Causes and Consequences of Algal Blooms in the Tidal Fresh James River

Home , Phytodetritus

Virginia Commonwealth University VCU Scholars Compass

Theses and Dissertations Graduate School

2014

Joseph Wood Virginia Commonwealth University

Follow this and additional works at: https://scholarscompass.vcu.edu/etd

Part of the Life Sciences Commons

Downloaded from https://scholarscompass.vcu.edu/etd/3429

This Dissertation is brought to you for free and open access by the Graduate School at VCU Scholars Compass. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of VCU Scholars Compass. For more information, please contact [email protected].

Causes and Consequences of Algal Blooms in the Tidal Fresh James River

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy at Virginia Commonwealth University.

Joseph D. Wood

M.S. Biology Virginia Commonwealth University, 2010 B.A. Philosophy, Virginia Polytechnic Institute and State University, 2007

Director: Dr. Paul Bukaveckas, Professor, Integrative Life Sciences

Virginia Commonwealth University Richmond, Virginia May, 2014

Causes and Consequences of Algal Blooms in the Tidal Freshwater James River Joseph Wood

Dissertation Committee Dr. Paul Bukaveckas (Major Advisor) Dr. Leigh McCallister Dr. Daniel McGarvey Dr. Stephen P. McIninch, Dr. Michael Pace (University of Virginia)

Dissertation Committee ...... ii ACKNOWLEDGEMENTS ...... viii PREFACE ...... ix Chapter 1: Increasing severity of phytoplankton nutrient limitation following reductions in point source inputs to the tidal freshwater segment of the James River Estuary...... 1 ABSTRACT ...... 1 INTRODUCTION ...... 2 METHODS ...... 4 Site Description ...... 4 Monitoring Data ...... 5 Bioassay experiments ...... 6 RESULTS ...... 7 Seasonal and Inter-annual Variation in CHLa and Nutrients ...... 7 Bioassay Experiments ...... 8 DISCUSSION ...... 10 Tables and Figures ...... 16 Table 1-1. Statistical analyses of light and nutrient effects on phytoplankton growth in bioassay experiments performed at two sites in the tidal freshwater James River during 2012. Effect size is the natural-log transformed ratio of phytoplankton growth (rPOC) under nutrient-enriched (+PN) and ambient nutrient (Control) concentrations at three light levels. ANCOVA was used to test for light limitation, nutrient limitation (+PN treatment) and their interactive effect (L*N) on algal growth (as POC). ns denotes non-significant p values (>0.05)...... 16 Table 1-2. Statistical analysis comparing C-based phytoplankton growth rates (rPOC) in bioassays

receiving single nutrient (PO4, NH4, NO3, Urea) and combined nutrient (+PN = PO4, NH4, NO3) additions relative to Controls (ambient nutrients). Experiments were performed monthly in 2012 at a near-shore site located in the tidal fresh James River (ns denotes non-significant p values >0.05). .. 17 Figure 1-1. Map of the James River Estuary showing the CHLa maximum in the tidal freshwater segment and the locations of main channel (JMS75) and near-shore (Rice Pier) sampling sites for bioassay experiments. Annual average CHLa concentrations are for 2005-2010 based on monthly measurements by the Virginia Department of Environmental Quality for the Chesapeake Bay Program...... 18 Figure 1-2. Seasonal variation in temperature, Freshwater Replacement Time (FRT), CHLa, POC,

TN, DIN, TP and PO4 and DIN:TP (as molar) at station JMS75 located in the CHLa maximum of the James River Estuary...... 19 Figure 1-3. Phytoplankton responses to light and nutrient amendments (as POC ±SE) during monthly bioassay experiments performed at a near shore site in the tidal fresh James River. Dashed

iii

lines represent initial POC levels, hollow symbols represent final concentrations under ambient nutrient conditions (Control) and dark symbols represent final concentrations under nutrient enriched conditions...... 20 Figure 1-4. Mean phytoplankton growth rates (as C; ±SE) among experimental bioassays receiving various forms of N addition (top) and additions of N, P and PN combined (bottom). Data from Controls (ambient nutrients) are shown in both the upper and lower panels for comparison to treatments...... 21 Figure 1-5. Relationships between DIN uptake and PON production observed in monthly bioassay experiments in the James River Estuary. Data points are means of three replicates (±SE); slopes (±SE) and R2 derived by Model II regression...... 22 Figure 1-6. C:N of biomass production under varying light and nutrient conditions in bioassay experiments with phytoplankton from the tidal freshwater James River (data pooled across experiments performed monthly during May-October 2012). Error bars are standard error...... 23 Figure 1-7. (Top) C-based phytoplankton growth rates under ambient (Control) and nutrient-enriched (+PN) conditions. (Bottom) The severity of nutrient limitation (ratio of ambient to nutrient enriched growth rates) and freshwater replacement time (FRT)...... 24

Figure 1-8. Long-term trends in DIN (top panel), PO4 (middle panel) and CHLa (bottom panel) in the tidal freshwater segment of the James River Estuary. Nutrient data are monthly values for June- September at two stations: a low-CHLa site in the upper, constricted section (JMS99) and a high- CHLa site in the broad, shallow section (JMS75). Lines depict summer-average values for each year. CHLa data are year-round monthly values at JMS75 (all data from VA DEQ Chesapeake Bay Monitoring Program)...... 25 Chapter 2: Inter-specific differences in feeding habits and exposure to the cyanotoxin Microcystin among fish and shellfish from the James River Estuary, Virginia...... 26 ABSTRACT ...... 26 INTRODUCTION ...... 27 METHODS ...... 28 Site Description ...... 28 Sample Collection ...... 29 Sample Analysis ...... 30 Consumer Diets & Feeding ...... 31 Rangia Toxicity Assays ...... 33 RESULTS ...... 33 Cyanobacteria ...... 33 MC in Fish and Shellfish ...... 34 Toxicity Assays ...... 35

DISCUSSION ...... 36 Tables and Figures ...... 42 Table 2-1. Extraction efficiencies from spiked samples observed in this and previously published studies using the ELISA Microcystin assay...... 42 Table 2-2. Body size and average Microcystin concentrations in tissues of fish and shellfish from the tidal freshwater James River Estuary. Also shown is the percentage of individuals with measurable Microcystin levels in muscle and liver/viscera and CHLa concentrations in stomach contents of fishes...... 43 Table 2-3. Biomass-specific clearance rates of CHLa by the wedge clam (Rangia cuneata) incubated at ambient (river in situ) and standard temperatures using clams and seston obtained from the tidal freshwater James River...... 44 Table 2-4. CHLa and Microcystin concentrations in suspended and sedimented particulate matter from the James River Estuary (data normalized to dry mass ...... 45 Figure 2-1. Seasonal patterns in phycocyanin fluorescence, Microcystin, and genes specific to Microcystis and Mycrocystin (mcyD) in the tidal freshwater segment of the James River Estuary during 2012...... 46 Figure 2-2. Seasonal variation in tissue concentrations of Microcystin in fish and shellfish from the tidal freshwater segment of the James River Estuary during May 2012 to March 2013 (mean ± SE; some error bars not visible)...... 47 Figure 2-3. Seasonal variation in feeding rates and Microcystin content in tissues of the common wedge clam (Rangia cuneata) in the tidal freshwater segment of the James River Estuary during May 2012 to March 2013 (mean ± SE; some error bars not visible)...... 48 Figure 2-4. Microcystin content in liver tissues of fish and shellfish in relation to algal contributions to their diet as indicated by CHLa in gut contents and stable C isotopic signatures of consumer tissues. Data points are mean and SE based on monthly collections during May-October 2012. Microcystin concentrations for the wedge clam (Rangia) are based on analysis of viscera. Regression lines are based on Model II regression analysis...... 49 Figure 2-5 Rangia clearance rates at controlled Microcystin concentrations...... 50 Figure 2-7 Relationships between water and tissue concentrations of the cyanotoxin Microcystin in fish and shellfish from the James River Estuary ...... 52 Chapter 3: Top Down Controls, Trophic pathways and the fate of algal production in the tidal freshwater James River ...... 53 ABSTRACT ...... 53 INTRODUCTION ...... 54 METHODS ...... 57 Study Site and Data Collection...... 57 Mesocosm Experiments ...... 62

RESULTS ...... 65 CHLa and PON fluxes ...... 65 Stable Isotope Analysis ...... 66 Mesocosm Experiments ...... 67 DISCUSSION ...... 70 Tables and Figures ...... 76 Table 3-1 Data sources used for analysis of top-down effects in comparison to other CHLa and PON fluxes...... 76 Table 3-2 Isotopic signatures, trophic level, biomass and phytoplankton support for common consumers of the tidal fresh James River ...... 77 Table 3-3 Mesocosm results for water quality parameters from June 2013. Nutrient, Microcystin and MYC (Microcystis specific genes) values represent single values observed at the end of the experiments while CHLa, POC and PON represent values (mean ± SE) from combined 3 day intervals (n = 5). Zooplankton estimates represent average and standard error values for monitoring which occurred on day 6 and day 15...... 78 Table 3-4 Mesocosm results for water quality parameters from July 2013. Nutrient, Microcystin and MYC (Microcystis specific genes) values represent single values observed at the end of the experiments while CHLa, POC and PON represent values (mean ± SE) from combined 3 day intervals (n = 5). Zooplankton estimates represent average and standard error values for monitoring which occurred on day 6 and day 15...... 79 Table 3-5 ANCOVA results for mesocosm experiments performed in June and July of 2013...... 80 Figure 3-1: Daily ecosystem respiration as a function of net primary production in the James River Estuary during 2012- 2013. The Y intercept of the regression line is interpreted as the amount of respiration supported by allochthonous carbon (del Giorgio and Peters 1994)...... 81 Figure 3-2 Fish as consumers of CHLa in the Tidal Freshwater James River. Data shown are (top) CHLa in fish gut contents, (2nd) per capita CHLa ingestion for a range of gut turnover rates, (3rd) fish biomass and density and (bottom) population-scale estimates of CHLa ingestion...... 82 Figure 3-3. Fish as recyclers of N in the Tidal Freshwater James River. Data shown are (top) PON in fish gut contents, (2nd) per capita PON ingestion for a range of potential gut turnover rates, (3rd) fish abundance and (bottom) population-scale estimates of PON ingestion...... 83 Figure 3-4. Abundance, per capita clearance rates and CHLa and PON ingestion by dominant zooplankton and benthic filter-feeders (Rangia) of the tidal freshwater James River. Zooplankton abundances are based on bi-monthly collections during March-September 2013. Rangia abundance is the average for 2001-2010 based on annual surveys by CBP...... 84 Figure 3-5. (Top) CHLa production in the Tidal Fresh James River compared to losses via respiration, grazing and advection (export). (Bottom) Phytoplankton demand for DIN compared to inputs via respiration (bacterial re-mineralization), consumer-mediated N cycling (grazing), and

advection (external inputs). Data shown are average daily values for March-November of 2012 and 2013. Grazing losses include contributions by zooplankton, benthic filter-feeders and fish...... 85 Figure 3-6 Relationships between muscle δ13C ‰ and CHLa contribution to diet (top) and δ13C ‰ to δ 15N‰. Dashed lines represent trophic enrichment levels from basal resources of autochthonus (seston samples during high CHLa and low TSS concentrations) and allochthonous forms of organic matter (Seston samples during low CHLa and high TSS concentrations)...... 86 Figure 3-7 Correlation matrices using principal components analyses of independent variables for mesocosm experiment performed in June and July 2013 ...... 87 Figure 3- 8 Chlorophyll, Nutrient, Zooplankton and Microcystin/Microcystis gene copies results for various treatment and nutrient loading rates in June Mesocosm experiments. Chlorophyll and zooplankton results represent averages across the course of the experiment while nutrient, Microcystin and Microcystis results represent final values...... 88 Figure 3- 9 Chlorophyll, Nutrient, Zooplankton and Microcystin/Microcystis gene copies results for various treatment and nutrient loading rates in July mesocosm experiments. Chlorophyll and zooplankton results represent averages across the course of the experiment while nutrient, Microcystis and Microcystin results represent final values...... 89 Figure 3-10 Nutrient effect sizes calculated as the natural logarithm of nutrient enriched: control CHLa concentrations (top) and DIN uptake rates (bottom) for nutrient limitation bioassays using algal cultures from mesocosms with variable grazer treatments and nutrient loading rates...... 90 Synthesis ...... 92 Literature Cited ...... 93

vii

ACKNOWLEDGEMENTS The author wishes to thank many different people for their various contributions to this project. First I’d like to acknowledge the tremendous amount of time and support from my primary adviser Dr. Paul Bukaveckas on this project. I am grateful to the many contributors to field and lab work which include: David Elliott (zooplankton characterization), KC Filippino (Urea analyses), Dominique Grimm (zooplankton characterization), Dave Hopler (boat captain and fish expertise), Will Isenberg (water quality monitoring), William Lee (nutrient analyses), Aaron Porter (Genetics, Microcystis & C&N Analyses), Sarah Sinclair-Govoni (Bioassays), Spencer Tassone (Mesoscosms, Rangia toxicity) and Ryan Weaver (Bioassays, Fish Disection, Rangia Grazing). I’d also like to thank Jameson Brunkow and Chuck Frederickson of James River Association for taking us out to sample Rangia and sediment. I’d like to thank Anne Schlegel and Arthur Butt of the Virginia Department of Environmental Quality for preliminary review of our reports and data analyses. I’d also like to thank the VCU Rice Center, the Virginia Department of Environmental quality for funding the majority of this research. Finally, I would like to thank my wife, Elizabeth and my parents, for their love and support throughout this project.

viii

PREFACE This dissertation includes 3 chapters which focus on algal bloom of the tidal fresh James River. The first chapter describes nutrient and light limitation assays performed on algal cultures and draw conclusions about long-term patterns in nutrient limitation by comparing results with a previous study . This chapter also describes the influence of riverine discharge upon nutrient limitation in a point-source dominated estuary. This chapter was published in Estuaries and Coasts (Wood and Bukaveckas 2014).

The second chapter presents the first comprehensive assessment of the occurrence of the cyanotoxin Microcystin in water and biota of the James River. Data presented in this chapter show that bivalve grazing declines in the presence of Microcystin in the water. The chapter also describes feeding habits in fish as a predictor for inter-specific differences in Microcystin accumulation in their tissues. The work presented in this chapter was published in Environmental Science & Technology (Wood et al. 2014).

The third chapter describes the fate of algal carbon in the James River Estuary and the importance of autochthonous and allochthonous sources of organic matter in supporting production of higher trophic levels. Here I draw upon ecosystem metabolism data (NPP and R), abundance and grazing estimates for primary consumers and estimates of advective losses of chlorophyll and external inputs of nitrogen to place ‘top-down’ effects in the broader context of factors influencing chlorophyll and nitrogen fluxes in the James. . This chapter also describes results from mesocosm experiments used to assess the influences of grazers on chlorophyll, nutrients and Microcystis. This work will be submitted in the summer of 2014 to the journal Ecosystems.

Chapter 1: Increasing severity of phytoplankton nutrient limitation following reductions in point source inputs to the tidal freshwater segment of the James River Estuary.

ABSTRACT We conducted monthly bioassay experiments to characterize light and nutrient use efficiency of phytoplankton communities from the chlorophyll-a maximum located in the tidal freshwater region of the James River Estuary. Bioassay results were interpreted in the context of seasonal and inter-annual variation in nutrient delivery and biomass yield using recent and long- term data. Bioassay experiments suggest that nutrient limitation of phytoplankton production has increased over the past 20 years coinciding with reductions in point source inputs and estuarine dissolved nutrient concentrations. Despite increasing nutrient stress, chlorophyll concentrations have not declined due to more efficient nutrient usage. Greater CHLa yield (per unit of N and P) may be due to feedback mechanisms by which the presence of toxin-producing cyanobacteria inhibits grazing by benthic and pelagic filter-feeders. Seasonal patterns in nutrient limitation indicate that phytoplankton in the James respond to variations in inflow concentrations of dissolved nutrients. This association gives rise to an atypical pattern whereby the severity of nutrient limitation diminishes with low discharge in late summer due to minimal dilution of local point sources inputs by riverine discharge. We suggest that this may be a common feature of estuaries located in proximity to urbanized areas.

INTRODUCTION Effective nutrient management strategies for eutrophic waterbodies require an understanding of the relationship between nutrient loading and algal production (Cloern 2001, Wagner et al. 2011). Quantitative depictions of the nutrient load-algal biomass relationship are the basis for determining nutrient allocations (caps) to protect designated uses (swimability, fishability, etc.) against impairments arising from eutrophication (Cerco and Cole 1993, Havens and Schelske 2001, Borah et al. 2006, Carstensen et al. 2011). The complexities which influence the relationship between nutrient supply and phytoplankton production pose a challenge to the implementation of this approach. This relationship is influenced by physical forces such as variations in temperature, water residence time and underwater light availability (Sterner et al. 1997, Sellers and Bukaveckas 2003, Borsuk et al. 2004, Lucas et al. 2009, Ochs et al. 2013), as well as biotic processes such as grazing and nutrient re-generation (Bronk et al. 1998, Vanni et al. 2006, Hall et al. 2007). Limiting resources such as nutrients and light vary spatially and temporally, as do constraints on the ability of phytoplankton to exploit these resources (Monbet 1992, Bukaveckas et al. 2011a). In estuaries, discharge is an important variable that influences both nutrient supply and demand; the former because loading rates are related to riverine inputs, and the latter because advective losses (washout) affect biomass accumulation (Rudek et al. 1991, Borsuk et al. 2004, Murrell et al. 2007, Lucas et al. 2009). The effects of discharge on biomass yield may be viewed as a constraint on resource use efficiency whereby under high discharge conditions, short water residence time limits opportunities for phytoplankton to convert dissolved inorganic nutrients into algal biomass. Our limited knowledge of the interactive effects of light, nutrient and residence time conditions leads to uncertainty in establishing nutrient reduction targets for preventing harmful effects from algal blooms (Strayer et al. 2008, Bukaveckas et al. 2011a, Kratina et al. 2012).

Nutrient limitation arises from the imbalance between supply and demand; this imbalance affects both the severity and form (e.g., N vs. P) of limitation (Conley et al. 2009, Ptacnik et al. 2010). External nutrient supply is dictated by hydrologic loading (the ratio of watershed runoff to the surface area of the receiving water body) as well as anthropogenic activities in the watershed which determine concentrations in runoff (Lewis et al. 1999; Baron et al. 2013). Land use practices affect not only the quantity but also the relative ratios and forms of N and P. Nitrogen is generally thought to be limiting in coastal waters (Howarth and Marino 2006),

2 though co-limitation by N and P is commonly reported (Elser et al. 2007). Forms of N differ in their bioavailability - uptake of dissolved inorganic fractions (e.g., nitrate and ammonia) is well- known, whereas utilization of dissolved organic nitrogen has only recently been appreciated (Mulholland et al. 2009, Bradley et al. 2010, Filippino et al. 2011). Algal nutrient demand is a product of their growth rates and stoichiometry (C:N:P) both of which are affected by light conditions (Sterner et al. 1997, Hall et al. 2004, Brauer et al. 2012). At saturating light intensities, nutrients are more efficiently converted into algal biomass and it has also been shown that production per unit of N or P (e.g., CHLa:P or C:P of particulate matter) is greater under more favorable light conditions (Sterner et al. 1997, Mette et al. 2012)). Underwater irradiance is determined by incident solar radiation, light attenuation and the depth of the mixed layer, which in well-mixed estuaries is the overall depth.

The presence and severity of nutrient limitation is often inferred by comparing nutrient ratios in the environment (e.g., DIN:TP) to Redfield values (e.g., Ptacnik et al. 2010). This approach relies on the assumption that measured nutrient concentrations reflect availability which may be problematic given differences in lability among various nutrient fractions (Beardall 2001). Bioassay experiments allow direct measurement of nutrient enrichment on uptake rates, growth rates and stoichiometry (Tamminen and Andersen 2007, Ren et al. 2009). This approach has long been used to measure nutrient limitation and quantify relationships between nutrient supply and algal biomass yield in diverse settings such as lakes, rivers and estuaries (Lean and Pick 1981, Elser et al. 1988). In addition to providing qualitative information on which nutrients are limiting, they can be used to measure the severity of nutrient limitation based on the ratio of phytoplankton growth rates at ambient vs. enriched nutrient levels (Tamminen and Andersen 2007). A limitation of this approach is that bioassays isolate phytoplankton from external and certain regenerated sources of nutrients (e.g., sediment nutrient fluxes) such that nutrient deficiency may be artificially enhanced during the experiment. To minimize this, as well as “bottle effects” arising from colonization of surfaces, bioassays are typically run as dilution experiments and over short intervals (e.g., 48-72 h). Being of small scale, these experiments can be replicated to characterize spatial and temporal variation in nutrient limitation (e.g., Fisher et al 1999) and can be used to assess the effects of light conditions on nutrient use efficiency (e.g., Koch et al. 2004, Whalen and Benson 2007). A third, perhaps under-utilized, value to bioassay experiments is in establishing algal stoichiometric

3 properties to assess growth efficiency (i.e., CHLa or C yield per unit of N and P) under varying nutrient and light conditions (Gowen et al. 1992, Edwards et al. 2003).

In this study, we examine seasonal patterns in nutrient and light limitation of phytoplankton communities at an estuarine CHLa maximum. A region of elevated CHLa extends over 40 km within the tidal freshwater segment of the James River Estuary where annual average CHLa concentrations are among the highest in the Chesapeake Bay region (Figure 1-1). Longitudinal CHLa maxima have been reported in other estuaries, and in some cases, attributed to hydrodynamic retention whereby particulate matter in surface (seaward) currents is entrained in deeper, landward currents (North and Houde 2001, 2003). This is not the case in James where elevated CHLa is attributed to high growth rates (positive water column NPP) in the region where the channel transitions from a riverine to an estuarine morphometry (Bukaveckas et al. 2011b). It is hypothesized that shallow conditions release phytoplankton from light limitation and allow for greater utilization of nutrients from the upper watershed and local point source inputs. A recent mass balance analyses supports this view by showing high rates of inorganic nutrient assimilation within the CHLa maximum (Bukaveckas and Isenberg 2013). In this paper, we present results from bioassay experiments in which light and nutrient conditions were manipulated to determine the form and severity of resource constraints on phytoplankton growth in the CHLa maximum. We also analyze a 3-year time series of weekly CHLa measurements to make inferences about the relative importance of nutrient availability and water residence time in influencing seasonal patterns of nutrient limitation. Lastly, we consider whether the severity of nutrient limitation has changed in response to nutrient load reductions during the past 20 years by comparing our results to bioassay experiments previously performed at this site (Fisher et al. 1999).

METHODS

Site Description The James River Estuary is the southern-most of the five major sub-estuaries of Chesapeake Bay. Its principal tributary, the James River, is the third largest tributary of Chesapeake Bay by discharge and nutrient load. The tidal fresh segment of the estuary (salinity < 0.5) extends 115 km from the Fall Line (at Richmond, VA) to the confluence with the Chickahominy River. This segment experiences a large tidal prism (~60 cm) relative to average

4 depth (~3.0 m) creating a vertically and laterally well-mixed system. Nutrient loads to this segment are due to its small surface area (52 km2), large contributing area (watershed = 26,165 km2) and direct point source inputs from the Richmond metropolitan area (Bukaveckas and Isenberg 2013). P loads are principally (~80%) from riverine inputs which are transported in particulate form and trapped during high discharge events. For N, watershed and local point sources contribute approximately equally, though the latter dominate with respect to dissolved inorganic fractions and during low discharge periods. Photic depths are typically ~1 m and are relatively uniform throughout the tidal fresh segment (Bukaveckas et al. 2011b). Despite this, there are large differences in light conditions between the upper, constricted segment, where the deeper channel (>3 m) results in low average underwater irradiance, and the broader channel of the lower segment, where shallow depths (< 2 m) result in greater light availability. Water residence time in summer ranges from 5 to 25 d (mean = 15.6 d for May-October 2012). Estimates presented here are based on the date-specific freshwater replacement time (FRT) method (Alber and Sheldon 1999) using river discharge data from USGS gauges located near the Fall Line on the James (#2037500) and Appomattox Rivers (#2041650).

Monitoring Data We conducted ~weekly monitoring of CHLa and nutrient concentrations during July 2010 to December 2012 at a station located within the CHLa maximum (Chesapeake Bay Program designation: JMS75; Figure 1-1). The station is located 55 km below the Fall Line in the wide, shallow portion of the tidal fresh segment. Water samples were collected at a depth of 1 m for analysis of particulate matter (CHLa, TSS, POC, PON) and nutrient concentrations (TN,

NH4, NO3. TP, PO4). Samples for CHLa, TSS, POC and PON were filtered through Whatman GF/A glass filters (0.5 μm nominal pore size). Filters for CHLa analyses were extracted for 18 h in buffered acetone and analyzed on a Turner Design TD-700 Fluorometer (Arar and Collins 1997). TSS was determined gravimetrically using pre-weighed, pre-combusted filters. Filters for POC and PON analysis were dried at 60 C for 48 hours, fumed with HCl to remove inorganic carbon and analyzed on a Perkin–Elmer CHN analyzer. Concentrations of total nitrogen (TN), nitrate (NO3) ammonium (NH4), total phosphorus (TP) and phosphate (PO4) were determined using a Skalar segmented flow analyzer using standard methods (APHA 1992). Urea concentrations (bioassays only) were measured on an Astoria Pacific autoanalyzer using the colorimetric monoxime method of Price and Harrison (1987).

Bioassay experiments Experiments were performed monthly from May to October 2012 using water obtained from two locations within the CHLa maximum: a main channel site (JMS75) and the nearby Research Pier at the VCU Rice Center (Figure 1-1). The two sites were included in the design to test for differences in the severity of nutrient limitation between near-shore and main channel habitats. Water from these sites was obtained in conjunction with the weekly monitoring program and returned immediately to the lab. Bioassays comprised a 150 mL solution in a 250 mL Erlenmeyer flask containing 50% raw water and 50% filtered water (0.5 μm Whatman GF/A glass filter). Bioassays were diluted in order to reduce algal densities below equilibrium to measure algal growth responses (Sterner and Grover 1998). Six nutrient treatments were performed (Control, +NH4, +NO3, + Urea, +PO4, +PN) using water obtained from the near-shore site; only the Control (no nutrients added) and +PN treatments were performed at the main channel site. Each treatment was replicated three times. Enrichments entailed the addition of -1 -1 0.125 mg L of NO3, NH4, or Urea and 0.1 mg L .of PO4. Combined treatments (+PN) received -1 -1 0.125 mg L each of NO3 and NH4 and 0.1 mg PO4 L . Nutrient additions approximately doubled ambient concentrations increasing DIN from 0.10-0.15 mg L-1 to 0.25-0.3 mg L-1, Urea -1 -1 -1 -1 from 0.10-0.20 mg L to 0.20-0.30 mg L , and PO4 from 0.05-0.10 mg L to 0.15-0.20 mg L .

Bioassays were incubated on a shaker table at ambient (river) temperature inside a Conviron growth chamber for 48 h. Control and +PN treatments were incubated at three light levels (3, 6 and 12 E m-2 d-1) to assess light effects on phytoplankton growth, nutrient uptake and stoichiometry. These values represent the average daily irradiance experienced by phytoplankton circulating through the entire water column over depths ranging to 1, 2 and 4 m taking into account typical summer solar radiation (~40 E m-2 d-1; Fisher et al. 2003) and -1 underwater light attenuation (mean kd = 3.14 ± 0.33 m ) measured monthly in conjunction with bioassay experiments (Gosselain et al. 1994). The lowest light level represented ambient conditions in the upper, constricted section of the tidal fresh segment (e.g., near station JMS99; Figure 1-1). The higher light levels represented ambient conditions in the broad, shallow reach near JMS75. Light conditions within the incubator were modified by shade cloth and proximity to light sources and verified with a Li-Cor photometer.

Initial and final concentrations of CHLa, POC and nutrients were determined using the same analytical methods as for monitoring samples. Phytoplankton growth rates (r) were

6 calculated as the slope of the natural logarithms of POC as a function of time. Some other studies have used CHLa to calculate growth rates, however in preliminary experiments we observed changes in POC:CHLa during incubation in response to varying light exposure. POC- based growth rates were used to calculate effect sizes as the natural logarithm of treatment:control (Koch et al. 2004). An analysis of covariance (ANCOVA) was used to test for interactions between light and nutrient effects. Light saturation effects were assessed by comparing the fit of linear and non-linear (log and tangential) models of phytoplankton growth as a function of irradiance for individual experiments. Nitrogen use efficiency was assessed based PON production per unit of DIN uptake and from changes in the ratio of C:N in particulate matter. The relationship between DIN uptake and PON production was analyzed using Model II regression analysis since both parameters are measured with error. Statistical analyses were performed using Microsoft Excel and JMP Pro 10.

RESULTS

Seasonal and Inter-annual Variation in CHLa and Nutrients A 3-year time series of weekly monitoring at JMS75 is presented to place the 2012 bioassay experiments in the broader context of seasonal and inter-annual variation in phytoplankton abundance and nutrient availability. Data from this station located in the CHLa maximum showed well-defined bloom periods in each year corresponding to seasonal patterns in water temperature and residence time (Figure 1-2). The periods of elevated CHLa (>20 µg L-1) were associated with water temperature >15oC and freshwater replacement time >10 d whereas intervening periods (~November-April) were characterized by low CHLa (<10 µg L-1), low water temperature, and short FRT. Periods with elevated CHLa corresponded to peaks in TP (>0.10 mg L-1) and low DIN (<0.15 mg L-1). This resulted in low DIN:TP ratios during summer months (molar ratio <5). Declines in CHLa during Fall were often associated with storm events which had a pronounced flushing effect (e.g. Tropical Storm Lee, September 2011). Overall, CHLa was positively correlated with FRT in the 3-year time series (N = 100, R2 = 0.47, p < 0.001). Inter-annual differences in July-August means for CHLa were large (>2-fold) relative to inter-annual differences in TP and TN (CV = 6% and 13%, respectively). This resulted in a 3- fold range of variation in CHLa yield relative to TP (310-990 wt:wt) and TN (50-115 wt:wt). POC concentrations tracked seasonal patterns in CHLa, but like TP and TN, exhibited less inter-

7 annual variability during bloom periods (CV = 12%). C:N ratios of particulate matter were similar during July-August in all three years (means = 6.1-6.3) and comparable to the Redfield ratio (6.6; Redfield 1958).

Bioassay Experiments Phytoplankton exhibited statistically significant positive responses to light in 11 of 12 experiments, and to nutrient additions in 11 of 12 experiments performed at the two sites (Table 1-1). Significant interaction effects were detected for 3 of 6 experiments from each site in which their was a synergistic effect from the combination of enhanced light and nutrient enrichment. No significant differences in phytoplankton growth rates were observed between the main channel and near-shore sites in either Control or nutrient-enriched bioassays. Nutrient enrichment effects on growth rates were observed throughout the range of light intensities but larger responses were typically observed at the highest light intensity (Figure 1-3). Average effect sizes (natural-log transformed ratios of nutrient-enriched to Control growth rates) were 0.26 ± 0.06, 0.31 ± 0.07 and 0.41 ± 0.07 at 3, 6, and 12 E m-2 d-1, respectively. We compared the fit of linear and non-linear models relating phytoplankton growth (as POC) to irradiance and found that the non-linear (saturating) model provided a better fit in 3 of the 6 experiments at each site (June, July and August; Figure 1-3).

Forms of nutrient limitation differed among the monthly experiments (Figure 1-4, Table 1-2). The combined P and N addition resulted in significantly higher growth rates relative to controls in 11 of 12 experiments (excluding September, Main Channel site). Higher growth rates in response to P addition were not observed in any of the 6 experiments (performed at near-shore site only). Interpretation of N effects was somewhat dependent on the form of N tested. In June, all three forms of N (NO3, NH4 and Urea) resulted in significantly higher growth rates relative to Controls. Growth rates were not significantly different among the three N treatments indicating that phytoplankton were capable of exploiting all three forms of N. In August, additions of NH4 and Urea stimulated growth rates relative to Controls, whereas NO3 did not. In September, bioassays receiving NO3, exhibited significantly higher growth rates relative to Controls and to those receiving NH4 and Urea. Overall, these findings suggest that phytoplankton in the tidal freshwater segment of the James were consistently stimulated by the combined addition of N and P, and in a few cases responded to the addition of N alone.

Initial and final concentrations of DIN, PON and POC in the bioassays were used to derive two metrics of N use efficiency: one which considered the production of particulate N in relation to DIN assimilation, and a second, particulate C production per unit of particulate N production. In enriched bioassays, maximal rates of DIN uptake were ~0.15-0.20 mg L-1 d-1 such that DIN pools were depleted to ~20% of initial (starting) concentrations during the 48 h incubations.(Figure 1-5). Rates of DIN assimilation were significantly correlated with PON production in each of the monthly experiments (data pooled for both sites; R2 = 0.51 to 0.82). The slope of this relationship is an indicator of N use efficiency as it represents the proportion of assimilated N which is converted to particulate N. Highest N use efficiency was observed in June (32%) and August (29%) corresponding to the months when responses to N addition alone were observed. In other months, N use efficiency was lower (10-19%) with the exception of October (32%). Overall, the low proportion of N retained in the particulate fraction indicates that the bulk of assimilated N entered the DON pool. We analyzed variation in algal biomass yield (as C) per unit of N by comparing C:N ratios among bioassays under various light and nutrient conditions (Figure 1-6). C:N ratios increased by 2-fold in response to higher light levels in both Control and nutrient-enriched treatments. Highest C:N ratios (~12) were observed at high light levels (12 E m-2 d-1) and ambient nutrient concentrations. At low light levels (3 E m-2 d-1) C:N ratios were not significantly different between Control and +PN treatments, and were similar to Redfield (~6).

Growth rates at ambient nutrient concentrations ranged from 0.1 to 0.4 d-1 (mean = 0.23 d-1) and corresponded to an average doubling time of 3 d (Figure 1-7). Nutrient-saturated growth rates were higher (range = 0.2 to 0.6 d-1; mean = 0.44 d-1) and corresponded to an average doubling time of 1.6 d. Incubation temperature ranged from 19 to 30 Co and explained less than 10% of the variation in growth rates. Strong responses to nutrient enrichment were observed in May and June when phytoplankton growth rates at ambient nutrient concentrations were ~25% of nutrient-enriched growth rates. Weaker responses to nutrient enrichment were measured during August-September when growth rates at ambient nutrient concentrations were ~75% of nutrient-enriched growth rates. Seasonal patterns in the severity of nutrient limitation followed trends freshwater replacement time. Greater severity of nutrient limitation was associated with short FRT in May and June (5-10 d) with weaker responses to nutrient limitation occurring during periods of longer FRT (15-20 d) in late summer. To assess changes in the nutrient status

9 of the James, we compared current nutrient conditions with historical data from the CBP monthly monitoring. At the site where bioassay experiments were performed (JMS75), average summer values of DIN declined from 0.45 mg L-1 (1990-1996) to 0.25 mg L-1 (1997-2012) while -1 -1 PO4 declined from 0.022 mg L to 0.013 mg L during the same period (Figure 1-8). The incidence of very low DIN concentrations (< 0.10 mg L-1) has increased in recent years (e.g., >50% of summer monthly measurements during 2007-2012). Declining nutrient concentrations in the region of the CHLa maximum have led to the development of pronounced longitudinal gradients in nutrient availability as indicated by differences in concentration between stations JMS99 and JMS75 (Figure 1-8). Largest differences were observed during 2000-2010 with the exception of 2003, a year with high summer discharge and low CHLa.

DISCUSSION Our most striking finding was the prevalence of nutrient limitation (observed in 11 of 12 experiments), given that a prior study concluded that phytoplankton in the tidal fresh James were exclusively light limited (Fisher et al. 1999). The prior study was conducted at the same station (JMS75) and included monthly experiments performed during the same time of year (May- October 1993) showing no detectable response to nutrient addition. A number of methodological differences complicate direct comparisons including use of different response variables (CHLa vs. POC), incubation length (6-8 d vs. 48 h) and light exposure (8-30 vs. 3-12 E m-2 d-1; Fisher et al. 1999, this study; respectively). Two of these differences (our use of lower irradiances and shorter incubations) would be expected to diminish the likelihood of observing nutrient -1 limitation. Also, our nutrient additions were smaller (DIN = 0.125-0.250 mg L , PO4 = 0.1 mg -1 -1 -1 L ) than those used by Fisher et al. (DIN = 0.35 mg L , PO4 = 0.155 mg L ). Therefore it is unlikely that differences in methodology biased our results in favor of finding nutrient limitation.

The bioassay results suggest that the severity of nutrient limitation has increased during the 20-year interim between the two studies. Two potential mechanisms could account for this: an increase in water clarity and/or a reduction in nutrient availability. An analysis of diffuse attenuation coefficients (kd) measured monthly at this site during May-October of 1994-2010 revealed no long-term trends in water clarity (N = 93, R2 < 0.10, p = 0.46; data from VA DEQ Chesapeake Bay Program). Average attenuation values measured in conjunction with our

-1 bioassay experiments (kd = 3.14 ± 0.33 m ) were similar to the long-term average at this location -1 (kd = 3.49 ± 0.10 m ). In contrast, summertime DIN and PO4 concentrations have declined at the site where bioassay experiments were performed with present values being less than half of those measured during the earlier bioassay study. Interpretation of the long-term nutrient data is complicated by changes in analytical methodology after 1994 (Marshall et al. 2009). However, our assessment of nutrient availability is based on a comparison between two sites showing that nutrient concentrations at the station where bioassay experiments were performed (JMS75) have declined relative to an upstream station (JMS99). Concentration differences between the two sites became apparent after 2000. We attribute the lower nutrient concentrations at JMS75 to high rates of assimilatory uptake in the region of the CHLa maximum. The effects of assimilatory uptake in depleting nutrients have likely been enhanced by reductions in point source inputs. Our recent mass balance analyses showed that point source inputs to this segment of the James have declined by one-third (TN) and one-half (TP) since the early 1990’s (Bukaveckas and Isenberg 2013). Together, the monitoring and bioassay data suggest that reductions in nutrient loads have fostered a shift toward greater nutrient limitation of phytoplankton while nutrient concentrations have declined to levels not previously seen in the 25-year record.

Despite the increasing severity of nutrient limitation, CHLa levels in the James were unchanged during this period (1990-present; Figure 1-8). CHLa yields per unit of N and P are higher now (2010-2012 - CHLa:TP = 310-990, CHLa:TN = 50-115) in comparison to data from throughout the 1990’s (CHLa:TP ~200, CHLa:TN ~30) indicating that increased nutrient use efficiency has compensated for declines in nutrient availability. Carstensen et al. (2011) have similarly reported a doubling of the CHLa:TN ratio in recent decades for estuaries from various regions, including the saline reaches of Chesapeake Bay. They attribute this trend to factors which include rising temperature and CO2 as well as potential reductions in grazing due to increasing abundance of cyanobacteria and other harmful algae. Our results show that this phenomenon extends to the tidal freshwater reaches of the estuary and is evident for both TN and TP. Our data show that CHLa yield is higher during periods of long residence time as indicated by a significant positive relationship between CHLa:TN and FRT in our weekly monitoring data (N = 100, R2 = 0.34, p < 0.001). However, there were no significant trends in water inputs at this site, either in the long –term data (1899-2011; p = 0.39) or during the period spanning the

11 bioassay studies (1990-2011, p = 0.50) that would suggest that higher CHLa yield could be attributed to reduced flushing. Higher CHLa yield may be indicative of shifts in phytoplankton community composition favoring species with greater nutrient use efficiency. However, the two major groups contributing to phytoplankton biomass in the James (chlorophytes and diatoms) have dominated throughout this period. The abundance of cyanobacteria has been increasing over the past 20 years (Marshall et al. 2009) though their proportional contribution to biomass remains low (e.g., 5% in 2012; H. Marshall, pers. comm.) Thus there are no shifts among major phytoplankton groups that could be linked to CHLa yield, though within-group replacement by more nutrient-efficient species could account for this trend. Historical changes in fisheries may also play a role through their effects on zooplankton abundance and consumer-mediated nutrient recycling (Havens et al. 2001, Vanni et al. 2006, Caraco et al. 2006).

A second finding arising from our bioassay experiments is that increasing ratios of ambient to nutrient-enriched growth rates indicate that the severity of nutrient limitation declines with decreasing discharge in late summer. As declining discharge is linked to reduced riverine nutrient loads, a weakening of nutrient limitation appears counter-intuitive. Current paradigms on seasonal patterns of nutrient limitation originate from lake studies where the severity of limitation increases during longer residence time in summer due to diminishing external nutrient inputs and progressive depletion of nutrients from the euphotic epilimnion (Schindler 1977, Wetzel 2001). Similar explanations have been invoked for estuaries. For example, a recent study of the New and Neuse River estuaries reported that CHLa increased over a range of short residence times, but at longer residence times, CHLa declined (Peierls et al. 2012). The shift from positive to negative slope in the CHLa-residence time relationship was attributed to biotic processes exerting a greater influence on phytoplankton biomass during long residence time. These included increases in the severity of nutrient limitation as well as higher losses due to grazing.

For estuaries located in proximity to urban areas, local point source inputs are a key factor influencing nutrient availability. In the James, direct point source inputs account for a large proportion of the external nutrient load, particularly for dissolved inorganic fractions during summer (e.g., 93% of NH4 and 75% of NO3 and PO4 ; Bukaveckas and Isenberg 2013). We contend that phytoplankton in the James are responding to changes in the concentration of nutrients in inflow, rather than to loading rates associated with riverine fluxes. N yields from the 12

James watershed are low among east coast rivers (Boyer et al. 2002; Howarth et al. 2006) such that periods of elevated discharge are characterized by inputs of relatively N-poor waters. Nutrient concentrations in point source inputs are orders of magnitude higher, and during low river discharge, these would be subject to smaller dilution effects from river inflow, thereby resulting in higher inflow concentrations. The influence of point sources on estuarine nutrient concentrations was apparent at our upstream sampling station (JMS99) where summer DIN and

PO4 concentrations were consistently higher than those observed at the bioassay site (JMS75; Figure 1-8). We suggest that in late summer, the severity of nutrient limitation declines due to higher nutrient concentrations of inflow and that this may be a common feature among estuaries which receive substantial point source inputs. We cannot discount the possibility that acceleration in the rate of internal nutrient recycling could increase nutrient supply in late summer, though we lack data on seasonal variation in grazing rates and sediment nutrient fluxes to test this hypothesis. An important implication of these findings is that nutrient limitation of phytoplankton in the tidal fresh James is principally determined by local point source nutrient inputs, and the extent to which these are diluted by watershed (riverine) runoff. We suggest that in estuaries where local point sources account for a large fraction of inputs, the expected relationship of increasing nutrient stress during low discharge and long water residence time may be reversed.

Lastly we consider the effects of light and nutrient conditions on phytoplankton stoichiometry and nutrient use efficiency. We derived two metrics of efficiency: the proportion of assimilated DIN which was converted to particulate N, and the C biomass yield per unit of particulate N. The proportion of DIN uptake retained in the particulate fraction was variable but low (10-31%) indicating that 70-90% of DIN was transferred to the DON pool during the 48 h experiment. This finding is consistent with a number of field and laboratory studies reporting that a large fraction of DIN uptake is subsequently released as DON (e.g., 25-40% in Bronk et al. 1994). DON release has been attributed to light and salinity effects, phytoplankton physiological condition, and grazing (Hu and Smith Jr. 1998, Ward & Bronk 2001, Bronk et al. 2010). Transfer of DIN to the DON pool would potentially shunt more N into the microbial food web (Stepanauskas et al. 1999, Seitzinger et al. 2002a, 2002b, Wiegner et al. 2006). A related study of bacterial communities in the James revealed higher cell densities, greater proportion of live cells and shifts in community composition of active taxa in the region of the CHLa maximum

(Franklin et al. 2013). A number of studies have shown that NH4 is the primary source of N for phytoplankton uptake and DON production (e.g., Bronk & Ward 1999; Bradley et al. 2010) though other studies have reported that uptake rates for NO3 exceeded those for NH4 (Parker et al. 2012). In our bioassay experiments we observed similar uptake rates for NH4 (mean = 0.070 -1 -1 -1 -1 ± .001 mg L d ) and NO3 (mean = 0.062 ± .003 mg L d ); urea uptake rates were lower and more variable (mean = 0.045 ± .014 mg L-1 d-1). These results suggest that all three forms of N play a role in supporting phytoplankton production and DON release in the tidal fresh segment of the James.

Carbon production per unit of N varied in response to experimental treatments with two- fold higher C:N ratios observed under high light conditions. This finding is consistent with studies in other freshwater systems showing higher C:P ratios of primary producers under more favorable light conditions (e.g., Sterner et al 1997). Our higher light treatments (6-12 E m-2 d-1) were representative of underwater irradiance in the region where the CHLa maximum occurs and suggest that high phytoplankton production in this zone can be attributed in part to greater biomass yield per unit of N or P. Under these light conditions, C:N was double that of the Redfield ratio – a result which has implications for estimating algal N demand based on C production. In a prior paper we argued that external nitrogen loads could account for only 20% of phytoplankton demand based on measured production and Redfield ratios (Bukaveckas and Isenberg 2013). Using the higher C:N ratios derived empirically from the bioassay experiments we find that external inputs could account for 40% of algal demand, with 60% supported by internal recycling. We observed both linear and non-linear responses to light availability among the monthly bioassay experiments suggesting that phytoplankton are fully or partially released from light limitation at irradiances representative of light conditions in the shallower segment of the estuary (near JMS75). These results are supportive of our earlier hypothesis (Bukaveckas et al. 2011b) that the location of the CHLa maximum is linked to the morphometry of the channel whereby shallow conditions result in greater average water column irradiance and higher nutrient utilization.

In summary, our prior work has shown that the tidal freshwater segment of the James Estuary is a hot spot for phytoplankton production and nutrient retention owing to favorable conditions of light, residence time and nutrient supply. In the present study we report that reductions in point sources inputs have resulted in lower concentrations of dissolved inorganic 14 nutrients and a strengthening of longitudinal nutrient gradients in this segment of the estuary. Seasonal patterns in nutrient limitation suggest that phytoplankton in the James are responsive to local point sources inputs and the extent to which these are diluted by riverine discharge. Reduced riverine inputs during late summer result in a weakening of nutrient stress due to higher inflow concentrations. Increasing nutrient limitation in the James is a promising first step toward oligotrophication. However, the increase in biomass yield, as indicated by higher CHLa:TP and CHLa:TN, has outweighed the effects of declining nutrient availability such that phytoplankton abundance remains unchanged. Achieving reductions in the magnitude and duration of blooms may depend on internal nutrient cycles and their influence on the nutrient load-algal biomass relationship.

Tables and Figures

Table 1-1. Statistical analyses of light and nutrient effects on phytoplankton growth in bioassay experiments performed at two sites in the tidal freshwater James River during 2012. Effect size is the natural-log transformed ratio of phytoplankton growth (rPOC) under nutrient-enriched (+PN) and ambient nutrient (Control) concentrations at three light levels. ANCOVA was used to test for light limitation, nutrient limitation (+PN treatment) and their interactive effect (L*N) on algal growth (as POC). ns denotes non-significant p values (>0.05).

Nutrient Effect Sizes at variable light Light vs. Nutrients (ANCOVA) 3 6 12 Light Nutrients L * N Site Month E m-2 d-1 E m-2 d-1 E m-2 d-1 (p) (p) (p) Near May 0.01 0.05 0.25 0.003 0.045 0.036 Shore June 0.71 0.76 0.93 0.002 <.0001 0.024 July 0.19 0.36 0.30 0.024 0.001 ns August 0.25 0.34 0.41 <.0001 <.0001 0.003 September 0.39 0.33 0.21 0.011 0.001 ns October 0.21 0.16 0.26 0.008 0.003 ns Main May 0.30 0.21 0.40 <.0001 <.0001 ns Channel June 0.53 0.67 0.85 0.003 <.0001 0.008 July 0.42 0.45 0.55 0.017 <.0001 ns August 0.12 0.29 0.17 0.001 0.019 ns September 0.06 0.09 0.13 <.0001 ns ns October -0.02 0.04 0.48 <.0001 0.001 <.0001

Table 1-2. Statistical analysis comparing C-based phytoplankton growth rates (rPOC) in bioassays receiving single nutrient (PO4, NH4, NO3, Urea) and combined nutrient (+PN = PO4, NH4, NO3) additions relative to Controls (ambient nutrients). Experiments were performed monthly in 2012 at a near-shore site located in the tidal fresh James River (ns denotes non-significant p values >0.05).

Nutrient Form growth rate comparisons (Tukey's HSD)

+PO4 +NH4 +NO3 +Urea +PN Month (p) (p) (p) (p) (p) May ns ns ns ns <.0001 June ns <.0001 <.0001 <.0001 <.0001 July ns ns ns ns 0.001 August ns 0.04 ns 0.014 <.0001 September ns ns 0.004 ns 0.019 October ns ns ns ns 0.003

Fall Line

Chickahominy River

VCU Rice Pier

JMS99

Appomattox JMS75 River

Figure 1-1. Map of the James River Estuary showing the CHLa maximum in the tidal freshwater segment and the locations of main channel (JMS75) and near-shore (Rice Pier) sampling sites for bioassay experiments. Annual average CHLa concentrations are for 2005-2010 based on monthly measurements by the Virginia Department of Environmental Quality for the Chesapeake Bay Program.

40 40 Temperature 30 FRT 20 20

10 Freshwater Temperature Temperature (C)

0 0 Replacement (d)Time 1507/13/2010 10/21/2010 1/29/2011 5/9/2011 8/17/2011 11/25/2011 3/4/2012 6/12/2012 9/20/2012 12/29/20126

CHLa )

POC )

1 - 100 4 -

50 2

POC (mg (mg POC L CHLa(µg L

0 0 1.57/13/2010 10/21/2010 1/29/2011 5/9/2011 8/17/2011 11/25/2011 3/4/2012 6/12/2012 9/20/2012 12/29/2012 TN DIN

) 1.0

1 -

0.5 N (mg L N(mg

0.0 0.157/13/2010 10/21/2010 1/29/2011 5/9/2011 8/17/2011 11/25/2011 3/4/2012 6/12/2012 9/20/2012 12/29/2012 TP

POPO44

) 1

- 0.10

P (mg P (mg L 0.05

0.00 10007/13/2010 10/21/2010 1/29/2011 5/9/2011 8/17/2011 11/25/2011 3/4/2012 6/12/2012 9/20/2012 12/29/2012

100

1 DIN:TP DIN:TP (molar)

0 Jul-10 Oct-10 Jan-11 May-11 Aug-11 Nov-11 Mar-12 Jun-12 Sep-12 Dec-12

Figure 1-2. Seasonal variation in temperature, Freshwater Replacement Time (FRT), CHLa, POC, TN,

DIN, TP and PO4 and DIN:TP (as molar) at station JMS75 located in the CHLa maximum of the James River Estuary.

E m-2 d-1 E m-2 d-1 Figure 1-3. Phytoplankton responses to light and nutrient amendments (as POC ±SE) during monthly bioassay experiments performed at a near shore site in the tidal fresh James River. Dashed lines represent initial POC levels, hollow symbols represent final concentrations under ambient nutrient conditions (Control) and dark symbols represent final concentrations under nutrient enriched conditions.

Figure 1-4. Mean phytoplankton growth rates (as C; ±SE) among experimental bioassays receiving various forms of N addition (top) and additions of N, P and PN combined (bottom). Data from Controls (ambient nutrients) are shown in both the upper and lower panels for comparison to treatments.

Figure 1-5. Relationships between DIN uptake and PON production observed in monthly bioassay experiments in the James River Estuary. Data points are means of three replicates (±SE); slopes (±SE) and R2 derived by Model II regression.

Figure 1-6. C:N of biomass production under varying light and nutrient conditions in bioassay experiments with phytoplankton from the tidal freshwater James River (data pooled across experiments performed monthly during May-October 2012). Error bars are standard error.

Figure 1-7. (Top) C-based phytoplankton growth rates under ambient (Control) and nutrient-enriched (+PN) conditions. (Bottom) The severity of nutrient limitation (ratio of ambient to nutrient enriched growth rates) and freshwater replacement time (FRT).

1.2

1.0

0.8

) -1 0.6

L (mgDIN 0.4

0.2

0.0 1989 1992 1995 1998 2001 2004 2007 2010 0.10 JMS99

0.08 JMS75

)

-1 0.06

(mg L (mg

4 0.04 PO 0.02

0.00 1251989 1992 1995 1998 2001 2004 2007 2010

100

) 75 -1

L (µg CHLa 25

0 1990 1993 1996 1999 2002 2005 2008 2011

Chapter 2: Inter-specific differences in feeding habits and exposure to the cyanotoxin Microcystin among fish and shellfish from the James River Estuary, Virginia.

ABSTRACT The cyanotoxin, Microcystin (MC), is known to accumulate in the tissues of diverse aquatic biota although factors influencing exposure, such as feeding habits and seasonal patterns in toxin production, are poorly known. We analyzed seasonal variation in the Microcystin content of primary and secondary consumers, and used dietary analysis (gut contents and stable isotopes) to improve understanding of cyanotoxin transport in food webs. Periods of elevated toxin concentration were associated with peaks in the abundance of genes specific to Microcystis and MC toxin production (mcyD). Peak toxin levels in consumer tissues coincided with peak MC concentrations in seston. However, toxins in tissues persisted in over-wintering populations suggesting that potential health impacts may not be limited to bloom periods. Inter-specific differences in tissue MC concentrations were related to feeding habits and organic matter sources as pelagic fishes ingested a greater proportion of algae in their diet, which resulted in greater MC content in liver and muscle tissues. Sediments contained a greater proportion of allochthonous (terrestrial) organic matter and lower concentrations of MC, resulting in lower toxin concentrations among benthic detritivores. Among shellfish, the benthic suspension feeder Rangia cuneata (wedge clam) showed seasonal avoidance of toxin ingestion due to low feeding rates during periods of elevated MC. Among predators, adult Blue Catfish had low MC concentrations, whereas Blue Crabs exhibited high levels of MC in both muscle and viscera.

INTRODUCTION Harmful algal blooms are a growing worldwide concern (Anderson et al. 2008, O’ Neil et al. 2012 and Michalak et al. 2013). Some harmful algae produce secondary metabolites that act as toxins and therefore pose threats to human health and aquatic biota (De Figueiredo et al. 2004). Microcystin (MC), a hepatotoxin, has received considerable attention due to its widespread occurrence in freshwaters and deleterious effects on humans and aquatic biota (Ueno et al. 1996, Best et al. 2001, WHO 2003, Bláha et al. 2004, Malbrouck et al. 2006, Prieto et al. 2007, Ibelings et al. 2008, Graham et al. 2010, Paerl et al. 2013). Human exposure occurs through drinking water, recreational contact, or fish consumption (Poste et al. 2011). MC is water stable and resistant to boiling, thus posing a threat to drinking water supplies and fish consumption. Human exposure to MC raises concerns regarding impairment of designated uses (e.g., swimmable and fishable). The World Health Organization (WHO) has issued guidelines for drinking water (1 µg L-1), recreational contact (low and moderate risk = 4 and 20 µg L-1, respectively) and consumption (0.04 µg kg-1 body weight d-1) (WHO 2003). MC concentrations are typically highest in liver and viscera (Garcia et al. 2010, Papadimitriou et al. 2010, Donald et al. 2011); shellfish may therefore pose a greater threat for human exposure because consumable portions include non-muscle tissues.

Cyanotoxins most frequently occur in warm, shallow, algal-rich waters that receive large anthropogenic nutrient loads (Moisander et al. 2009, Garcia et al. 2010). Genes for production of MC are found in Microcystis as well as others forms of cyanobacteria (Rolland et al 2005, Ginn et al. 2009, Donald et al 2011). MC accumulates in the tissues of a diverse group of organisms including fish, insects, crustaceans, bivalves, amphibians, birds, and mammals. (Wilson et al. 2008, Gérard et al. 2009, Poste et al. 2011, Garcia et al. 2010, Papadimitriou et al. 2010, Acuña et al. 2012). Exposure is thought to occur primarily through consumption though little is known about the factors that contribute to variable exposure and toxin contamination among consumers (Kozlowsky-Suzuki et al. 2012). Field studies have demonstrated that long- lived species exposed to recurrent blooms accumulate the toxin and thus present a contamination source to higher-order consumers (Watanabe et al. 1997 and Ibelings et al. 2005). Prior studies have been unable to establish direct relationships between diet variables and MC concentrations in consumers. This has been attributed to complex processes regulating toxin exposure, elimination and accumulation (Deblois et al. 2011).

Information regarding the spread of cyanotoxins into food webs is needed in order to assess implications for human health and aquatic resources. A trophic perspective that considers sources of organic matter (autochthonous vs. allochthonous) and feeding habitats (pelagic vs. benthic) may provide a useful framework for identifying risks to living resources and human exposure. Dietary information based on gut content analysis and C isotopic signatures of consumer tissues provides a basis estimating algal contributions to consumer diets and their exposure to cyanotoxins. As part of a broader effort to assess harmful algal blooms and associated impairments in the James River Estuary, we (1) characterized seasonal variation in the MC content of suspended and sedimented organic matter, and in the tissues of pelagic and benthic fishes, (2) tracked variation in toxin producers by monitoring a cyanobacteria-specific pigment (phycocyanin) and the abundance of genes specific to Microcystis and MC toxin production (mcyD), and (3) established linkages between toxin production and consumer exposure using information on dietary habits obtained by gut content and stable isotope analysis.

METHODS

Site Description The James River Estuary is the southernmost of the five major sub-estuaries of Chesapeake Bay. The tidal fresh segment (salinity < 0.5 ppt) is well-mixed (vertically and laterally) owing to fluvial inputs (freshwater replacement time = 5-25 d) and a large tidal prism (60 cm) relative to depth (mean ~3 m) (Chapter 1, Wood and Bukaveckas 2014). It shares a number of features in common with systems where cyanotoxins have been reported including large anthropogenic nutrient loads, elevated CHLa and presence of cyanobacteria (Marshall et al. 2008, Bukaveckas et al. 2011a, Bukaveckas and Isenberg 2013). During low discharge conditions (May-October), elevated CHLa is observed in the region where the James transitions from a deep, riverine channel to a broad, estuarine channel. Shallow depths provide favorable light conditions allowing phytoplankton to exploit proximal nutrient inputs from riverine and point sources (Chapter 1, Wood and Bukaveckas 2014, Bukaveckas and Isenberg 2013). In this section, annual average CHLa exceeds 20 μg L-1 with summer concentrations commonly above 50 μg L-1. Prior work has documented the presence of MC in other sub-estuaries of Chesapeake

Bay (Tango et al. 2008); this study presents the first regular monitoring of MC in water and the first comprehensive assessment of MC in biota for the Chesapeake region.

Sample Collection Near-surface (1 m) water samples were collected ~weekly from May through November 2012 at three sites, all located within the region of elevated CHLa (Figure 1-1). Two of the locations were main channel sites, which are also long-term monitoring stations for the Chesapeake Bay Program (CBP; JMS69, JMS75); one was a near-shore site located at the VCU Rice Center Research Pier. Water samples were analyzed for MC, CHLa and genetic markers. Total suspended solids (TSS) were also measured so that MC and CHLa could be normalized to dry mass when comparing suspended and sedimented material. Phycocyanin was monitored continuously (15 min) at the Rice Research Pier using a YSI 6600 multiparameter data sonde (0.5 m depth) equipped with an in vivo fluorescence sensor (Yellow Springs, OH) calibrated every two weeks. Triplicate surficial sediment samples (0-2 cm) were collected monthly from each of three near-shore sites (depth < 2 m) located near JMS75. Sediments from these sites were typical for the area being dominated by silty-sandy deposits. Surficial sediments were analyzed for stable isotopes, MC and CHLa.

Our sampling targeted abundant and ecologically important taxa in the James. The tidal fresh James River has large resident fish populations of Gizzard Shad (Dorosoma cepedianum), Threadfin Shad (D. petenense), and Blue Catfish (Ictalurus furcatus), as well as transient populations of Atlantic Menhaden (Brevoortia tyrannus) (Jenkins and Burkhead 1994). Their feeding habits include pelagic filter-feeding (Atlantic Menhaden, Threadfin Shad and young Gizzard Shad), benthic detritivory (juvenile Blue Catfish, adult Gizzard Shad) and piscivory (adult Blue Catfish). The common wedge clam (Rangia cuneata) is the dominant benthic filter- feeder based on annual surveys during 2001-2010 (CBP). It is considered to be a generalist suspension feeder ingesting a mixture of algal and non-algal suspended matter (Wong et al. 2010). We also measured MC in Blue Crabs as they represent the most likely pathway for human exposure. Blue Crabs (Callinectes sapidus) are predators and scavengers found in tidal fresh waters from May (as juveniles) through October.

Fish were collected for analysis of tissues (MC and stable isotopes) and gut contents (CHLa and organic matter content). Approximately 10-15 individuals were obtained per month

(May-October 2012) for each of 6 size-taxa groups: Gizzard Shad (adult and YOY), Threadfin Shad, Atlantic Menhaden and Blue Catfish (<20 and 20-40 cm total length). Size classes were used to assess potential effects of ontogenetic shifts in feeding on toxin content. Additional samples of water and fish were collected in March 2013 to assess toxin levels in over-wintering populations. Fish were obtained by electrofishing (low and high frequency) along multiple transects located in proximity to the water monitoring locations (JMS75, Rice Pier). Fish were euthanized according to IACUC protocols (VCU AD#20042). Blue Crabs were obtained from multiple crab pots deployed in proximity to the Rice Pier. Rangia were collected from a near- shore site in proximity to JMS75 using oyster tongs. Clams were held in deionized water for 48 h prior to dissection to allow for clearance of gut contents. Muscle and liver (fish) or viscera (crabs, clams) were surgically removed. For juvenile fishes, individuals were occasionally pooled (2-3 per sample) to obtain sufficient material for MC analysis. A total of 395 paired measurements (muscle and liver/viscera) of tissue MC concentrations were made among all taxa during the study.

Sample Analysis Filters for CHLa analyses (Whatman GF/A 0.5 μm nominal pore size) were extracted for 18 h in buffered acetone and analyzed on a Turner Design TD-700 Fluorometer (Arar and Collins 1997). TSS was determined gravimetrically using pre-weighed, pre-combusted filters. MC samples were analyzed using the high sensitivity ADDA ELISA Kit (detection limit 0.05 µg L-1; Abraxis; Warminster, PA). The assay measures numerous forms of MC using polyclonal antibodies with concentrations reported in MC-LR equivalents. To release MC from cells, water samples were thawed and refrozen two times (as recommended by the manufacturer), and then microwaved and sonicated to improve exaction efficiency (Silva-Stenico et al. 2009). Salinity was < 0.15 ppt for all water samples analyzed for MC. To extract MC from tissues and sediment, samples were dried at 60oC for 48 h, ground with a mortar and pestle, and extracted in 75% aqueous methanol for 24 h. (Wilson et al. 2008, Garcia et al. 2010). Extracts were centrifuged and supernatant collected. Subsamples were diluted with deionized water such that samples to be run on the ELISA plate contained <5% methanol. Plates were read on an ELISA plate reader at 450 nm. For each 96-well plate, six standards were run in duplicate to derive plate-specific standard curves. Average recovery from spiked samples was similar to previously published values (Table 2-1).

Quantitative PCR (qPCR) was performed on weekly samples collected at JMS75 and targeted conserved regions of the 16S rRNA gene associated with the genus Microcystis and the mcyD gene for toxin production (Rinta-Kanto et al. 2009). Filters (0.45 µm polycarbonate) were stored frozen (-20oC) until whole-community DNA could be extracted using the Mo Bio PowerWater DNA Isolation Kit (Carlsbad, CA). qPCR was conducted on a Bio-Rad CFX96 Real-Time system using SsoAdvanced SYBR Green qPCR Supermix (BioRad, Hercules, CA). Extracted DNA was amplified using the following primer sets: CYAN 108F (Urbach et al. 1992) and 377R (Nübel et al. 1997) for cyanobacterial 16s rRNA; MICR 184F and 431R (Neilan et al. 1997) for Microcystis-specific 16s rRNA; mcyD F2 and R2 (Kaebernick et al. 2002) for the MC- LR producing strains. All reactions were carried out in a total volume of 20 µL, which contained 10 µL of Bio-Rad Sso Advanced SYBR Greener Supermix, 40 µM of each primer, DNA suspension, and molecular-grade water. q-PCR programs for all genes included an initial 4 min at 95oC. Microcystis reactions utilized 55 cycles consisting of 95oC for 30 s, 56oC for 60 s and 72oC for 20 s while the mcyD assays were run for 50 cycles of 95oC for 20 s, 55.5oC for 30 s, and 72oC for 30 s. A melt curve was analyzed for each assay to ensure appropriate gene amplification and checked on a 1.5% agarose gel using ethidium bromide staining. Standards curves were created in duplicate using 10-fold serial dilutions of genomic DNA obtained from cultures of M. aeruginosa (UTEX LB 2388). A non-template control was run in duplicate for each assay.

Consumer Diets & Feeding

Fish As our interest was in fish exposure to algal toxins, we did not perform a detailed gut contents analysis, but rather measured the CHLa content of gut materials as an indicator of algal contributions to their diet. CHLa is known to degrade during passage through the digestive system (Friedland et al. 2005) and therefore our estimates are likely to be conservative with respect to the contribution of phytoplankton and phytodetritus to fish diets. Contents from the entire gut were removed for determination of wet weight, dry weight and CHLa content. Material for CHLa analysis was extracted overnight in 90% buffered acetone and analyzed on a TD-700 fluorometer (Arar and Collins 1997). Feeding habits of fishes were also evaluated through stable C isotope analyses of muscle tissues. C isotopic signatures are widely used to assess the relative contribution of autochthonous and allochthonous sources of organic matter to consumer diets (Peterson et al. 1987, Forsberg et al 1993, Finlay 2001, Martineau et al. 2004,

Hoffman et al. 2008, Davis et al. 2013). To aid in the interpretation of the fish stable isotope data, we characterized differences in δ13C of autochthonous and allochthonous organic matter using seston samples collected during our weekly monitoring. Samples (N=32) collected during high-discharge, low CHLa conditions were used to represent allochthonous dominated C isotopic signatures and samples collected during low-discharge, high CHLa conditions were used to represent isotopic signatures with greater autochthonous contributions. For seston and fish tissue samples, we followed preparation protocols of, and submitted samples to, the UC-Davis Stable Isotope Lab. Tissue samples were analyzed using a PDZ Europa ANCA-GSL elemental analyzer interfaced to a PDZ Europa 20-20 isotope ratio mass spectrometer. POM and sediment samples were analyzed using an Elementar Vario EL Cube or Micro Cube elemental analyzer interfaced to a PDZ Europa 20-20 isotope ratio mass spectrometer. The long term standard deviation for 13C is 0.2 ‰. Relationships between dietary variables (gut content and stable isotope analysis) and MC concentrations in consumer tissues were analyzed using Model II regression analysis for data pooled by species across months.

Rangia Grazing As it was not practical to sample Rangia gut contents, we measured their grazing rates in order to relate toxin content of tissues to exposure. Experimental design followed Wong et al. whereby we monitored changes in CHLa (at 0, 2 and 4 h) in 20 L mesocosms with and without clams. In the presence of clams, concentrations of suspended particulates decline faster (linear) such that differences in the slopes of regression lines (concentration vs. time) between mesocosms with and without clams can be used to estimate the clearance rate. This rate is a theoretical value representing the volume of water cleared of particulate material based on the mass removed by consumers and average concentration in the water (Coughlan 1969). Water with natural seston was collected from the Rice Pier for use in the grazing experiments. Clams were obtained in proximity to one of the water monitoring sites (near JMS69) for monthly trials performed during March-October 2012. Clams were kept overnight for acclimation to experimental conditions that included two temperature treatments: a standard reference of 20oC and the ambient river temperature (which ranged from 14o to 32o C). Each mesocosm contained a similar mass of clams (3-10 individuals depending on size; average body mass of 2.6 g ind-1; range = 0.5 to 5.0 g ind-1) and final results were normalized to soft-tissue body mass. Six mesocosms were used for each temperature treatment (3 with, 3 without clams). Mesocosms

32 were kept under low light conditions (to minimize phytoplankton growth during the experiment) and equipped with a circulating pump to maintain particulates in suspension. CHLa was measured by the same methods as for the weekly monitoring.

Rangia Toxicity Assays Experiments were performed in May, June and September of 2013 to measure the effect of Microcystin upon Rangia Clearance rates (due to the theft of oyster tongs we were unable to collect clams during July & August). Methods followed the same principles as previous estimates of Rangia clearance rates but used 50% ambient river water and 50% filtered river water. We found these dilutions increased the rate of grazing and allowed for a better opportunity to observe Microcystin effects. These experiments were also performed in 3 L buckets with 1 individual per bucket to improve replication capacity. Experiments were performed in triplicate at 0, .4, .8, 1.6 and 3.2 µg L-1 dissolved Microcystin. Aqueous dissolved Microcystin was added at the beginning of the experiment and water samples were collected at the end of the experiment to determine mesocosm toxin concentrations. Toxin concentrations were within 10% intended concentrations and never significantly different between clam and no- clam containers for all experiments. We also monitored Rangia grazing upon Pamunkey River water, which it typically lower in CHLa and Microcystin, to make observations regarding the influence of dietary Microcystin upon grazing rates. Pamunkey River Water was used at ambient concentrations in attempt to obtain similar levels of POC and CHLa.

RESULTS

Cyanobacteria We characterized cyanobacteria abundance based on continuous measurement of phycocyanin and weekly measurement of MC concentrations and the abundance of genes specific to Microcystis and MC (mcyD). Phycocyanin monitoring revealed six peaks occurring at ~monthly intervals from May through October (Figure 2-1). The first peak (May) was associated with low toxin concentrations (<0.10 µg L-1) and low abundance of Microcystis and mcyD genes. Subsequent peaks in phycocyanin were associated with higher toxin levels and increased Microcystis gene abundance. The July and August peaks were noteworthy for exhibiting peaks in both Microcystis and mcyD gene abundances and highest toxin concentrations. Peak MC concentrations occurred on July 17 (0.92 µg L-1) and August 28 (0.78

µg L-1; average values for three sampling locations). Inter-site variability in MC concentrations was low (Figure 2-1; average SE = 0.03 µg/L). By November 27, MC was undetectable at all stations (data not shown). Phycocyanin fluorescence (daily average values) was a useful predictor of MC concentrations in the James (R2 = 0.63, p < 0.01).

MC in Fish and Shellfish Seasonal variation in toxin content among fish and shellfish generally followed patterns of MC concentration in the water column (Figure 2-2; Table 2-2). Crab viscera contained on average 6-fold higher levels of MC than muscle (all months), with similar seasonal patterns observed for both. Tissue concentrations were low but detectable during the first sampling (May); concentrations in August were ~3-fold greater and coincided with peak water column concentrations. Among fishes for which we were able to resolve seasonal patterns (juvenile and adult Gizzard Shad, Threadfin Shad), highest MC concentrations were observed in July-August. Young-of-the-year (YOY) Gizzard Shad exhibited an early and exceptionally high peak in liver MC concentrations (~1 µg g-1 DM) in comparison to other fishes (<0.3 µg g-1 DM). Across all taxa, the proportion of fish with measurable MC concentrations in liver tissue generally increased during the succeeding peaks in cyanobacteria abundance with highest incidence observed in August (94%) and September (83%). Samples obtained in March 2013 showed no detectable concentrations in water, whereas tissue content was low but measureable (in livers, 67% of individuals among all taxa; mean = 0.044 ± 0.007 µg MC g DW-1). Wedge clams (Rangia) exhibited seasonal patterns of toxin content similar to those observed in fish and Blue Crabs although tissue concentrations were ~10-fold lower and peak values occurred later (August-September; Figure 2-3). The rise in tissue concentrations coincided with increasing water column concentrations of MC and a decline in feeding rates. Biomass-specific clearance rates were highest before the onset of elevated MC (March-May mean = 0.19 ± 0.03 L g-1 h-1) and lowest in Summer (Jun-Sep = 0.07 ± 0.01 L g-1 h-1). Partial recovery in clearance rates was observed in Fall (Oct-Nov = 0.11 ± 0.03 L g-1 h-1) after the decline of MC concentrations in water. Clams incubated at standardized (20oC) and in situ (river) temperatures exhibited similar seasonal patterns in clearance rates (Table 2-3). Clearance rates were significantly correlated with water MC concentrations (R2 = 0.66, p<0.01); a non-linear model depicted rapid declines in clearance when MC exceeded 0.2 µg L-1.

Inter-specific variation in MC content among primary consumers was related to dietary habits as inferred from gut contents and stable isotope analyses (Figure 2-4). CHLa concentrations in fish gut contents and δ13C values of muscle tissue were both significant predictors of inter-specific differences in liver MC (R2 = 0.82, p = 0.012 and R2 = 0.62, p = 0.038, respectively). Intra-specific (month-to-month) variation in the two dietary metrics was generally low (SE < 25% of mean) with the exception of YOY Gizzard Shad, which also showed the greatest variability in liver MC concentrations. Highest liver MC concentrations and lowest δ13C values were observed among planktivores (Threadfin Shad, YOY Gizzard Shad, and Atlantic Menhaden). Two of these taxa (excluding Menhaden) also exhibited high concentrations of CHLa in their gut contents. By comparison, the benthic detritivores (Blue Catfish, adult Gizzard Shad) showed lower liver MC concentrations, less depleted δ13C values and lower CHLa in their gut contents. These findings were consistent with observed differences in CHLa, MC, and δ13C signatures between suspended and sedimented materials. Suspended particulate matter collected during peak algal abundance was depleted in δ13C (-28.5 ± 0.2 ‰) compared to samples collected during high discharge (-26.8 ± 0.1 ‰) and from surficial sediments (-26.7 ± 0.1 ‰). Suspended particulate matter contained 200-fold greater CHLa (1,650 µg g-1 DM) than sedimented materials (7 µg g-1 DM) and was 4 orders of magnitude higher in MC (16,100 µg kg-1 DM and 0.4 µg kg-1 DM, respectively). MC concentrations in sediment varied seasonally with peak values in September corresponding to highest sediment CHLa content (Table 2-4).

Toxicity Assays Rangia Clearance rates responded to dissolved Microcystin additions in May and June experiments while in September clearance rates were low (< 0.02 L g DW-1 h-1) for all treatments (Figure 2-5). In May and June clearance rates exhibited a non-linear response to Microcystin concentrations and these results suggest a 50% loss of clearance rates at Microcystin (MC50) concentrations of 0.4 and 1.2 µg L-1 respectively.

Comparisons between 50% James River water clearance rates and Pamunkey River water clearance rates must be considered in the context of water quality differences between treatments. CHLa concentrations in 50% diluted James River treatments were significantly greater than Pamunkey River concentrations in May, June and September (Figure 2-6). In

35 contrast, Microcystin concentrations were low for both Pamunkey and James River water samples in April, May and June but significantly higher in James River water in September. While no significant difference occurred between sites in Microcystin concentrations, CHLa- specific clearance rates were either greater in James River water (April) or not significantly different between sites (May & June). In September when Microcystin was significantly greater in James River mesocosms, clearance rates were higher for Pamunkey River water.

DISCUSSION We found widespread cyanotoxin contamination among the dominant fish and shellfish of the tidal freshwater segment in the James River Estuary. The ubiquitous presence of MC in water (>90% of samples) and biota (~67% of individuals) was surprising given that cyanobacteria are a minor component of the phytoplankton community. During the period of study, cyanobacteria accounted for 6% of biomass (peak = 9%) with diatoms (76%) and chlorophytes (16%) accounting for the bulk (mean for weekly samples; H. Marshall, ODU, pers. comm.). Although the tidal-fresh James exhibits some of the common features of eutrophic systems (shallow depth, high CHLa, elevated nutrient loads), it is actively mixed by tidal forces, which negate the advantages of buoyancy that favor some cyanobacteria (e.g., Microcystis) (Paerl 1988, Johnk et al. 2008). Strong tidal mixing likely accounts for the low spatial variability observed in MC concentrations among the three sites. Our findings show that even in systems with low cyanobacteria abundance (<10% of algal biomass) and moderate levels of MC (~1 µg L-1), toxin exposure may be widespread among consumers. Avoidance of cyanobacteria has been documented in a wide range of consumers (Lampert 1987), but these mechanisms may be inefficient or non-advantageous when cyanobacteria constitute a minor component of food resources.

An important and novel finding from this study is that variable levels of MC in consumer tissues can be linked to feeding habits and dietary exposure. Pelagic-feeding, planktivorous fishes had higher levels of CHLa in their diet as well as higher concentrations of MC in their tissues compared to benthic-feeding detritivores. These differences reflect the orders of magnitude lower concentrations of CHLa and MC in sediments in comparison to suspended particulate matter. Lower sediment concentrations are likely due to dilution of settling phytodetritus by the much larger pool of sedimented material that is largely terrestrial in origin. Factors in addition to dilution by watershed-derived particulate matter may also contribute to low

MC concentrations in sediment. The ratio of CHLa to MC in suspended particulate matter was ~80:1, whereas the corresponding value for surficial sediments was >1,000, suggesting post- depositional biodegradation of MC (Grützmacher et al. 2010). It has also been reported that MC can adsorb to clay particles and thereby become resistant to conventional extraction procedures. Rinta-Kanto et al. attributed the lack of Microcystin in Lake Erie sediments to this mechanism, though other studies have reported high values in sediments using similar extraction techniques (Schauss et al. 2002, Chen et al. 2008) ELISA-based results may be indicative of bioavailable (vs. total) MC in sediment as this test detects free MCs and metabolized MCs that cross-react with antibodies by immuno-affinity.

Inferences regarding algal contributions to consumer diets based on CHLa analysis of gut contents were supported by stable isotope results. Carbon isotopic signatures of surficial sediments and tissues of benthic-feeding fishes were similar to seston samples collected during high discharge events when contributions from terrestrial sources are greatest. Isotopic signatures of planktivorous fishes were more similar to seston samples collected during periods when autochthonous contributions were greater (high CHLa:TSS). Thus both lines of evidence (gut contents and stable isotopes) show that toxin exposure is linked to diet whereby autochthonous sources comprise a greater proportion of the diets of pelagic fishes and lead to greater MC content in tissues. The benefit of linking toxin content in tissues with dietary habits is in providing a framework for assessing threats to aquatic biota and pathways of human exposure. In the James, species at greatest risk for cyanotoxin exposure include Atlantic Menhaden, YOY Gizzard Shad, Threadfin Shad and anadromous shad (Alosa spp.; not sampled in this study). Benthic feeders such as juvenile Blue Catfish, adult Gizzard Shad and Atlantic Sturgeon experience lower exposure.

The benthic filter-feeder Rangia exhibited low tissue MC concentrations comparable to benthivorous fishes. Rangia is considered a generalist suspension feeder but δ13C values suggest that autochthonous organic matter comprises a small fraction of their diet. When feeding, Rangia extend their siphons above the sediment-water interface and may ingest sedimented and recently re-suspended particulate matter. Prior work has shown that benthic sources of particulate matter can be important to bivalves such as oysters (Marchais et al. 2013). Low toxin content of clams may also reflect avoidance of MC ingestion due to seasonally-variable feeding rates. For the average body size of individuals used in our experiments, clearance rates were 37

12.0 L ind-1 d-1 during periods when MC concentrations were low (Spring and Fall) and 4.0 L ind-1 d-1 when MC concentrations were elevated (June-September). Taking into account their density in this segment of the James (mean = 30 ind m-2 based on CBP annual benthic surveys for 2001-2010) and the average depth (1.3 m), this seasonal decline in feeding corresponds to a drop in water-column filtration rates from 28% to 10% d-1. This drop does not appear to be a temperature effect as it was observed in clams incubated at both ambient (river) and standard (20oC) temperatures (Table 2-3). Results from controlled toxicity assays experiments (Figure 2- 5) indicate this negative effect upon grazing can be reproduced in the laboratory with dissolved Microcystin. The combination of these results suggests the presence of Microcystin results in declining feeding rates for these bivalves. Results associated with food quality (figure 2-6) support the view that dietary Microcystin may also lead to reduced grazing rates, though other aspects of food quality (Ger et al. 2010) cannot be ruled out with these experiments. This question merits further investigation as it has direct implications for assessing harmful bloom effects on valuable ecosystem services provided by filter-feeding bivalves (Ermgassen et al. 2013).

Our data on toxins in secondary consumers is limited to Blue Crabs and adult (>40 cm) Blue Catfish. Adult, piscivorous catfish exhibited lower MC concentrations (0.026 µg g-1 DM) relative to smaller, detritivorous catfish (<20 cm = 0.086 µg g-1 DM; 20-40 cm = 0.093 µg g-1 DM) consistent with prior studies showing that MC does not bioaccumulate (Kozlowsky-Suzuki et al. 2012). In contrast, Blue Crabs exhibited the highest incidence of muscle contamination among all consumers (72% of individuals) and 4- to 10- fold higher toxin concentrations than their prey (Rangia (Garcia et al. 2010)). Our estimates of MC in Blue Crab muscle tissue (0.020 µg g-1 DM) were similar to those previously reported by Garcia et al. 2010 for a eutrophic Louisiana estuary (0.021 µg g-1 DM). Laboratory studies by Dewes et al. 2006 on an estuarine burrowing crab (Chasmagnathus) demonstrated that tissue concentrations exceeding 0.013 µg MC g-1 induced physiological and biochemical imbalances. These findings suggest that toxins may have detrimental effects on James River Blue Crab populations. As reported in other studies, we observed lower toxin levels in muscle relative to liver and viscera (Wilson et al. 2008, Garcia et al. 2010, Papadimitriou et al. 2010). This has positive implications for human exposure, though it should be noted that human consumption of Blue Crabs may include non- muscle tissues. Also, apex predators such as Osprey and Bald Eagles are known to consume

38 non-muscle tissues of fish. To assess implications for human health, we compared MC concentrations in Blue Crab muscle to WHO tolerable daily intake (TDI) guidelines for human consumption (0.04 µg kg-1 body mass d-1) WHO 2003. Following the calculations of Garcia et al. 2012, and using our monthly crab MC concentrations, we found that for a serving size of 300 g wet mass, MC ingestion corresponded to 31 to 150% of the TDI guideline (mean = 73% for all months). This serving size exceeded the TDI guideline in all months (mean = 114% of TDI) when we included a small proportion of viscera (10%) in the consumed portion due to the higher concentrations of MC in hepato-pancreas tissues. However, it should be noted that the long-term average shellfish consumption would be substantially lower than the serving size used for this analysis and therefore unlikely to exceed TDI guidelines over a lifetime of exposure. At present, little is known regarding acute vs. chronic exposure to MC in humans or other biota to allow an assessment of their relative risks.

Seasonal variation in tissue MC concentrations was apparent in all consumers; in some cases, these tracked seasonal patterns in water MC concentrations (Figure 2-7). Statistically significant correlations between water and tissue concentrations were observed in Threadfin Shad and Wedge Clams but not among Catfish, Gizzard Shad or Blue Crabs. Intra-specific variation in tissue MC concentrations was small in relation to inter-specific differences with the exception of YOY gizzard shad. Gizzard Shad shift their feeding habits from planktivory as juveniles to greater reliance on detritivory as adults (Schauss et al. 2002). We observed decreasing CHLa in gut contents and lower liver MC concentrations in succeeding months suggesting that ontogenic shifts in feeding reduced their exposure to MC. For all consumers, measurable toxin levels persisted in tissues during periods when MC was not detected in the water column (March 2013 sampling). Ozawa et al. 2003 similarly reported measurable levels of MC in freshwater snails during fall and winter following a spring cyanobacterial bloom. These findings suggest that although the toxin is known to be metabolized (Pires et al. 2004), health effects may occur outside of bloom periods when toxins are produced.

Microcystin content of planktivorous fish tissues, which are high relative to other fish species, are low when considered in the context of non-selective ingestions rates (Chapter 3). If we consider a high-end estimates of total Microcystin in threadfin shad (liver concentrations * body mass) and the rate of Microcystin ingestion (Microcystin content of suspended particulate matter * low end suspended particulate matter ingestion rates (Chapter 3)) it is apparent that the 39 amount of toxin present in fish biomass is small (<15%) relative to daily ingestion rates. This would also imply that there is nearly twice as much Microcystin present in the stomach of fish than the rest of the organism. This result considers many different estimates which all have complex errors associated with them but given the conservative approach we feel that it may indicate a few possible insights regarding toxin accumulation. First, it may suggest planktivorous fish are selectively avoiding microcystis and thus we are over-estimating ingestion of Microcystin for these organisms. A second potential explanation is that the majority of Microcystin being ingested is not accumulating. This could be due to lack of assimilation or a high rate of metabolism of Microcystin (Friedland et al. 2005). Previous studies have suggested that the level of Microcystin within fish tissues exhibit complex non-linear relationships (Deblois et al. 2008). In addition to monitoring toxin levels, we used a variety of approaches to track the abundance of cyanobacteria. Prior work has shown that phycocyanin can predict the occurrence of elevated MC.(Urbach et al. 1992, Marion et al. 2012) We found that continuous monitoring of phycocyanin was useful for discerning bloom events at fine-scale (daily) resolution and in providing real-time information to trigger sampling activities. The genetics analysis allowed us to bridge the phycocyanin and toxin results by identifying bloom events associated with peaks in the number of gene copies specific to Microcystis and toxin production (mcyD). The July and August peaks (as delineated by phycocyanin) were associated with highest toxin levels as well as highest abundance of Microcystis and mcyD, whereas the May, June and October peaks exhibited low toxin levels and low abundance of mcyD. An exception was the September peak during which MC concentrations were high, but Microcystis and mcyD abundance was low. Elevated toxin levels may indicate persistence of the toxin in the water column from preceding peaks. There was a strong correlation between gene copies of Microcystis and mcyD (R2 = 0.83, p < 0.01) suggesting that either the former contributed directly to the latter, or that variation in Microcystis abundance was synchronized to that of other toxin-producers. On average, the number of copies of the toxin-coding gene was 20% of Microcystis gene copies. This proportion was higher (40%) during periods of elevated toxin concentration (July and August), and below 10% in other months. Thus variation in toxin concentrations was principally driven by cyanobacteria abundance (which varied by orders of magnitude) and secondarily by variation in the proportion of toxin-producing strains (which varied by 3-fold).

In summary, we surveyed the dominant consumers comprising the James River food web to assess seasonal and inter-specific variation in toxin exposure. Peak occurrence of toxin contamination (% of individuals) and peak tissue concentrations were observed in months with highest toxin levels in the water column. Low but measurable toxin levels persisted in over- wintering populations suggesting that potential impacts on living resources may not be confined to bloom periods. Tissue MC concentrations varied 10-fold among species and revealed variable levels of exposure associated with differences in feeding habitats (benthic vs. pelagic) and dietary sources (autochthonous vs. allochthonous). Measurements of CHLa in gut contents and stable C isotopes allowed us to characterize food sources supporting secondary production and to explicitly link feeding habits with cyanotaoxin exposure. Highest MC levels were found among pelagic, planktivorous fishes with lower concentrations measured in benthic detritivores. Pelagic feeders were subject to greater toxin exposure due to a greater proportion of autochthonous organic matter in their diets. For the suspension-feeding wedge clam, exposure was mitigated by low feeding rates in summer when MC was present in the water column. Among secondary consumers, linkages between feeding and exposure were less clear as adult Blue Catfish exhibited lower toxin levels than their prey, whereas Blue Crabs had higher levels of toxin. Elevated toxin levels in Blue Crabs raise concern for detrimental effects on their populations and human health impacts.

Tables and Figures

Table 2-1. Extraction efficiencies from spiked samples observed in this and previously published studies using the ELISA Microcystin assay.

Material Recovery (%) Source Water 94 + 6% This study Sediment 90 + 11% This study water/seston 91 - 106% Cong et al. 2006 92 - 111% Wang et al. 2007 92% Wilson et al. 2008 Blue Crab Muscle 104 + 7% This study 50% Garcia et al. 2010 Blue Crab Viscera 67 + 7% This study 98% Garcia et al. 2010 Fish Liver 74 + 15% This study Fish Muscle 89 + 12% This study Fish Tissues 65% Wilson et al. 2008 68% Ernst et al. 2005 68% Ibelings et al. 2005 Rangia Muscle 84 + 9% This study 64 + 1% Rangia Viscera This study

Table 2-2. Body size and average Microcystin concentrations in tissues of fish and shellfish from the tidal freshwater James River Estuary. Also shown is the percentage of individuals with measurable Microcystin levels in muscle and liver/viscera and CHLa concentrations in stomach contents of fishes. Length Wet Mass Gut Contents CHLa Average Microcystin in Average Microcystin in Muscle % of individauls with (cm) (g) (µg g DW-1) Viscera/Liver (µg g DW-1) (µg g DW-1) Microcystin Species Mean ± SE N Mean ± SE N Mean ± SE Max Mean ± SE Max Muscle Viscera Atlantic Menhaden 14.8 ± 1.0 48.6 ± 0.6 23 76.6 ± 6.8 23 0.143 ± .039 0.90 0.0039 ± .003 0.07 9% 74% Juvenille Gizzard Shad 10.8 ± 0.5 17.6 ± 1.7 43 207 ± 24 32 0.247 ± .101 3.20 0.0025 ± .001 0.03 7% 79% Threadfin Shad 9.1 ± 0.2 9.7 ± 0.5 70 170 ± 21 42 0.197 ± .024 0.62 0.0053 ± .001 0.03 12% 81% Blue Catfish (0-20 cm) 15.9 ± 0.3 38.3 ± 1.5 97 38.3 ± 4.2 54 0.126 ± .010 1.26 0.0047 ± .003 0.06 5% 68% Blue Catfish (20-40 cm) 27.3 ± 0.6 196 ± 15 128 31.9 ± 3.9 58 0.100 ± .024 1.10 0.0004 ± .0002 0.01 5% 68% Adult Gizzard Shad 30.9 ± 0.5 356 ± 14 115 38.8 ± 5.5 60 0.038 ± .009 0.29 0.0019 ± .0007 0.02 21% 38% Rangia 6.2 ± 0.1 98.5 ± 0.5 - - 90 0.025 ± .002 0.10 0.0058 ±.0012 0.07 15% 78% Blue Crab 13.7 ± 0.3 129 ± 6 - - 36 0.133 ± .023 0.58 0.0200 ± .0096 0.34 72% 100%

Table 2-3. Biomass-specific clearance rates of CHLa by the wedge clam (Rangia cuneata) incubated at ambient (river in situ) and standard temperatures using clams and seston obtained from the tidal freshwater James River. Initial POC Initial CHLa Ambient Standard

Date mg L-1 µg L-1 CR (L g-1 DW h-1) oC CR (L g-1 DW h-1) oC p

19-Mar-12 0.45 31.17 0.25 ± 0.03 21.5 0.25 ± 0.03 21.5 ns 19-Apr-12 0.68 27.14 0.19 ± 0.07 21.0 0.19 ± 0.07 21.0 ns 5-Jun-12 0.78 25.02 0.14 ± 0.07 23.5 0.15 ± 0.07 19.8 ns 22-Jun-12 0.84 16.11 0.06 ± 0.04 27.1 0.07 ± 0.04 19.8 ns 20-Jul-12 1.46 24.31 0.07 ± 0.02 31.5 0.09 ± 0.02 20.0 ns 23-Aug-12 1.19 53.31 0.07 ± 0.01 29.5 0.01 ± 0.01 20.0 0.008 20-Sep-12 1.18 37.71 0.07 ± 0.02 26.5 0.06 ± 0.01 20.0 ns 18-Oct-12 1.05 40.81 0.08 ± 0.04 20.0 0.11 ± 0.04 20.0 ns 30-Nov-12 0.69 30.26 0.14 ± 0.05 14.0 0.05 ± 0.02 20.0 ns 15-Mar-13 0.59 18.66 0.24 ± 0.02 20.3 - - -

Table 2-4. CHLa and Microcystin concentrations in suspended and sedimented particulate matter from the James River Estuary (data normalized to dry mas

Figure 2-1. Seasonal patterns in phycocyanin fluorescence, Microcystin, and genes specific to Microcystis and Mycrocystin (mcyD) in the tidal freshwater segment of the James River Estuary during 2012.

Figure 2-2. Seasonal variation in tissue concentrations of Microcystin in fish and shellfish from the tidal freshwater segment of the James River Estuary during May 2012 to March 2013 (mean ± SE; some error bars not visible).

Figure 2-3. Seasonal variation in feeding rates and Microcystin content in tissues of the common wedge clam (Rangia cuneata) in the tidal freshwater segment of the James River Estuary during May 2012 to March 2013 (mean ± SE; some error bars not visible).

0.4

0.3 Gizzard

DW) Shad 1 - YOY Gizzard Shad YOY Atlantic Menhaden 0.2 Blue Cat Blue Cat (0-20 cm) (0-20 cm) Threadfin Threadfin Shad Shad Blue Cat (20-40 cm)

0.1 Atlantic Liver Microcystin (µg g (µg Microcystin Liver Blue Cat Menhaden Adult Gizzard (20-40 cm) Shad 2 R2 = 0.82 R = 0.62 Adult Gizzard p = 0.04 Rangia Shad p = 0.01 0.0 0 50 100 150 200 250-29 -28 -27 -26 -25 -24 CHLa of Gut Contents (µg g-1 DW) Muscle δ 13C

Figure 2-4. Microcystin content in liver tissues of fish and shellfish in relation to algal contributions to their diet as indicated by CHLa in gut contents and stable C isotopic signatures of consumer tissues. Data points are mean and SE based on monthly collections during May-October 2012. Microcystin concentrations for the wedge clam (Rangia) are based on analysis of viscera. Regression lines are based on Model II regression analysis.

Figure 2-5 Rangia clearance rates at controlled Microcystin concentrations.

50 Pamunkey * 40 James

) 1 - 30 *

20 CHLa (µg (µg L CHLa

0 0.3 Apr May Jun Sep pre-bloom post-bloom

) 1 - 0.2

0.1 Microcystin (µg L

0.0 Apr May Jun Sep 0.3

) *

1 *

1 1 - 0.2

0.1 Clerance ( LRate DW g

0.0 Apr May Jun Sep

Figure 2-6. Results from dietary exposure study including CHLa concentrations in water from Pamunkey and James River Water (top), Microcystin concentrations in water from the two sources (middle) and clearance rates of clams (bottom).

Figure 2-7 Relationships between water and tissue concentrations of the cyanotoxin Microcystin in fish and shellfish from the James River Estuary

Chapter 3: Top Down Controls, Trophic pathways and the fate of algal production in the tidal freshwater James River

ABSTRACT To assess the importance of top-down effects on chlorophyll-a (CHLa) concentrations in the James Estuary, we derived estimates of grazing rates for the dominant consumers inclusive of zooplankton, benthic filter-feeders and planktivorous fish. We also used stable isotope analysis to quantify algal contributions to consumer diets. Lastly, we performed 2 mesocosm experiments to manipulate grazer abundance and assess their influence on algal and nutrient dynamics. Our results suggest that grazers consume only a small fraction (15%) of daily CHLa production but play an important role in N cycling through ingestion of allochthonous organic matter. The dominant fate of phytoplankton production (74%) was in supporting microbial decomposition with a small fraction being exported downstream (<5%). Stable isotope (SI) analysis allowed us to distinguish two trophic pathways: a benthic omnivore pathway that was largely supported by allochthonous (terrestrial) inputs, and a pelagic planktivore pathway that was more reliant on autochthonous production. SI results suggest that 70% of metazoan biomass in the James is supported by allochthonous organic matter inputs. Mesocosm experiments suggest that planktivorous fish have qualitative influences phytoplankton communities whereby Micrcocystis is inhibited through zooplankton mediated trophic cascades.

INTRODUCTION Reductions in nutrient loading have had variable success in mitigating the symptoms of eutroiphication in estuaries. A few systems have quickly responded and returned to more oligotrophic conditions while the more common result is a lack of biotic response or a significant lag between the time of nutrient reductions and biotic response (Greening and Janicki 2006, Orth et al. 2010, Greening et al. 2011, Boynton et al. 2013, Verdonschot et al. 2013). This lack of algal response is evident by globally increasing CHLa:TN ratios across a wide range of estuarine ecosystems (Carstensen et al. 2011). When systems do exhibit recovery they frequently have non-linear convoluted trajectories with altered baselines (Duarte et al. 2009). These trajectories have been attributed to several different environmental impacts including invasive species, loss of submerged aquatic vegetation and legacy effects of sedimented nutrients (Duarte et al. 2009). Sedimented nutrients may be re-mineralized over long periods of time and thus curtail recovery. These impediments and delays to recovery make it challenging to connect current management actions with verifiable results. Understanding the role of these internal drivers is critical to forecasting responses as well as restoring eutrophic water bodies.

Primary consumers are an important component of food webs because of their ability to control standing stocks of primary producers and influence trophic transfer efficiency. Consumers can potentially influence phytoplankton in three ways: first, by directly consuming algae (i.e., ‘grazing’), second, through consumer-mediated nutrient recycling (Vanni et al. 2005,2006), and third through selective feeding which may favor certain groups of algae (e.g., cyanobacteria) and/or influence trophic relationships (e.g., trophic cascades; Schauss et al. 2000). A few studies have suggested that grazing fish may exacerbate cyanobacteria blooms due to their ability to survive gut passage and assimilate nutrients from the digestive tract (Friedland et al. 2003, Jančula et al. 2008). Thus, grazer effects on biomass may be negative (suppression of CHLa through grazing), or positive (fueling algal blooms through nutrient re-generation) and their effects on community composition unpredictable. Selectivity effects are difficult to predict because primary consumers comprise diverse feeding habits that include consumption of both phyto- and zooplankton as well as allochthonous organic matter . (Vanni et al. 2005, Wilson and Xenopoulos 2012). These complications preclude a clear understanding of the strength and direction (positive or negative) of top-down effects from grazers at primary trophic levels.

In addition to their influences upon trophic status, primary consumers play an important role in determining the flow of energy through food webs. The flow of energy in fluvial aquatic systems has received significant attention over the last 50 years with the development of conceptual models such as the river continuum concept (Vannote et al. 1980), the flood pulse concept (Junk et al. 1989), and the riverine productivity model (Thorp and Delong 1994, 2002). The Riverine Productivity Model (RPM) suggests that while terrestrial organic matter dominates ecosystem metabolism, the majority of higher trophic level secondary production is supported by autochthonous production. This conceptual model is supported by the fact that aquatic primary producers are more nutritious due to their lack of structural molecules (e.g., cellulose and lignin; Sarkanen and Ludwig 1971, Renaud et al. 1999). While several studies have found support for the RPM, other studies have shown that terrestrial organic matter supports aquatic food webs (Kendall et al. 2001, Hoffman et al. 2007, Roach 2013). The importance of terrestrial organic matter in these systems has usually been attributed to a lack of autochthonous production associated with high levels of turbidity and light limitation or, advective (washout) effects. Studies that are able to address questions related to the energy flow through food webs are uncommon in part because of the comprehensive data requirements (Benke et al 2011). These studies require an understanding of population size as well as ingestion rates for the dominant consumers of an ecosystem. While data collection for these studies is extensive, results provide insights about the structure and function of food webs.

The tidal fresh segment of the James River Estuary represents a system with high algal production, which also receives substantial inputs of terrestrial organic matter due to a large watershed area. The food web of the James has likely undergone historical changes, in part due to fisheries introductions, but there is little recent or historical data on grazing rates. The most recent study (Bukaveckas et al. 2011a) reported grazing rates by zooplankton of ~5% CHLa/d, which were low in comparison to other estuaries (typically ~30%/d). This study did not consider other potential consumers such as benthic filter-feeders and planktivorous fish. It has been argued for Chesapeake Bay that loss of oyster populations has resulted in historically low grazing rates and that this may preclude attainability of lower CHLa even with reductions in nutrient loads. While oysters do not occur in the tidal freshwater James, other consumers may regulate algal abundance. Recent work has suggested that internal recycling provides 60% of the nutrients sustaining algal blooms in the James (Bukaveckas & Isenberg 2013; Wood &

Bukaveckas 2014), though the contribution of consumer-mediating re-cycling is unknown. Lack of information on grazing rates diminishes our ability to forecast CHLa under various environmental conditions and nutrient loading scenarios.

Here we present a study of top-down effects in the James which considers the role of primary consumers as grazers (i.e., in removing CHLa) as well as their importance to internal nutrient cycling. The assessment is based on population-level estimation of CHLa and PON ingestion by the dominant primary consumers which include planktovorous and benthic fishes, benthic filter-feeders (Wedge clams) and zooplankton. Consumer effects are evaluated in the context of algal production and mortality (senescence and washout) and in relation to other factors influencing N fluxes. For nutrient recycling, we focused on nitrogen, as our recent studies have shown that phytoplankton in the tidal fresh segment are N- or co- (N-P) limited (Chapter 1). We also present stable isotope analysis of the food web which was used to quantify autochthonous and allochthonous contributions to metazoan production. Lastly we used mesocosms to conduct experiments in which grazer effects on primary producers could be evaluated across a gradient of nutrient conditions.

METHODS

Study Site and Data Collection This analysis of top-down effects is based on data collected from the lower half of the tidal fresh segment (designated JMSTF1; as per CBP segmentation scheme). This is the portion of the tidal-fresh segment between Hopewell, VA and the confluence with the Chickahominy which is characterized by persistent, elevated CHLa concentrations in summer months (Bukaveckas et al. 2011). This segment includes three of the VCU River Lab’s weekly sampling locations, which are also monthly sampling sites for the CBP monitoring (JMS75, JMS69 and JMS56). I collaborated with co-authors to collect data for this chapter which include CHLa and PON in fish gut contents (monthly in 2012), fish abundance (3 surveys in 2012-2013), Rangia grazing rates (monthly in 2012), and zooplankton abundance (bi-monthly in 2013) (Table 3-1). Other data included in this analysis include daily river discharge values which were used to determine CHLa export (from USGS gauges on the James and Appomattox Rivers), primary production and respiration rates which were used to derive CHLa production and algal N demand (using diel oxygen data from VCU Rice Center) and annual benthic surveys of Rangia abundance (from CBP). Finally, this chapter includes results from 2 mesocosm experiments performed with variable nutrient loading rates and grazer treatments (June and July 2013).

Consumption of CHLa and PON by grazers In order to determine the amount of CHLa and PON consumed by grazers in this system we estimated per capita grazing rates and population abundance of the most common grazers in this system. The grazers chosen for this analysis include several planktivorous and benthivorous fish, Wedge Clams (Rangia cuneata) and zooplankton. These grazers were chosen based upon abundance in this system and expected trophic position (primary consumers).

Fish Fish abundance was estimated using low and high frequency electrofishing. Five to ten randomly located 500 m transects were sampled in each of three areas: City Point, Rice Center, and Tar Bay (all within 1 km of JMS75). The transects encompassed both main channel and shallow water habitats with separate transects for high vs. low frequency electrofishing. Low frequency electrofishing was used to estimate Blue Catfish abundance; high frequency electrofishing was used for all other species. The electrofishing boat traversed the 500 m transect

equipped with a fixed 5-m cross-section to delineate the sampling area. During high frequency electrofishing, all fish within the reference area were collected for identification. For low frequency electrofishing, catfish appearing within the 5-m span were counted and assigned to one of three size categories (<20 cm, 20-40 cm, and > 40 cm). For each transect, 10 randomly selected fish of each species were measured for length and weight. Sampling was conducted in September 2012, June 2013 and August 2013. Fish abundances were estimated based on sampling area (2500 m2 per transect) and volume (the product of transect area and depth). As previously described in chapter 2 I analyzed gut contents of numerically-dominant fish species of the tidal-fresh James: Gizzard Shad (YOY and adult), Threadfin Shad, Atlantic Menhaden and juvenile Blue Catfish (<20 and 20-40 cm). Here we also present carbon and nitrogen analyses of dietary materials. C and N content was determined on dried and acid-fumed samples using a Perkin-Elmer CHN Analyzer. We derived monthly estimates of CHLa and PON consumption based on fish abundance, gut contents analysis and assumed gastric evacuation rates. We used the PON content of gut materials to estimate fish N consumption. For CHLa, w did not use the CHLa content of gut materials to directly estimate consumption because prior work has shown appreciable degradation (loss) of CHLa from the foregut to the hindgut (Friedland et al. 2005). Instead, we used the measured organic matter content of dietary materials (C%) along with the measured CHLa:C ratio of suspended particulate matter in the James. The CHLa:C ratio was derived from the slope of a regression relating paired measurements of CHLa and POC using bi-monthly data from three stations (JMS75, JMS69 and JMS56) obtained during March-October 2012-13 (CHLa:C = 12.2 ± 0.8 µg:mg; N = 108, R2 = 0.70, p<0.0001). The CHLa content of suspended particulate matter was appreciably higher than that of gut contents from planktivorous fishes (0.17 – 0.23 µg:mg) thereby resulting in higher estimates of the amount of CHLa consumed than those derived from CHLa in gut contents. For benthivorous fishes we substituted the CHLa:C ratio of surficial sediments (0.27 µg:mg) to estimate CHLa ingestion. In addition to fish abundance and gut contents data, estimation of ingestion rates requires knowledge of gut clearance or turnover times, which are known to vary among species and food conditions. We considered a range of gastric evacuation rates (4-18 d-1; Shepherd and Mills 1996, Salvatore 1987, Haskell et al. 2013, van der Lingen 1998) to derive a potential range of CHLa consumption and N recycling rates.

Grazing by Zooplankton and Rangia Wedge clams (Rangia cuneata) are the dominant benthic filter-feeders in the tidal freshwater zone of the James and other major sub-estuaries of Chesapeake Bay (Gerritsen et al. 1994; Cerco & Noel 2010). We derived estimates of CHLa and PON consumption by Rangia as a point of comparison for similar data from fish. Rangia grazing estimates were based on measured clearance rates derived using clams and food sources obtained from the James and abundance estimates for the James provided by the CBP (Table 1). Methods for determining Rangia clearance rates were described in chapter 2. We used the long-term (2001-2010) annual average abundance and the average of the monthly measured clearance rates to derive the average CHLa and PON ingestion rate by Rangia. Zooplankton consumption of CHLa and PON was determined based on measured abundance (Table 3-1) and previously-published species-specific per capita clearance rates. Zooplankton abundance in the James was measured bi-monthly during March-September 2013 (total = 13 collections). Triplicate vertical net tows were collected at a single station (JMS75) and retained for subsequent identification and enumeration (within 1 week of collection). For nauplii, copepods and cladocerans, samples were obtained with a 64 µm mesh size, 0.5 m diameter net. Dead animals, as indicated by aniline blue staining, were not included in final abundance estimates used to calculate grazing (Elliott et al. 2009; 2010). For rotifers, a 20 L depth-integrated sample was passed through a 20 µm mesh net and preserved in Lugol's until analysis. Clearance rates were obtained from the literature for the dominant species occurring in the James: rotifers = 0.38 ml ind-1 d-1 (Sierszen & Frost 1990), Bosmina = 1.63 ml ind-1 d-1 (Sierszen & Frost 1990), Eurytemora adults and copepodids = 9.67 ml ind-1 d-1 (Sierszen & Frost 1990), and copepod nauplii = 2.00 ml ind-1 d-1 (Bogdan & Gilbert 1984).

Comparisons to Other Fluxes To place top-down effects in the broader context of factors regulating CHLa and N in the James, we compared losses due to grazing with other input and output fluxes. We compared CHLa production rates to losses via respiration, grazing and downstream export (advective loss). Over the course of a growing season, CHLa production by phytoplankton should balance losses due to respiration, grazing and advection, assuming no net change in the CHLa content of sediments. Determination of whether these independent estimates of production and loss are in balance provides a means for assessing our understanding of sources and fate of phytoplankton

production in this system. For N, we compared phytoplankton DIN demand against external inputs (from the watershed and direct point source inputs) and internal re-generation via microbial- and consumer- mediated nutrient recycling. CHLa production (µg L-1 d-1) was derived from measurements of daily NPP and an empirically-determined CHLa:C ratio for the James (see above). Phytoplankton DIN demand was similarly derived from NPP and the Redfield C:N ratio. Daily primary production was calculated from diel oxygen data measured at the VCU Rice Center (near JMS75) during March- November of 2012 and 2013. Use of a single-station diel method to derive metabolism estimates is potentially complicated by tidal influences on local O2 concentrations. After de-trending the diel data for a 24-h cycle we did not find that tidal intervals explained a significant fraction of the residual variation and therefore concluded that a single-station method was appropriate for this site. Methods of computation were previously described by Bukaveckas et al. (2011) and follow those used by Caffrey (2003, 2004) for National Estuarine Research Reserve sites. Dissolved O2 concentrations recorded at 15-min intervals were smoothed to 30-min averages for flux analyses.

We derived daily NPP from the sum of daytime O2 fluxes, daily R from the sum of nighttime O2 fluxes extrapolated to 24 h, and daily GPP from NPP plus R occurring during daytime. Derivation of NPP and R from diel data requires correction for atmospheric exchange. Air–water

O2 fluxes were calculated assuming that atmospheric exchange varied only in response to dissolved O2 saturation. This approach uses a fixed exchange coefficient which yields a −2 −1 potential range of −0.5 to +0.5 g O2 m h for 0–200% saturation. We had previously compared these fixed estimates to those corrected for variable wind speed (Marino & Howarth 1993) and −2 −1 found that the latter yielded a similar range of exchange values (−0.3 to +0.3 g O2 m h ) for wind speeds observed at this site. O2-based NPP and R values were converted to C assuming a photosynthetic and respitory quotient of 1.0. Determination of respitory losses of CHLa requires partitioning total respiration into fractions supported by autochthonous production and allochthonous inputs. We used the method of del Giorgio and Peters (1994) whereby the y-intercept for the regression of R vs. NPP is taken to represent the fraction of R that is supported by allochthonous inputs (i.e., R at NPP=0). The regression was based on pooled 2012 and 2013 data and yielded an estimate for allochthonous-R -2 -1 2 of 2.0 ± 0.19 mg O2 m d (R = 0.51, p < 0.0001) (Figure 3-1). When subtracted from total-R -1 -1 -1 -1 (mean = 5.6 mg O2 L d ) this provided an estimate of autochthonous-R of 3.6 mg O2 L d .

This value was used along with the previously-referenced C:CHLa ratio to estimate daily loss of CHLa due to respiration. Internal re-generation of DIN via microbial decomposition was derived similarly except that we used the C:N content of autochthonous (Redfield = 6.6) and allochthonous (C:N = 15.1) of organic matter in combination with autochthonous-R and allochthonous-R. The allochthonous C:N ratio was determined from seston samples collected during periods of high discharge and low CHLa at stations JMS75, JMS69 and JMS56. We estimate that autochthonous contributions to these samples were less than 10% based on POC and C:CHLa. Lastly, we considered export losses of CHLa via advection downstream and DIN inputs from external sources. DIN loads to the tidal fresh segment include riverine inputs at the Fall Line as well as local point source discharges; these were previously estimated to be 0.125 mg DIN L-1 d-1 (Bukaveckas & Isenberg 2013). CHLa export losses were derived as the net difference between input and output fluxes. We used discharge data from USGS gauges on the James and Appomattox Rivers (#2037500 and #2041650, respectively) and measured CHLa concentrations at a station located above our study reach (JMS99) to derive input fluxes. We used the average CHLa concentrations of three stations located within our study reach (JMS75, JMS69 and JMS56) in combination with the discharge data to estimate output losses. Concentrations were measured ~weekly at all stations during March-November of 2012 and 2013. We interpolated daily values for concentration based on regressions relating CHLa to discharge. Daily input and output fluxes were summed and converted to average daily volumetric values (µg L-1 d-1) that represent the net gain or loss of CHLa from the study reach.

Stable Isotope Analysis In combination with C isotopic values for seston and consumers presented in chapter 2, we also obtained N15 isotopic ratios to determine trophic levels of consumers. We used these results to estimate autochthonous and allochthonous contrinutions to metazoan production. We analyzed seston samples from 2011-2012 with high and low CHLa:TSS ratios for C and N isotopic composition. Seston samples with high CHLa:TSS ratios (mean = 2.28 ± 0.14 µg CHLa : mg TSS) represent on average 60% carbon contribution from phytoplankton based on C:CHLa ratios. Seston samples with low CHLa:TSS (mean = 0.09 ± 0.02 µg CHLa : mg TSS) represent on average 10% carbon contribution from phytoplankton. A simple mixing model was then used to determine isotopic signatures of phytoplankton and terrestrial end-members. In addition

to the consumer isotopic values presented in Chapter 2, here I also consider Zooplankton and Blue Crabs. Zooplankton were collected from the VCU Rice Pier in July and August of 2013 and picked for stable isotope analysis. Zooplankton larger than 64 µm and smaller than 64 µm were submitted separately. Zooplankton isotopic signatures did not show variation between size classes and as a result the data was grouped. Muscle tissues of blue crabs were collected during the 2012 field season. Trophic level estimates were determined by considering the number of δ15N trophic enrichments between primary energy sources and consumers. Due to non-trivial variation in δ15N of end members, we used species specific interpolations of energy sources to estimate basal δ15N. To estimate N levels of composite energy sources we considered the intersection between 2 lines. The first line represents species-specific consumer tissues and estimates of food they consume (1 trophic enrichment below). The second line represents both basal end members and thus potential composite values. The δ15N level at the intersection of these two points was used to deduce trophic levels ((δ15N of Consumer - δ15N at intersection)/3.4 δ15N). Previous studies have suggested δ15N enrichment may occur in the sediments as a result of de-nitrification (Seitzinger et al. 2006). To avoid associated trophic level inflation, we used sediment based δ15N for our terrestrial δ15N end-member. In order to determine proportion autochthonous and allochthounous support of consumers we used estimated end members with SIAR, a bayesian isotopic mixing model (http://cran.r-project.org/web/packages/siar/index.html, Parnell et al. 2010). The average trophic fractionation factors used for C and N were set at 0.4 ± 0.2 % and 3.4 ± 1.2 % respectively (McCutchan et al. 2003).

Mesocosm Experiments We performed 2 mesocosm experiments (June and July 2013) to determine the effects of consumers at variable nutrient loading rates upon phytoplankton growth and nutrient limitation. We also monitored zooplankton abundance, Microcystis gene copies and Microcystin concentrations over the course of 15 days. Sixteen circular tanks (1.8 m in diameter) were filled from the James River to a depth of 1 m. Tanks were covered to provide light conditions of 4.8 E m2 d which would be consistent with a 2.5 meter deep water column at typical James River light attenuation. Treatments included Control, Rangia (5 ind. tank-1), planktivorous fish (1 YOY Gizzard Shad, 2 Threadfin) and sediment (10 8-inch buckets). Sediment was collected from near shore areas of the James River using an Ekman sampler and placed at the bottom of the tanks.

The amount of sediment added was intended to represent a water volume to sediment surface area ratio equivalent to a 2.5 m depth. The highest nutrient loading rate (+0.12 mg DIN L-1 d-1,

+ 0.03 mg PO4) for mesocosm experiments was chosen to represent the average volumetric loading rates to the James River (Bukaveckas & Isenberg 2013, Wood & Bukaveckas 2014). Each treatment was performed at 4 different nutrient loading rates which included no nutrient -1 -1 addition, +0.04/0.01, +0.08/0.02, 0.12/0.03 mg DIN/PO4 L d . DIN additions consisted of

65% NH4 and 35% NO3 which is representative of DIN inputs to this system (Bukaveckas & Isenberg 2012). Nutrient additions were performed every three days immediately following sample collection. CHLa, POC, PON, and TSS were monitored every 3 days while zooplankton abundance, nutrients (TN, TP, NH4, NO3, and PO4), Microcystis gene copies, and Microcystin were monitored initially and on days 6 and 15. CHLa, POC, PON and TSS analyses follow methodology described in Chapter 1 while Microcystin toxin analysis and Microcystis genetic analyses follow methodology described in Chapter 2. Zooplankton were collected by filtering 2 L of mesocosm water from a (2-L Kemmerer water sampler) through a 23 µm mesh sieve and preserving in alcohol. Zooplankton were counted following previously described methods. An analysis of covariance (ANCOVA) was used to test for significant nutrient, grazer and interaction effects. Statistical analyses were performed using JMP Pro 10. At the conclusion of both experiments, samples were collected from each Control tank, -1 -1 -1 -1 the low (0/0 mg DIN/PO4 L d ) Shad tank, the high (0.12/0.03, mg DIN/PO4 L d ) Shad tank and the low Rangia tank to determine nutrient limitation status. For this analysis we used the design described in Chapter 1. Bioassay cultures comprised a 150 mL solution in a 250 mL Erlenmeyer flask containing 50% unfiltered and 50% filtered mesocosm water (0.5 μm Whatman GF/A glass filter). Cultures were diluted in order to reduce algal densities below equilibrium to measure algal growth responses (Sterner and Grover 1998). Control and nutrient treatments were replicated twice for each tank monitored. Enrichments represented 0.125 mg L-1 each of -1 NO3 and NH4 and 0.1 mg PO4 L . Cultures were incubated on a shaker table at ambient (river) temperature inside a Conviron growth chamber for 48 h. Cultures were incubated at 10 E m-2 d-1 which represents the average daily irradiance experienced by phytoplankton circulating through the entire water column at a depth of 1.5 m taking into account typical summer solar radiation -2 -1 (~40 E m d ; Fisher et al. 2003) and underwater light attenuation during summer (mean kd = 3.14 ± 0.33 m-1 based on monthly measurements made in conjunction with bioassay

experiments). Light conditions within the incubator were modified by shade cloth and proximity to light sources and verified with a Li-Cor photometer. CHLa, POC, PON and nutrient analysis was performed on all initial and final bioassays per previously described analytical methods.

RESULTS

CHLa and PON fluxes

Fish Our analysis of fish abundance and their diet shows that planktivorous fishes, particularly Threadfin Shad, were the dominant consumers of CHLa among fishes in the tidal fresh James (Figure 3-2). Planktivorous fish had higher CHLa content in their diet, while benthivorous fish had low CHLa in their diet (Chapter 2). Threadfin Shad was the most abundant taxa accounting for almost half of total fish density (Figure 3-2). Standard errors derived from fish density estimates across the three sampling periods were small ranging 9 to 22%. Threadfin Shad accounted for 65% of the total CHLa consumption by fish despite their small individual size (6.5 g ind-1). YOY Gizzard Shad were the second most important contributor accounting for 25% of fish CHLa consumption. The combined CHLa consumption by benthivorous fishes was less than 3% of the total, despite accounting for 90% of fish biomass. Though not important as consumers of CHLa, benthivorous fishes, particularly adult Gizzard Shad, accounted for a large proportion of PON ingestion (Figure 3-3). N content of diet materials was similar among all fishes (mean = 36.2 mg PON g DW-1; range = 31.1 to 49.1 mg PON g DW-1) and therefore relative contributions to N recycling were largely determined by per capita ingestion rates (a function of body size) and numerical abundance. The large body size of adult Gizzard Shad (mean = 386 g ind-1) yielded a proportionally large mass of gut contents and the highest per capita PON ingestion rate. For the range of gut turnover times considered (4 - 18 d-1), we estimate that adult Gizzard Shad consume 0.33 to 1.48 mg PON ind-1 d-1. Adult Gizzard Shad were also the most abundant group by biomass (3.9 g m-3) accounting for 60% of total fish biomass. This resulted in a population-scale estimate of PON consumption for adult Gizzard Shad that was highest among the 6 taxa-size groups.

Grazing by Rangia and Zooplankton Rotifers and wedge clams (Rangia) were the dominant consumers of CHLa and PON among invertebrates (Figure 3-4). Rotifers (principally Brachionus calyciflorus) were important consumers despite their low per capita grazing rates (0.38 ml ind-1 d-1) due to their high densities (mean = 154,000 ± 13,500 ind m-3) which yielded consumption estimates of 1.1 µg CHLa L-1 d-1 and 0.02 mg PON L-1 d-1. Rangia abundance was orders of magnitude lower (mean = 10 ± 2 ind

m-3) but their clearance rates were correspondingly higher (7,400 ml ind-1 d-1) which resulted in similar consumption rates (1.4 µg CHLa L-1 d-1 and 0.021 mg PON L-1 d-1) to rotifers. Other zooplankton (Bosmina, Eurytemora, copepod nauplii) did not contribute appreciably to CHLa or PON consumption. Combined rates of CHLa consumption for invertebrates (2.7 µg L-1 d-1) were 8 times higher than those for fish (0.36 µg L-1 d-1) even when maximal gut clearance rates were used for fish. The combined PON consumption by invertebrates (0.045 mg L-1 d-1) was also higher than that of fish (0.027 mg L-1 d-1) though only by a factor of 2 due to the large contributions by benthivorous fish.

Grazing in Relation to Other Sources and Sinks Daily average CHLa production in the James was 20.1 µg L-1 d-1 during March- November of 2012 and 2013 (Figure 3-5). The dominant fate of this production was microbial decomposition with respitory losses accounting for 14.8 µg L-1 d-1 (74% of production). By comparison, losses due to the combined grazing of fish, zooplankton and Rangia were 3.1 µg L-1 d-1 (15% of production). Export losses were also small (0.9 µg L-1 d-1) corresponding to 4% of production. The sum of all three losses accounted for 93% of the estimated CHLa production. For N, we estimated daily phytoplankton demand was 0.290 mg L-1 d-1. The combined PON ingestion rate for all grazers was 0.072 mg L-1 d-1 suggesting that consumer-mediated recycling could contribute 25% of daily phytoplankton N demand (assuming all ingested PON was excreted in available form). By comparison, microbial-mediated N cycling from a combination of allochthonous and autochthonous sources (0.291 mg L-1 d-1) could potentially account for 100% of daily phytoplankton N demand. External inputs of DIN (0.125 mg L-1 d-1) were equivalent to 39% of daily phytoplankton demand.

Stable Isotope Analysis Seston samples with large algal contributions (high CHLa:TSS) were depleted in δ13C‰ (mean. = -28.5 ± 0.2 ‰) and enriched in δ15N‰ (mean = 6.2 ± 0.2) relative to seston collected during periods of low algal production (δ13C‰ =26.8 ± 2.0, δ15N‰ = 2.0 ± 0.4) and to sediment samples (δ13C‰ 26.7 ± 0.1 ‰, δ15N =5.4 ± 0.1‰) (Figure 3-6). Sediment samples and high TSS:CHLa exhibited similar δ13C‰ estimates but δ15N‰ differed by 3.4 ‰δ15N. Using proportion of algal carbon and δ13C‰ of these two groups we derived the autochthonous

13 15 13 15 (δ CAuto. = -29.8‰, δ NAuto. = 8.8‰) and allochthonous (δ CAlloch. = -26.5‰, δ NAlloch. =1.2‰) end members which were used in the Bayesian isotopic mixing model (SIAR).

Planktivorous fish which exhibited higher CHLa contribution to diet had more depleted δ13C‰ values (δ13C =27.8 ± 0.3‰) which were similar to suspended particulate organic matter during algal blooms (Table 3-2). These fish had an average trophic level of 1.6 (Table 3-2). Benthivorous fish, which exhibited lower CHLa contribution to diet, had enriched δ13C signatures and were more similar to sediment and SPOM during low algal abundance. Benthivorous fish had an average trophic level of 2.3. Blue crabs and Rangia were similar to benthivorous fish in isotopic signature and trophic level (Table 3-2). These signatures were similar to our projected phytoplankton end-member and had an average trophic level of 1.0. SIAR results indicate that on average 60% and 95% of planktivorous fish and zooplankton biomass were supported by phytoplankton while only 10% of other species (Blue Catfish, adult Gizzard Shad, Rangia and Blue Crab) were supported by phytoplankton. Blue crabs (4.1%) and adult Gizzard Shad (2.3%) received the smallest autochthonous contribution. Biomass weighted averages of autochthonous contributions indicate 27% of metazoan biomass (with the exception of blue crabs of which biomass was not estimated) was supported by phytoplankton (Table 3-2). Of the biomass fraction supported by autochthonous material, Rangia and zooplankton contributed 90%.

Mesocosm Experiments We present results of two mesocosm experiments performed in June and July 2013 to describe potential influences of grazers (Fish and Rangia) upon phytoplankton communities, zooplankton abundance and nutrient dynamics. Principal component analysis using correlation matrices was performed on variables for June and July results to assess overall patterns of variation among treatments (Figure 3-7). Variables included POC, PON, DON, DIN, TP, CHLa, Copepods, Rotifers, Microcystin and Microcystis gene copies. The combination of principal components 1 and 2 explained 69% and 70% of the variation for June and July data sets respectively. Results indicate that both nutrient loading and treatment effects affected conditions within the tanks. In June, the largest variation occurred between tanks with Shad and Control tanks. In this experiment, axis 1 was positively related to CHLa and negatively related to Copepods, while nutrient and zooplankton contributed to both axes. In July, tanks with grazers

(Shad and Rangia) were clustered together and tanks without grazers clustered together (Control and Sediment). Axis 1 was related to CHLa and Microcystis gene copies while copepods were strongly related to axis 2 while other variables contributed to both axes. For this experiment, differences were greatest at the 2 higher nutrient loading rates (Figure 3-7).

CHLa & Nutrient Results CHLa responded to nutrient loading rates in both experiments (Table 3-3 and Table 3-4). ANCOVA analyses indicated nutrients had a positive effect upon CHLa in both months (Table 3-5). Significant grazer treatment effects also occurred in both experiments whereby CHLa concentrations in tanks with shad (June: 69 ± 11, July: 109 ± 9 µg L-1) were significantly higher than all other tanks (Table 3-5, Figure 3-8, Figure 3-9). CHLa concentrations were variable throughout the 15 day experiments (average coefficient of variation = 57%) but values were significantly larger at higher nutrient loading rates (Table 3-3 and Table 3-4).

NOx, NH4, TN, PO4 and TP all showed significant response to nutrient loading rates. DIN accumulation in mesocosms was typically lower than expected values based on the amounts added DIN addition rates across treatments; this difference was greatest in tanks without grazers

(Control & Sediment) (Figure 3-8, Figure 3-9). In June, tanks with shad accumulated more NH4 than Control tanks (Table 3-3). In July, Rangia and Shad tanks accumulated significantly more

DIN and NOx than Control tanks (Table 3-4). In contrast, Shad and Rangia tanks did not accumulate more PO4 than Control tanks in either month.

Response of Zooplankton communities In the June experiment, tanks were filled during a period of high zooplankton abundance and as a result initial abundance values were 4, 2 and 2 times higher than average annual abundance of Rotifers, Copepod Nauplii and Cladocerans respectively. Initially, Rotifers were abundant relative to Copepods and Cladocerans (Table 3-3). After the experiment rotifer abundance declined on average by 75% in all treatments but did not vary by treatment (Table 3- 5). In contrast, Copepods and Cladocerans increased on average by 6 and 30 times respectively. Copepods were the only group to experience a response to nutrient loading (negative). Cladocerans were the only group to experience a significant grazer effect whereby Shad tanks (5 ± 1 ind L-1) were depleted relative to Control tanks (112 ± 17 ind L-1).

In the July experiment, initial Rotifer and Cladoceran concentrations were less than half of June experiments while Copepod nauplii abundance was nearly 4 times higher than June (Table 3-4). On average, Rotifer abundance was similar between initial and final values, while Cladoceran and Copepod concentrations increased by 10 and 2 fold respectively. None of the zooplankton groups experienced nutrient loading effects in July. Copepods did decline in response to Shad (all Shad tanks has 0 copepods on day 15) (Table 3-5). To consider the net effects of changes in zooplankton community structure on phytoplankton, we estimated zooplankton community clearance rates by applying literature estimates of grazing rates with measured abundances as was performed for James River abundances. Clearance rates for tanks without shad (mean = 22 ± 2% d-1) were high in both experiments in comparison with values derived based on abundances in the James (7 ± 3 % d-1). This result occurred in part because tanks were filled during a period of high zooplankton abundance but also due to increasing zooplankton abundance throughout the experiments. Shad tanks experienced lower clearance rates (mean = 10 ± 1% d-1) though these levels were still higher than values derived for the James.

Micorycstis and Microcystin In June, initial Microcystin concentrations (0.08 µg L-1) and Microcystis gene copies (72 copies mL-1) were low (Table 3-3, Figure 3-8). These parameters stayed low in all treatments over the course of the June experiment with Microcystin never exceeding 0.1 µg L-1. Despite, these low levels we detected a significant difference between Control and Shad tanks where tanks with shad had lower Microcystin concentrations. In July, initial Microcystin concentrations (0.08 µg L-1) were also low but Microcystis gene copies (1800 copies mL-1) were large relative to June. Both Microcystin and Microcystis gene copies exhibited significant interaction effects associated with nutrient loading and treatment effects (Table 3-5, Figure 3-9). In tanks with shad, Microcystin (average = 0.12 ± 0.03 µg L-1) and Microcystis gene copies (253 ± 154 copies mL-1) were low at all nutrient loading rates. In Control tanks, Microcystin and Microcystis gene copies increased with increasing nutrient loading rates (Figure 3-9). At the two higher nutrient loading rates, Microcystin and Microcystis gene copies in tanks without grazers were respectively 2 and 3 orders of magnitude greater than tanks with shad. Microcystin and Microcystis gene copies normalized per unit CHLa (µg Microcystin:µg CHLa) were also higher in tanks without grazers as these tanks had lower CHLa concentrations. Tanks with Rangia

exhibited intermediate values for both Microcystin and Microcystis, though these were not significantly different from Control or Shad tanks (Figure 3-9).

Nutrient Limitation Bioassays The severity of nutrient limitation, as indicated by the nutrient effect size (natural logarithm of [CHLaamb]:[CHLa+PN]) was lowest in tanks at higher nutrient loading rates (Table 3-3, Table 3-4). In tanks with no nutrient additions, the severity of nutrient limitation was reduced in the presence of grazers (Rangia and Shad treatments). Patterns in nutrient limitation followed differences in nutrient availability which were variable for different treatments. Initial DIN concentrations for bioassays from Control tanks at lower nutrient loading rates were lower than initial concentrations in tanks with grazers as a result of higher DIN values at the end of the experiments (Table 3-5). DIN uptake was negative in bioassay cultures from the two lowest nutrient loading Control treatments while DIN uptake rates of 0.06 ± 0.01 and 0.08 ± 0.01 mg N L-1 d-1 were observed in the bioassay Control treatments of the lowest nutrient loaded Rangia and Shad tanks (Figure 3-10). In the nutrient enriched bioassays from Control mesocosms that did not receive nutrient loading, 90% of available DIN was assimilated.

DISCUSSION Our findings suggest weak top-down effects of grazers via algal consumption and N recycling when considered in the context of other fluxes in the tidal fresh James River. The combined grazing rate by zooplankton, benthic filter-feeders and fish was small (3.1 µg L-1 d-1) corresponding to 15% of average daily CHLa production. Assigning error to this estimate of CHLa consumption is problematic given the difficulties in deriving propagated errors from multiple sources. For fish, potential sources of error include intra-specific variation in algal contributions to diet, variation in fish abundance and uncertainty regarding gut turnover rates. Our largest source of uncertainty was the gut turnover time for which we lack empirical data specific to this study and therefore incorporated a large range of potential values (~4-fold) to assess their effect on CHLa consumption rates. However, our results were robust in that CHLa ingestion by fish was low (0.36 µg L-1 d-1) even when a maximal turnover rate was used (18 d-1). Our data suggest that the other two sources of error were small. Intra-specific variation in algal contributions to diet was low as indicated by the small standard error (<20%) relative to the means derived from monthly values for each of the species. Similarly, standard errors for fish

abundance estimates were within 20% of the mean derived for the three sampling periods. Our three fish samplings were conducted over a period of 12 months and therefore would not be representative of longer-term variability in fish abundance that might be associated with year-to- year variation in recruitment or survivorship. There is little comparative data with which to assess the accuracy of our fish abundance estimates except that our estimate of blue catfish density (511 ind. ha-1) showed agreement with an estimate derived through mark-recapture methods (681 ind. ha-1; R. Greenlee, VDGIF, pers. comm..). Our estimates of adult Gizzard Shad biomass (117 kg ha-1) was also within the range of previously published values (20-400 kg ha-1) (Shostell and Bukaveckas 2004, Vanni et al. 2005).

The dominant fate of algal production in the tidal fresh segment of the James was in supporting microbial decomposition with a small fraction entering the grazer food chain directly and a smaller fraction lost through downstream export (Figure 3-5). Our results are dependent on derived estimates of the proportion of ecosystem respiration which is driven by allochthonous organic matter. Though statistical uncertainty in this estimate is small (<10%) we lack empirical data to assess the validity of this method. Our estimates of the proportion of respiration driven by allochthony and autochthony yield estimates of CHLa losses which in combination with grazer and advection losses account for 93% of algal production in this system. Advective and grazing losses were measured independently of algal respiration thus providing support for our results. Finally, there is potential error associated with our assumed algal C:N ratio which was used to estimate algal N demand. For this ratio, we considered three potential values including the slope of C:N for riverine seston samples (4.9), a C:N ratio associated with bioassays performed in 2012 (9.9), as well as the Redfield ratio (6.7, presented in results). While other C:N estimates would have influenced the proportion of N supported by advection and grazers, all of these estimates would still yield the conclusion that microbial decomposition is the dominant source of recycled N.

Grazers were somewhat more important for their role in recycling nitrogen than CHLa, potentially contributing 25% of daily phytoplankton demand. Their greater contribution to N recycling vs. direct consumption of CHLa stems from the fact that phytoplankton account for a relatively small proportion of particulate matter in the James. Non-algal materials have low CHLa content but are relatively rich in N; therefore grazers ingesting a mixed diet will consume algal and non-algal N. This was most apparent in benthivorous fishes whose detritus-based diet 71

was low in CHLa but comparable to planktivores in N content. Because benthivorous fishes are important in the James (90% of fish biomass), fish accounted for a greater proportion of grazer PON ingestion (37%) than for CHLa (11%). Rates of N recycling by Gizzard Shad in the James (0.0033 - 0.015 mg N L-1 d-1) were similar to values reported in mid-western reservoirs (Shostell and Bukaveckas 2004: 0.0025, Vanni et al. 2006: 0.0024 - 0.0375 mg N L-1 d-1). Despite the similarity in N recycling rates, the proportion of phytoplankton N demand met by Gizzard Shad is much smaller in the James (1-5%) in comparison with these other systems (Shostell and Bukaveckas = 43%, Vanni et al. 2006 = 27-67%). This is primarily due to high levels of primary production and algal nitrogen demand in the James relative to these other systems.

Thus our findings suggest that top-down effects, though weak, favor eutrophic conditions in the James by supporting algal blooms through the recycling of N while removing a relatively small proportion of CHLa through ingestion. Mesocosm experiments further support the idea that grazers favor eutrophic conditions in that tanks with shad exhibited higher CHLa concentrations and higher DIN concentrations. While the proportion of algal demand which is met by consumer recycling is small, these fluxes of nitrogen are large (64%) when compared with external nutrient loads (Bukaveckas and Isenberg 2013). Declines in external nutrient loading rates in this system (Wood & Bukaveckas 2014) and other systems (Carstensen et al. 2011) have resulted in minimal changes in phytoplankton abundance. Microbial- and consumer- mediated N cycling may explain the lack of response to reductions in external nutrient loads.

Our analysis of consumer abundance and grazing suggests weak effects via removal of algal biomass and somewhat stronger effects through nutrient cycling. These findings were supported by results from mesocosm experiments which showed greater algal abundance in tanks with grazers. Grazer effects were observed at all levels of nutrient addition but were greatest in mesocosms receiving low nutrient loads. Thus the experiments suggest that reductions in external nutrient loads to the James will be accompanied by increasing importance of consumer- mediated nutrient recycling which will offset or delay recovery. In addition, the experimental findings suggest that consumers influence other aspects of trophic state and specifically, the occurrence of harmful cyanobacteria. We observed lower abundance of Microcystis in the presence of fish. Microcystis gene copies and Microcystin concentrations both declined significantly in the presence of shad relative to tanks without grazers across nutrient loading rates. Given that CHLa levels were elevated in tanks with shad, the proportion of algae which 72

was composed of Microcystis experienced even stronger declines. This result was unexpected given other studies have suggested Microcystis can transverse the gut of fish and potentially benefit from gut passage by assimilating nutrients (Friedland et al. 2005, Jančula et al. 2008). We speculate that the most likely mechanism to explain the negative effect of shad upon Microcystis is through a trophic cascade associated with zooplankton. Previous studies have suggested that Microcystin deters zooplankton grazing (Davis and Gobler 2011) and thus we suggest in the presence of planktivorous fish, zooplankton populations (and thus zooplankton grazing) are reduced leading to a competitive disadvantage for toxic strains of Microcystis. Most previous studies have emphasized the exacerbation of harmful algal blooms associated with fish consumption, but here we suggest planktivorous fish may negatively influence Microcystis through trophic cascades. Our analysis of Microcystin and Microcystis gene copies in the July mesocosm experiment also suggests that nutrient loading in this system may have qualitative influences upon algal communities. The positive response of Microcystis to nutrient loading supports previous studies which suggest that Microcystis is more common under eutrophic conditions (Moisander et al. 2009, Poste et al. 2011).

The extent to which zooplankton grazing by Rangia and planktivorous fish influences other phytoplankton species is unknown however stable isotope analysis suggests these species occupy the 1.5th trophic level (figure 3-6) and thus receive similar contributions from phytoplankton and zooplankton. These effects could result in reduced net grazing pressure upon phytoplankton due to reductions in zooplankton grazing. Zooplankton exhibit low abundance in this system relative to other estuaries, which we have previously suggested may be due to large contributions of inorganic material to seston (Bukaveckas et al 2011a). These results suggest grazing pressures by fish and clams may also act to constrain zooplankton populations in this system.

Our analysis of stable isotopes in seston, sediment and consumers in combination with biomass estimates indicate that 29% of monitored metazoan biomass is supported by algal production in this system. This is striking given that algal production is high in this system and therefore the prevalence of terrestrial organic matter in supporting higher trophic levels cannot be attributed to a deficit of autochthonous production (Wood and Bukaveckas et al. 2014). The most abundant species (adult Gizzard Shad, Rangia & Blue Catfish) were largely supported by allochthonous organic matter. Planktivorous fish received larger contributions from 73

phytoplankton, but even these fish only received 50-60% of their biomass from algal production. Descriptions of basal sources for metabolic pathways described by the Riverine Productivity Model (Thorp and Delong 1994, 2002) suggest that the majority of microbial respiration in riverine systems is supported by allochthonous carbon but that metazoan biomass is primarily supported by phytoplankton production. Our results contradict this model on both of these points. First, our data suggest the majority (>70%) of microbial respiration in this system is supported by algal material (Figure 3-5). Our estimates are sensitive to our calculation of allochthonous respiration, though even the lower bound estimate for this rate would suggest that autochthonous respiration accounts for over 60% of ecosystem respiration. Secondly, our data suggests only 29% of metazoan biomass is supported by algal production (Table 3-2, Figure 3- 6). Other recent studies have argued terrestrial carbon can account for large proportions of metazoan biomass in riverine food webs in situations where turbidity limits algal production or during high discharge periods which contribute large allochthonous loads (Hoffman et al. 2007, Zeug and Winemiller 2008, Cole and Solomon 2012, Roach 2013). Our results, for a system with high production and dominated by nutritious forms of algae (greens and diatoms; Todd Egerton, pers. comm..), suggests that terrestrial support of food webs can be important even in systems where with high levels of autochthonous production. We suggest the high level of metazoan biomass supported by terrestrial material in this system may be a product of hydrodynamics whereby tidal mixing acts to maintain terrestrial organic matter in suspension, even during periods of low external inputs. Our seston samples collected during peak algal biomass periods were comprised of 40% allochthonous material. In a system with such a large proportion of seston composed of non-algal material, obtaining a diet high in algal material may be difficult. Zooplankton exhibited high algal contribution (96%) indicating strong selectivity for autochthonous organic matter, whereas other suspensions feeders including pelagic fishes and benthic bi-valves relied on allochthonous organic matter to a greater extent. A second factor contributing to higher than expected allochthony in this system was the dominance of benthivorous fishes (juvenile Blue Catfish and adult Gizzard Shad) which feed upon sediment that is dominated by terrestrial organic matter and contribute a large fraction of metazoan biomass.

In summary, this study found that community grazing in the tidal fresh James River consumes only a small fraction of daily algal production and the majority of algal production is

decomposed by microbes. Furthermore, fish and other grazers contribute to eutrophic conditions by recycling N from autochthonous and allochthonous organic matter. Planktivorous fish may also cause trophic cascades which reduce zooplankton grazing and in turn reduce competitive advantages of species that deter zooplankton grazing such as Microcystis. Stable isotope analyses suggest that the over 70% of metazoan biomass was supported by terrestrial organic matter despite high rates of primary production. While previously reported increases in nutrient limitation is a positive sign for this system (Wood & Bukaveckas 2014), the capacity for consumers and microbes to recycle nutrients may curtail recovery associated with nutrient reductions. We suggest that even with large reductions to external nutrient loading rates, improved trophic status may not be apparent for several years or even decades following onset due to recycling of large nutrient legacies.

Tables and Figures

Table 3-1 Data sources used for analysis of top-down effects in comparison to other CHLa and PON fluxes.

Category Data Source Discharge USGS (Stations #2037500, #2041650) Advective losses Water Chemistry VCU River Lab Water Quality Monitoring Zooplankton Clearance Rates Sierzen & Frost (1990), Bogdan & Gilber Zooplankton Abundance David Elliott (co-author) Rangia Clearance Rates This Study Rangia Abundance Dan Dauer Lab, Chesapeake Bay Program Grazer Ingestion Fish Gut Contents This Study Salvatore 1987, Shepherd and Mills 1994, Fish Gut Clearance Rates van der Lingen 1998,Haskell et al. 2013, Fish Abundance This Study Dissolved Oxygen Data VCU Rice Pier Data Metabolism C:CHLa & C:N Stoichiometry VCU River Lab Water Quality Monitoring

Table 3-2 Isotopic signatures, trophic level, biomass and phytoplankton support for common consumers of the tidal fresh James River

Sample N 13C ‰ 15N ‰ Trophic Biomass % Autocthonous Level (g DW m-3) Seston (10% algal) 28 -26.8 ± .06 2.0 ± .36 - - Seston (60% algal) 28 -28.5 ± .20 6.2 ± .22 - - Sediment 18 -26.7 ± .07 5.4 ± .13 - - Zooplankton 6 -29.3 ± .43 10.1 ± .49 0.97 0.2 -0.8 96.0% Rangia 30 -25.8 ± .16 11.6 ± .09 1.86 2.33 23.6% Blue Crab 20 -25.2 ± .31 11.7 ± .28 1.96 - 4.1% Threadfin Shad 29 -28.2 ± .19 11.2 ± .26 1.42 0.06 68.3% Menhaden 15 -27.8 ± .39 12.6 ± .25 1.88 0.02 69.1% YOY Gizzard Shad 28 -27.3 ± .19 10.6 ± .40 1.39 0.03 43.7% Adult Gizzard Shad 28 -25.1 ± .32 11.9 ± .27 2.06 0.98 2.3% Blue Catfish 61 -25.7 ± .08 14.0 ± .14 2.55 0.37 8.7%

Table 3-3 Mesocosm results for water quality parameters from June 2013. Nutrient, Microcystin and MYC (Microcystis specific genes) values represent single values observed at the end of the experiments while CHLa, POC and PON represent values (mean ± SE) from combined 3 day intervals (n = 5). Zooplankton estimates represent average and standard error values for monitoring which occurred on day 6 and day 15.

Loading Rate Treatment CHLa NOx NH4 DIN TN PO4 TP Rotifers Copepods Cladocerans MC myc copies Nutrinet Limitation (µg L-1 d-1) µg L-1 mg L-1 mg L-1 mg L-1 mg L-1 mg L-1 mg L-1 ind L-1 ind L-1 ind L-1 µg L-1 copies mL-1 Effect Size (d-1) Initial N/A 42.5 ± 1.2 0.15 0.05 0.20 0.51 0.05 0.06 632 ± 104 3 ± 1 2 ± 0 0.08 72 nd 0.00 N, 0.00 P Control 5.4 ± 2.0 0.02 0.02 0.04 0.27 0.04 0.02 98 ± 3 52 ± 37 106 ± 68 0.09 0.03 2.07 Rangia 6.5 ± 2.3 0.06 0.01 0.07 0.27 0.06 0.04 149 ± 78 36 ± 4 17 ± 10 0.07 0.02 1.71 Shad 24.7 ± 4.9 0.11 0.09 0.20 0.48 0.06 0.04 161 ± 90 9 ± 9 6 ± 0 0.05 0.00 0.73 Sediment 7.8 ± 2.5 0.10 0.02 0.12 0.24 0.05 0.01 131 ± 59 61 ± 42 101 ± 68 0.07 0.01 1.11 0.04 N, 0.01 P Control 13.1 ± 3.1 0.02 0.03 0.05 0.39 0.10 0.10 114 ± 36 29 ± 28 156 ± 87 0.08 0.00 1.12 Rangia 17.8 ± 2.2 0.02 0.01 0.03 0.31 0.10 0.09 82 ± 10 27 ± 2 53 ± 25 0.07 0.00 nd Shad 35.7 ± 11.7 0.37 0.07 0.44 0.69 0.17 0.17 98 ± 7 4 ± 1 3 ± 1 0.05 1.11 nd Sediment 15.5 ± 4.7 0.05 0.01 0.06 0.29 0.07 0.05 101 ± 28 2 ± 1 80 ± 55 0.06 0.00 nd 0.08 N, 0.02 P Control 17.5 ± 7.2 0.65 0.02 0.67 0.90 0.27 0.27 210 ± 138 16 ± 13 110 ± 62 0.08 0.11 0.72 Rangia 20.1 ± 16.1 0.66 0.04 0.70 1.01 0.26 0.28 233 ± 78 30 ± 13 60 ± 30 0.07 0.00 nd Shad 45.4 ± 20.2 1.00 0.03 1.04 1.31 0.30 0.34 107 ± 10 3 ± 0 8 ± 2 0.07 0.00 nd Sediment 40.2 ± 18.3 0.18 0.00 0.18 0.45 0.14 0.15 142 ± 134 3 ± 3 8 ± 5 0.08 0.00 nd 0.12 N, 0.03 P Control 25.8 ± 11.3 1.24 0.02 1.26 1.45 0.43 0.44 105 ± 1 10 ± 2 75 ± 37 0.07 0.00 0.08 Rangia 39.7 ± 10.0 1.22 0.03 1.25 1.50 0.48 0.46 151 ± 14 14 ± 0 98 ± 39 0.09 0.00 nd Shad 59.9 ± 34.1 2.32 0.03 2.35 2.66 0.58 0.68 256 ± 62 7 ± 0 5 ± 5 0.07 0.00 0.59 Sediment 27.5 ± 8.7 1.20 0.02 1.22 1.47 0.40 0.40 119 ± 7 3 ± 1 77 ± 69 0.09 4.66 0.31

Table 3-4 Mesocosm results for water quality parameters from July 2013. Nutrient, Microcystin and MYC (Microcystis specific genes) values represent single values observed at the end of the experiments while CHLa, POC and PON represent values (mean ± SE) from combined 3 day intervals (n = 5). Zooplankton estimates represent average and standard error values for monitoring which occurred on day 6 and day 15.

3 Loading Rate Treatment CHLa NOx NH4 DIN TN PO4 TP Rotifers Copepods Cladocerans MC myc 10 Nutrinet Limitation -1 -1 (mg L d ) µg L-1 mg L-1 mg L-1 mg L-1 mg L-1 mg L-1 mg L-1 ind L-1 ind L-1 ind L-1 µg L-1 copies mL-1 Effect Size (d-1) Initial N/A 52.6 ± 2.8 0.10 0.02 0.12 0.58 0.04 0.04 276 ± 64 11 ± 2 1 ± 0 0.07 1.9 nd 0.00 N, 0.00 P Control 32.0 ± 18.3 0.00 0.01 0.01 0.35 0.03 0.01 375 ± 349 8 ± 8 6 ± 5 0.08 0.0 2.26 Rangia 17.2 ± 4.7 0.00 0.03 0.19 0.58 0.06 0.03 174 ± 155 16 ± 11 10 ± 2 0.04 0.0 0.03 Shad 89.0 ± 16.2 0.10 0.02 0.19 0.77 0.05 0.06 234 ± 209 1 ± 1 1 ± 1 0.18 0.2 0.68 Sediment 16.5 ± 6.3 0.60 0.03 0.29 0.45 0.08 0.01 189 ± 142 43 ± 35 5 ± 1 0.00 0.2 0.26 0.04 N, 0.01 P Control 46.1 ± 11.6 0.10 0.09 0.04 0.48 0.08 0.12 291 ± 249 11 ± 1 13 ± 2 0.39 14.6 1.40 Rangia 79.5 ± 21.3 0.24 0.29 0.53 1.03 0.13 0.18 206 ± 78 22 ± 11 28 ± 22 0.19 1.4 nd Shad 113.0 ± 34.3 0.66 0.20 0.60 1.08 0.18 0.14 235 ± 208 0 ± 0 18 ± 18 0.16 0.0 nd Sediment 94.5 ± 15.0 1.20 0.13 0.15 0.60 0.09 0.09 361 ± 302 54 ± 45 29 ± 17 0.33 0.7 nd 0.08 N, 0.02 P Control 96.1 ± 18.7 0.12 0.07 0.12 0.88 0.17 0.28 414 ± 302 62 ± 57 15 ± 2 0.52 94.5 0.66 Rangia 88.0 ± 51.7 0.54 0.06 0.86 1.67 0.26 0.41 442 ± 6 13 ± 7 9 ± 3 0.24 161.6 nd Shad 99.9 ± 24.0 0.79 0.10 0.89 1.92 0.20 0.33 399 ± 358 0 ± 0 1 ± 1 0.07 0.7 nd Sediment 77.5 ± 23.6 1.37 0.06 0.48 1.13 0.18 0.22 471 ± 148 46 ± 43 4 ± 2 0.47 32.4 nd 0.12 N, 0.03 P Control 101.0 ± 12.6 0.26 0.03 0.63 1.59 0.31 0.53 197 ± 48 17 ± 13 8 ± 5 1.84 270.3 0.08 Rangia 108.9 ± 28.2 0.14 0.01 1.33 2.28 0.41 0.61 863 ± 457 13 ± 8 10 ± 2 0.55 124.0 nd Shad 131.1 ± 17.3 0.38 0.10 1.43 2.57 0.33 0.60 220 ± 100 0 ± 0 2 ± 2 0.08 0.1 0.00 Sediment 76.1 ± 8.6 0.95 0.02 0.97 2.13 0.24 0.37 279 ± 155 87 ± 85 5 ± 5 3.25 13.6 0.27

Table 3-5 ANCOVA results for mesocosm experiments performed in June and July of 2013.

Month Parameter Nutrient Grazer Effect Interaction p Control Sediment Rangia Shad p June 13' CHLa 0.0011 b b b a 0.0192 ns

NH4 ns b b b a 0.003 0.013

NOx 0.0002 - - - - ns ns DIN 0.0002 - - - - ns ns TN <0.0001 - - - - ns ns

PO4 <0.0001 - - - - ns ns TP <0.0001 ab b ab a 0.037 ns rotifer ns - - - - ns ns copepod 0.0098 (-) - - - - ns ns cladoceran ns a ab ab b 0.011 ns MC 0.0256 a a a b 0.0292 0.0098 myc ns - - - - ns ns July 13' CHLa < 0.0001 b b b a 0.0103 ns

NH4 ns b ab a ab 0.027 ns Nox <0.0001 b ab a a 0.016 ns DIN <0.0001 b ab a a 0.003 ns TN <0.0001 c bc ab a 0.003 ns

PO4 <0.0001 ab b a ab 0.031 ns TP <0.0001 b ab a a 0.013 ns rotifer ns - - - - ns ns copepod ns a a a b 0.048 ns cladoceran ns - - - - ns ns MC 0.0085 ab a ab b 0.023 ns myc 0.0037 a ab ab b 0.032 0.0293

Figure 3-1: Daily ecosystem respiration as a function of net primary production in the James River Estuary during 2012- 2013. The Y intercept of the regression line is interpreted as the amount of respiration supported by allochthonous carbon (del Giorgio and Peters 1994).

300

CHLa in Diet

1 -

200

µg CHLa CHLa µg gDW 100

0 15,000 Per Capita CHLa

Ingestion Rate

1 -

d 10,000

- µg ind.µg 5,000 Range of Gut Turnover Times

0 0.05 6 Fish Abundance

) 0.04

)

3 - Density - 4 0.03 Biomass 0.02

2 Biomassm (g Densitym (ind 0.01

0.00 0 0.3 Population

CHLa Ingestion

1 -

d 0.2 Rate

- µg L µg 0.1

0.0 Blue Cat Blue Cat Gizzard Gizzard Threadfin Atlantic (<20 cm) (20-40 cm) Shad Shad Shad Menhaden YOY Figure 3-2 Fish as consumers of CHLa in the Tidal Freshwater James River. Data shown are (top) CHLa in fish gut contents, (2nd) per capita CHLa ingestion for a range of gut turnover rates, (3rd) fish biomass and density and (bottom) population-scale estimates of CHLa ingestion.

Figure 3-3. Fish as recyclers of N in the Tidal Freshwater James River. Data shown are (top) PON in fish gut contents, (2nd) per capita PON ingestion for a range of potential gut turnover rates, (3rd) fish abundance and (bottom) population-scale estimates of PON ingestion.

Figure 3-4. Abundance, per capita clearance rates and CHLa and PON ingestion by dominant zooplankton and benthic filter-feeders (Rangia) of the tidal freshwater James River. Zooplankton abundances are based on bi-monthly collections during March-September 2013. Rangia abundance is the average for 2001-2010 based on annual surveys by CBP.

Figure 3-5. (Top) CHLa production in the Tidal Fresh James River compared to losses via respiration, grazing and advection (export). (Bottom) Phytoplankton demand for DIN compared to inputs via respiration (bacterial re-mineralization), consumer-mediated N cycling (grazing), and advection (external inputs). Data shown are average daily values for March-November of 2012 and 2013. Grazing losses include contributions by zooplankton, benthic filter-feeders and fish.

Figure 3-6 Relationships between muscle δ13C ‰ and CHLa contribution to diet (top) and δ13C ‰ to δ 15N‰. Dashed lines represent trophic enrichment levels from basal resources of autochthonus (seston samples during high CHLa and low TSS concentrations) and allochthonous forms of organic matter (Seston samples during low CHLa and high TSS concentrations).

Figure 3-7 Correlation matrices using principal components analyses of independent variables for mesocosm experiment performed in June and July 2013

-1 -1 PO4 Loading Rate (mg L d ) 0 0.01 0.02 0.03 0 0.01 0.02 0.03 80 a 2.5

2.0 b

)

) 1

60 1

- - 1.5 40

1.0 DIN DIN (mgL CHLa CHLa (µg L 20 0.5

0 0.0 0.8 Control 400

Rangia )

c 1 d -

) 0.6 300

1 Shad - Sediment

0.4 expected 200

(mg L 4

PO 0.2 100 RotifersL (Ind.

0.0 0

120 300

)

1 )

e -

1 f -

80 200

40 100

CopepodsL (Ind. Cladocerans (Ind. L 0 0

0.10 6

) h )

1 g

- - 4

0.05

myc (copies mL(copiesmyc MicrocystinL (µg 0.00 0 0 0.04 0.08 0.12 0 0.04 0.08 0.12 DINLoading Rate (mg L-1 d-1)

Figure 3- 8 Chlorophyll, Nutrient, Zooplankton and Microcystin/Microcystis gene copies results for various treatment and nutrient loading rates in June Mesocosm experiments. Chlorophyll and zooplankton results represent averages across the course of the experiment while nutrient, Microcystin and Microcystis results represent final values.

-1 -1 PO4 Loading Rate (mg L d )

0 0.01 0.02 0.03 0 0.01 0.02 0.03 2.0 160 a b

) 1.5 )

1 120

1 -

80 1.0

DIN L (mgDIN 0.5 CHLa CHLa (µg L 40

0 0.0

0.6 1,500 )

c 1 d

) 1

- 0.4 1,000

(mg L 4

0.2 500

PO RotifersL (Ind.

0.0 0

200 60

)

- 1 - e f 150 40 100 20

CopepodsL (Ind. Cladocerans (Ind. L 0 0

4 300

)

- 1

- g h 3 Control 200 Rangia

2 Shad copiesml

Sediment 3 100

1 Expected

MicrocystinL (µg myc(10 0 0 0 0.04 0.08 0.12 0 0.04 0.08 0.12 DINLoading Rate (mg L-1 d-1)

Figure 3- 9 Chlorophyll, Nutrient, Zooplankton and Microcystin/Microcystis gene copies results for various treatment and nutrient loading rates in July mesocosm experiments. Chlorophyll and zooplankton results represent averages across the course of the experiment while nutrient, Microcystis and Microcystin results represent final values.

Figure 3-10 Nutrient effect sizes calculated as the natural logarithm of nutrient enriched: control CHLa concentrations (top) and DIN uptake rates (bottom) for nutrient limitation bioassays using algal cultures from mesocosms with variable grazer treatments and nutrient loading rates.

Synthesis This dissertation provides insights about the drivers of algal blooms and the consequences they have upon ecosystem services for the Tidal Fresh James River. Major external drivers of algal blooms in this system include light, hydrology and nutrient loading rates. These factors help explain why this reach is so productive, patterns of limitation, and nutrient uptake rates. Declines in external nutrient loads have led to an increased nutrient limitation which is a positive sign for eutrophication recovery. However, consideration of current external forcing conditions cannot alone explain current levels of algal production. This is evident by the declines of nutrient loading rates without corresponding declines in CHLa levels. We suggest this disconnect between external forcing and current levels of algal abundance is driven by the accumulation of external nutrient loads over the past century and current internal recycling of this material. The wide shallow channel of this system has previously been shown to dissipate energy and retain large nutrient and sediment loads during high flow events. As a consequence, large legacies have likely accumulated and these nutrient legacies are made bioavailable by detritivorous fish as well as microbial communities. As a result reductions in current external nutrient loads have not resulted in corresponding proportional declines in algal production. While there is limited ability to deal with nutrient legacies, understanding recycling mechanisms is necessary to project reasonable restoration goals. The internal recycling of these large deposits of sedimented nutrients currently plays a significant role in supporting current algal production, and as external loads decline, recycled nutrients may become more important. In addition to conclusions about drivers of algal production, this work helps characterize the structure and pattern of the James River food web. This system which has undergone large changes due to nutrient loading, sedimentation and invasive species, and is now dominated by Gizzard Shad, Catfish and Rangia wedge clams, which are all supported primarily by terrestrial carbon. This work also characterizes the harmful algae Microcystis and its influence upon the food web. While Microcystis does not make up a large portion of the algal biomass, the toxin it produces is inconspicuous throughout summer and produces toxins which accumulate in many different varieties of biota and inhibit the feeding of Rangia. Nutrient loading and zooplankton grazing seem to favor Microcystis and the production of microcystin. These results suggest anthropogenic nutrient and sediment loading can have temporally extended effects on eutrophication and legacies may delay recovery of ecosystem services.

Literature Cited Acuña, S.; Baxa, D.; Teh, S. Sublethal dietary effects of Microcystin producing Microcystis on threadfin shad, Dorosoma petenense. Toxicon 2012, 60, 1191–1202.

Alber, M., and J. E. Sheldon 1999. Use of a date-specific method to examine variability in the flushing times of Georgia Estuaries. Estuarine, Coastal and Shelf Science 49: 469-482.

Anderson, D. M.; Burkholder, J. M.; Cochlan, W. P.; Glibert, P. M.; Gobler, C. J.; Heil, C. A.; Kudela, R. M.; Parsons, M. L.; Rensel, J. E. J.; Townsend, D. W.; Trainer, V. L.; Vargo, G. A. Harmful algal blooms and eutrophication: Examining linkages from selected coastal regions of the United States. Harmful Algae 2008, 8, 39–53.

APHA, AWWA, and WPCF, 1992. Standard Methods for the Examination of Water and Wastewater, 18th ed. American Public Health Association, Washington, D.C. USA.

Arar, E. J.; Collins, G. B. Method 445.0: In vitro determination of chlorophyll a and pheophytin a in marine and freshwater algae by fluorescence; United States Environmental Protection Agency, Office of Research and Development, National Exposure Research Laboratory, 1997.

Baron, J. S., E. K. Hall, B. T. Nolan, J. C. Finlay, E. S. Bernhardt, J. A. Harrison, F. Chan, and E. W. Boyer 2013. The interactive effects of excess reactive nitrogen and climate change on aquatic ecosystems and water resources of the United States. Biogeochemistry 114: 71-92.

Beardall, J., E. Young and S. Roberts. 2001. Approaches for determining phytoplankton nutrient limitation. Aquatic Sciences 63:44–69.

Bernardi, R. de; Giussani, G. Are blue-green algae a suitable food for zooplankton? An overview. Hydrobiologia 1990, 200-201, 29–41.

Benke, A. 2011. Secondary Production, Quantitative Food Webs, and Trophic Position. Nature Education Knowledge 3(10):26

Best, J. H.; Eddy, F. B.; Codd, G. A. Effects of purified Microcystin-LR and cell extracts of Microcystis strains PCC 7813 and CYA 43 on cardiac function in brown trout (Salmo trutta) alevins. Fish Physiology and Biochemistry 2001, 24, 171–178.

Bláha, L.; Kopp, R.; Šimková, K.; Mareš, J. Oxidative Stress Biomarkers are Modulated in Silver Carp (Hypophthalmichthys molitrix Val.) Exposed to Microcystin-Producing Cyanobacterial Water Bloom. Acta Veterinaria Brno 2004, 73, 477.

Bogdan, K.G. and J.J. Gilber. 1984. Body size and food size in fresh-water zooplankton. Proceedings of the national academy of sciences of the united states of America - biological sciences. 81:20, 6247-6431.

Borah, D. K., G. Yagow, A. Saleh, P. L. Barnes, W. Rosenthal, E. C. Krug and L. M. Hauck. 2006. Sediment and nutrient modeling for TMDL development and implementation. Transactions of the Asabe 49:967–986.

Borsuk, M. E., C. A. Stow and K. H. Reckhow. 2004. Confounding effect of flow on estuarine response to nitrogen loading. Journal of Environmental Engineering 130:605–614.

Boyer, E. W., C. L. Goodale, N. A. Jaworsk and R. W. Howarth. 2002. Anthropogenic nitrogen sources and relationships to riverine nitrogen export in the northeastern USA. Biogeochemistry 57:137–169.

Boynton. W.R, C.L.S. Hodgkins, C.A. O’Leary, E.M. Baily, A.R. Bayard and L.A. Wainger. 2013. Multi- decade Responses of a Tidal Creek System to Nutrient Load Reductions: Mattawoman Creek, Maryland USA. Estuaries and Coasts. DOI 10.1007/s12237-013-9690-4.

Bradley, P. B., M. W. Lomas and D. A. Bronk. 2010. Inorganic and organic nitrogen use by phytoplankton along Chesapeake Bay measured using a flow cytometric approach. Estuaries and Coasts 33:971-984.

Brauer, V. S., Stomp, and J. Huisman. 2012. The nutrient-load hypothesis: patterns of resource limitation and community structure driven by competition for nutrients and light. The American Naturalist 179:721– 740.

Bronk, D. A., P. M. Gilbert, and B. B. Ward. 1994. Nitrogen uptake, dissolved organic nitrogen release, and new production. Science 265:1843-1846.

Bronk, D. A., P. M. Glibert, T. C. Malone, S. Banahan, and E. Sahlsten. 1998. Inorganic and organic nitrogen cycling in Chesapeake Bay: autotrophic versus heterotrophic processes and relationships to carbon flux. Aquatic Microbial Ecology 15:177–189.

Bukaveckas, P. A., A. Macdonald, A. K. Aufdenkampe, J. H. Chick, J. E. Havel, R. E. Schultz, T. Angradi, D. W. Bolgrien, T. M. Jicha, and D. Taylor. 2011a. Phytoplankton abundance and contributions to suspended particulate matter in the Ohio, Upper Mississippi and Missouri Rivers. Aquatic Sciences 73:419-436.

Bukaveckas, P. A., L. E. Barry, M. J. Beckwith, V. David, and B. Lederer 2011b. Factors determining the location of the chlorophyll maximum and the fate of algal production within the tidal freshwater James River. Estuaries and Coasts 34:569-582.

Bukaveckas, P. A., and W. N. Isenberg 2013. Loading, transformation and retention of nitrogen and phosphorus in the tidal freshwater James River (Virginia). Estuaries and Coasts 36:1219-1236.

Caffrey, J.M. 2003. Production, respiration and net ecosystem metabolism in U.S. estuaries. Environmental Monitoring and Assessment 81: 207–219.

Caffrey, J.M. 2004. Factors controlling net ecosystem metabolism in U.S. estuaries. Estuaries 27: 90–101

Caraco, N. F., J. J. Cole, and D. L. Strayer 2006. Top-down control from the bottom: Regulation of eutrophication in a large river by benthic grazing. Limnology and Oceanography 51: 664-670.

Carstensen, J., M. Sánchez-Camacho, C. M. Duarte, D. Krause-Jensen and N. Marbà. 2011. Connecting the Dots: Responses of Coastal Ecosystems to Changing Nutrient Concentrations. Environmental Science and Technology 45:9122-9132.

Cerco, C., and T. Cole. 1993. 3-Dimensional Eutrophication Model of Chesapeake Bay. Journal of Environmental Engineering-Asce 119:1006-1025.

Chen, W.; Song, L.; Peng, L.; Wan, N.; Zhang, X.; Gan, N. Reduction in Microcystin concentrations in large and shallow lakes: Water and sediment-interface contributions. Water Research 2008, 42, 763–773.

Cloern, J. E. 2001. Our evolving conceptual model of the coastal eutrophication problem. Marine Ecology Progress Series 210:223–253.

Cole, J. and C. Solomon. 2012. Terrestrial support of zebra mussels and the Hudson River food web: A multi-isotope, Bayesian analysis. Limology and Oceanography 57, 2 1802-1815.

Conley, D. J., H. W. Paerl, R. W. Howarth, D. F. Boesch, S. P. Seitzinger, K. E. Havens, C, Lancelot and G. E. Likens. 2009. Controlling eutrophication: nitrogen and phosphorus. Science 323:1014–1015.

Coughlan, J. The estimation of filtering rate from the clearance of suspensions. Marine Biology 1969, 2, 356–358.

Davis, T. W., and C. J. Gobler. 2011. Grazing by Mesozooplankton and Microzooplankton on Toxic and Non-toxic Strains of Microcystis in the Transquaking River, a Tributary of Chesapeake Bay. Journal of Plankton Research 33(3):415.

Davis, A. M., M. L. Blanchette, B. J. Pusey, T. D. Jardine, and R. G. Pearson. 2012. Gut content and stable isotope analyses provide complementary understanding of ontogenetic dietary shifts and trophic relationships among fishes in a tropical river. Freshwater Biology 57: 2156-2172.

Dawson, R. M. The toxicology of Microcystins. Toxicon 1998, 36, 953–962.

De Figueiredo, D. R.; Azeiteiro, U. M.; Esteves, S. M.; Gonçalves, F. J. M.; Pereira, M. J. Microcystin- producing blooms—a serious global public health issue. Ecotoxicology and Environmental Safety 2004, 59, 151–163.

Deblois, C. P., Aranda-Rodriguez, R., Giani, A., & Bird, D. F. 2008. Microcystin accumulation in liver and muscle of tilapia in two large Brazilian hydroelectric reservoirs. Toxicon, 51(3), 435–448.

Deblois, C. P.; Giani, A.; Bird, D. F. Experimental model of Microcystin accumulation in the liver of Oreochromis niloticus exposed subchronically to a toxic bloom of Microcystis sp. Aquatic Toxicology 2011, 103, 63–70. del Giorgio, P. A. and Peters, R. H. 1994. Patterns in Planktonic P:R Ratios in Lakes: Influence of Lake Trophy and Dissolved Organic Carbon. Limnology and Oceanography, 39(4), 772–787.

Dewes, L. J.; Sandrini, J. Z.; Monserrat, J. M.; Yunes, J. S. Biochemical and physiological responses after exposure to Microcystins in the crab Chasmagnathus granulatus (Decapoda, Brachyura). Ecotox. Environ. Safe. 2006, 65, 201–208.

Donald, D. B.; Bogard, M. J.; Finlay, K.; Leavitt, P. R. Comparative effects of urea, ammonium, and nitrate on phytoplankton abundance, community composition, and toxicity in hypereutrophic freshwaters. Limnology and Oceanography 2011, 56, 2161–2175. 95

Duarte, C. M., Conley, D. J., Carstensen, J., & Sánchez-Camacho, M. 2009. Return to Neverland: Shifting Baselines Affect Eutrophication Restoration Targets. Estuaries and Coasts, 32(1), 29–36.

Edwards, V. R., P. Tett, and K.J. Jones. 2003. Changes in the yield of chlorophyll a from dissolved available inorganic nitrogen after an enrichment event—applications for predicting eutrophication in coastal waters. Continental Shelf Research 23:1771–1785.

Elliott D.T. and K.W. Tang. 2009. Simple staining method for differentiating live and dead marine zooplankton in field samples. Limnology and Oceanography, 7 585-595.

Elliott D.T., C.K. Harris and K.W. Tang. 2010. Dead in the water: The fate of copepod carcasses in the York River estuary, Virginia. Limnology and Oceanography, 55:1821-1834.

Elser, J. J., M. E. S. Bracken, E. E. Cleland, D. S. Gruner, W. S. Harpole, H. Hillebrand, J. T. Ngai, E. W. Seabloom, J. B. Shurin, and J. E. Smith. 2007. Global analysis of nitrogen and phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecology Letters 10:1135-1142.

Elser, J. J., M. M. Elser, N. A. MacKay, amd S. R. Carpenter. 1988. Zooplankton-mediated transitions between N- and P-limited algal growth. Limnology and Oceanography 33:1–14.

Ermgassen, P. S. E. Z.; Spalding, M. D.; Grizzle, R. E.; Brumbaugh, R. D. Quantifying the Loss of a Marine Ecosystem Service: Filtration by the Eastern Oyster in US Estuaries. Estuaries and Coasts 2013, 36, 36–43.

Filippino, K. C., M. R. Mulholland, P. Bernhardt, G. Boneillo, R. Morse, M. Semcheski, H. G. Marshall, N. Love, Q. Roberts and D. Bronk. 2011. The bioavailability of effluent-derived organic nitrogen along an estuarine salinity gradient. Estuaries and Coasts 34:269–280.

Finlay, J. C. 2001. Stable-carbon-isotope ratios of river biota: Implications for energy flow in lotic foodwebs. Ecology 82: 1052-1064.

Fisher, T., A. Gustafson, G. Radcliffe, K. Sundberg and J. Stevenson. 2003. A long-term record of photosynthetically available radiation (PAR) and total solar energy at 38.60 degrees N, 78.2 degrees W. Estuaries 26:1450–1460.

Fisher, T., A. Gustafson, K. Sellner, R. Lacouture, L. Haas, R. Wetzel, R. Magnie, D. Everitt, B. Michaels and R. Karrh. 1999. Spatial and temporal variation of resource limitation in Chesapeake Bay. Marine Biology 133:763–778.

Forsberg, B. R., C. A. R. M. Araujo-Lima, L. A. Matinelli, R. Victoria, and J. A. Bonassi 1993. Autotrophic carbon sources for fish of the central Amazon. Ecology 74: 643-652.

Franklin, R. B., C. Luria, L. S. Ozaki, and P. A. Bukaveckas. 2013. Community composition and activity state of estuarine bacterioplankton assessed using differential staining and metagenomic analysis of 16S rDNA and rRNA. Aquatic Microbial Ecology 69:247-261.

Friedland, K. D.; Ahrenholz, D. W.; Haas, L. W. Viable gut passage of cyanobacteria through the filter- feeding fish Atlantic menhaden, Brevoortia tyrannus. Journal of Plankton Research 2005, 27, 715–718.

Garcia, A. C.; Bargu, S.; Dash, P.; Rabalais, N. N.; Sutor, M.; Morrison, W.; Walker, N. D. Evaluating the potential risk of Microcystins to blue crab (Callinectes sapidus) fisheries and human health in a eutrophic estuary. Harmful Algae 2010, 9, 134–143.

Ger, K. A.; Teh, S. J.; Baxa, D. V.; Lesmeister, S.; Goldman, C. R. The effects of dietary Microcystis aeruginosa and Microcystin on the copepods of the upper San Francisco Estuary. Freshwater Biology 2010, 55, 1548–1559.

Gérard, C.; Poullain, V.; Lance, E.; Acou, A.; Brient, L.; Carpentier, A. Influence of toxic cyanobacteria on community structure and Microcystin accumulation of freshwater molluscs. Environmental Pollution 2009, 157, 609–617.

Gerritson, J., A. F. Holland and D.E. Irvine. Suspension-feeding bivalves and the fate of primary production - an estuarine model applied to Chesapeake Bay. Estuaries.17:403-416.

Ginn, H. P.; Pearson, L. A.; Neilan, B. A. Hepatotoxin Biosynthesis and Regulation in Cyanobacteria - the Putative Involvement of Nitrogen and Iron Homeostasis Mechanisms. Chiang Mai J. Sci. 2009, 36, 200–223.

Gosselain, V., J.-P. Descy, and E. Everbecq. 1994. The phytoplankton community of the River Meuse, Belgium: Seasonal dynamics and the possible incidence of zooplankton grazing. Hydrobiologia 289: 179–191.

Gowen, R.J., P. Tett and K.J. Jones. 1992. Predicting marine eutrophication: the yield of chlorophyll from nitrogen in Scottish coastal waters. Marine Ecology Progress Series 85:153-161.

Graham, J. L.; Loftin, K. A.; Meyer, M. T.; Ziegler, A. C. Cyanotoxin Mixtures and Taste-and-Odor Compounds in Cyanobacterial Blooms from the Midwestern United States. Environ. Sci. Technol. 2010, 44, 7361–7368.

Greening, H., & Janicki, A. 2006. Toward Reversal of Eutrophic Conditions in a Subtropical Estuary: Water Quality and Seagrass Response to Nitrogen Loading Reductions in Tampa Bay, Florida, USA. Environmental Management, 38(2), 163–178.

Greening, H. S., Cross, L. M., & Sherwood, E. T. 2011. A Multiscale Approach to Seagrass Recovery in Tampa Bay, Florida. Ecological Restoration, 29(1), 82–93.

Grützmacher, G.; Wessel, G.; Klitzke, S.; Chorus, I. Microcystin Elimination During Sediment Contact. Environ. Sci. Technol. 2010, 44, 657–662.

Hall, S. R., M. A. Leibold, D. A. Lytle and V. H. Smith. 2004. Stoichiometry and planktonic grazer composition over gradients of light, nutrients, and predation risk. Ecology 85: 2291–2301.

Hall, S. R., M. A. Leibold, D. A. Lytle and V.H. Smith. 2007. Grazers, producer stoichiometry, and the light: nutrient Hypothesis revisited. Ecology 88:1142–1152.

Haskell, C.A., K.F. Tiffan and D.W. Rondrof. 2013. The Effects of Juvenile American Shad Planktivory on Zooplankton Production in Columbia River Food Webs. Transactions of the American Fisheries Society.

Havens, K. E., and C.L. Schelske. 2001. The importance of considering biological processes when setting total maximum daily loads (TMDL) for phosphorus in shallow lakes and reservoirs. Environmental Pollution 113:1–9.

Havens, K. E., J. Hauxwell, A. C. Tyler, S. Thomas, K. J. McGlathery, J. Cebrian, I. Valiela, A. D. Steinman and S. J. Hwang. 2001. Complex interactions between autotrophs in shallow marine and freshwater ecosystems: implications for community responses to nutrient stress. Environmental Pollution 113:95–107.

Hoffman, J. C., D. A. Bronk, and J. E. Olney 2008. Organic matter sources supporting lower food web production in the tidal freshwater portion of the York River estuary. Estuaries and Coasts 31: 898-911.

Howarth, R. W., and R. Marino. 2006. Nitrogen as the limiting nutrient for eutrophication in coastal marine ecosystems: Evolving views over three decades. Limnology and Oceanography 51:364-376.

Howarth, R. W., D. P. Swaney, E. W. Boyer, R. Marino, N. Jaworski and C. Goodale. 2006. The influence of climate on average nitrogen export from large watersheds in the Northeastern United States. Biogeochemistry 79:163–186.

Hu, S., and W.O. Smith Jr. 1998. The effects of irradiance on nitrate uptake and dissolved organic nitrogen release by phytoplankton in the Ross Sea. Continental Shelf Research 18: 971–990.

Ibelings, B. W.; Bruning, K.; de Jonge, J.; Wolfstein, K.; Pires, L. M. D.; Postma, J.; Burger, T. Distribution of Microcystins in a lake foodweb: No evidence for biomagnification. Microb. Ecology 2005, 49, 487–500.

Ibelings, B. W.; Havens, K. E. Cyanobacterial toxins: a qualitative meta-analysis of concentrations, dosage and effects in freshwater, estuarine and marine biota. In Cyanobacterial Harmful Algal Blooms: State of the Science and Research Needs; Hudnell, H. K., Ed.; Springer-Verlag Berlin: Berlin, 2008; Vol. 619, pp. 675–732

Jančula, D., M. Míkovcov, Z. Adámek, and B. Maršálek. 2008. Changes in the photosynthetic activity of Microcystis colonies after gut passage through Nile tilapia (Oreochromis niloticus) and silver carp (Hypophthalmichthys molitrix). Aquaculture Research 39: 311–314..

Jenkins, R. E.; Burkhead, N. M. Freshwater fishes of Virginia.; American Fisheries Society, 1994.

Junk, W. J., P. B. Bayley, and R. E. Sparks. 1989. The flood pulse concept in river-floodplain systems. Pages 110–127 in D. P. Dodge (editor). Proceedings of the International Large River Symposium. Volume 106. Canadian Special Publications of Fisheries and Aquatic Sciences. Department of Fisheries and Oceans, Ottawa, Ontario.

Johnk, K. D.; Huisman, J.; Sharples, J.; Sommeijer, B.; Visser, P. M.; Stroom, J. M. Summer heatwaves promote blooms of harmful cyanobacteria. Global. Change Biology. 2008, 14, 495–512.

Kaebernick, M.; Dittmann, E.; Börner, T.; Neilan, B. A. Multiple Alternate Transcripts Direct the Biosynthesis of Microcystin, a Cyanobacterial. Appl. Environ. Microbiology 2002, 68, 449–455.

Kendall, C., S. R. Silva, and V. J. Kelly. 2001. Carbon and nitrogen isotopic compositions of particulate organic matter in four large river systems across the United States. Hydrological Processes 15:1301– 1346.

Koch, R. W., D. L. Guelda, and P. A. Bukaveckas. 2004. Phytoplankton growth in the Ohio, Cumberland and Tennessee rivers, USA: inter-site differences in light and nutrient limitation. Aquatic Ecology 38:17– 26.

Kozlowsky-Suzuki, B.; Wilson, A. E.; Ferrão-Filho, A. da S. Biomagnification or biodilution of Microcystins in aquatic foodwebs? Meta-analyses of laboratory and field studies. Harmful Algae 2012, 18, 47–55.

Kratina, P., H. S. Greig, P. L. Thompson, T. S. A. Carvalho-Pereira and J. B. Shurin. 2012. Warming modifies trophic cascades and eutrophication in experimental freshwater communities. Ecology 93:1421– 1430.

Lampert, W. Laboratory Studies on Zooplankton-Cyanobacteria Interactions. N. Z. J. Mar. Freshw. Res. 1987, 21, 483–490.

Lean, D. R. S. and F. R. Pick. 1981. Photosynthetic response of lake plankton to nutrient enrichment: A test for nutrient limitation. Limnology and Oceanography 26:1001–1019.

Lewis, W. M., Melack, J. M., McDowell, W. H., McClain, M., and J.E. Richey. 1999. Nitrogen yields from undisturbed watersheds in the Americas. Biogeochemistry 46:149–162.

Lucas, L. V., J. K. Thompson, and L.R. Brown. 2009. Why are diverse relationships observed between phytoplankton biomass and transport time? Limnology and Oceanography 54:381–390.

Malbrouck, C.; Kestemont, P. Effects of Microcystins on Fish. Environmental Toxicology & Chemistry 2006, 25, 6.

Marchais, V.; Schaal, G.; Grall, J.; Lorrain, A.; Nerot, C.; Richard, P.; Chauvaud, L. Spatial Variability of Stable Isotope Ratios in Oysters (Crassostrea gigas) and Primary Producers Along an Estuarine Gradient (Bay of Brest, France). Estuaries and Coasts 2013, 36, 808–819.

Marino, R. and R.W. Howarth. 1984. Atmospheric oxygen-exchange in the hudson river - dome measurements and comparison with other natural-waters. Estuaries 16:433-445.

Marion, J. W.; Lee, J.; Wilkins, J. R.; Lemeshow, S.; Lee, C.; Waletzko, E. J.; Buckley, T. J. In Vivo Phycocyanin Flourometry as a Potential Rapid Screening Tool for Predicting Elevated Microcystin Concentrations at Eutrophic Lakes. Environ. Sci. Technol. 2012, 46, 4523–4531.

Marshall, H. G., M. F. Lane, K. K. Nesius and L. Burchardt. 2009. Assessment and significance of phytoplankton species composition within Chesapeake Bay and Virginia tributaries through a long-term monitoring program. Environmental Monitoring and Assessment 150: 143–155.

Martineau, C., W. F. Vincent, J.-J. Frenette, and J. J. Dodson 2004. Primary consumers and particulate organic matter: Isotopic evidence of strong selectivity in the estuarine transition zone. Limnology and Oceanography 49: 1679-1686.

McCutchan, J. H., Lewis, W. M., Kendall, C., & McGrath, C. C. 2003. Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos, 102(2), 378–390.

Mette, E. M., M. J. Vanni, J. M. Newell, and M. J. González 2011. Phytoplankton communities and stoichiometry are interactively affected by light, nutrients, and fish. Limnology and Oceanography 56:1959–1975.

Michalak, A. M.; Anderson, E. J.; Beletsky, D.; Boland, S.; Bosch, N. S.; Bridgeman, T. B.; Chaffin, J. D.; Cho, K.; Confesor, R.; Daloğlu, I.; DePinto, J. V.; Evans, M. A.; Fahnenstiel, G. L.; He, L.; Ho, J. C.; Jenkins, L.; Johengen, T. H.; Kuo, K. C.; LaPorte, E.; Liu, X.; McWilliams, M. R.; Moore, M. R.; Posselt, D. J.; Richards, R. P.; Scavia, D.; Steiner, A. L.; Verhamme, E.; Wright, D. M.; Zagorski, M. A. Record- setting algal bloom in Lake Erie caused by agricultural and meteorological trends consistent with expected future conditions. PNAS 2013, 110, 6448–6452.

Moisander, P. H.; Ochiai, M.; Lincoff, A. Nutrient limitation of Microcystis aeruginosa in northern California Klamath River reservoirs. Harmful Algae 2009, 8, 889–897.

Monbet, Y. 1992. Control of phytoplankton biomass in estuaries: A comparative analysis of microtidal and macrotidal estuaries. Estuaries 15:563-571.

Mulholland, M. R., R. E. Morse, G. E. Boneillo, P. W. Bernhardt, K. C. Filippino, L. A. Procise, J. L. Blanco-Garcia, H. G. Marshall, T. A. Egerton, W. S. Hunley, K. A. Moore, D. L. Berry and C. J. Gobler. 2009. Understanding causes and impacts of the Dinoflagellate, Cochlodinium polykrikoides, blooms in the Chesapeake Bay. Estuaries and Coasts, 32:734–747.

Murrell, M., J. Hagy, E. Lores, and R. Greene. 2007. Phytoplankton production and nutrient distributions in a subtropical estuary: Importance of freshwater flow. Estuaries and Coasts 30: 390–402.

Neilan, B. A.; Jacobs, D.; DelDot, T.; Blackall, L. L.; Hawkins, P. R.; Cox, P. T.; Goodman, A. E. rRNA sequences and evolutionary relationships among toxic and nontoxic cyanobacteria of the genus Microcystis. Int. J. Syst. Bacteriol. 1997, 47, 693–697.

North, E. W., and E.D. Houde. 2001. Retention of White Perch and Striped Bass larvae: Biological- Physical interactions in Chesapeake Bay estuarine turbidity maximum. Estuaries 24:756–769.

North, E. W., and E.D. Houde. 2003. Linking ETM physics, zooplankton prey, and fish early-life histories to striped bass Morone saxatilis and white perch M. americana recruitment. Marine Ecology Progress Series 260:219–236.

Nübel, U.; Garcia Pichel, F.; Muyzer, G. PCR primers to amplify 16S rRNA genes from cyanobacteria. Appl. Environ. Microbiol. 1997, 63, 3327–3332.

O’Neil, J. M.; Davis, T. W.; Burford, M. A.; Gobler, C. J. The rise of harmful cyanobacteria blooms: The potential roles of eutrophication and climate change. Harmful Algae 2012, 14, 313–334.

100

Ochs, C. A., O. Pongruktham, and P. V. Zimba. 2013. Darkness at the break of noon: Phytoplankton production in the Lower Mississippi River. Limnology and Oceanography 58:555-568.

Orth, R. J., Williams, M. R., Marion, S. R., Wilcox, D. J., Carruthers, T. J. B., Moore, K. A., Batiuk, R. A. 2010. Long-Term Trends in Submersed Aquatic Vegetation (SAV) in Chesapeake Bay, USA, Related to Water Quality. Estuaries and Coasts, 33(5), 1144–1163.

Ozawa, K.; Yokoyama, A.; Ishikawa, K.; Kumagai, M.; Watanabe, M. F.; Park, H. D. Accumulation and depuration of Microcystin produced by the cyanobacterium Microcystis in a freshwater snail. Limnology 2003, 4, 131–138.

Paerl, H. W. Nuisance Phytoplankton Blooms in Coastal, Estuarine, and Inland Waters. Limnology and Oceanography 1988, 33, 823–847.

Paerl, H. W.; Otten, T. G. Harmful Cyanobacterial Blooms: Causes, Consequences, and Controls. Microb Ecol 2013, 65, 995–1010.

Papadimitriou, T.; Kagalou, I.; Bacopoulos, V.; Leonardos, I. D. Accumulation of Microcystins in water and fish tissues: An estimation of risks associated with Microcystins in most of the Greek Lakes. Environmental Toxicology 2010, 25, 418–427.

Parker, A. E., V. E. Hogue, F. B. Wilkerson, and R. C. Dugdale. 2012. The effect of inorganic nitrogen speciation on primary production in the San Francisco Estuary. Estuarine, Coastal and Shelf Science 104- 105:91-101.

Parnell, A.C., Inger R., Bearhop, S. & Jackson, A.L. 2010. Source partioning using stable isotopes: coping with too much variation. PLoS ONE, 5(3),

Peierls, B. L., N. S. Hall and H.W. Paerl. 2012. Non-monotonic responses of phytoplankton biomass accumulation to hydrologic variability: A comparison of two coastal plain North Carolina estuaries. Estuaries and Coasts, 35:1376–1392.

Peterson, B. J., and B. Fry 1987. Stable isotopes in ecosystem studies. Annual Review of Ecology and Systematics 18: 293-320.

Pires, L. M. D.; Karlsson, K. M.; Meriluoto, J. A. O.; Kardinaal, E.; Visser, P. M.; Siewertsen, K.; Donk, E. V.; Ibelings, B. W. Assimilation and depuration of Microcystin–LR by the zebra mussel, Dreissena polymorpha. Aquatic Toxicology 2004, 69, 385–396.

Poste, A. E.; Hecky, R. E.; Guildford, S. J. Evaluating Microcystin Exposure Risk through Fish Consumption. Environ. Sci. Technol. 2011, 45, 5806–5811.

Price, N.M., and P.J. Harrison. 1987. Comparison of methods for the analysis of dissolved urea in seawater. Marine Biology 94:307–317.

Prieto, A. I.; Pichardo, S.; Jos, Á.; Moreno, I.; Cameán, A. M. Time-dependent oxidative stress responses after acute exposure to toxic cyanobacterial cells containing Microcystins in tilapia fish (Oreochromis niloticus) under laboratory conditions. Aquatic Toxicology 2007, 84, 337–345.

101

Ptacnik, R., T. Andersen and T. Tamminen. 2010. Performance of the Redfield Ratio and a family of nutrient limitation indicators as thresholds for phytoplankton N vs. P limitation. Ecosystems, 13:1201– 1214.

Redfield, A. C. 1958. The biological control of chemical factors in the environment. American Scientist 46:230A–221.

Ren, L., N. Rabalais, R. Turner, W. Morrison, and W. Mendenhall. 2009. Nutrient limitation on phytoplankton growth in the Upper Barataria Basin, Louisiana: Microcosm bioassays. Estuaries and Coasts 32: 958–974.

Renaud S.M, L.V. Thinh and D.L. Parry. 1999 The gross composition and fatty acid composition of 18 species of tropical Australian microalgae for possible use in mariculture. Aquaculture 170:147–159.

Rinta-Kanto, J. M.; Konopko, E. A.; DeBruyn, J. M.; Bourbonniere, R. A.; Boyer, G. L.; Wilhelm, S. W. Lake Erie Microcystis: relationship between Microcystin production, dynamics of genotypes and environmental parameters in a large lake. Harmful Algae 2009, 8, 665–673.

Roach K.A. Environmental factors affecting incorporation of terrestrial material into large river food webs. Freshwater Science, 32(1):283-298. 2013

Rolland, A.; Bird, D. F.; Giani, A. Seasonal changes in composition of the cyanobacterial community and the occurrence of hepatotoxic blooms in the eastern townships, Québec, Canada. Journal of Plankton Research 2005, 27, 683–694.

Rudek, J., H.W. Paerl, M. A. Mallin and P.W. Bates. 1991. Seasonal and hydrological control of phytoplankton nutrient limitation in the lower Neuse River Estuary, North Carolina. Marine Ecology Progress Series 75:133-142.

Salvatore, S.R., N.D. Mundahl and T.E. Wissing. 1987. Effect of Water Temperature on Food Evacutation Rate and Feeding-activity of Age-0 Gizzard Shad.

Sarkanen, K. V., and C. H. Ludwig. 1971. Lignins: occurrence, formation, structure, and reactions. John Wiley and Sons, New York.

Schaus, M. H., and M. J. Vanni. 2000. Effects of Gizzard Shad on Phytoplankton and Nutrient Dynamics: Role of Sediment Feeding and Fish size. Ecology 81: 1701–1719.

Schaus, M. H.; Vanni, M. J.; Wissing, T. E. Biomass-dependent diet shifts in omnivorous gizzard shad: implications for growth, food web, and ecosystem effects. Transactions of the American Fisheries Society 2002, 131, 40–54.

Schindler, D. W. 1977. Evolution of phosphorus limitation in Lakes. Science 195:260–262.

Seitzinger, S., C. Kroeze, A. Bouwman, N. Caraco, F. Dentener, and R. Styles. 2002a. Global patterns of dissolved inorganic and particulate nitrogen inputs to coastal systems: Recent conditions and future projections. Estuaries and Coasts 25:640–655.

102

Seitzinger, S. P., R. W. Sanders, and R. Styles 2002b. Bioavailability of DON from natural and anthropogenic sources to estuarine plankton. Limnology and Oceanography 47:353-366.

Seitzinger, S., J.A. Harrison, J.K. Bohlke, A.F. Bouwman, R. Lowrance, B. Peterson, C. Tobias and G. Van Drecht. 2006. Denitrification across landscapes and waterscapes: A synthesis. Ecological Applications 16:2064-2090.

Sellers, T., and P. A. Bukaveckas. 2003. Phytoplankton production in a large, regulated river: A modeling and mass balance assessment. Limnology and Oceanography 48:1476-1487.

Shepherd, W.C. and E.L. Mills. 1996. Diel feeding, daily food intake, and Daphnia consumption by age-0 gizzard shad in Oneida Lake, New York. Transaction of the American Fisheries Society 125:411-421.

Shostell, J., and Bukaveckas, P. A. 2004. Seasonal and interannual variation in nutrient fluxes from tributary inputs, consumer recycling and algal growth in a eutrophic river impoundment. Aquatic Ecology, 38(3), 359–373.

Sierszen, M.E. and T.M. Frost. 1990. Effects of an experimental lake acidification on zooplankton feeding rates and selectivity. Canadian Journal of Fisheries and Aquatic Sciences. 47, 772-779.

Silva-Stenico, M. E.; Cantúsio Neto, R.; Alves, I. R.; Moraes, L. A. B.; Shishido, T. K.; Fiore, M. F. Hepatotoxin Microcystin-LR extraction optimization. Journal of the Brazilian Chemical Society 2009, 20, 535–542.

Stepanauskas, R., L. Leonardson, and L. J. Tranvik.1999. Bioavailability of wetland-derived DON to freshwater and marine bacterioplankton. Limnology and Oceanography 44: 1477-1485.

Sterner, R. W., and J. P. Grover. 1998. Algal growth in warm temperate reservoirs: kinetic examination of nitrogen, temperature, light and other nutrients. Water Research 32:3539-3548.

Sterner, R. W., J. J. Elser, E. J. Fee, S. J. Guildford and T. H. Chrzanowski. 1997. The light:nutrient ratio in lakes: The balance of energy and materials affects ecosystem structure and process. The American Naturalist 150:663–684.

Strayer, D. L., M. L Pace, N. F. Caraco, J. J. Cole, and S. E. G. Findlay. 2008. Hydrology and grazing jointly control a large-river food web. Ecology 89:12–18.

Tamminen, T., and T. Andersen. 2007. Seasonal phytoplankton nutrient limitation patterns as revealed by bioassays over Baltic Sea gradients of salinity and eutrophication. Marine Ecology Progress Series 340:121–138.

Tango, P. J.; Butler, W. Cyanotoxins in Tidal Waters of Chesapeake Bay. Northeastern Naturalist 2008, 15, 403–416.

Thorp, J. H., and M. D. Delong. 1994. The riverine productivity model: an heuristic view of carbon sources and organic processing in large river ecosystems. Oikos 70:305–308.

Thorp, J. H., and M. D. Delong. 2002. Dominance of autotrophic carbon in food webs of heterotrophic rivers. Oikos 96:543–550. 103

Ueno, Y.; Nagata, S.; Tsutsumi, T.; Hasegawa, A.; Watanabe, M. F.; Park, H.-D.; Chen, G.-C.; Chen, G.; Yu, S.-Z. Detection of Microcystins, a blue-green algal hepatotoxin, in drinking water sampled in Haimen and Fusui, endemic areas of primary liver cancer in China, by highly sensitive immunoassay. Carcinogenesis 1996, 17, 1317–1321.

Urbach, E.; Robertson, D.; Chisholm, S. Multiple Evolutionary Origins of Prochlorophytes Within the Cyanobacterial Radiation. Nature 1992, 355, 267–270.

Van der Lingen, C.D. 1998. Nitrogen excretion and absorption efficiencies of sardine Sardinops sagax fed phytoplankton and zooplankton diets. Marine Ecology Progress Series 175: 67-76.

Vanni, M. J., K. K. Arend, M. T. Bremigan, P. A. Soranno, D. B. Bunnell, J. E. Garvey, M. J. González, W. H. Renwick, and R. A. Stein. 2005. Linking Landscapes and Food Webs: Effects of Omnivorous Fish and Watersheds on Reservoir Ecosystems. BioScience 55: 155–167.

Vanni, M. J., A. M. Bowling, E. M. Dickman, R. S. Hale, K. A. Higgins, M. J. Horgan, L. B. Knoll, W. H. Renwick and R. A. Stein. 2006. Nutrient cycling by fish supports relatively more primary production as lake productivity increases. Ecology 87:1696–1709.

Verdonschot, P. F. M., B. M. Spears, C. K. Feld, S. Brucet, H. Keizer-Vlek, A. Borja, M. Elliott, M. Kernan, and R. K. Johnson. 2013. A comparative review of recovery processes in rivers, lakes, estuarine and coastal waters. Hydrobiologia 704: 453–474.

Vannote, R. L., G. Minshall, K. W. Cummins, J. R. Sedell, and C. E. Cushing. 1980. The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences 37:130–137.

Viergutz, C.; Linn, C.; Weitere, M. Intra- and interannual variability surpasses direct temperature effects on the clearance rates of the invasive clam Corbicula fluminea. Mar. Biol. 2012, 159, 2379–2387.

Wagner, T., P. A. Soranno, K. E. Webster and D. S. Cheruvelil. 2011. Landscape drivers of regional variation in the relationship between total phosphorus and chlorophyll in lakes. Freshwater Biology 56:1811–1824.

Ward, B. B., and D. A. Bronk. 2001. Net nitrogen uptake and DON release in surface waters: importance of trophic interactions implied from size fractionation experiments. Marine Ecology Progress Series 219:11-24.

Watanabe, M. F.; Park, H.-D.; Kondo, F.; Harada, K.; Hayashi, H.; Okino, T. Identification and estimation of Microcystins in freshwater mussels. Natural Toxins 1997, 5.

Wetzel, R.G. 2001. Limnology, 3rd ed. New York: Academic. 220-224.

Whalen, S. C. and P. M. Benson. 2007. Influence of nutrient reduction, light and light-nutrient interactions on phytoplankton biomass, primary production and community composition in the Neuse River, USA. Fundamental and Applied Limnology 168:257–270.

104

Wiegner, T. N., S. P. Seitzinger, P. M. Gilbert, and D. A. Bronk, 2006. Bioavailability of dissolved organic nitrogen and carbon from nine rivers in the eastern United States. Aquatic Microbial Ecology 43:277-287.

Wilson, A. E. W. A. .; Gossiaux, D. C. G. D. .; Höök, T. O. H. T. .; Berry, J. P. B. J. .; Landrum, P. F. L. P. .; Dyble, J. D. .; Stephanie, J.; Guildford, G. S. 2008. Evaluation of the human health threat associated with the hepatotoxin Microcystin in the muscle and liver tissues of yellow perch (Perca flavescens). Canadian journal of fisheries and aquatic sciences 65, 1487–1497.

Wilson, H.F. and M.A. Xenopoulos. 2011. Nutrient recycling by fish in streams along a gradient of agricultural land use. Global Change Biology 17: 130-139.

Wong, W. H.; Rabalais, N. N.; Turner, R. E. Abundance and ecological significance of the clam Rangia cuneata (Sowerby, 1831) in the upper Barataria Estuary (Louisiana, USA). 2010. Hydrobiologia, 651, 305–315.

Wood, J.D.; Bukaveckas, P.A. Increasing Severity of Phytoplankton Nutrient Limitation Following Reductions in Point Source Inputs to the Tidal Freshwater Segment of the James River Estuary. Estuaries and Coasts. DOI 10.1007/s12237-013-9756-3.

World Health Organization. Guidelines for Safe Recreational Water Environments; Coastal and Fresh Waters; Geneva, 2003; Vol. 1, p 219.

Zeug, S.C. and K.O. Winemiller. 2008. Evidence supporting the importance of terrestrial carbon in a large-river food web. Ecology:89, 1733-1743

105