A modeling study on methylmercury bioaccumulation and its controlling factors

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/ecolmodel

A modeling study on methylmercury bioaccumulation and its controlling factors Eunhee Kim a,b,∗ , Robert P. Mason c , Christine M. Bergeron d a

Chesapeake Biological Laboratory, University of Maryland Center for Environmental Science (UMCES), P.O. Box 38, Solomons, MD 20688-0038, USA b Department of Oceanography, Chonnam National University, 300 Yongbongdong, Gwangju, 500-757, Republic of Korea c Marine Sciences Department, University of Connecticut, Groton, CT 06340, USA d Department of Fisheries and Wildlife Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24060, USA

a r t i c l e

i n f o

a b s t r a c t

Article history:

The objectives of this study were: (1) to develop a methylmercury (MeHg) bioaccumula-

Received 2 March 2007

tion model using data from STORM (high bottom Shear realistic water column Turbulence

Received in revised form

Resuspension Mesocosms) experiments; and (2) to use the model as a diagnostic tool to

11 July 2008

examine an effect of sediment resuspension and other important factors on MeHg bioaccu-

Accepted 15 July 2008

mulation. There were four mesocosm experiments (1–4) conducted both in summer and fall.

Published on line 22 August 2008

Tidal resuspension (4 h on- and 2 h off-cycles) was simulated using the STORM facility at CBL, UMCES (Chesapeake Biological Laboratory, University of Maryland Center for Environmental

Keywords:

Science). The model results showed that changes in clam biomass had a great effect on phy-

Mercury

toplankton and zooplankton biomass, and consequently MeHg accumulation. In addition, it

Methylmercury

appeared that sediment resuspension played a role in transferring the enhanced sediment

Bioaccumulation model

MeHg into organisms inhabiting both water column and sediment.

Sediment resuspension

© 2008 Elsevier B.V. All rights reserved.

Mercury methylation Methylmercury demethylation

1.

Introduction

Estuaries play an important role in the global mercury cycle as they form the link between the terrestrial and ocean environments. Estuaries are often impacted by human activity and estuarine sediments also serve as the primary location for mercury (Hg) methylation (Heyes et al., 2006; Gilmour and Henry, 1991; Benoit et al., 1998). Studies have shown that the in situ production of methylmercury (MeHg) in sediments can be a significant source in estuarine environments (Mason et al., 1999; Hammerschmidt et al., 2004; Sunderland et al., 2004). Methylmercury production is mediated by sulfate reducing

bacteria (SRB) in anaerobic environments and is a persistent and highly toxic contaminant that readily accumulates in aquatic food chains. Adverse effects of MeHg on higher trophic level organisms include neurological, reproductive and behavioral effects (US EPA, 1997). Methylmercury accumulation into higher levels of food chains (e.g., fish) is of great concern for human health, as humans are principally exposed to MeHg by fish consumption (Clarkson, 1990; Fitzgerald and Clarkson, 1991). It has been shown that inorganic Hg, organic matter, and sulfide are the most important factors in controlling MeHg levels in surface sediments (Benoit et al., 1998; Mason and

∗ Corresponding author at: Department of Oceanography, Chonnam National University, 300 Yongbong-Dong, Gwangju, 500-757, Republic of Korea. Tel.: +82 10 6731 8123. E-mail address: [email protected] (E. Kim). 0304-3800/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ecolmodel.2008.07.008

268

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Lawrence, 1999; Conaway et al., 2003; Hammerschmidt and Fitzgerald, 2004). The mobility and bioavailability of Hg and MeHg depends on the nature and concentration of the binding phases in the sediment, which appear to be controlled by sediment redox state. Sediment resuspension is one process that can induce a change in sediment redox status, which can be an important factor in controlling Hg methylation in sediments, and can also affect MeHg bioavailability. Resuspension also provides an effective link for transport of chemicals between the sediment and the water column. Kim et al. (2006) found from two mesocosm experiments with muddy sediment from Baltimore Harbor, Maryland (USA), that sediment resuspension did not appear to directly impact Hg transformations and MeHg formation. However, it appeared that sediment resuspension could lead to changes in the association of Hg with binding phases, thereby influencing Hg methylation. Sediment organic content plays an important role in Hg methylation as it controls Hg distribution between particulate and dissolved phases, determining Hg bioavailability to SRB. It has been shown that Hg methylation was negatively correlated with the sediment Hg distribution coefficient (Kd ) and that sediment–water partitioning of Hg and MeHg was positively related to organic matter in surface sediments with low levels of acid volatile sulfide (AVS) (Hammerschmidt and Fitzgerald, 2004). There have been extensive efforts aimed at modeling Hg transport, speciation, and bioavailability in aquatic environments. In Hg models, various processes and biogeochemical reactions have been included (e.g., atmospheric deposition, diffusive flux, biogeochemical transformation, sorption processes with particles, resuspension, and biouptake) to simulate the Hg cycling within these aquatic ecosystems. Existing Hg models are well described elsewhere (Bale, 2000; Braga et al., 2000) and most bioaccumulation models for trace metals, including Hg, have been run under steady-state conditions (Fisher et al., 2000). Kinetic models have been developed to effectively and quantitatively separate uptake pathways of contaminants by aquatic organisms (Thomann et al., 1995; Morrison et al., 1997; Wang et al., 1998; Fisher et al., 2000; Roditi et al., 2000) but these previous bioaccumulation models did not include physically induced processes such as resuspension. Sediment resuspension may play a large role in transporting sediment Hg and MeHg back to the water column, resulting in increased bioaccumulation of Hg and MeHg into benthic and pelagic organisms. Additionally, studies have shown that primary production and plankton density can be important factors in determining bioaccumulation of contaminants, including Hg and MeHg (Winkels et al., 1998; Pickhardt et al., 2002; Chen and Folt, 2005) and these can also be impacted by the extent of sediment resuspension. While Hg models are mostly focused on equilibrium speci-

ation, few models exist that relate ecological factors (e.g., primary production) to the bioaccumulation of Hg and MeHg within aquatic systems, and especially for coastal environments. Given the above, the objectives of this study were: (1) to develop a MeHg bioaccumulation model using data from STORM (high bottom Shear realistic water column Turbulence Resuspension Mesocosms) experiments; and (2) to examine impacts of sediment resuspension and other important factors on MeHg bioaccumulation using the model as a diagnostic tool.

2.

Methods

2.1.

STORM experiments

The model described here was developed to examine the results of a series of mesocosm experiments where sediment resuspension was simulated in a shallow estuarine environment (1 m2 cylindrical tanks, 1 m deep). Four STORM experiments of approximately 1 month duration were conducted (Experiments 1–4) and the details of the experimental design and system have been described elsewhere (Kim, 2004; Kim et al., 2004, 2006; Bergeron, 2005; Porter et al., 2004a, 2004b, 2006). Briefly, muddy sediment from Baltimore Harbor (Maryland, USA) was collected and transferred to a fiberglass holding tank at CBL. The sediment was transferred into six cylindrical STORM tanks after defaunation. The sediment was thoroughly mixed and flattened. Filtered ambient water from the mouth of the Patuxent River (a tributary of the Chesapeake Bay, Maryland, USA) was added into the tanks without any disturbance of the sediment layer to a depth of 20 cm above the sediment surface. After a 2-week sediment equilibration period, to re-establish realistic porewater gradients (Porter, 1999; Porter et al., 2004a, 2004b, 2006), unfiltered ambient water was added to the tanks (as “source water”). All the experiments were run with two treatments and three replicates. Table 1 summarizes all the experiments with different treatments. In each treatment, water column turbulence intensities were similar and the water columns of both the resuspension (R) and non-resuspended (NR) sediment systems were well mixed (e.g., no stratification). However, high instantaneous bottom shear in the R systems was induced using a specific paddle design and this generated the resuspension of sediment, whereas bottom shear velocity was low in the NR systems and no sediment resuspension occurred (Crawford and Sanford, 2001; Porter et al., 2004a). Tidal resuspension (4 h on- and 2 h off-cycles) was maintained over the 4-week period. As seen in Table 1, hard clams, Mercenaria mer-

Table 1 – Experimental design of all the STORM experiments Treatments with three replicates Experiment 1 (July 2001) Experiment 2 (October 2001) Experiment 3 (July 2002) Experiment 4 (July 2003)

Resuspension (R) vs. non-resuspension (NR) Resuspension with clams (RC) vs. non-resuspension with clams (NRC) Resuspension with clams (RC) vs. resuspension with no clams (RNC) Resuspension with high density clams (RHC) vs. resuspension with low density clams (RLC)

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Fig. 1 – Model schematic: (a) carbon flow and (b) MeHg flow.

cenaria, were introduced into the sediment for some of the studies to investigate Hg and MeHg bioaccumulation. Additionally, Experiments 3 and 4 were aimed at examining the effects of clams on plankton productivity and accumulation of MeHg. The model developed here used data from Experiments 2 and 3 as inputs (Kim et al., 2004, 2006; Bergeron, 2005).

2.2.

Model structure

The basic structure of the model developed in this study is similar to that of Ashley (1998) and Chang (2001), with some modifications. The carbon model consists of six state variables in the water column and sediment, respectively (Fig. 1a). The model includes phytoplankton (PP) and two different size groups of zooplankton (ZP): ZP1 (>210 ␮m) and ZP2 (63–210 ␮m) in the water column. While ZP samples were collected using mesh sizes of 63 and 210 ␮m during the STORM experiments, only one group of ZP (>210 ␮m) was analyzed for Hg and MeHg. The two groups of ZP feed on PP, as the primary food, and on resuspended microphytobenthos (MPB), as it has been found that MPB could serve as a food source for ZP in estuarine environments (Kibirige and Perissinotto, 2003). The model also included predation on small ZP2 by ZP1, similar to other modeling studies (Verity, 2000; Griffin et al., 2001). Microphytobenthos and filter feeders (FF) (clams) are included in the sediment pool. Filter feeders consume PP, resuspended particulate organic carbon (RPOC), and MPB. There are two sediment layers in the model: the surface sediment (top 2 cm) and the deeper layer (below 2 cm). Particulate organic carbon (POC) in the water column was divided into two pools of carbon, RPOC and water column

269

particulate organic carbon (WPOC) (Fig. 1a). While RPOC was derived from sediment organic carbon (SPOC) during resuspension events, sources of WPOC were PP and ZP mortality, as well as bacterial uptake. Water column particulate organic carbon and PP sinking were connected to the SPOC pool. Both WPOC and SPOC consisted of living (e.g., bacteria) and non-living organic matter in the model. Dissolved organic carbon pool in the water column (WDOC) gained carbon from excretion of biota and degradation of both WPOC and RPOC and lost DOC to these fractions due to bacterial uptake. Particulate organic carbon degradation in both the water column and sediment was assumed to be a first-order reaction. Decay constants for POC were obtained from Wainright and Hopkinson (1997). Overall, the model included vertical carbon exchange between the water column and sediment such as (a) FF ingestion of PP, resuspended MPB and RPOC; (b) sinking of WPOC and PP; (c) resuspension/deposition of SPOC between the sediment and the water column; (d) porewater DOC (SDOC or PWDOC) diffusion between the water column and the sediment (Fig. 1a). The diffusive flux was modeled as a bi-direction flow system where the direction of flows was determined by concentration gradients between the water column and the sediment. Diffusion coefficients for MeHg and organic matter were obtained from the literature (Gill et al., 1999). After initiating the system with unfiltered Patuxent River water, there was no particle input in all the experiments (Experiments 1–4). In other words, the sediment and in situ plankton (derived from PP and ZP in the “source” water) were the direct sources of particles in the water column. There were, however, dissolved inputs as filtered seawater was daily exchanged at 10% of the total volume during the experiments. Water exchange was always done during the off-cycle to minimize particle loss, and it was modeled that losses of all the variables in the water column occurred only during the offcycles. Since 10% of the total water volume was flushed out every day, PP and ZP were also set to be lost at 10% of their biomass per day in the model. In addition, the model assumed negligible bioturbation/bioirrigation effects for simplicity as the sediments used in all the experiments were defaunated prior to the beginning of each experiment. Methylmercury flows and accumulation were incorporated into the carbon model (Fig. 1b). There were additional state variables (e.g., dissolved MeHg in the water column and porewater) and processes, including adsorption/desorption in the water column and the sediment as well as methylation/demethylation in the sediment (Fig. 1b). Gas exchange at the air–water interface was assumed to be negligible. The results of mercury stable isotope incubation experiments with both surface and deep waters of the Chesapeake Bay (Maryland, USA), whose salinity ranged from 6.5 to 34, suggested that photodemethylation was not dominant (Whalin et al., 2007). Thus, photodemethylation was assumed to be negligible in the model and biotic water column demethylation was also not included. All the state inventories in the sediments were normalized to the volume of the water column. All the standing stocks are in units of g m−3 and flows are in g m−3 h−1 . Initial PP biomass was estimated from the measurements of Chl a at the beginning of Experiment 2. A carbon to Chl a ratio

270

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

of 50:1 was assumed, as used in other studies (Bougis, 1976; Dagg and Wyman, 1983; Griffin et al., 2001; Harding et al., 2002). Dry weight of ZP and FF was obtained from published equations, based on length-weight relationships (White and Roman, 1992; Grizzle et al., 2001). The carbon content (dry weight) of ZP and FF was assumed to be 40% (Gorsky et al., 1988; Jerling and Wooldridge, 1995; Froneman, 2000; SiokouFrangou et al., 2002). Model simulations were carried out using STELLA II® for 576 and 672 h (24 and 28 days), mimicking Experiments 2 and 3, respectively. The time step used in the model was 0.25 h with the Rung-Kutta Type II integration method. This method gives a more accurate estimate for the change in stocks during a given time step, compared to Euler’s approximation approaches. Detailed calculation steps are described in detail elsewhere (Stella Manual).

2.3.

Model formulation

2.3.1.

Carbon flow model

The model formulation and equations are given in more detail in Kim (2004). Briefly, each state variable is calculated as the difference between all the gain terms minus all the loss terms at each time step, as summarized in Appendices A and B. Additionally, all the parameters used in the model are presented in Appendix C. Phytoplankton growth is based on light, temperature, and nutrients in the model. Although light availability is an important factor in photosynthesis, the light influence on PP growth was modeled here as only the duration of day light (10 h per day), not light intensity. Phytoplankton growth rate was modeled using the following equation and was a function of temperature (modified from Angelini and Petrere, Jr., 2000):

PP growth rate = Gmax e(0.0693×Temp) where Gmax = maximal growth rate of PP (h−1 ), ◦ Temp = temperature data ( C). The effect of nutrients on growth rate was expressed as a Michaelis–Menten equation: nutrient =

nut K + nut

where nut = nutrient data (nitrate and nitrite), K = halfsaturation constant. Thus, the overall PP production in the model was expressed as P = PP growth rate × Nutrient × Light × PP (Appendix A). The measurements of temperature and nutrients every 2–3 days from Experiments 2 and 3 were used in this model. Similar approaches to model PP growth have been used by Griffin et al. (2001) and Darrow et al. (2003). Microphytobenthos growth was modeled in a similar fashion as PP in the water column. Although a carbon stock of resuspended MPB was not included explicitly in the model, consumption of resuspended MPB by ZP and FF was included. Studies have shown that ZP feed on resuspended MPB (Baillie and Welsh, 1980; de Jonge and van Beusekom, 1992). Kim et al. (2004) found in the STORM experiments that MPB was not substantially resuspended to the water column. Similarly, a mesocosm experiment by Sloth et al. (1996) found that a small amount of MPB (<2%) was transported to the water column for a 2-h resuspension period.

After the model was calibrated and compared with the data in Experiment 2, it was assumed that 5% of MPB was resuspended and consumed by both ZP and FF. For organisms gaining carbon through food ingestion, carbon consumption was modeled using the following equation (Appendix A): grazing = ZP (or FF) × FR × PP × AE × f (× ˛ for FF) where ZP (or FF) = zooplankton (or filter feeder) biomass (g C m−3 ), FR = weight-specific filtration rate (often referred as clearance rate, m3 h−1 g C−1 ), PP = phytoplankton (or microphytobenthos) biomass (g C m−3 ), AE = carbon assimilation efficiency, f = fraction of diet from a particular source, ˛ = multiplier for clam feeding. Filtration rate was represented using the equation modified from White and Roman (1992): FR = FRmax (1 − e−0.009×Temp ) where FR = weight-specific filtration rate (m3 h−1 g C−1 ), Temp = temperature. A different FRmax was used for ZP1 and ZP2 as carbonspecific ingestion rate increased with temperature but decreased with body size (White and Roman, 1992; Griffin et al., 2001). Carbon-specific FRmax values used in the model fell within the range of filtration rates for ZP found in the Chesapeake Bay (White and Roman, 1992). Similarly, FRmax for FF used in the model was within the range of filtration rates found in Grizzle et al. (2001). Porter et al.’s (2007) clam gape experiments showed that clams feed about 62% of the time regardless of resuspension. Thus, in the model, the multiplier of 0.62 (˛) was included to take the non-constant clam feeding into account. Another modeling study also used a similar value (0.67; Padilla et al., 1996). Similarly, it has been found that filter feeders such as zebra mussels filter about 50–67% of the time (Morton, 1969; Walz, 1978). In addition, filter feeders may cease filtering when algal concentration decreases below the threshold level of approximately 0.025 mg C L−1 (Riisgård et al., 2003). Thus, for improved model stability, herbivores were set to stop filtering when PP concentration reached below the threshold level.

2.3.2.

Modeling resuspension

Physical processes (i.e., resuspension (erosion, E) and deposition, D) were modeled using the following equation (Sanford and Halka, 1993; Chang, 1999): DE = Ws C where DE = deposition rate (g m−2 h−1 ), Ws = settling speed (m h−1 ), C = particle concentration in the water column (g m−3 ). When erosion occurs, it is assumed that the deposition rate equals the erosion rate at equilibrium. Then, ER = DE, so ER = Ws × Ceq . In the model, a similar approach to Wainright and Hopkinson (1997) was adopted, such that resuspension cycles were modeled using a resuspension timing parameter (either 1 or 0). In their model, resuspension and deposition did not occur simultaneously (e.g., when resuspension time was 1, deposition time was 0). However, in this model, continuous deposition was assumed. Modeling studies have shown that a continuous deposition assumption results in better agreement

271

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Fig. 2 – Comparison of model outputs and experimental data: (a) biomass in the water column; (b) biomass in the sediment; (c) MeHg in biota (water column); and (d) MeHg concentration in biota (sediment).

of models with data (Sanford and Halka, 1993; Sanford and Chang, 1997). Thus, in the model, the erosion was expressed as ER = Ws /water depth × Resp T × Ceq , where Resp T is the resuspension timing parameter as described above, and Ceq was obtained from the POC data in the water column (Fig. 2a). The overall equation with all the flows between SPOC1 and RPOC is described in Appendix A.

used to model MeHg accumulation in PP: MeHgPP (t) = MeHgPP (t − dt) + (MeHgI DUPP +MeHgDOC DUPP − MeHg SPP − MeHg EPP − MeHg MPP − MeHg GRPP−ZP1 − MeHg GRPP−ZP2 − MeHg GRPP−FF − MeHg OUTPP ) × dt

2.3.3.

Modeling bioaccumulation of MeHg

As seen in Fig. 1b, each state variable was represented as the time-variable MeHg inventory in response to each carbon state variable. As mentioned earlier, there were two additional state variables, dissolved MeHg and dissolved porewater MeHg. Since dissolved MeHg is mainly associated with DOC (Mason et al., 1999; Ravichandran, 2004), two fractions of dissolved MeHg were defined in the model: (1) dissolved MeHg bound to DOC (MeHgDOC ); and (2) MeHg bound to inorganic species, or as the free MeHg ion (MeHgI ), in both water and porewater. These fractions have different bioavailability to PP. The fractions of each species were calculated based on initial concentrations of MeHg and DOC, assuming that the species were at thermodynamic equilibrium. Methylmercury accumulation into PP was only from the dissolved phase. It was assumed that MeHg bound to DOC was available to PP, but with a lower uptake rate. Thus, the following equation was

where MeHgPP = MeHg concentration in phytoplankton (g m−3 ), MeHgI DUPP = dissolved MeHgI transfer to phytoplankton (g m−3 h−1 ), MeHgDOC DUPP = dissolved MeHgDOC transfer to phytoplankton (g m−3 h−1 ), MeHg SPP = MeHg transfer to SPOC by sinking (g m−3 h−1 ), MeHg EPP = MeHg excretion (g m−3 h−1 ), MeHg MPP = MeHg transfer to WPOC by mortality (g m−3 h−1 ), MeHg GRPP-ZP1,ZP2 = MeHg transfer to zooplankton by grazing on phytoplankton (g m−3 h−1 ), MeHg GRPP-FF = MeHg transfer to filter feeder by feeding on phytoplankton (g m−3 h−1 ), MeHg OUTPP = MeHg loss by water exchange (g m−3 h−1 ). Dissolved MeHg uptake into PP was modeled using the following equation: MeHgI DUPP = UR1 ×

PP × MeHgI MC

where UR1 = uptake rate (m3 h−1 cell−1 ), MC = mass of cell (g cell−1 ).

272

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

A similar equation was used for MeHgDOC (Appendix B). The uptake rate of MeHg into PP was estimated from experimental data in Mason et al. (1996). It was assumed that PP were spherical with a radius of 5 ␮m. Most filter feeders have been found to be able to retain particles >5 ␮m with maximum efficiency (Young et al., 1996; Grizzle et al., 2001). Similar equations were used for MeHg bioaccumulation into ZP and FF (Appendix B).

2.3.4.

Sorption processes for MeHg

Chang (1999) developed a 1D numerical time and depth dependent water quality model in order to include reversible sorption processes for a hydrophobic organic contaminant (HOC). Most models to date have assumed equilibrium partitioning between the dissolved and particulate phases but this is unrealistic for strongly bound contaminants such as HOCs and Hg. Chang (1999) used a linear-reversible model to simulate sorption processes of pyrene as Rads = k1 × Cd − k−1 × Cs = −Rdes , where Rads = the net adsorption rate, k1 = adsorption rate constant, Cd = dissolved concentration of pyrene, k−1 = desorption rate constant, Cs = particulate concentration of pyrene, Rdes = the net desorption rate. At equilibrium, Rads = 0, then, k1 /k−1 = Cs /Cd = Kd , where Kd is the distribution coefficient. A similar approach was used for sorption processes of MeHg in this model. Appendix B shows how sorption processes for MeHg are expressed in the model. As an example, adsorption and desorption between MeHgI and RPOC were modeled using the following equations: MeHgI ADSRPOC = ADS R2 × RPOC × MeHgI ; MeHgI DESRPOC = DES R2 × MeHgRPOC × (1 − fMeHgDOC ) where ADS R2 = adsorption rate constant 2 (m3 g−1 h−1 ), RPOC = RPOC concentration (g m−3 ), MeHgI = MeHgI concentration (g m−3 ), DES R2 = desorption rate constant 2 (h−1 ), MeHgRPOC = MeHgRPOC concentration (g m−3 ), Fraction as MeHg = 1 − fMeHgDOC . The desorption rate constant was obtained from Hintelmann and Harris (2004) and the organic-carbon based Kd from the actual data obtained in the mesocosm experiment (Kim et al., 2004). The adsorption rate constant was calculated using the formulation outlined by Chang (1999) above.

2.3.5. Modeling methylation/demethylation in the sediment In the model, sediment methylation/demethylation processes were treated as pseudo-first order reactions. Sediment total Hg (THg) concentration was used as a constant source to the MeHg pool in this model. As mentioned earlier, the rate constants were obtained from the measurements made in Experiment 2 (Kim et al., 2006). Thus, the model includes methylation and demethylation in the sediment using the following equation: d[MeHg] = k1 [Hg] − k2 [MeHg] dt where [MeHg] = sediment MeHg concentration (MeHgSPOC1 , MeHgSPOC2 ) (g m−3 ), [Hg] = sediment THg concentration (g m−3 ), k1 = methylation rate (h−1 ), k2 = demethylation rate (h−1 ).

No attempt was made to model how resuspension changed methylation/demethylation rates over time as there were uncertainties in parameterizing, amongst others, microbial activity, sulfate reduction rate, and the bioavailable fraction of Hg.

2.3.6.

Sensitivity analysis

A sensitivity analysis was performed to examine the effect of changes in key parameters on the state variables of interest in the model (Jorgensen, 1994). The following equation was used for the sensitivity analysis (Simas et al., 2001): S=

V × 100 V

where S = sensitivity of each state variable to a chosen parameter, V = state variable under initial condition (base case), V = variation in the state variable due to changes in the chosen parameter. One parameter (Experiment 2) was changed at ±20% of its default value while keeping all other parameters the same as the initial condition. The resulting changes for each stock were averaged for the model run time and compared with the average of the base case.

2.4.

Model applications

2.4.1.

Effects of filter feeders

After model calibration based on the Experiment 2 data, the model was run with conditions of Experiments 3 and 4 (Tables 1 and 2). The model was run with 3 scenarios (i.e., the same FF biomass as Experiment 3, half the FF biomass, and without FF). To allow comparison with data, the model was run with inputs used from the data in Experiment 3 (Bergeron, 2005). The initial biomass of PP, ZP and MPB were used from the data in Experiment 3. The same FF biomass was used because of the same density of FF in both Experiments 2 and 3. Average water temperature was higher in Experiment 3, compared to Experiment 2. Initial concentrations of dissolved MeHg, DOC, POC and nutrient data were also obtained from Experiment 3 data. The organic-carbon based Kd for MeHg was used based on measurements from Experiment 3.

2.4.2. Model simulations with different organic carbon contents The model was applied to field situations with some modifications required to run the model for a longer period (May to October). Details are described elsewhere (Kim, 2004). Briefly, inputs for nutrient and temperature were obtained from the average data for the mainstem Chesapeake Bay (http://www.chesapeakbay.net) (Chesapeake Bay Program data for stations CB5.1, CB5.2, and CB6.4). These three stations showed a range of sediment organic matter content, high (CB5.2, 12% OM), medium (CB6.4, 6% OM), and low (CB5.1, 3% OM). The stations chosen here were within the mesohaline-polyhaline regions of the Bay, as salinity in the STORM experiments with clams was within a range between the two regions. Sediment resuspension was modeled in a similar way to the model developed for the STORM experiments. The field data suggested that total suspended solids (TSS) concentrations for the three stations were relatively con-

273

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Table 2 – A summary of the water column data from Experiments 1–4 (Kim et al., 2004; Bergeron, 2005) Temperature (◦ C)

TSS (mg L−1 )

Chl a

Part. THg (nmol g−1 )

Part. MeHg (pmol g−1 )

Diss. THg (pM)

Diss. MeHg (pM)

E1

R NR

25 ± 1.2 25 ± 1.3

150 ± 27 10 ± 0.2

24 ± 2 13 ± 0.9

2.3 ± 0.1 1.1 ± 0.05

11 ± 2.0 34 ± 5.0

5.5 ± 1.0 5.5 ± 1.0

0.3 ± 0.2 0.3 ± 0.1

E2

RC NRC

20 ± 1.9 20 ± 1.9

63 ± 22 4.5 ± 0.6

6.7 ± 0.3 3.6 ± 0.1

2.3 ± 0.2 1.4 ± 0.05

6.0 ± 1.0 26 ± 5.0

8.0 ± 0.5 6.0 ± 0.3

0.2 ± 0.05 0.2 ± 0.05

E3

RC RNC

26 ± 0.02 26 ± 0.01

130 ± 46 49 ± 16

6.9 ± 3.1 19 ± 8.5

2.4 ± 0.3 2.3 ± 0.2

6.4 ± 0.8 9.4 ± 5.6

9.7 ± 3.8 9.2 ± 6.5

0.3 ± 0.1 0.4 ± 0.07

E4

RHC RLC

24 ± 0.1 24 ± 0.08

90 ± 54 52 ± 20

12 ± 8.2 23 ± 8.2

2.3 ± 0.3 2.3 ± 0.6

14 ± 9.8 15 ± 14

5.1 ± 0.8 4.0 ± 1.2

0.3 ± 0.2 0.3 ± 0.1

Data are shown as the average and standard deviation of three replicate tanks.

stant over time, averaging 6.9 ± 1.4 mg L−1 . Thus, TSS in the water column was used as a constant, after converting to POC concentration. It should be noted that TSS concentration was much lower than that of the R tanks during the resuspension period but comparable to TSS of the NR tanks in the STORM experiments (Kim et al., 2004; Bergeron, 2005). Compared to the model run with the STORM experiment data, this model application was run with a 50% decrease in loss terms for ZP and FF (e.g., excretion, respiration, and mortality) because zooplankton (ZP1 and ZP2) and filter feeders (FF) were modeled as feeding only on phytoplankton (PP). Such a reduction in the loss terms was necessary to compensate for the dramatic biomass reduction in ZP and FF due to food limitation during the longer period of model running, and to help the model stability. While the initial biomass of ZP1 and ZP2 was kept the same as the previous model, half the FF biomass was used in the model application, a value which falls within the range typically found in the field (Grizzle et al., 2001). Additionally, predation terms for ZP, FF, and MPB were added to simulate consumption by higher trophic level organisms in the field situation even though those organisms were not included in the model application. In this model, the Hg methylation rate constants were obtained from the relationship between % organic matter content and Hg methylation rate. Hammerschmidt and Fitzgerald (2004) found that Hg methylation potential (% day−1 ) in nearshore marine sediment collected from Long Island Sound, USA, was negatively correlated with % organic matter content. Thus, based on the relationship found by Hammerschmidt and Fitzgerald (2004), the Hg methylation rate constants were estimated and used for the model. The estimated rate constants were 0.67 × 10−3 , 2.3 × 10−3 , and 3.1 × 10−3 h−1 , corresponding to organic carbon content of 12, 6, and 3%, respectively. These rate constants used in the model were similar to those from Experiments 1 and 2 (Kim et al., 2006) and fell within the range of the same magnitude found by others for various ecosystems (Kim et al., 2004 and references therein; Heyes et al., 2006). The initial concentrations of THg and MeHg concentrations were used based on field data collected from the Chesapeake Bay (Kim, 2004; Heyes et al., 2006). Since there were no concurrent porewater MeHg measurements with sediment MeHg concentrations, the initial porewater MeHg concentration was estimated using the relationship between distribution coefficients (Kd ) and % organic

matter content in the sediment at each organic carbon content (Bloom et al., 1999; Hammerschmidt and Fitzgerald, 2004).

3.

Results and discussion

3.1.

Model outputs calibrated with observations

The results from Experiment 2 were used as the framework for the calibration and validation of the model, and the various parameters were adjusted within their reasonable environmental limits to obtain the best fit for the carbon section of the model. The MeHg concentrations were then compared to the actual data. The model results and comparison with the data are discussed below. As can be seen in Figs. 2a and b, the modeled biomass compared well overall with the measured data in the water column, and the benthic environment, over time (Figs. 2a and b). The oscillation of PP biomass in the model is due to the lack of growth at night, and the removal of PP by feeding and other processes. Filter feeder biomass slightly decreased with time in the model. The modeled MPB biomass slightly increased over time, which showed a similar pattern with the data (Fig. 2b). Even though there were no data for FF biomass changes during Experiment 2, ash-free dry weight (AFDW) measurements of selected clams at the beginning and the end of the experiment showed that clams did not grow substantially during the course of the experiment (Porter et al., 2007). This lack of growth is likely due to food limitation during the experiment, as the model-derived effective food accumulation rate constant, estimated based on filtration rate, temperature, suspended organic carbon content (biomass), clam feeding time, assimilation efficiency, and the fraction of seston as food, is of the same order as the respiration and elimination rate constants. The modeled and measured concentrations of MeHg in biota are shown in Figs. 2c and d. Modeled MeHg concentration in PP reached a plateau after the first week and decreased slightly toward the end of the model run (Fig. 2c). As only bulk suspended matter was collected and analyzed for MeHg, it is not possible to compare the model results with the actual data. Modeled MeHg in both small (ZP2) and larger ZP (ZP1) increased gradually over time. The modeled MeHg concentration in ZP1 agreed very well with the measured

274

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

concentrations. Model results showed that the MeHg concentration in FF and MPB increased slightly over time (Fig. 2d). Modeled MeHg accumulation in ZP and FF was mainly from food ingestion (>95%), and not from dissolved MeHg uptake. Other studies have also shown that aquatic invertebrates accumulated contaminants mainly from food ingestion, and this is especially true for MeHg (Luoma et al., 1992; Wang et al., 1998; Lawrence and Mason, 2001; Chang and Reinfelder, 2002; Tsui and Wang, 2004). Overall, the agreement between the model and data for both carbon and MeHg suggest that the model provides a good simulation of the mesocosm dynamics. The results of sensitivity analysis are presented in Table 3. The results showed that PP growth rate was a highly sensitive parameter to both PP and ZP biomass but less so to FF

biomass. Similarly, filtration rates of FF had a great impact on both PP and ZP biomass. As seen in Table 3, changing PP growth rate resulted in changing MeHg concentrations in PP and ZP but the extent was much less than changes in biomass. Nonetheless, it was apparent that MeHg burden decreased, as biomass increased (bio-dilution). The results suggested that PP growth rate was not the only parameter influencing MeHg concentrations in biota. Changing dissolved MeHg uptake rate for PP (±20%) resulted in about a 17% change in PP body burden and subsequently a 10% change in ZP MeHg concentrations. In addition, varying Hg methylation in the sediment resulted in about 10% changes in PP and ZP body burden. The link between MeHg concentration in the sediment and biota in the water column was due to sediment resuspension. The trans-

Table 3 – Model sensitivity analysis (results are shown in percent) PP

ZP1

ZP2

FF

PP growth rate +20% −20%

47.1 −52.2

39.0 −48.6

918 −65.1

4.5 −3.0

ZP1 filtration rate +20% −20%

−3.6 2.4

48.7 −31.6

−7.6 6.1

−0.2 0.1

ZP2 filtration rate +20% −20%

−6.2 3.8

−6.3 4.7

71.3 −43.1

−0.2 0.2

FF filtration rate +20% −20%

−32.7 37.0

−35.6 131

−50.9 276

0.7 −1.0

Methylation rate +20% −20%

NC NC

NC NC

NC NC

NC NC

Diss. MeHg uptake rate +20% −20%

NC NC

NC NC

NC NC

NC NC

MeHgPP

MeHgZP1

MeHgZP2

MeHgFF

PP growth rate +20% −20%

−18.0 29.1

−19.9 32.1

−21.2 39.2

−3.8 2.8

ZP1 filtration rate +20% −20%

NC NC

2.2 −2.2

1.3 NC

0.1 −0.1

ZP2 filtration rate +20% −20%

NC NC

2.1 −1.2

2.1 NC

0.2 −0.1

FF filtration rate +20% −20%

2.2 −1.9

10.5 −8.4

13.1 −9.4

1.0 −0.8

Methylation rate +20% −20%

11.2 −11.2

10.0 −10.0

9.8 −9.8

2.0 −2.0

Diss. MeHg uptake rate +20% −20%

17.0 −17.5

9.9 −10.2

10.0 10.3

0.2 −0.2

NC: no change.

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

ferred sediment MeHg desorbed and became available for PP uptake.

3.2.

Model application

3.2.1.

Effects of filter feeders

As mentioned earlier, to examine the impact of filter feeders on the system dynamics, the model was run with the parameters used in Experiment 3 and with the differing clam densities used in Experiments 3 and 4 (Tables 1 and 2). This allowed an assessment of the impact of different clam densities on MeHg dynamics. Fig. 3a and b shows the modeled biomass with data from Experiment 3. In contrast to the model results of PP biomass in the early stage of the run, the model did not predict the later PP bloom in the observed data (Fig. 3a). Similarly, the model results for ZP biomass showed better agreement with the data in the early stage of the model run (Fig. 3b). Overall, the model showed that PP biomass substantially decreased, as FF biomass increased. A similar pattern was observed for ZP biomass. This was ascribed to PP being more available to ZP as FF biomass decreased (i.e., reduction in FF filtering activity). Thus, the effect of changes in FF biomass on ZP appeared to be indirect, given that FF grazing on ZP was not considered in the model. Additionally, there was a time lag between PP and ZP peaks during the model runs probably because the grazing rate by ZP was slower than the PP growth rate, and thus the response time to changes in PP was slower.

275

Similarly, other modeling studies have shown that benthic filter feeders (e.g., zebra mussels) had a great impact on PP biomass, and this is expected at low biomass levels due to grazing by filter feeders (Padilla et al., 1996; Caraco et al., 1997; Descy et al., 2003). Additionally, high zebra mussel density resulted in a decrease in small zooplankton such as rotifers especially in summer when mussels actively filter (Viroux, 2000). Benthic filter feeders may not have similar impacts on different plankton communities. Padilla et al. (1996) found from their modeling study that mussel impact was greater on large PP than on small PP. It was likely that small PP compensated for grazing losses by enhanced growth due to increasing nutrient cycling and water clarity. As a result, ZP that consumed mainly small PP was less affected by mussels. Although changes in FF biomass showed a great impact on PP and ZP biomass, MeHg concentrations in PP and ZP were affected to a lesser degree by varying FF biomass (Fig. 3c and d). Overall, the relationship between benthic filter feeders and plankton can be more complicated than modeled here due to differences in size classes of prey and feeding preferences of ZP and FF, as well as other factors. Phytoplankton biomass may not compensate by enhanced growth if there are nutrientlimited conditions and/or lower temperature. Filtration rate of FF can be affected by temperature, size of FF, hydrodynamic processes, and food availability (Riisgård, 1998; Grizzle et al., 2001; Riisgard et al., 2003; Newell, 2004), and may influence MeHg accumulation in plankton as well as benthic filter feeders.

Fig. 3 – Effects of changes in FF biomass on: (a) PP biomass; (b) ZP1 biomass; (c) MeHg in PP; and (d) MeHg in ZP1.

276

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Fig. 4 – Effects of sediment organic matter contents and the subsequent Hg methylation rates on: (a) sediment MeHg; (b) particulate MeHg in the water column; and (c) dissolved MeHg in the water column. The methylation rate constants used in the model were 0.67 × 10−3 , 2.3 × 10−3 , and 3.1 × 10−3 h−3 , corresponding to sediment organic matter content of 12, 6, and 3%, respectively.

3.2.2. Effects of changes in sediment Hg methylation on MeHg bioaccumulation As mentioned above, the model was applied under conditions of 3, 6, and 12 % sediment organic matter and the model was run for a longer duration (May to October) to simulate conditions in the Chesapeake Bay. Sediment, particulate, and dissolved MeHg concentrations are presented in Fig. 4. As seen in Fig. 4a and c, MeHg concentrations at 3% organic matter (Hg methylation rate constant, 3.1 × 10−3 h−1 ) were higher, compared to 6 and 12% organic matter with corresponding Hg methylation rate constants, 2.3 × 10−3 and 0.67 × 10−3 h−1 , respectively. Thus, it was apparent that higher Hg methylation rate constants resulted in increasing sediment MeHg concentration and subsequently resulted in the transfer of the increased MeHg sediment to the water column via sediment resuspension. However, it should be noted here that although the model for 3, 6, and 12% organic matter cases used different Hg methylation rate constants, sediment Hg concentration, methylation and demethylation rate constants were kept constant over the experimental period. Therefore, the results showed a similar pattern that sediment MeHg concentration was relatively constant over time, regardless of changes in sediment organic content. In the model, once MeHg contaminated sediments are resuspended to the

water column, sediment MeHg desorbs, is ingested directly by benthic-pelagic organisms, or re-settles to the sediment surface. As seen in Fig. 4b, particulate MeHg concentration in the water column increased with enhanced sediment MeHg (higher Hg methylation rate constant). It appeared that a small fraction of particulate MeHg desorbed (Fig. 4c) and the change in dissolved MeHg concentration over time was closely linked to MeHg uptake by PP. This will be further discussed in the following section. Fig. 5 shows the resulting MeHg concentrations in PP, ZP1, and FF with varying sediment organic matter content and the subsequent Hg methylation rate constants. As seen in Fig. 5a, MeHg concentration in PP showed a similar pattern with dissolved MeHg concentration in the water column (Fig. 4c). In addition, MeHg concentrations in ZP1 and FF increased with decreasing % organic matter (Fig. 5b and c). The overall results showed that changes in MeHg concentrations in the water column were closely linked to varying sediment MeHg due to sediment resuspension. Thus, the model results suggested that resuspension could play an important role in transferring sediment MeHg to the water column, enhancing the dissolved concentration and therefore increasing subsequent MeHg accumulation into benthic-pelagic organisms.

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

277

Fig. 5 – Effects of sediment organic matter contents and the subsequent Hg methylation rates on: (a) MeHg in PP; (b) MeHg in ZP1; and (c) MeHg in FF. The methylation rate constants used in the model were 0.67 × 10−3 , 2.3 × 10−3 , and 3.1 × 10−3 h−1 , corresponding to sediment organic matter content of 12, 6, and 3%, respectively.

3.2.3. Effects of sediment resuspension on MeHg bioaccumulation In order to compare the model results with/without sediment resuspension, the model was run with no sediment resuspension at 3% organic matter (achieved by shutting down all the flows in/out of the RPOC and MeHg in RPOC pools). The results are presented in Fig. 6. While sediment MeHg were similar between resuspension and non-resuspension, particulate and dissolved MeHg were substantially higher with sediment resuspension (Fig. 6a). The overall average of MeHg concentration increased by 25 and 72% for particulate and dissolved MeHg, respectively. As seen in Fig. 6b, PP biomass did not significantly change with sediment resuspension. However, MeHg in PP showed a significant change between sediment resuspension and non-resuspension (p < 0.05), averaging 178 (R) to 53 pmol g C−1 (NR). Similarly, MeHg in ZP1 increased significantly to about 67% with sediment resuspension (p < 0.05) (Fig. 6c). Interestingly, MeHg in FF slightly increased (7%) under non-resuspension conditions. This change reflected a lack of FF feeding on RPOC in the model, which resulted in a decrease in FF biomass (23%) (Fig. 6d). In addition, the results showed that a 25% increase in ZP1 biomass with non-resuspension was likely a response to the decrease in FF biomass, and associated feeding capacity. This suggested that PP became more available to ZP1 as FF

feeding on PP decreased and a similar result was found for ZP2 (data not shown). Thus, it appeared that there was a complex relationship between prey and predation under the given feeding conditions. Nonetheless, the overall results showed that sediment resuspension could play a substantial role in reintroducing MeHg in sediments to the water column and food chains, as suggested by others (Sunderland et al., 2004; Sager, 2002). In addition, this provides insight into why MeHg concentration in biota, especially in the upper levels of food chains, may remain high even though Hg loadings to environments may be reduced. Despite the model’s promising results and the clear demonstration that it could be used as a diagnostic tool to examine many different aspects of MeHg cycling in shallow ecosystems, this model, as with all models, has some limitations that are a function of the degree of parameterization of the variables, and the lack of detail in some aspects of the model framework. Thus, these results may not be directly extrapolated to different ecosystems. The model was initially developed to mimic Experiment 2 conditions, which did not include benthic deposit feeders and higher trophic level organisms. Additionally, to improve the model results, bioturbation effects on in situ MeHg production need to be included. There has been evidence that bioturbation can increase Hg methylation in sediments (Hammerschmidt et al.,

278

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Fig. 6 – Comparison of model outputs with sediment resuspension (R) and non-resuspension (NR) at 3% sediment organic matter content: (a) MeHg in the sediment and the water column; (b) PP MeHg and biomass; (c) ZP1 MeHg and biomass; and (d) FF biomass and MeHg.

2004; Benoit et al., 2006). Impacts of Hg methylation on THg and MeHg bioaccumulation into deposit feeders can be critical as they are potentially exposed to THg and MeHg from both the porewater/overlying water and the sediments (solid phase), although Lawrence and Mason (2001) concluded that sediment and settling organic matter were more important sources of MeHg to amphipods than either porewater or overlying water. Sager (2002) suggested that continued Hg release, resuspension and/or bioturbation of buried sediments with enhanced Hg can contribute to elevated Hg levels in the higher trophic organisms. Similarly, it was evaluated that most of the MeHg in PP of Long Island Sound was due to sedimentary flux (diffusion, advection, and resuspension) and this could contribute to MeHg accumulation in higher trophic levels (Hammerschmidt et al., 2004). In addition, the model used both methylation/ demethylation rate constants as fixed values over time. As environmental factors, such as temperature, can play a role in controlling microbial activity involved in Hg methylation and MeHg demethylation, a better parameterization of both methylation and demethylation rate constants is necessary, especially for a longer duration model simulation, where seasonal effects could be important. In the model

applications, the effects of sulfide levels on Hg methylation were not included but the extension of the model to include such interactions would improve the predictive power of the model. Organic matter content in sediments was appears to be the principal control on MeHg production in low sulfide (<0.01 ␮M) sediments (Hammerschmidt and Fitzgerald, 2004). However, the effects of sulfide are more important than organic content in higher sulfide sediments, with apparent inhibition of methylation at levels above 0.1 ␮M (Benoit et al., 2003). Although the modeling study has the limitations stated above, it has provided insight into some potential management concerns. In many estuarine systems, changes in nutrient loadings are occurring, either increasing as a result of either enhanced nutrient inputs, or in some locations, decreasing as a resultant of nutrient reduction programs. These changes will lead to changes on sediment organic content, which could affect MeHg production by controlling the bioavailability of Hg in the sediment, as indicated above. For example, a reduction in nutrient loadings could therefore inadvertently lead to an enhancement of MeHg concentration in the sediments, as suggested by Hammerschmidt and Fitzgerald (2004).

279

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

4.

in the mesocosms, it could be expanded to include better parameterization of Hg methylation/MeHg demethylation as well as upper trophic levels over a longer simulation period.

Summary

The derived model produced comparable results with observed data for both biomass and MeHg burden under varying scenarios. The modeling results suggested that dissolved MeHg in the water column was increased by elevated sediment MeHg transported to the water column via sediment resuspension. As a result, MeHg burden in plankton increased. Benthic filter feeders, with the dominant biomass, however, were less affected than plankton dynamics in the water column, likely as a result of the limited duration of the simulations. Model outputs were highly sensitive to phytoplankton growth and the filtration rate of filter feeders. Changes in filter feeder biomass had a great impact on plankton biomass but had less of an impact on MeHg burden in plankton, though it was still an important parameter. While the model provides a reasonable simulation of the conditions

State variables

Acknowledgement We would like to thank Dr. Elka Porter, Heather Soulen, Melissa Bonner, Sandra Fernandes, Matt Reardon and others who were involved in the STORM experiments for their help. This work formed part of the Ph.D. dissertation of Eunhee Kim at the University of Maryland and was partially supported by a grant from the Hudson River Foundation (Grant No. 009-01A), and was also partially supported by Grant No. R824850-01-0 from the US EPA STAR Program.

Appendix A. Carbon-based state variables and equations

Description

Equation PP(t) = PP(t − dt) + (PPP − SPP − RPP − EPP − MPP − GRPP-ZP1 – GRPP-ZP2 – GRPP-FF – OUTPP ) × dt

Phytoplankton (PP) PPP

PP production

PP growth × Nutrient × Light × PP

SPP

PP sinking

PP × SPP R

RPP

PP respiration

PP RPP R

EPP

PP excretion

PP EPP R

MPP

PP mortality

PP MPP R

GRPP-ZP

PP grazing by ZP1 and ZP2

If PP ≥ 0.025 then PP × ZP × FRZP1

GRPP-FF

PP grazing by FF

If PP ≥ 0.025 then PP × FF × FRFF × AE2 × ˛ × fPP

OUTPP

Loss due to water exchange >210 ␮m

If resuspension = 0 then PP × 0.1 × 1/8a else 0

(ZP2)

× AE1 × fPP

else 0

else 0

Zooplankton1 (ZP1)

ZP1(t) = ZP1(t − dt) + (GRPP-ZP1 + GRZP2-ZP1 + GRMPB-ZP1 − RZP1 − EZP1 − MZP1 – OUTZP1 ) × dt

GRZP2-ZP1

ZP2 grazing by ZP1

ZP2 × FRZP1 × ZP1 × AE3 × fZP2

GRMPB-ZP1

MPB grazing by ZP1

ZP1 × FRZP1 × 5% MPB × AE1 × fMPB

RZP1

ZP1 respiration

ZP1 RZP1 R

EZP1

ZP1 excretion

ZP1 EZP1 R

MZP1

ZP1 mortality

ZP1 MZP1 R

OUTZP1

ZP1 loss

If resuspension = 0 then ZP1 × 0.1 × 1/8 else 0

Zooplankton2 (ZP2)

63–210 ␮m

ZP2(t) = ZP2(t − dt) + (GRPP-ZP2 + GRMPB-ZP2 − GRZP2-ZP1 − RZP2 − EZP2 − MZP2 − OUTZP2 ) × dt

GRMPB-ZP2

MPB grazing by ZP2

ZP2 × FRZP2 × 5% MPB × AE1 × fMPB

RZP2

ZP2 respiration

ZP2 RZP2 R

EZP2

ZP2 excretion

ZP2 EZP2 R

MZP2

ZP2 mortality

ZP2 MZP2 R

280

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix A (Continued ) State variables OUTZP2

Description ZP2 loss

Equation If resuspension = 0 then ZP2 × 0.1 × 1/8 else 0 WPOC(t) = WPOC(t − dt) + (MPP + MZP1 + MZP2 + BUWDOC − SWPOC − DEGWPOC − OUTWPOC ) × dt

Water column particulate organic carbon (WPOC) BUWDOC

Bacterial Uptake

WDOC BU R1

SWPOC

WPOC sinking

WPOC SWPOC R

DEGWPOC

WPOC Degradation

WPOC DEGWPOC R

OUTWPOC

WPOC Loss

If resuspension = 0 then WPOC × 0.1 × 1/8 else 0 WDOC(t) = WDOC(t − dt) + (EPP + EZP1 + EZP2 + EFF + INWDOC + DEGRPOC + DWPOC ± DFWDOC-PW1DOC − BUDOC − OUTWDOC ) × dt

Water column dissolved organic carbon (WDOC) EFF

FF excretion

FF EFF R

INWDOC

WDOC input from incoming water

DOC conc. in incoming water × 0.1 × 1/24b

DEGRPOC

RPOC degradation

RPOC × DEGRPOC R

DFWDOC-PW1DOC

Diffusive flux between WDOC and PW1DOC

− DCDOC × (WDOC − PW1DOC)

OUTWDOC

WDOC Loss

If resuspension = 0 then WDOC × 0.1 × 1/8 else 0 RPOC(t) = RPOC(t – dt) + (ER − DE − GRRPOC-FF − DEGRPOC − OUTRPOC ) × dt

Resuspended particulate organic carbon (RPOC) ER

Erosion

Ws /water depth × Resp T × Ceq

DE

Deposition

Ws /water depth × RPOC

GRRPOC-FF

RPOC grazing by FF

FF × FRFF × RPOC × AE4 × ˛ × fRPOC

OUTRPOC

RPOC loss

If resuspension = 0 then RPOC × 0.1 × 1/8 else 0

Sediment particulate organic carbon (S1POC)

Top 2 cm

S1POC(t) = S1POC(t − dt) + (SPP + SWPOC + DE + BUPW1DOC + MMPB + BDFF − DEGS1POC − ER − BS1POC ) × dt

BUPW1DOC

Bacterial uptake

PW1DOC × BU R2

MMPB

MPB mortality

MPB × MMPB R

BDFF

FF biodeposit

FF × BDFF R

DEGS1POC

S1POC degradation

S1POC × DEGSPOC R

BS1POC

S1POC burial

S1POC × BS1POC R

Below 2 cm

SPOC2(t) = (BSPOC1 + BUPW2DOC − DEGSPOC2 ) × dt

BUPW2DOC

Bacterial uptake

PW2DOC × BU R2

DEGS2POC

S2POC degradation

S2POC × DEG RSPOC

Sediment particulate organic carbon (S2POC)

281

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix A (Continued ) State variables Porewater dissolved organic carbon (PW1DOC)

Description Top 2 cm

PW1DOC(t) = PW1DOC(t – dt) + (DEGS1POC + EMPB ± DFWDOC-PW1DOC ± DFPW1DOC-PW2DOC – BUPW1DOC ) × dt

EMPB

MPB excretion

MPB × EMPB R

DFPW1DOC-PW2DOC

Diffusive flux between PW1DOC and PW2DOC Below 2 cm

− DC × (PW1DOC − PW2DOC)

Porewater dissolved organic carbon (PW2DOC) Filter feeder (FF)

b

PW2DOC(t) = PW2DOC(t − dt) + (DEGS2POC ± DFPW1DOC-PW2DOC – BUPW2DOC ) × dt FF(t) = FF(t − dt) + (GRPP-FF + GRRPOC-FF + GRMPB − EFF − BDFF − RFF ) × dt

GRMPB

MPB grazing by FF

FF × FRFF × 5% MPB × AE1 × fMPB × ˛

RFF

FF respiration

FF × RRFF MPB(t) = MPB(t − dt) + (PMPB − RMPB − EMPB − MMPB − GRMPB-ZP1 − GRMPB-ZP2 − GRMPB-FF ) × dt

Microphytobenthos (MPB)

a

Equation

PMPB

MPB production

MPB growth × Nutrient × Light × MPB

RMPB

MPB respiration

MPB × RMPB R

EMPB

MPB excretion

MPB × EMPB R

MMPB

MPB mortality

MPB × MMPB R

Off-cycle hours per day. Conversion factor, day to hour.

Appendix B. MeHg state variables and equations

State variables

Description

Equation MeHgPP (t) = MeHgPP (t − dt) + (MeHgI DUPP + MeHgDOC DUPP − MeHg SPP − MeHg EPP − MeHg MPP − MeHg GRPP-ZP1 − MeHg GRPP-ZP2 − MeHg GRPP-FF − MeHg OUTPP ) × dt

MeHg in PP (MeHgPP )

MeHgI diffusive uptake by PP

DUPP R1 × (PP/MC) × MeHgI

MeHgDOC DUPP

MeHgDOC diffusive uptake by PP

DUPP R2 × (PP/MC) × MeHgDOC

MeHg SPP

MeHgPP sinking

SPP /PP × MeHgPP

MeHg EPP

MeHgPP excretion

EPP /PP × MeHgPP

MeHg MPP

MeHgPP mortality

MPP /PP ×× MeHgPP

MeHg GRPP-ZP1

MeHgPP grazing by ZP1

GRPP-ZP1 /PP × MeHgPP × AE ratio1

MeHg GR PP-ZP2

MeHgPP grazing by ZP2

GRPP-ZP2 /PP × MeHgPP × AE ratio1

MeHg GR PP-FF

MeHgPP grazing by FF

GRPP-FF /PP × MeHgPP × AE ratio2

MeHgI

DUPP

282

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix B (Continued ) State variables MeHg OUTPP

Description MeHgPP loss

Equation OUTPP /PP × MeHgPP MeHgZP1 (t) = MeHgZP1 (t − dt) + (MeHgI DUZP1 + MeHgDOC DUZP1 + MeHg GRPP-ZP1 + MeHg GRZP2-ZP1 + MeHg GRMPB-ZP1 − MeHg EZP1 − MeHg MZP1 − MeHg OUTZP1 ) × dt

MeHg in ZP1 (MeHgZP1 )

MeHgI DUZP1

MeHgI diffusive uptake by ZP1

DUZP R × ZP1 × MeHgI

MeHgDOC DUZP1

MeHgDOC diffusive uptake by ZP1

DUZP R × ZP1 × MeHgDOC

MeHg GRZP2-ZP1

MeHgZP2 grazing by ZP1

GRZP2-ZP1 /ZP2 × MeHgZP2 × AE ratio3

MeHg GRMPB-ZP1

MeHgMPB grazing by ZP1

GRMPB-ZP1 /(MPB × 0.05) × MeHgMPB × AE ratio1

MeHg EZP1

MeHgZP1 excretion

EZP1 /ZP1 × MeHgZP1

MeHg MZP1

MeHgZP1 mortality

MZP1 /ZP1 × MeHgZP1

MeHg OUTZP1

MeHgZP1 loss

OUTZP1 /ZP1 × MeHgZP1 MeHgZP2 (t) = MeHgZP2 (t − dt) + (MeHgI DUZP2 + MeHgDOC DUZP2 + MeHg GRPP-ZP2 + MeHg GRMPB-ZP2 − MeHg GRZP2-ZP1 − MeHg EZP1 − MeHg MZP2 − MeHg OUTZP2 ) × dt

MeHg in ZP2 (MeHgZP2 )

MeHgI DUZP2

MeHgI diffusive uptake by ZP1

DUZP R × ZP2 × MeHgI

MeHgDOC DUZP2

MeHgDOC diffusive uptake by ZP1

DUZP R × ZP2 × MeHgDOC

MeHg GRMPB-ZP2

MeHgMPB grazing by ZP2

GRMPB-ZP2 /(MPB × 0.05) × MeHgMPB × AE ratio1

MeHg EZP2

MeHgZP2 excretion

EZP2 /ZP2 × MeHgZP2

MeHg MZP2

MeHgZP2 mortality

MZP2 /ZP2 × MeHgZP2

MeHg OUTZP2

MeHgZP1 loss

OUTZP2 /ZP2 × MeHgZP2 MeHgWPOC (t) = (t − dt) + (MeHg MPP + MeHg MZP1 + MeHg MZP2 + MeHgI ADSWPOC + MeHgDOC ADSWPOC – MeHg SWPOC − MeHg DEGWPOC − MeHgI DESWPOC − MeHgDOC DESWPOC − MeHg OUTWPOC ) × dt

MeHg in WPOC (MeHgWPOC )

MeHgI ADSWPOC

MeHgI adsorption to WPOC

ADS R1 × WPOC × MeHgI

MeHgDOC ADSWPOC

MeHgDOC adsorption to WPOC

ADS R1 × WPOC × MeHgDOC

MeHg SWPOC

MeHgWPOC sinking

SWPOC /WPOC × MeHgWPOC

MeHg DEGWPOC

MeHgWPOC degradation

DEGWPOC /WPOC × MeHgWPOC

MeHgI DESWPOC

Desorption to MeHgI

DES R1 × MeHgWPOC (1 − fMeHgDOC )

MeHgDOC DESWPOC

Desorption to MeHgDOC

DES R1 × MeHgWPOC fMeHgDOC

MeHg OUTWPOC

MeHgWPOC loss

OUTWPOC /WPOC × MeHgWPOC

Water column dissolved MeHg (inorganic) MeHgI

MeHgI (t) = MeHgI (t − dt) + (MeHgI IN + MeHgI DESWPOC + MeHgI DESRPOC ± DFMeHgI -PW1MeHgI − MeHgI DUPP − MeHgI DUZP1 − MeHgI DUZP2 − MeHgI DUFF − MeHgI DUMPB − MeHgI ADSWPOC − MeHgI ADSRPOC − OUTMeHgI ) × dt

283

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix B (Continued ) State variables

Description

Equation

MeHgI IN

MeHgI input from incoming water

MeHgI conc. in incoming water × 0.1 × 1/24

MeHgI DESRPOC

MeHgRPOC desorption

DES R2 × MeHgRPOC × (1 − fMeHgDOC )

MeHgI DUFF

MeHgI diffusive uptake by FF

DUFF R × FF × MeHgI

MeHgI DUMPB

MeHgI Diffusive Uptake by MPB

DUMPB R1 × (MPB/MC) × MeHgI

MeHgI ADSRPOC

MeHgI adsorption to RPOC

ADS R2 × RPOC × MeHgI

DFMeHgI -PW1MeHgI

Diffusive flux between overlying water and porewater

− DCMeHgI × (MeHgI − PW1MeHgI )

MeHgI OUT

MeHgI loss

If resuspension = 0 then MeHgI × 0.1 × 1/8 else 0 MeHgDOC (t) = MeHgDOC (t − dt) + (MeHgDOC IN + MeHg DESWPOC + MeHg DESRPOC + MeHgDOC DEGWPOC + MeHgDOC DEGRPOC + MeHgDOC EPP + MeHgDOC EZP1 + MeHgDOC EZP2 + MeHgDOC EFF ± MeHgDOC DFPW1MeHgDOC − MeHgDOC DUPP − MeHgDOC DUZP1 − MeHgDOC DUZP2 − MeHgDOC DUFF − MeHgDOC DUMPB − MeHgDOC ADSWPOC − MeHgDOC ADSRPOC − MeHgDOC OUT) × dt

Water column dissolved MeHg bound to DOC (MeHgDOC )

INMeHgDOC

MeHgDOC Input from incoming water

MeHgDOC conc. in incoming water × 0.1 × 1/24

MeHg DESMeHgRPOC

MeHgRPOC Desorption

DES R2 × MeHgRPOC fMeHgDOC

MeHg DEGRPOC

MeHgRPOC Degradation

DEGRPOC /RPOC × MeHgRPOC

MeHg EFF

FF Excretion

EFF /FF × MeHgFF

MeHgDOC DFPW1MeHgDOC

Diffusive Flux between overlying water and porewater

DFWDOC-PW1DOC /(WDOC – PW1DOC) × (MeHgDOC –MeHgPW1DOC )

MeHgDOC DUFF

MeHgDOC Diffusive Uptake by FF

DUFF R × FF × MeHgDOC

MeHgDOC DUMPB

MeHgDOC Diffusive Uptake by MPB

DUMPB R2 × (MPB/MC) × MeHgDOC

MeHgDOC ADSRPOC

Adsorption to RPOC

ADS R2 × RPOC × MeHgDOC

MeHgDOC OUT

MeHgDOC Loss

If resuspension = 0 then MeHgDOC × 0.1 × 1/8 else 0 MeHgRPOC (t) = MeHgRPOC (t – dt) + (MeHgSPOC1 ER + MeHgI ADSRPOC + MeHgDOC ADSRPOC − MeHg GRRPOC-FF − MeHgRPOC DE − MeHg DEGRPOC − MeHgI DESRPOC − MeHgDOC DESRPOC − MeHg OUTRPOC ) × dt

MeHg in RPOC (MeHgRPOC )

MeHgRPOC ER

MeHgRPOC Erosion

ER/SPOC1 × MeHgSPOC1

MeHg GRRPOC-FF

MeHgRPOC Grazing by FF

GRRPOC-FF /RPOC × MeHgRPOC × AE ratio 4

MeHgRPOC DE

MeHgRPOC Deposition

DE/RPOC × MeHgRPOC

284

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix B (Continued ) State variables MeHg OUTRPOC

Description MeHgRPOC Loss

Equation OUTRPOC /RPOC × MeHgRPOC MeHgS1POC (t) = MeHgS1POC (t − dt) + (MeHg SPP + MeHg SWPOC + MeHgRPOC DE + MeHg MMPB + MES1POC + MeHgPW1I ADSS1POC + MeHgPW1DOC ADSS1POC − MeHgS1POC ER − DMS1POC − MeHg DEGS1POC − MeHg BS1POC − MeHgS1POC DESPW1I − MeHgS1POC DESPW1DOC ) × dt

MeHg in S1POC (MeHgS1POC )

MeHg MMPB

MeHgMPB mortality

MMPB /MPB × MeHgMPB

MES1POC

Methylation to MeHgS1POC

ME R1 × THg1

MeHgPW1I ADSS1POC

MeHgPW1I adsorption to S1POC

ADS R3 × S1POC × MeHgPW1I

MeHgPW1DOC ADSS1POC

MeHgPW1DOC adsorption to S1POC

ADS R3 × S1POC × MeHgPW1DOC

DMS1POC

MeHgS1POC Demethylation1

DM R × MeHgS1POC

MeHg DEGS1POC

MeHgS1POC degradation

DEGS1POC /S1POC × MeHgS1POC

MeHg BS1POC

MeHgS1POC burial

BS1POC /S1POC × MeHgS1POC

MeHgS1POC DESPW1I

Desorption to MeHgPW1I

DES R3 × MeHgS1POC × (1 − fMeHgPW1DOC )

MeHgS1POC DESPW1DOC

Desorption to MeHgPW1DOC

DES R3 × MeHgS1POC × fMeHgPW1DOC MeHgS2POC (t) = MeHgS2POC (t − dt) + (MeHg BS1POC + MES2POC + MeHgPW2I ADSS2POC + MeHgPW2DOC ADSS2POC − DMS2POC – MeHg DEGS2POC – MeHgS2POC DESPW2I − MeHgS2POC DESPW2DOC ) × dt

MeHg in S2POC (MeHgS2POC )

ME2

Methylation2 to MeHgS2POC

ME R2 × THg2

MeHgPW2I ADSS2POC

MeHgPW2I Adsorption to S2POC

ADS R3 × S2POC × MeHgPW2I

MeHgPW2DOC ADSS2POC

MeHgPW2DOC Adsorption to S2POC

ADS R3 × S2POC × MeHgPW2DOC

DMS2POC

MeHgS2POC Demethylation

DM R × MeHgS2POC

MeHg DEGS2POC

MeHgS2POC Degradation

DEGS2POC /S2POC × MeHgS2POC

MeHgS2POC DESPW2I

Desorption to MeHgPW2I

DES R3 × MeHgS2POC × (1− fMeHgPW2DOC )

MeHgS2POC DESPW2DOC

Desorption to MeHgPW2DOC

DES R3 × MeHgS2POC × fMeHgPW2DOC MeHgPW1I (t) = MeHgPW1I (t − dt) + (MeHgS1POC DESPW1I ± MeHgDOC DFMeHgPW1I ± MeHgPW1I DFMeHgPW2I − MeHgPW1I ADSS1POC ) × dt

Porewater1 dissolved MeHgI (MeHgPW1I ) MeHgPW1I DFMeHgPW2I

Porewater1 dissolved MeHg bound to DOC (MeHgPW1DOC )

Diffusive flux between porewater MeHgI

DCMeHgI × (MeHgPW1I − MeHgPW2I )

MeHgPW1DOC (t) = MeHgPW1DOC (t − dt) + (MeHg DEGS1POC + MeHgS1POC DESPW1DOC ± MeHgDOC DFPW1MeHgDOC ± MeHgPW1DOC DFPW2MeHgDOC + MeHg EMPB − MeHgPW1DOC ADSS1POC ) × dt

285

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix B (Continued ) State variables

Description

Equation

MeHgPW1DOC DFPW2MeHgDOC

Diffusive flux between porewater MeHgDOC

DFPW1DOC-PW2DOC /(PW1DOC − PW2DOC) × (MeHgPW1DOC − MeHgPW2DOC )

MeHg EMPB

MPB excretion

EMPB /MPB × MeHgMPB

Porewater2 dissolved MeHgPW2I (MeHgPW2I )

MeHgPW2I (t) = MeHgPW2I (t − dt) + (MeHgS2POC DESPW2I ± MeHgPW1I DFMeHgPW2I − MeHgPW2I ADSS2POC ) × dt

Porewater2 dissolved MeHg bound to DOC (MeHgPW2DOC )

MeHgPW2DOC (t) = MeHgPW2DOC (t – dt) + (MeHg DEGS2POC + MeHgS2POC DESPW2DOC ± MeHgPW1DOC DFPW2MeHgDOC − MeHgPW2DOC ADSS2POC ) × dt

MeHg in FF (MeHgFF )

MeHgFF (t) = MeHgFF (t − dt) + (MeHgI DUFF + MeHgDOC DUFF + MeHg GRPP-FF + MeHg GRRPOC-FF + MeHg GRMPB-FF − MeHg EFF ) × dt

MeHg GRMPB-FF

EFF /FF × MeHgFF

MeHgMPB Grazing by FF

MeHgMPB (t) = MeHgMPB (t − dt) + (MeHgI DUMPB + MeHgDOC DUMPB − MeHg GRMPB-ZP1 − MeHg GRMPB-ZP2 − MeHg GRMPB-FF − MeHg MMPB − MeHg EMPB ) × dt

MeHg in MPB (MeHgMPB )

Appendix C. Parameters used in the model simulations

Parameter

Value

Unit

Reference

Phytoplankton (PP) Maximum growth rate (Gmax )

0.01

h−1

Calibration

Excretion rate (ER)

4.2 × 10−4

h−1

Calibration

Mortality rate (MR)

2.1 × 10−4

h−1

Ashley (1998)

Respiration rate (RR)

2.1 × 10−3

h−1

Ashley (1998)

Sinking rate (SR)

1.3 × 10−3

h−1

Bienfang (1981)

Half saturation constant for nutrients

0.24

␮mol L−1

Schnoor (1996)

Carbon assimilation efficiency for ZP (AE1)

0.7

Unitless

Halvorsen et al. (2001)

Carbon assimilation efficiency for FF (AE3)

0.8

Unitless

Grizzle et al. (2001)

Fraction of PP ingestion by ZP1 (fzp1 )

0.9

Unitless

Assumed

Fraction of PP ingestion by ZP2 (fzp2 )

0.95

Unitless

Assumed

Fraction of PP ingestion by FF (fFF )

0.6

Unitless

Assumed

0.19

m3 h−1 C g−1

Calibration

Maximum filtration rate for ZP2

0.23

m3

Excretion rate

1.7 × 10−3

h−1

Kiorbe et al. (1985)

Mortality rate

8.7 × 10−4

h−1

Calibration

Respiration rate

1.7 × 10−3

h−1

Kiorbe et al. (1985)

Fraction of ZP2 ingestion by ZP1

0.05

Unitless

Assumed

Zooplankton (ZP) Maximum filtration rate (Fmax ) for ZP1

h−1

C g−1

Calibration

286

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix C (Continued ) Parameter ZP2 AE for ZP1 (AE2)

Value

Unit

Reference

0.3

Unitless

Assumed

Maximum filtration rate for FF

0.015

m3 h−1 C g−1

Calibration

Excretion rate

2.5 × 10−4

h−1

Calibration

Respiration rate

2.5 × 10−4

h−1

Calibration

Biodeposition rate

1.3 × 10−4

h−1

Grizzle (2001)

Maximum growth rate

0.003

h−1

Calibration

Excretion rate

4.2 × 10−4

h−1

Calibration

Mortality rate

2.1 × 10−4

h−1

Ashley (1998)

Respiration rate

2.1 × 10−3

h−1

Ashley (1998)

Fraction of MPB ingestion by ZP1

0.05

Unitless

Assumed

Fraction of MPB ingestion by ZP2

0.05

Unitless

Assumed

Fraction of MPB ingestion by FF

0.1

Unitless

Assumed

9.2 × 10−3

h−1

Wainright and Hopkinson (1997)

0.011

h−1

Calibration

Filter feeder (FF)

Microphytobenthos (MPB)

Particulate organic carbon in the water column (WPOC) Degradation rate (DRWPOC ) Sinking rate (SRWPOC )

Dissolved organic carbon in the water column (WDOC) Bacterial uptake rate (BURWDOC )

0.011

h−1

Calibration

Diffusion coefficient

7.2 × 10−7

m2 h−1

Gill et al. (1999)

0.019

h−1

Calibration

RPOC AE for FF

0.2

Unitless

Assumed

Fraction of RPOC ingestion by FF

0.3

Unitless

Assumed

RPOC degradation rate

9.2 × 10−3

h−1

Wainright and Hopkinson (1997)

1.66

m h−1

Data from Experiment 2

SPOC degradation rate

1.0 × 10−4

h−1

Wainright and Hopkinson (1997)

Burial rate

8.33 × 10−7

h−1

Uptake rate by PP

2.0 × 10−12

m3 h−1 cell−1

Calibration

Uptake rate by ZP

1.2 × 10−4

m3 h−1 C g−1

Calibration

Uptake rate by FF

1.0 × 10−4

m3

h−1

C g−1

Calibration

Uptake rate by MPB

1.0 × 10−12

m3

h−1

cell−1

Mason et al. (1996)

Adsorption rate 1

1.7 × 10−3

m3 h−1 C g−1

Adsorption rate 2

0.00150

m3

Desoprtion rate

0.00180

h−1

Organic carbon based Kd Fraction of MeHgDOC

DOC in the pore water (PW DOC) Bacterial uptake Resuspended organic carbon (RPOC)

Deposition rate Sediment organic carbon (SPOC)

Dissolved MeHg bound to DOC (MeHgDOC)

h−1

C g−1

Calculated using Kd and desorption rate Calculated using Kd and desorption rate Hintelmann and Harris (2004)

0.83

m3

C g−1

Data from Experiment 2

0.99

Unitless

Data from Experiment 2

Uptake rate by PP

1.0 × 10−11

m3 h−1 cell−1

Calibration

Uptake rate by ZP

1.2 × 10−4

m3

Dissolved MeHgI h−1

C g−1

Calibration

287

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Appendix C (Continued ) Parameter

Value

Unit

Uptake rate by FF

1.0 × 10−4

m3 h−1 C g−1

Calibration

Uptake rate by MPB

5.0 × 10−12

m3 h−1 cell−1

Calibration

Diffusion coefficient

4.7 × 10−6

m2

Methylation rate 1

2.0 × 10−3

h−1

Data from Experiment 2

Methylation rate 2

1.4 × 10−3

h−1

Data from Experiment 2

Demethylation rate

0.7

h−1

Data from Experiment 2

Total Hg concentration 1

9.8 × 10−3

g m−3

Data from Experiment 2

Total Hg concentration 2

9.1 × 10−3

g m−3

Data from Experiment 2

Organic carbon based Kd

0.010

m3

Adsorption rate

1.5 × 10−3

m3 h−1 C g−1

Calculated

0.14

h−1

Calculated

h−1

Reference

Gill et al. (1999)

Sediment MeHg (MeHgSPOC )

Desorption rate

references

Angelini, R., Petrere, M.J., 2000. A model for the plankton system of the Broa reservoir, Sao Carlos, Brazil. Ecological Modeling 126 (2–3), 131–137. Ashley, J.T.F., 1998. Habitat use and trophic status as determinants of hydrophobic organic contaminant bioaccumulation within shallow systems. PhD Thesis, University of Maryland, College Park. Baillie, P.W., Welsh, B.L., 1980. The effect of tidal resuspension on the distribution of intertidal epipelic algae in an estuary. Estuarine, Coastal and Shelf Science 10, 165–180. Bale, A.E., 2000. Modeling aquatic mercury fate in clear lake, Calif. Journal of Environmental Engineering-Asce 126 (2), 153–163. Benoit, J.M., Gilmour, C.C., Mason, R.P., Riedel, G.S., Riedel, G.F., 1998. Behavior of mercury in the Patuxent River estuary. Biogeochemistry 40 (2–3), 249–265. Benoit, J., Gilmour, C.C., Heyes, A., Mason, R., Miller, C., 2003. Geochemical and biological controls over methylmercury production and degradation in aquatic systems. In: Chai, Y., Braids, O.C. (Eds.), Biochemistry of Environmental Important Trace Elements. American Chemical Society, Washington, DC, pp. 262–297. Benoit, J.M., Shull, D.H., Robinson, P., Ucran, L.R., 2006. Infaunal burrow densities and sediment monomethyl mercury distributions in Boston Harbor, Massachusetts. Marine Chemistry 102 (1–2), 124–133. Bergeron, C.M., 2005. The impact of sediment resuspension on mercury cycling and the bioaccumulation of methylmercury into benthic and pelagic organisms. MS Thesis, University of Maryland, College Park. Bienfang, P.K., 1981. Sinking rates of heterogeneous temperate phytoplankton population. Journal of Plankton Research 3, 235–250. Bloom, N.S., Gill, G.A., Cappellino, S., Dobbs, C., McShea, L., Driscoll, C., Mason, R., Rudd, J., 1999. Speciation and cycling of mercury in Lavaca Bay, Texas, sediments. Environmental Science & Technology 33 (1), 7–13. Bougis, P., 1976. Marine Plankton Ecology. North-Holland, Amsterdam, p. 355. Braga, M.C.B., Shaw, G., Lester, I.N., 2000. Mercury modeling to predict contamination and bioaccumulation in aquatic

C g−1

Bloom et al. (1999)

ecosystems. Reviews of Environmental Contamination and Toxicology 164, 69–92. Caraco, N.F., Cole, J.J., Raymond, P.A., Strayer, D.L., Pace, M.L., Findlay, S.E.G., Fischer, D.T., 1997. Zebra mussel invasion in a large, turbid river: phytoplankton response to increased grazing. Ecology 78, 588–602. Chang, C.-W.A., 2001. Modeling the fate and trophic transfer of chemical contaminants in an urban estuary. MS Thesis, University of Maryland, College Park. Chang, M.-L., 1999. Modeling the effects of resuspension and deposition on early diagenesis of nutrients and contaminants. PhD Thesis, University of Maryland, College Park. Chang, S.I., Reinfelder, J.R., 2002. Relative importance of dissolved versus trophic bioaccumulation of copper in marine copepods. Marine Ecology-Progress Series 231, 179–186. Chen, C.Y., Folt, C.L., 2005. High plankton densities reduce mercury biomagnification. Environmental Science & Technology 39, 115–121. Clarkson, T.W., 1990. Human health risks from methylmercury in fish. Environmental Toxicology and Chemistry 9, 957–961. Conaway, C.H., Squire, S., Mason, R.P., Flegal, A.R., 2003. Mercury speciation in the San Francisco Bay Estuary. Marine Chemistry 80, 199–225. Crawford, S.M., Sanford, L.P., 2001. Boundary shear velocities and fluxes in the MEERC experimental ecosystems. Marine Ecology-Progress Series 210, 1–12. Dagg, M.J., Wyman, K.D., 1983. Natural ingestion rates of the copepods Neocalanus plumchrus and N. cristatus calculated from gut contents. Marine Ecology-Progress Series 13, 37–46. Darrow, B.P., Walsh, J.J., Vargo, G.A., Masserini, R.T., Fanning, K.A., Zhang, J.Z., 2003. A simulation study of the growth of benthic microalgae following the decline of a surface phytoplankton bloom. Continental Shelf Research 23 (14–15), 1265–1283. de Jonge, V.N., van Beusekom, J.E.E., 1992. Contribution of resuspended microphytobenthos to total phytoplankton in the ems estuary and its possible role for grazers. Netherlands Journal of Sea Research 30, 91–105. Descy, J.-P., Everbecq, E., Gosselain, V., Viroux, L., Smitz, J.S., 2003. Modelling the impact of benthic filter-feeders on the composition and biomass of river plankton. Freshwater Biology 48, 404–417. ˜ Fisher, N.S., Stupakoff, I., Sanudo-Wilhelmy, S., Wang, W.X., Teyssié, J.L., Fowler, S.W., Crusius, J., 2000. Trace metals in

288

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

marine copepods: a field test of a bioaccumulation model coupled to laboratory uptake kinetics data. Marine Ecology-Progress Series 194, 211–218. Fitzgerald, W.F., Clarkson, T.W., 1991. Mercury and monomethylmercury – present and future concerns. Environmental Health Perspectives 96, 159–166. Froneman, P.W., 2000. Feeding studies on selected zooplankton in a temperate estuary, South Africa. Estuarine and Coastal Marine Science 51, 543–552. Gilmour, C.C., Henry, E.A., 1991. Mercury methylation in aquatic systems affected by acid deposition. Environmental Pollution 71, 131–169. Gill, G.A., Bloom, N.S., Cappellino, S., Driscoll, C.T., Dobbs, C., McShea, L., Mason, R., Rudd, J.W.M., 1999. Sediment-water fluxes of mercury in Lavaca Bay, Texas. Environmental Science & Technology 33 (5), 663–669. Gorsky, G., Dallot, S., Sardou, J., Fenaux, R., Carrie, C., Palazzoli, I., 1988. C and N composition of some Northwestern Mediterranean zooplankton and micronekton species. Journal of Experimental Marine Biology and Ecology 124 (2), 133–144. Griffin, S.L., Herzfeld, M., Hamilton, D.P., 2001. Modelling the impact of zooplankton grazing on phytoplankton biomass during a dinoflagellate bloom in the Swan River Estuary, Western Australia. Ecological Engineering 16, 373–394. Grizzle, R.E., Bricelj, V.M., Shumway, S.E., 2001. Physiological ecology of Mercenaria mercenaria. In: Kraeuter, J.N., Castagna, M. (Eds.), Biology of the Hard Clam. Elsevier Health Sciences. Halvorsen, E., Pedersen, O.P., Slagstad, D., Tande, K.S., Fileman, E.S., Batten, S.D., 2001. Microzooplankton and mesozooplankton in an upwelling filament off Galicia: modelling and sensitivity analysis of the linkage and their impact on the carbon dynamics. Progress in Oceanography 51, 499–513. Hammerschmidt, C.R., Fitzgerald, W.F., 2004. Geochemical controls on the production and distribution of methylmercury in near-shore marine sediments. Environmental Science & Technology 38 (5), 1487–1495. Hammerschmidt, C.R., Fitzgerald, W.F., Lamborg, C.H., Balcom, P.H., Visscher, P.T., 2004. Biogeochemistry of methylmercury in sediments of long island sound. Marine Chemistry 90 (1–4), 31–52. Harding, L.W.J., Mallonee, M.E., Perry, E.S., 2002. Toward a predictive understanding of primary productivity in a temperate, partially stratified estuary. Estuarine and Coastal Marine Science 55, 437–463. Heyes, A., Mason, R.P., Kim, E., Sunderland, E., 2006. Mercury methylation in estuaries: insights from using measuring rates using stable mercury isotopes. Marine Chemistry 102 (1), 134–147. Hintelmann, H., Harris, R., 2004. Application of multiple stable mercury isotopes to determine the adsorption and desorption dynamics of Hg (II) and MeHg to sediments. Marine Chemistry 90 (1–4), 165–173. Jerling, H.L., Wooldridge, T.H., 1995. Feeding of two mysid species on plankton in a temperate estuary. Journal of Experimental Marine Biology and Ecology 188, 243–259. Jorgensen, S.E., 1994. Fundamentals of Ecological Modelling. Elsevier, Amsterdam, The Netherlands. Kibirige, I., Perissinotto, R., 2003. The zooplankton community of the Mpenjati Estuary, a South Africa temporarily open/closed system. Estuarine, Coastal and Shelf Science 58 (4), 727–741. Kim, E., 2004. The importance of physical mixing and sediment chemistry in mercury and methylmercury biogeochemical cycling and bioaccumulation within shallow estuaries. PhD Thesis, University of Maryland, College Park. Kim, E., Mason, R.P., Porter, E.T., Soulen, H.L., 2004. The effect of resuspension on the fate of total mercury and methyl

mercury in a shallow estuarine ecosystem: a mesocosm study. Marine Chemistry 86, 121–137. Kim, E., Mason, R.P., Porter, E.T., Soulen, H.L., 2006. The impact of resuspension on sediment mercury dynamics, and methylmercury production and fate: a mesocosm study. Marine Chemistry 102, 300–315. Kiorbe, T., Mohlenberg, F., Riisgard, H.U., 1985. In situ feeding rates of planktonic copepods: a comparison of four methods. Journal of Experimental Marine Biology and Ecology 88, 67–81. Lawrence, A.L., Mason, R.P., 2001. Factors controlling the bioaccumulation of mercury and methylmercury by the estuarine amphipod Leptocheirus plumulosus. Environmental Pollution 111 (2), 217–231. Luoma, S.N., et al., 1992. Determination of selenium bioavailability to a benthic bivalve from particulate and solute pathways. Environmental Science & Technology 26 (3), 485–491. Mason, R.P., Reinfelder, J.R., Morel, F.M.M., 1996. Uptake, toxicity, and trophic transfer of mercury in a coastal diatom. Environmental Science & Technology 30, 1835–1845. Mason, R.P., Lawrence, A.L., 1999. Concentration, distribution, and bioavailability of mercury and methylmercury in sediments of Baltimore Harbor and Chesapeake Bay, Maryland, USA. Environmental Toxicology and Chemistry 18 (11), 2438–2447. Mason, R.P., Lawson, N.M., Lawrence, A.L., Leaner, J.J., Lee, J.G., Sheu, G.R., 1999. Mercury in the Chesapeake Bay. Marine Chemistry 65 (1–2), 77–96. Morrison, H.A., Gobas, F.A., Lazar, R., Whittle, D.M., Haffner, G.D., 1997. Development and verification of a benthic/pelagic food web bioaccumulation model for PCB congers in Western Lake Erie. Environmental Science & Technology 31 (11), 3267–3273. Morton, B., 1969. Studies on the biology of Dreissena polymorpha Pall II. Correlation of the rhythms of adductor activity, feeding, digestion and excretion. Proceedings of Malacological Society of London 38, 401–414. Newell, R.I.E., 2004. Ecosystem influences of natural and cultivated populations of suspension-feeding bivalve molluscs: a review. Journal of Shellfish Research 23 (1), 51–61. Padilla, D.K., Adolph, S.C., Cottingham, K.L., Schneider, D.W., 1996. Predicting the consequences of dreissenid musels on a pelagic food web. Ecological Modeling 85, 129–144. Pickhardt, R.C., Folt, C.L., Chen, C.Y., Klaue, B., Blum, J.D., 2002. Algal blooms reduce the uptake of toxic methylmercury in freshwater food webs. Proceedings of the National Academy of Science of the United States of America 99 (7), 4419–4423. Porter, E.T., 1999. Physical and biological scaling of benthic-pelagic coupling in experimental ecosystem studies. PhD dissertation Thesis, University of Maryland, College Park. Porter, E.T., Sanford, L.P., Gust, G., Porter, F.S., 2004a. Combined water-column mixing and benthic boundary-layer flow in mesocosms: Key for realistic benthic-pelagic coupling studies. Marine Ecology-Progress Series 271, 43–60. Porter, E.T., Cornwell, J.C., Sanford, L.P., 2004b. Effect of oysters (Crassostrea virginica) and bottom shear velocity on benthic-pelagic coupling and estuarine water quality. Marine Ecology-Progress Series 271, 61–75. Porter, E.T., Owens, M.S., Cornwell, J.C., 2006. Effect of sediment manipulation on the biogeochemistry of experimental sediment systems. Journal of Coastal Research 22, 1539–1551. Porter, E.T., Porter, F.S., Mason, R.P. and Matthew, R., 2007. Design of a bivalve gape monitor and measurement of feeding by Mercenaria mercenaria (L.) under induced tidal resuspension. Marine Biology, in review. Ravichandran, M., 2004. Interactions between mercury and dissolved organic matter – a review. Chemosphere 55 (3), 319–331.

e c o l o g i c a l m o d e l l i n g 2 1 8 ( 2 0 0 8 ) 267–289

Riisgård, H.U., 1998. Filter feeding and plankton dynamics in a Danish Fjord: a review of the importance of flow, mixing and density-driven circulation. Journal of Environmental Management 53, 195–207. Riisgård, H.U., Kittner, C., Seerup, D.F., 2003. Regulation of opening state and filtration rate in filter-feeding bivalves (Cardium edule, Mytilus eduliis, Mya arenaria) in response to low algal concentration. Journal of Experimental Marine Biology and Ecology 284, 105–127. Roditi, H.A., Fisher, N.S., Sanudo-Wilhelmy, S.A., 2000. Field testing a metal bioaccumulation model for zebra mussels. Environmental Science & Technology 34 (13), 2817–2825. Sager, D.R., 2002. Long-term variation in mercury concentrations in estuarine organisms with changes in release into Lavaca Bay, Texas. Marine Pollution Bulletin 44 (8), 807–815. Sanford, L.P., Halka, J.P., 1993. Assessing the paradigm of mutually exclusive erosion and deposition of mud, with examples from upper Chesapeake Bay. Marine Geology 114 (1–2), 37–57. Sanford, L.P., Chang, M.-L., 1997. The bottom boundary condition for suspended sediment deposition. Journal of Coastal Research 25, 3–17, Special issue. Schnoor, J.L., 1996. Environmental Modeling, first ed. Wiley-Interscience. Simas, T.C., Ribeiro, A.P., Ferreira, J.G., 2001. Shrimp – a dynamic model of heavy-metal uptake in aquatic macrofauna. Environmental Toxicology and Chemistry 20 (11), 2649–2656. Siokou-Frangou, I., Bianchi, M., Christaki, U., Christou, E., Giannakourou, A., Gotsis, O., Ignatiades, L., Pagou, K., Pitta, P., Psarra, S., Souvermezoglou, E., Van Wambeke, F., Zervakis, V., 2002. Carbon flow in the planktonic food web along a gradient of oligotrophy in aegean sea (mediterranean sea). Journal of Marine Systems 33, 335–353. Sloth, N.P., Riemann, B., Nielsen, L.P., Blackburn, T.H., 1996. Resilience of pelagic and benthic microbial communities to sediment resuspension in a coastal ecosystem, Knebel, Vig, Denmark. Estuarine, Coastal, and Shelf Science 42, 405–415. Stella Manual. http://www.iseesystems.com/index.aspx. Sunderland, E.M., Gobas, F.A.P.C., Heyes, A., Branfireun, B.A., Bayer, A.K., Cranston, R.E., Parsons, M.B., 2004. Speciation and bioavailability of mercury in well-mixed estuarine sediments. Marine Chemistry 90 (1–4), 91–105.

289

Thomann, R.V., Mahony, J.D., Mueller, R., 1995. Steady-state model of biota sediment accumulation factor for metals in two marine bivalves. Environmental Toxicology and Chemistry 14 (11), 1989–1998. Tsui, M.T.K., Wang, W.X., 2004. Uptake and elimination routes of inorganic mercury and methylmercury in Daphnia magna. Environmental Science & Technology 38, 808–816. US EPA, 1997. Mercury Study Report to Congress, US EPA Office of Air, Washington, DC. Verity, P.G., 2000. Grazing experiments and model simulations of the role of zooplankton in Phaeocystis food webs. Journal of Sea Research 43, 317–343. Viroux, L., 2000. Dynamique du metazooplancton en milieu fluvial. PhD Thesis, University of Namur, Namur, Belgium. Wainright, S.C., Hopkinson Jr., C.S., 1997. Effects of sediment resuspension on organic matter processing in coastal environments: a simulation model. Journal of Marine Systems 11, 353–368. Walz, N., 1978. The energy balance of the freshwater mussel Dreissena polymorpha pallas in laboratory experiments and in lake constance. I. Pattern of activity, feeding and assimilation efficiency. Archives of Hydrobiology (Supplement) 55, 83–105. Wang, W.X., Stupakoff, I., Gagnon, C., Fisher, N.S., 1998. Bioavailability of inorganic and methylmercury to a marine deposit feeding polychaete. Environmental Science & Technology 32 (17), 2564–2571. Whalin, L., Kim, E.-H., Mason, R., 2007. Factors influencing the oxidation, reduction, methylation and demethylation of mercury species in coastal waters. Marine Chemistry 107 (3), 278–294. White, J.R., Roman, M.R., 1992. Seasonal study of grazing by metazoan zooplankton in the mesohaline Chesapeake Bay. Marine Ecology-Progress Series 86, 251–261. Winkels, H.J., Blom, G., Korrnenberg, S.B., Lijklema, L., 1998. Dilution of riverine heavy metal input concentrations by suspension of sediments and algal growth in the Ijsselmeer. Water Research 32 (10), 2931–2940. Young, B.L., Padilla, D.K., Schneider, D.W., Hewett, S.W., 1996. The importance of size-frequency relationships for predicting ecological impact of zebra mussel populations. Hydrobiologia 332, 151–158.