroinvertebrates: a procedure based on habitat-specific sampling, family level identification, and a biotic index. Australian Journal of Ecology 20: 122-129.

Chessman, B., I., Growns, J. Currey, and N. Plunkett-Cole. 1999. Predicting diatom communities at the genus level for the rapid biological assessment of rivers. Freshwater Biology 41: 317-331.

Christie, W. J. 1974. Changes in the fish species composition of the Great Lakes. Journal of the Fisheries Research Board of Canada 31: 827-854.

256

Christman, C. S., and D. M. Dauer. 2003. An approach for identifying the causes of benthic degradation in Chesapeake Bay. Environmental Monitoring and Assessment 81: 187-197.

Conroy, J.D., D. D. Kane, D. M. Dolan, W. J. Edwards, M. N. Charlton, and D. A. Culver. 2004a. Recent increases in Lake Erie plankton biomass: roles of external phosphorus loading and dreissenid mussels. 57 manuscript pp. Journal of Great Lakes Research, in review.

Conroy, J. D., W. J. Edwards, R. A. Pontius, D. D. Kane, H. Zhang, J. F. Shea, J. N. Richey, and D. A. Culver. 2004b. Differential excretion by dreissenid taxa: implications for western Lake Erie. 52 manuscript pp. Freshwater Biology, in review.

Cornelius, F. C., K. M. Muth, and R. Kenyon. 1995. Lake Trout rehabilitation in Lake Erie: a case history. Journal of Great Lakes Research 21 (Supplement 1): 65-82.

Coulas, R. A., H. J. MacIsaac, and W. Dunlop. 1998. Selective predation on an introduced zooplankter (Bythotrephes cederstroemi) by lake herring (Coregonus artedii) in Harp Lake, Ontario. Freshwater Biology 40: 343-355.

Cristescu, M. E. A., P. D. N. Hebert, J. D. S. Witt, H. J. MacIsaac, and I. A.Grigorovich. 2001. An invasion history for Cercopagis pengoi based on mitochondrial gene sequences. Limnology and Oceanography 46: 234-239.

Culver, D. A., M. M. Boucherle, D. J. Bean, and J. W. Fletcher. 1985. Biomass of Great Lakes zooplankton from length-weight regressions. Canadian Journal of Fisheries and Aquatic Sciences 42: 1380-1390.

Culver, D. A., J. D. Conroy, W. J. Edwards, D. D. Kane, R. A. Pontius, J. N. Richey, J. F. Shea, and H. Zhang. 2003. Do dreissenid mussels cause the Lake Erie “Dead Zone?” Paper presented at the 2003 Aquatic Sciences meeting of the American Society of Limnology and Oceanography, Salt Lake City, Utah.

Culver, D. A., and L. Wu. 1997. Relative importance of predation and competition in the seasonal dynamics of zooplankton: results from larval fish rearing ponds. Archiv für Hydrobiologie Special Issues and Advances in Limnology 49: 27-35.

Davis, C.C. 1954. A preliminary study of the plankton of the Cleveland Harbor area, Ohio. III. The zooplankton, and general ecological considerations of phytoplankton and zooplankton production. Ohio Journal of Science 54: 388-408.

Davis, C. C. 1964. Evidence for the eutrophication of Lake Erie from phytoplankton records. Limnology and Oceanography 9: 275-283.

257

Delorme, L. D. 1982. Lake Erie oxygen: the prehistoric record. Canadian Journal of Fisheries and Aquatic Sciences 39: 1021-1029.

DeMott, W.R. 1989. The role of competition in zooplankton succession. pp. 195-252. in U. Sommer (ed.). Plankton ecology: succession in plankton communities. Springer-Verlag. U.S.A.

DeMott, W. R., and F. Moxter. 1991. Foraging on cyanobacteria by copepods: responses to chemical defenses and resource abundance. Ecology 72: 1820-1834.

DeMott, W. R., Q.-X. Zhang, and W. W. Carmichael. 1991. Effects of toxic cyanobacteria and purified toxins on the survival and feeding of a copepod and three species of Daphnia. Limnology and Oceanography 36: 1346-1357.

DePinto, J. V., T.C. Young, and L.M. McIlroy. 1986. Great Lakes water quality improvement. Environmental Science and Technology 20: 754-759.

Devault, D. S., and D. C. Rockwell. 1986. Preliminary results of the 1978-79 Lake Erie Intensive Study – phytoplankton. Unpublished Report. Great Lakes National Program Office, EPA, Chicago, Illinois.

Dillon, P. J., and F. H. Rigler. 1974. The phosphorus-chlorophyll relationship in lakes. Limnology and Oceanography 19: 767-773.

Dixit, S. S., J. P. Smol, J. C. Kingston, and D. F. Charles. 1992. Diatoms-powerful indicators of environmental change. Environmental Science and Technology 26: 22-33.

Dobson, H. F., M. Gilbertson, and P. G. Sly. 1974. A summary and comparison of nutrients and related water quality in Lake Erie, Ontario, Huron, and Superior. Journal of the Fisheries Research Board of Canada 31: 731-738.

Dodson, S. I., T. Hanazato, and P. R. Gorski. 1995. Behavioral responses of Daphnia pulex exposed to carbaryl and Chaoborus kairomone. Environmental Toxicology and Chemistry 14: 43-50.

Dodds, W. K., V. H. Smith, and K. Lohman. 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 865-874.

Dolan, D. M., A. K. Yui, and R. D. Geist. 1981. Evaluation of river load estimation methods for total phosphorus. Journal of Great Lakes Research 7: 207-214.

258

Dolan, D. M. 1993. Point source loadings of phosphorus to Lake Erie: 1986-1990. Journal of Great Lakes Research 19: 212-223.

Downing, J. A., and F. H. Rigler. 1984. A Manual of the Methods for the Assessment of Secondary Production in Fresh Waters. IBP Handbook #17. Blackwell Scientific Publishers, Oxford.

Dumitru, C., W. G. Sprules, and N. D. Yan. 2001. Impact of Bythotrephes longimanus on zooplankton assemblages of Harp Lake, Canada: an assessment based on predator consumption and prey production. Freshwater Biology 46: 241-251.

Edmondson, W. T., and J. T. Lehman. 1981. The effect of changes in the nutrient income on the condition of Lake Washington. Limnology and Oceanography 26: 1-29.

Edmondson, W.T., and A. H. Litt. 1982. Daphnia in Lake Washington. Limnology and Oceanography 27: 272-293.

Epplett, T. D., S. Gewurtz, R. Lazar, and G. D. Haffner. 2000. Seasonal dynamics of PCBs in the plankton of Lake Erie. Journal of Great Lakes Research 26: 65-73.

Elster, H. J. 1974. History of limnology. Mitteilungen Vereinigung für Theoretische und Angewandte Limnologie 13: 101-120.

Engle, V. D., and J. K. Summers. 1999. Refinement, validation, and application of a benthic condition index for Northern Gulf of Mexico estimates. Estuaries 22: 624-635.

Evans, M. S., and D. W. Sell. 1983. Zooplankton sampling strategies for environmental studies. Hydrobiologia 99: 215-223.

Faber, D. J., and E. G. Jermolajev. 1966. A new copepod genus in the plankton of the Great Lakes. Limnology and Oceanography 11: 301-303.

Fausch, K. D., J. R. Karr, and P. R. Yant. 1984. Regional application of an Index of Biotic Integrity based on stream fish communities. Transactions of the American Fisheries Society 113: 39-55.

Fernandez-Casalderrey, A., M.D. Ferrando, and E. Andreu-Moliner. 1994. Effect of sublethal concentrations of pesticides on the feeding behavior of Daphnia magna. Ecotoxicology and Environmental Safety 27: 82-89.

Fish, C. J. 1960. Limnological survey of eastern and central Lake Erie, 1928-1929. U.S. Fish. Wildl. Serv., Spec. Sci. Rpt. Fish. No. 334. 198 p.

259

Fisher, S. W., D. C. Gossiaux, K. A. Bruner, and P. F. Landrum. 1993. Investigations of the toxicokinetics of hydrophobic contaminants in the zebra mussel ( Dreissena polymorpha). pp. 465-490 in Zebra mussels: biology, impacts and control. T. F. Nalepa, and D. W. Schloesser (eds.). Lewis Publishers. Boca Raton, Florida.

Fogg, G.E. 1991. Tansley Review No. 30. The phytoplanktonic ways of life. New Phytologist 118: 191-232.

Fore, L.S., J.R. Karr, and R.W. Wisseman. 1996. Assessing invertebrate responses to human activities: evaluating alternative approaches. Journal of the North American Benthological Society 15: 212-231.

Fraser, A.S. 1987. Tributary and point source total phosphorus loading to Lake Erie. Journal of Great Lakes Research 13: 659-666.

Frost, P. C., and D. A. Culver. 2001. Spatial and temporal variability of phytoplankton and zooplankton in western Lake Erie. Journal of Freshwater Ecology 16: 435- 443.

Gannon, J. E. 1971. Two counting cells for the enumeration of zooplankton micro- Crustacea. Transactions of the American Microscopical Society 90: 486-490.

Gannon, J. E. 1972. A contribution to the ecology of zooplankton Crustacea of Lake Michigan and Green Bay. Ph.D. thesis, University of Wisconsin.

Gannon, J. E. 1975. Horizontal distribution of crustacean zooplankton along a cross- lake transect in Lake Michigan. Journal of Great Lakes Research 1: 79-91.

Gannon, J. E., and A. M. Beeton. 1971. The decline of the large zooplankter, Limnocalanus macrurus Sars (Copepoda: Calanoida), in Lake Erie. Proceedings of the 14th Conference on Great Lakes Research 27-38.

Gannon, J. E., and R. S. Stemberger 1978. Zooplankton (especially crustaceans and rotifers) as indicators of water quality. Transactions of the American Microscopical Society 97: 16-35.

Garton, D. W., D.J. Berg, and R.J. Fletcher. 1990. Thermal tolerances of the predatory cladocerans Bythotrephes cederstroemi and Leptodora kindti- relationship to seasonal abundance in western Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 47: 731-738.

260

Gliwicz, Z. M. 1980. Filtering rates, food size selection, and feeding rates in cladocerans-another aspect of interspecific competition in filter-feeding zooplankton. pp. 282-291 in Evolution and ecology of zooplankton communities. W.C. Kerfoot, (ed.). University Press of New England. Hanover, New Hampshire

Gliwicz, Z. M., and W. Lampert. 1990. Food thresholds in Daphnia species in the absence and presence of blue-green filaments. Ecology 7: 691- 702.

Gliwicz, Z. M., and E. Siedlar. 1980. Food size limitation and algae interfering with food collection in Daphnia. Archiv für Hydrobiologie 88: 155-177.

Gliwicz, M. J., and A. Sieniawska. 1986. Filtering activity of Daphnia in low concentrations of a pesticide. Limnology and Oceanography 31: 1132-1138.

Glooschenko, W.A., J. E. Moore, and R. A. Vollenweider. 1974. Spatial and temporal distribution of chlorophyll a and pheopigments in surface waters of Lake Erie. Journal of the Fisheries Research Board of Canada 31: 265-274.

Gopalan, G., D. A. Culver, L. Wu, and B. K. Trauben. 1998. The effect of recent ecosystem changes on the recruitment of young-of-year fish in western Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 55: 2572-2579.

Great Lakes Fisheries Commission. 2001. Minutes, 2001 Annual Meeting of the Lake Erie Committee (With Attachments). 440 pages. Great Lakes Fishery Commission, 2100 Commonwealth Blvd., Ann Arbor, MI 48105.

Grigorovich, I. A., H. J. MacIsaac, I. K. Rivier, N. V. Aladin, and V. E. Panov. 2000. Comparative biology of the predatory cladoceran Cercopagis pengoi from Lake Ontario, Baltic Sea, and Caspian Sea. Archiv für Hydrobiolgie 149: 23-50.

Gunderson, L. H. 2000. Ecological resilience - in theory and application. Annual Review of Ecology and Systematics 31: 425-439.

Hall, D. J. 1964. An experimental approach to the dynamics of a natural population of Daphnia galeata mendotae. Ecology 45: 94-112.

Hall, S. R., N. K. Pauliukonis, E. L. Mills, L. G. Rudstam, C. P. Schneider, S. J. Larry, and F. Arrhenius. 2003. A comparison of total phosphorus, chlorophyll a, and zooplankton in embayment, nearshore, and offshore habitats of Lake Ontario. Journal of Great Lakes Research 29: 54-69.

Hanazato, T. 2001. Pesticide effects on freshwater zooplankton: an ecological perspective. Environmental Pollution 112: 1-10.

261

Hanazato, T., and S. I. Dodson. 1995. Synergistic effects of low oxygen concentration, predator kairomone, and a pesticide on the cladoceran Daphnia pulex. Limnology and Oceanography 40: 700-709.

Haney, J. F., and D. J. Hall. 1973. Sugar-coated Daphnia: a preservation technique for Cladocera. Limnology and Oceanography 18: 331-333.

Hartgers, E. M., E. H. W. Heugens, and J. W. Deneer. 1999. Effect of lindane on the clearance rate of Daphnia magna. Archives of Environmental Contamination and Toxicology 36: 399-404.

Hartig, J. H., M. A. Zarull, T. B. Reynoldson, G. Mikol, V. A. Harris, R. G. Randall, and V. W. Cairns. 1997. Quantifying targets for rehabilitating degraded areas of the Great Lakes. Environmental Management 21: 713-723.

Havens, K. E. 1993. An experimental comparison of the effects of two chemical stressors on a freshwater zooplankton assemblage. Environmental Pollution 84: 245-251.

Havens, K. E. 1998. Size structure and energetics in a plankton food web. Oikos 81: 346- 358.

Havens, K.E. 1999. Comparative analysis of lake plankton structure vs. function. Aquatic Sciences 61: 150-167.

Hawkins, B. E., and M. S. Evans. 1979. Seasonal cycles of zooplankton biomass in southeastern Lake Michigan. Journal of Great Lakes Research 5: 256-263.

Hebert P. D. N., B. W. Muncaster, and G. L. Mackie. 1989. Ecological and genetic studies on Dreissena polymorpha (Pallas): a new mollusk in the Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 46: 1587-1591.

Hoffman, J. C., M. E. Smith, and J. T. Lehman. 2001. Perch or plankton: top-down control of Daphnia by yellow perch (Perca flavescens) or Bythotrephes cederstroemi in an inland lake? Freshwater Biology 46: 759-775.

Hubschman, J. H. 1960. Relative daily abundance of planktonic crustacea in the island region of Lake Erie. Ohio Journal of Science 60: 335-340.

Hutchinson, G. E. 1957. Treatise on Limnology, Volume 1. John Wiley and Sons. New York.

262

Jeppesen, E., J. P. Jensen, S. Amsinck, F. Landkildehus, T. Lauridsen, and S. F. Mitchell. 2002. Reconstructing the historical changes in Daphnia mean size and planktivorous fish abundance in lakes from the size of Daphnia ephippia in the sediment. Journal of Paleolimnology 27: 133-143.

Johannsson, O. E., C. Dumitru, and D. M. Graham. 1999a. Estimation of zooplankton mean length for use in an index of fish community structure and its application to Lake Erie. Journal of Great Lakes Research 25: 179-186.

Johannsson, O. E., D. M. Graham, D. W. E. Einhouse, and E. L. Mills. 1999b. Historical and recent changes in the Lake Erie zooplankton community and their relationship to ecosystem function. pp. 169-196. in State Of Lake Erie (SOLE)- Past, Present, and Future. Munawar, M., T. Edsall, and I. F. Munawar (eds.). Backhuys Publishers, Leiden, The Netherlands.

Johengen, T. H., O. E. Johannsson, G. L. Pernie, and E. S. Millard. 1994. Temporal and seasonal trends in nutrient dynamics and biomass measures in Lakes Michigan and Ontario in response to phosphorus control. Canadian Journal of Fisheries and Aquatic Sciences 51: 2570-2578.

Jones, G. J., and W. Korth. 1995. In-situ production of volatile odor compounds by river and reservoir phytoplankton populations in Australia. Water Science and Technology 31: 145-151.

Kane, D. D., D. A. Culver, M. M. Munawar, and S. I. Gordon. 2003a. A Planktonic Index of Biotic Integrity (P-IBI) for Lake Erie: development, validation, and application opportunities. pp.128-129. 46th Conference on Great Lakes Research/ 10th World Lake Conference- Abstracts. 2003 June 22-26. International Association for Great Lakes Research, Ann Arbor, Michigan.

Kane, D.D., E. M. Haas, and D. A. Culver. 2003b. The characteristics and potential ecological effects of the exotic crustacean zooplankter Cercopagis pengoi (Cladocera: Cercopagidae), a recent invader of Lake Erie. Ohio Journal of Science 103: 79-83.

Kane, D. D., J.E. Gannon, and D.A. Culver. 2004. The Status of Limnocalanus macrurus (Copepoda: Calanoida: Centropagidae) in Lake Erie. Journal of Great Lakes Research 30: 22-30.

Karr, J. R. 1981. Assessment of biotic integrity using fish communities. Fisheries 6: 21- 27.

263

Karr, J. R. 1991. Biological integrity: a long-neglected aspect of water resource management. Ecological Applications 1: 66-84.

Karr, J. R., and E. W. Chu. 1997. Biological monitoring and assessment: Using multimetric indices effectively. USEPA 235-R97-001.

Karr, J. R., and D. R. Dudley. 1981. Ecological perspective on water quality goals. Environmental Management 5: 55-68.

Karr, J.R., K.D. Fausch, P.L. Angermeier, P.R. Yant, and I.J. Schlosser. 1996. Assessing biological integrity in running waters: a method and its rationale. Illinois Natural History Survey, Special Publication 5.

Kerfoot, W.C., C. Levitan, and W.R. DeMott. 1988. Daphnia-phytoplankton interactions: density-dependent shifts in resource quality. Ecology 69: 1806-1825.

Kershner, M.W., and J. Hopkins. 2003. Development and evaluation of an index of biotic integrity for the offshore fish assemblage of Lake Erie. Final Report for LEQI 01-09. 38 pp.

Kratz, T. K., T. M. Frost, and J. J. Magnuson. 1987. Influences from spatial and temporal variability in ecosystems: long-term zooplankton data from lakes. American Naturalist 129: 830-846.

Krieger, K. A., D.W. Schloesser, B. A. Manny, C. E. Trisler, S. E. Heady, J. J. H. Ciborowski, and K. M. Muth. 1996. Recovery of burrowing mayflies (Ephemeroptera: Ephemeridae: Hexagenia) in Western Lake Erie. Journal of Great Lakes Research 22: 254-263.

Krylov, P. I., Bychenkov, D. E., Panov, V. E., Rodionova, N. V., and Telesh, I. V. 1999. Distribution and seasonal dynamics of the Ponto-Caspian invader Cercopagis pengoi (Crustacea, Cladocera) in the Neva Estuary (Gulf of Finland). Hydrobiologia 393: 227-232.

Lam, D.C.L., and W.M. Schertzer. 1987. Lake Erie thermocline model results: comparison with 1967-1982 data and relati on to anoxic occurrences. Journal of Great Lakes Research 13: 757-769.

Lampert, W. 1981. Inhibitory and toxic effects of blue-green algae on Daphnia. Internationale Revue der Gesamten Hydrobiolgie 66: 285-298.

Lampert, W. 1982. Further studies on the inhibitory effect of the toxic blue-green Microcystis aeruginosa on the filtering rate of zooplankton. Archiv für Hydrobiolgie 95: 207-220.

264

Lampert, W. 1989. The adaptive significance of diel vertical migration of zooplankton. Functional Ecology 3: 21-27.

Lampert, W. 1997. Zooplankton research: the contribution of limnology to general ecological paradigms. Aquatic Ecology 31: 19-27.

Lange, C., and Cap, R. 1986. Bythotrephes cederstroemi (Schodler) (Cercopagidae, Cladocera)- a new record for Lake Ontario. Journal of Great Lakes Research 12: 142-143.

Larsson, P., L. Okla, and G. Cronberg. 1998. Turnover of polychlorinated biphenyls in an oligotrophic and an eutrophic lake in relation to internal lake processes and atmospheric fallout. Canadian Journal of Fisheries and Aquatic Sciences 55: 1926-1937.

Lauridsen, T. L. and I. Buenk. 1996. Diel changes in the horizontal distribution of zooplankton in the littoral zone of two shallow eutrophic lakes. Archiv für Hydrobiologie 137: 161-176.

Leach, J. H., and R. C. Herron. 1992. A review of lake habitat classification. pp. 27-57 in The development of an aquatic habitat classification system for lakes. W.-D.N. Busch, and P. G. Sly (eds.). CRC Press, United States.

Lehman, J. T., and D. K. Branstrator. 1995. A model for growth, development, and diet selection by the invertebrate predator Bythotrephes cederstroemi. Journal of Great Lakes Research 21: 610-619.

Lehman, J.T., and C. E. Caceres. 1993. Food-web responses to species invasion by a predatory invertebrate: Bythotrephes in Lake Michigan. Limnology and Oceanography 38: 879-891.

Lindeman, R. L. 1942. The tophic-dynamic aspect of ecology. Ecology 23: 392-412.

Lorenzen, C. J. 1967. Determination of chlorophyll and pheo-pigments: spectrophotometric equations. Limnology and Oceanography 12: 343-346.

Lougheed, V. L., and P. Chow-Fraser. 2002. Development and use of a zooplankton index of wetland quality in the Laurentian Great Lakes basin. Ecological Applications 12: 474-486.

Ludsin, S. A., M. W. Kershner, K. A. Blocksom, R. L. Knight, and R. A. Stein. 2001. Life after death in Lake Erie: nutrient controls drive fish species richness. Ecological Applications 11: 731-746.

265

Lund, J. W. G., C. Kipling, and E. D. LeCren. 1958. The inverted microscope method of estimating algal numbers and the statistical basis of estimations by counting. Hydrobiologia 11: 143-170.

MacIsaac, H. J., W. G. Sprules, O. E. Johannsson, and J. H. Leach. 1992. Filtering impacts of larval and sessile zebra mussels (Dreissena polymorpha) in western Lake Erie. Oecologia 92: 30-39.

MacIsaac, H. J., I. A. Grigorovich, J. A. Hoyle, N. D. Yan, and V. E. Panov. 1999. Invasion of Lake Ontario by the Ponto-Caspian predatory cladoceran Cercopagis pengoi. Canadian Journal of Fisheries and Aquatic Sciences 56: 1-5.

Magnuson, J. J. 1990. Long-term ecological studies and the invisible present. Bioscience 40: 495-501.

Magnuson, J. J., K. E. Webster, R. A. Assel, C. J. Bowser, P. J. Dillon, J. G. Eaton, H. E. Evans, E. J. Fee, R. I. Hall, L. R. Mortsh, D. W. Schindler, and F. H. Quinn. 1997. Potential effects of climate change on aquatic systems: Laurentian Great Lakes and Precambrian Shield Region. Hydrological Processes 11: 825- 871.

Makarewicz, J. C. 1993a. Phytoplankton biomass and species composition in Lake Erie, 1970 to 1987. Journal of Great Lakes Research 19: 258-274.

Makarewicz, J. C. 1993b. A lakewide comparison of zooplankton biomass and its species composition in Lake Erie, 1983-87. Journal of Great Lakes Research 19: 275-290.

Makarewicz, J.C., and P. Bertram. 1991. Evidence for the restoration of the Lake Erie ecosystem. Bioscience 41: 216- 223.

Makarewicz, J. C., P. Bertram, and T. W. Lewis. 1998. Changes in phytoplankton size- class abundance and species composition coinciding with changes in water chemistry and zooplankton community structure of Lake Michigan, 1983 to 1992. Journal of Great Lakes Research 24: 637-657.

Makarewicz, J. C., P. Bertram, T. W. Lewis, and E. H. Brown, Jr. 1995. A decade of predatory control of zooplankton species composition of Lake Michigan. Journal of Great Lakes Research 21: 620-640.

266

Makarewicz, J. C., I. A. Grigorovich, E. Mills, E. Damaske, M. E. Cristescu, W. Pearsall, M. J. LaVoie, R. Keats, L.Rudstam, P. Hebert, H. Halbritter, T. Kelly, C. Matkovich, and H. J. MacIsaac. 2001. Distribution, fecundity, and genetics of Cercopagis pengoi (Ostroumov) (Crustacea, Cladocera) in Lake Ontario. Journal of Great Lakes Research 27: 19-32.

Makarewicz, J. C., and G. E. Likens. 1979. Structure and function of the zooplankton community of Mirror Lake, New Hampshire. Ecological Monographs 49: 109- 127.

Marvin, C. H., M. N. Charlton, E. J. Reiner, T. Kolic, K. MacPherson, G. A. Stern, E. Braekevelt, J. F. Estenik, L. Thiessen, and S. Painter. 2002. Surficial sediment contamination in Lakes Erie and Ontario: a comparative analysis. Journal of Great Lakes Research 28: 437-450.

May, B., and J. E. Marsden. 1992. Genetic identification and implications of another invasive species of dreissenid mussel in the Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 49: 1501-1506.

Mazumder, A., D. R. S. Lean, and W. D. Taylor. 1992. Dominance of small filter feeding zooplankton in the Lake Ontario foodweb. Journal of Great Lakes Research 18: 456-466.

McGarigal, K., S. Cushman, and S. Stafford. 2000. Multivariate statistics for wildlife and ecology research. Springer-Verlag. New York, New York.

McNaught, D. C., and A. D. Hasler. 1966. Photoenvironments of planktonic Crustacea in Lake Michigan. Verhandlungen Internationale Vereinigung für Theoretische und Angewandte Limnologie 16: 194-203.

Meng, L., C. D. Orphanides, and J. C. Powell. 2002. Use of a fish index to assess habitat quality in Naragansett Bay, Rhode Island. Transactions of the American Fisheries Society 131: 731-742.

Microsoft Access. 2000. Microsoft Corporation.

Minitab Version 13.1. 2000. Minitab Incorporated.

Mills, E. L., D. M. Green, and A. Schiavone Jr. 1987. Use of zooplankton size to assess the community structure of fish populations in freshwater lakes. North American Journal of Fisheries Management 7: 369-378.

267

Mills, E. L., J. H. Leach, J. T. Carlton, and C. L. Secor. 1994. Exotic species and the integrity of the Great Lakes. Bioscience 44: 666-676.

Mills, E. L., and A. Schiavone Jr. 1982. Evaluation of fish communities through assessment of zooplankton populations and measures of lake productivity. North American Journal of Fisheries Management 2: 14-27.

Miltner, R. J., and E. T. Rankin. 1998. Primary nutrients and the biotic integrity of rivers and streams. Freshwater Biology 40: 145-158.

Minitab Version 13.1. 2000. Minitab Incorporated.

Mischke, C. C., and P. V. Zimba. 2001. Assessment of zooplankton size fractionation for monitoring fry and fingerling culture ponds. North American Journal of Aquaculture 63: 289-292.

Müller-Navarra, D. C., S. Güss, and H. Von Storch. 1997. Interannual variability of seasonal succession events in a temperate lake and its relation to temperature variability. Global Change Biology 3: 429-438.

Munawar, M., and I. F. Munawar. 1976. A lakewide study of phytoplankton biomass and its species composition in Lake Erie, April- December 1970. Journal of the Fisheries Research Board of Canada 33: 581-600.

Munawar, M., I. F. Munawar, R. Demott, H. Niblock, and S. Carou. 2002. Is Lake Erie a resilient ecosytem? Aquatic Ecosystem Health and Management 5: 79-93.

Muzinic, C. J. 2000. First record of Daphnia lumholtzi Sars in the Great Lakes. Journal of Great Lakes Research 26: 352-354.

Naiman, R. J., J. J. Magnuson, D. M. McKnight, J.A. Stanford, and J. R. Karr. 1995. Freshwater ecosystems and their management: a national initiative. Science 270: 584-585.

Neilson, M. A., D. S. Painter, G. Warren, R. A. Hites, I. Basu, D. V. C. Weseloh, D. M. Whittle, G. Christie, R. Barbiero, M. Tuchman, O. E. Johannsson, T. F. Nalepa, T. A. Edsall, G. Fleischer, C. Bronte, S. B. Smith, and P. C. Baumann. 2003. Ecological monitoring for assessing the state of the nearshore and open waters of the Great Lakes. Environmental Monitoring Assessment 88: 103-117.

Nicholls, K. H., D. W. Standen, G. J. Hopkins, and E. C. Carney. 1977. Declines in the near-shore phytoplankton of Lake Erie’s western basin since 1971. Journal of Great Lakes Research 3: 72-78.

268

Nicholls, K. H. 1997. Planktonic green algae in western Lake Erie: the importance of temporal scale in the interpretation of change. Freshwater Biology 38: 419-425.

Nygaard, G. 1949. Hydrobiological studies in some lakes and ponds. Part II- The quotient hypothesis and some new or little known phytoplankton organisms. Kgl. Danske. Vidensk. Selsk. Biol. Skrifter 7: 1-293.

Odum, E. P. 1969. The strategy of ecosystem development. Science 164: 262- 270.

Ohio Department of Natural Resources, Ohio Division of Wildlife, Lake Erie Fisheries Unit. 1999. Ohio’s Lake Erie Fisheries. 94 pp.

Ohio Department of Natural Resources (ODNR). 2003. Ohio’s Lake Erie Fisheries - 2002. Report prepared by the Lake Erie Fisheries Units, Division of Wildlife, ODNR for submission to the Lake Erie Committee, Great Lakes Fishery Commission. 92 pp.

Ohio EPA. 1988. Biological criteria for the protection of aquatic life. Volumes 1-3. Ohio EPA, Division of Water Quality Monitoring and Assessment. Columbus Ohio.

Ohio Lake Erie Commission. 1998. State of Ohio 1998- State of the Lake Report- Lake Erie Quality Index. 88 pp.

Ojaveer, H., L. A. Kuhns, R. P. Barbiero, and M. L. Tuchman. 2001. Distribution and population characteristics of Cercopagis pengoi in Lake Ontario. Journal of Great Lakes Research 27: 19-32.

Ojaveer, E., A. Lumberg, and H. Ojaveer. 1998. Highlights of zooplankton dynamics in Estonian waters (Baltic Sea). ICES Journal of Marine Science 55: 748-755.

Paerl, H.W. 1988. Nuisance phytoplankton blooms in coastal, estuarine, and inland waters. Limnology and Oceanography 33: 823-847.

Panek, J., D. M. Dolan, and J. H. Hartig. 2003. Detroit's role in reversing cultural eutrophication of Lake Erie, p. 79-90. In: Hartig, J. H. (ed.). Honoring Our Detroit River. Wayne St. Univ. Press, Detroit, MI.

Paterson, M. J., D. C. G. Muir, B. Rosenberg, E. J. Fee, C. Anema, and W. Franzin. 1998. Does lake size affect concentrations of atmospherically derived polychlorinated biphenyls in water, sediment, zooplankton, and fish? Canadian Journal of Fisheries and Aquatic Sciences 55: 544-553.

269

Persson, P.E. 1980. Sensory properties and analysis of 2-muddy odor compounds, geosmin and 2-methylisoborneol, in water and fish. Water Research 14: 1113- 1118.

Pinel-Alloul, B. 1995. Spatial heterogeneity as a multiscale characteristic of zooplankton community. Hydrobiologia 300/301: 17-42.

Porter, K.G., and J. D. Orcutt, Jr. 1980. Nutritional adequacy, manageability, and toxicity as factors that determine the food quality of green and blue-green algae for Daphnia. pp. 268-281 in: Evolution and ecology of zooplankton communities. W.C. Kerfoot, (ed.). University Press of New England. Hanover, New Hampshire.

Prescott, G.W. 1978. How to know the freshwater algae. McGraw-Hill. U.S.A.

Price, J. W. 1963. A study of the food habitats of some Lake Erie fish. Bulletin of the Ohio Biological Survey 2: 89 pp.

Quinn, F. H. 2002. Secular changes in Great Lakes water level seasonal cycles. Journal of Great Lakes Research 28: 451-465.

Rand, G. M., J. R. Clark, C. M. Holmes. 2001. The use of outdoor freshwater pond microcosms. III. Responses of phytoplankton and periphyton to pyridaben. Environmental Toxicology 16: 96-103.

Rathke, D. E., and G. McCrae. 1989. Appendix B, Volume III, Report of the Great Lakes Water Quality Board. International Joint Commission, Windsor, Ontario.

Rawson, D. S. 1956. Algal indicators of trophic lake types. Limnology and Oceanography 1: 18-25.

Resh, V. H., R. H. Norris, and M. T. Barbour. 1995. Design and implementation of rapid assessment approaches for water resource monitoring using benthic macroinvertebrates. Australian Journal of Ecology 20: 108-121.

Ricciardi, A., and H. J. MacIsaac. 2000. Recent mass invasion of the North American Great Lakes by Ponto-Caspian species. Trends in Ecology and Evolution 15: 62- 65.

Ricciardi, A., and J. B. Rasmussen. 1998. Predicting the identity and impact of future biological invaders: a priority for aquatic resource management. Canadian Journal of Fisheries and Aquatic Sciences 55: 1759-1765.

270

Richards, B. P., D. B. Baker, J. W. Kramer, and D. E. Ewing. 1996. Annual loads of herbicides in Lake Erie tributaries of Michigan and Ohio. Journal of Great Lakes Research 22: 414-428.

Rodhe, W. 1958. Primarproduktion und Seetypen. Verhandlungen Internationale Vereinigung für Theoretische und Angewandte Limnologie. 13: 121-141.

Roff, J. C. 1972. Aspects of the reproductive biology of the planktonic copepod Limnocalanus macrurus Sars. Crustaceana (Leiden) 22: 155-160.

Rusak, J. A., N. D. Yan, K. M. Somers, K. L. Cottingham, F. Micheli, S. R. Carpenter, T. M. Frost, M. J. Paterson, and D. J. McQueen. 2002. Temporal, spatial, and taxonomic patterns of crustacean zooplankton variability in unmanipulated north- temperate lakes. Limnology and Oceanography 47: 613-625.

Ryder, R. A. 1965. A method for estimating the potential fish production of north temperate lakes. Transactions of the American Fisheries Society 94: 214- 218.

Sameoto, D., P. Wiebe, J. Runge, L. Postel, J. Dunn, C. Miller, and S. Coombs. 2000. Collecting zooplankton. pp. 56-81 in R.P. Harris, P. H. Wiebe, J. Lenz, H. R. Skjooldal, and M. Huntley (eds.). ICES Zooplankton Methodology Manual. Academic Press, Great Britian.

SAS Version 8e. 1999-2001. The SAS Institute. Cary, North Carolina.

Schertzer, W.M., J. H. Saylor, F. M. Boyce, D. G. Robertson, and F. Rosa. 1987. Seasonal thermal cycle of Lake Erie. Journal of Great Lakes Research 13:468- 486.

Schindler, D. W. 1977. Evolution of phosphorus limitation in lakes. Science 195: 260- 262.

Schindler, D. W. 1987. Detecting ecosystem response to anthropogenic stress. Canadian Journal of Fisheries and Aquatic Sciences 44(Supplement 1): 6-25.

Schriver, P., J. Bogestrand, E. Jeppesen, and M. Sondergaard. 1995. Impact of submerged macrophytes on fish-zooplankton-phytoplankton interactions- large- scale enclosure experiments in a shallow eutrophic lake. Freshwater Biology 33: 255-270.

Selgeby, J. H. 1975. Life histories and abundance of crustacean zooplankton in the outlet of Lake Superior, 1971-1972. Journal of the Fisheries Research Board of Canada 32: 461-470.

271

Shannon, C.E. and W. Weaver. 1949. The Mathematical Theory of Communication. The University of Illinois Press, Urbana, Illinois.

Shannon, E.E., and P.O. Brezonik. 1972. Relationships between lake trophic state and nitrogen and phosphorus loading rates. Environmental Science and Technology 6: 719-725.

Shaw, M. A., and J. R. M. Kelso. 1992. Environmental factors influencing zooplankton species composition of lakes in north-central Ontario, Canada. Hydrobiologia 241: 141-154.

Simberloff, D., and B. Von Holle. 1999. Positive interactions of nonindigenous species: invasional meltdown? Biological Invasions 1: 21-32.

Smith, R. E. H., J. A. Furgal, and D. R. S. Lean. 1998. The short-term effects of solar ultraviolet radiation on phytoplankton photosynthesis and photosynthate allocation under contrasting mixing regimes in Lake Ontario. Journal of Great Lakes Research 24: 427-441.

Smith, V. H. 1979. Nutrient dependence of primary productivity in lakes. Limnology and Oceanography 24: 1051-1064.

Smith, V. H. 1983. Low nitrogen to phosphorus ratios favor dominance by blue-green algae in lake phytoplankton. Science 221: 669-671.

Sommer, U., U. Gaedke, and A. Schweizer. 1993. The first decade of oligotrophication of Lake Constance. II. The response of phytoplankton taxonomic composition. Oecologia 93: 276-284.

Sommer, U., Z. M. Gliwicz, W. Lampert, and A. Duncan. 1986. The PEG-model of seasonal succession of planktonic events in fresh waters. Archiv für Hyrgobiolgie 106: 433-471.

Soranno, P.A. 1997. Factors affecting the timing of surface scums and epilimnetic blooms of blue-green algae in a eutrophic lake. Canadian Journal of Fisheries and Aquatic Sciences 54: 1965-1975.

Sprules, W. G., H. P. Riessen, and E. H. Jin. 1990. Dynamics of the Bythotrephes invasion of the St. Lawrence Great Lakes. Journal of Great Lakes Research 16: 346-351.

Stapelton, H. M., J. Skubinna, and J. E. Baker. 2002. Seasonal dynamics of PCB and toxapherne bioaccumulation within a Lake Michigan food web. Journal of Great Lakes Research 28: 52-64.

272

StatSoft. 1998. Statistica Version 6.0. Tulsa, Oklahoma.

Stemberger, R. S. and C. Y. Chen. 1998. Fish tissue metals and zooplankton assemblages of northeastern U.S. lakes. Canadian Journal of Fisheries and Aquatic Sciences 55: 339-352.

Stemberger, R. S., and E. K. Miller. 1998. A zooplankton N:P ratio indicator for lakes. Environmental Monitoring and Assessment 51: 29-51.

Sterner, R. W. 1989. The role of grazers in phytoplankton successon. Pp. 107-170. in U. Sommer (ed.). Plankton ecology: succession in plankton communities. Springer- Verlag. U.S.A.

Stockner, J.G., and W. W. Benson. 1967. The succession of diatom assemblages in the recent sediments of Lake Washington. Limnology and Oceanography 12: 513- 532.

Stockwell, J. D., and W. G. Sprules. 1995. Spatial and temporal patterns of zooplankton biomass in Lake Erie. ICES Journal of Marine Science 52: 557-564.

Stockwell, J. D., P. Dutilleul, and W. G. Sprules. 2002. Spatial structure and the estimation of zooplankton biomass in Lake Erie. Journal of Great Lakes Research 28: 362-378.

Strøm, K. M. 1946. The ecological niche. Nature 157: 375.

Stuckey, R. L., and D. L. Moore. 1995. Return and increase in abundance of aquatic flowering plants in Put-In-Bay Harbor, Lake Erie, Ohio. Ohio Journal of Science 95: 262-266.

Sugiura, N., N. Iwami, Y. Inamori, O. Nishimura, and R. Sudo. 1998. Significance of attached cyanobacteria relevant to the occurrence of musty odor in Lake Kasumigaura. Water Research 32: 3549-3554.

Sugiura, N., B. Wei, and T. Maekawa. 2002. The discrimination of the response pattern of inter-phylum phytoplankton diversity to long-term eutrophication trends in Lake Kasumigaura, Japan. Aquatic Ecosystem Health and Management 5: 403- 410.

Surfer Version 6.01. 1993-95. Golden Software Incorporated.

Swackhammer, D. L., R. F. Pearson, and S. P. Schottler. 1998. Toxaphene in the Great Lakes. Chemosphere 37: 2545-2561.

273

Tansley, A. G. 1935. The use and abuse of vegetational terms. Ecology 16: 284-307.

Tatrai, I., J. Olah, G. Paulovits, K. Matyas, B. J. Kawieka, V. Jozsa, and F. Pekar. 1997. Biomass dependent interactions in pond ecosystems: responses of lower trophic levels to fish manipulation. Hydrobiologia 345: 117-129.

Therriault, T. W., I. A. Grigorovich, D. D. Kane, E. M. Haas, D. A. Culver, and H. J. MacIsaac. 2002. Range expansion of the exotic zooplankter Cercopagis pengoi (Ostroumov) into western Lake Erie and Muskegon Lake. Journal of Great Lakes Research 28: 698-701.

Tin, M. 1965. Comparison of some ratio estimators. Journal of the American Statistical Association 60: 294-307.

Toms, J. D., and J. D. Lesperance. 2003. Piecewise regression: a tool for identifying ecological thresholds. Ecology 84: 2034-2041.

Torke, B. G. 1975. The population dynamics and life histories of crustacean zooplankton at a deep-water station in Lake Michigan. Ph.D. thesis, University of Wisconsin.

Twiss, M. R., P. G. C. Campbell, J.-C. Auclair. 1996. Regeneration, recycling, and trophic transfer of trace metals by microbial food-web organisms in the pelagic surface waters of Lake Erie. Limnology and Oceanography 41: 1425-1437.

Uitto, A., E. Gorokhova, and P. Valipakka. 1999. Distribution of the non-indigenous Cercopagis pengoi in the coastal waters of the eastern Gulf of Finland. ICES Journal of Marine Science 56 (Supplement): 49-57.

USEPA. 1998. Lake and reservoir bioassessment and biocriteria: technical guidance document. EPA 841-B-98-007.

USEPA. 2004. http://www.epa.gov/glnpo/lakeerie/eriedeadzone.html

Utermöhl, H. 1958. Zur vervollkommung der quantitativen phytoplankton-methodik. Mitteilungen-Internationale Vereiningung für Limnologie 9: 1-38.

Uttormark, P.D., and J. P. Wall. 1975. Lake classification- a trophic characterization of Wisconsin lakes. EPA-660/3-75-033. U.S. Environmental Protection Agency, Washington D.C.

274

Vaga, R. M., D. A. Culver, and C. S. Munch. 1985. The fecundity ratio of large to small filter-feeding cladocerans as a function of inedible algal standing crop. Verhandlungen Internationale Vereinigung für Theoretische und Angewandte Limnologie 22: 3072- 3075.

Vallentyne, J. R., J. Shapiro, and A. M. Beeton. 1969. The process of eutrophication and criteria for trophic state determination. pp. 57-67 in Modelling the Eutrophication Process Proceedings of the Workshop. St. Petersburg, Florida.

Van Dam, R. A., C. Camilleri, and C. M. Finlayson. 1998. The potential of rapid assessment techniques as early warning indicators of wetland degradation: a review. Environmental Toxicology and Water Quality 13: 297-312.

Van Oosten, J. 1937. First records of the smelt, Osmerus mordax, in Lake Erie. Copeia 1937: 64-65.

Vanderploeg, H.A., T. F. Nalepa, D. J. Jude, E. L. Mills, K. T. Holeck, J. R. Liebig, I. A. Grigorovich, and H. Ojaveer. 2002. Dispersal and emerging ecological impacts of Ponto-Caspian species in the Laurentian Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 59: 1209-1228.

Vollenweider, R. A. 1968. Scientific fundamentals of the eutrophication of lakes and flowing waters, with particular reference to nitrogen and phosphorus as factors in eutrophication.OECD Rep. DAS/SCI/ 68.27 U. N. Organization for Economic and Cultural Development. Paris.

Watson, N. H. F. 1976. Seasonal distribution and abundance of crustacean zooplankton in Lake Erie, 1970. Journal of the Fisheries Research Board of Canada 33: 612- 621.

Watson, N. H. F., and G. F. Carpenter. 1974. Seasonal abundance of crustacean zooplankton and net plankton biomass of Lakes Huron, Erie, and Ontario. Journal of the Fisheries Research Board of Canada 31: 309-317.

Weber, C. A. 1907. Aufbau und vegetation der Moore Norddeutschlands. Bot. Jahrb . Beibl. 90:19-34.

Weisberg, S.B., J.A. Ranasinghe, L.C. Schaffner, R.J. Diaz, D.M. Dauer, and J.B. Frithsen.1997. An estuarine Benthic Index of Biotic Integrity for Chesapeake Bay. Estuaries 20: 149-158.

275

Weisgerber, K.M. 2000. Lower trophic level dynamics in western basin, Lake Erie: changes in biomass, clearance, and nutrient excretion rates in crustacean zooplankton versus zebra mussels, M.S. Thesis, The Ohio State University, Columbus, Ohio.

Wells, L. 1970. Effects of alewife predation on zooplankton populations in Lake Michigan. Limnology and Oceanography 15: 556-565.

Wen, Y. H., and R.H. Peters. 1994. Empirical models of phosphorus and nitrogen excretion rates by zooplankton. Limnology and Oceanography 39: 1669-1679.

Wetzel, R. G. 2001. Limnology: lake and river ecosystems. 3rd Edition. Academic Press. United States of America.

Wiener, N. 1948. Cybernetics, or Control and Communication in the Animal and the Machine. Massachusetts Institute of Technology Press, Cambridge, Massachusetts.

Winkler, G., J. J. Dodson, N. Bertrand, D. Thivierge, and W. F. Vincent. 2003. Trophic coupling across the St. Lawrence River estuarine transition zone. Marine Ecology Progress Series 251: 59-73.

Witt, J. D. S, P. D. N. Hebert, and W. B. Morton. 1997. Echinogammarus ischnus: Another crustacean invader in the Laurentian Great Lakes basin. Canadian Journal of Fisheries and Aquatic Sciences 54: 264-268.

Wright, S. 1955. Limnological survey of western Lake Erie. U.S. Fish Wildl. Serv., Spec. Sci. Rpt. Fish. No. 139, 341 pp.

Wu, J.-T. 1999. A generic index of diatom assemblages as bioindicator of pollution in the Keelung River of Taiwan. Hydrobiologia 397: 79-87.

Wu, L., and D. A. Culver. 1991. Zooplankton grazing and phytoplankton abundance: an assessment before and after invasion of Dreissena polymorpha. Journal of Great Lakes Research 17: 425-436.

Wu L., and D. A. Culver. 1992. Ontogenetic diet shift in Lake Erie age-0 yellow perch: a size related response to zooplankton density. Canadian Journal of Fisheries and Aquatic Sciences 49: 1932-1937.

Wu, L., and D. A. Culver. 1994. Daphnia dynamics in western Lake Erie: Regulation by algal limitation and young-of-year fish predation. Journal of Great Lakes Research 20:537-545.

276

Xu, F.-L., R. W. Dawson, S. Tao, J. Cao, and B.-G. Li. 2001. A method for lake ecosystem health assessment: an Ecological Modeling Method (EMM) and its application. Hydrobiologia 443: 159-175.

Yaksich, S. M., D. A. Melfi, D. B. Baker, and J. W. Kramer. 1985. Lake Erie nutrient loads, 1970-1980. Journal of Great Lakes Research 11: 117-131.

Yan, N.D. and T.W. Pawson. 1998. Seasonal variation in the size and abundance of the invading Bythotrephes in Harp Lake, Ontario, Canada. Hydrobiologia 361:157- 168.

Yu, R.-Q., and W.-X. Wang. 2002. Trace metal assimilation and release budget in Daphnia magna. Limnology and Oceanography 47: 495-504.

277