based on estuarine mixing behavior and the migration patterns of striped bass. Underlying each water column segment was placed two to fourteen sediment segments for a total of 30 water column segments and 120 sediment segments (see Farley etal., 1999 for details).
Although total PCB was used as a state variable in many previous model investigations (for example, the Great Lakes: Thomann and DiToro, 1983; New Bedford Harbor: Connolly, 1991), this approach was not considered adequate for this study. This is because PCBs represent a family of 209 possible compounds, with each compound containing one to ten chlorines on the biphenyl structure and exhibiting different physical-chemical and biochemical behavior. Modeling the transport, fate, and bioaccumulation of a large number of individual PCB compounds over decadal time periods in a 150 segment model, however, was not considered readily tractable. As a compromise solution, model calculations were performed for the five PCB homologue groups representing di- through hexa-chlorobiphenyl (CB) that contain the largest mass of PCBs. Description of PCB loads during the 19872002 simulation period is presented below and is followed by discussions of PCB transport and fate, and PCB bioaccumulation models.
PCB concentrations in the Lower Hudson are strongly linked to external loads from the Upper Hudson and downstream sources. For model applications, PCB loads from the Upper Hudson were specified on a monthly basis over the fifteen-year simulation period as follows. For January 1987-April 1991, PCB homologue loads were determined by adjusting the exponentially decreasing average annual load function given in Thomann et al. (1989) to monthly PCB homologue loads. This was accomplished by assigning 68 percent of the annual load to the March-April high flow period and 32 percent of the annual load to the remaining ten months. For May 1991-December 1998, monthly PCB homologue loads were calculated from measured concentrations (O'Brien and Gere, 1997) and estimated flows at Thompson Island Dam (Kilometer Point (km) 304; River Mile (RM) 189). Monthly PCB loadings for all years after 1998 were assumed to follow 1998 monthly loads (without the January 1998 high flow event). The resulting distribution of monthly-averaged loads from the Upper Hudson is given in Figure 25.2. Elevated loads for the early 1990s, which were estimated to be as high as 10-15 kilograms per day, were largely attributed to scouring of PCB contaminated sediments from an old water intake structure and PCB oil seeps through the fracturedbedrock underlying the Hudson Falls facility (Rhea, Connolly, and Haggard, 1997). Controls subsequently implemented at the Hudson Falls facility have been effective in reducing PCB loads to the river, and by the late 1990s, estimated loads to the Lower Hudson decreased to approximately 0.5 kilograms per day. These later loads are believed to be dominated by leaching of PCBs from contaminated sediments in the Upper Hudson (for example, Thompson Island Pool) (Garveyand Hunt, 1997).
All other PCB homologue loads were considered constant throughout the model simulation period. These included the Mohawk River (0.16 kg d-1), the New Jersey tributaries (0.12 kg d-1), New York City and New Jersey wastewater treatment plants
Figure 25.3. Schematic diagram of processes affecting the transport and fate of PCB homologues in water column and surface sediment segments.
Figure 25.3. Schematic diagram of processes affecting the transport and fate of PCB homologues in water column and surface sediment segments.
Aerobic Degradation or Reductive Dechlorination
(0.26 kg d-1), combined sewer overflows to New York Harbor (0.14 kg d-1), and direct atmospheric loads to the Lower Hudson and New York Harbor (0.03 kg d-1). Details of the calculations are given in Farley et al. (1999). The total PCB load from all these sources is denoted by the horizontal line in Figure 25.2 and is shown to be a significant portion of the load, particularly for the late 1990s.
pcb transport and fate modeling
The chemical transport and fate of PCBs in the Lower Hudson and New York Harbor is affected by hydrodynamic and sediment transport of dissolved and particulate PCBs, water column-pore water exchange, volatilization, and chemical transformations. To evaluate the overall effect of these processes on PCBs, we constructed mass conserva-tionequations for each water column and sediment segment based on the schematic diagram shown in Figure 25.3. For water column segments, the mass conservation equationfor total (that is, freely-dissolved, DOC-bound plus particulate) concentration for each PCB homologue is given as: dC,
= £ QjiCi - £ QjC + J2 Eij(Cj - Ci) + WCi - wsASimiH
+ wUi ASi msedi rsed
+ k'fASi (Cdis+DOC,ed - Cdis+DOCi) - k ASi Cdis - ki Vi Cdisi where the first term represents the change of the total PCB homologue mass in water column segment 'i' with time; the second and third terms represent the mass rate of PCB homologue flowing into and out of segment 'i', respectively; the fourth term represents PCB homologue entering or leaving segment 'i' by tidal dispersion; the fifth term represents PCB homologue input into segment 'i' from external sources (e.g., tributary input or wastewater discharge); the sixth term represents PCB homologue lossfrom the water column by settling; the seventh term represents PCB homologue gain from resuspension; the eighth term represents diffusive exchange between freely-dissolved and DOC-bound PCB homologue concentrations in the water column and pore waters; the ninth term represents the transfer of PCB homologue across the air-water interface (that is, volatilization); and the last term represents transformation losses from the water column (for example, by aerobic degradation). A complete listing of terms used in the equation is given in the appendix of this chapter.
A similar equation can be written for total (that is, freely-dissolved, DOC-bound plus particulate) concentrations for each PCB homologue in the surface sediment layer:
VsedidCdjf~ = WsASimi r; - WmASimsed, TSed,
- WbiASimSedi Tsedi - k'fASi (Cdis+DOC,edi
Cdis+DOCi) ksedfVsediCdissedi + kf ASi (Cdis+DOCdeep,edi - Cdis+DOC„di)
where the first term represents the change in the total PCB homologue mass in sediment segment 'i' with time; the second term represents the gain of PCB homologue by settling from the overlying water, the third and fourth terms represent the loss of PCB homologue by resuspension and burial into deeper sediments, respectively; the fifth term represent diffusive exchange of freely-dissolved and DOC-bound PCB homologue with the overlying water; and the sixth term represents transformation losses from the sediments (for example, by anaerobic dechlorination); and the last term represents diffusive exchange of freely-dissolved and DOC-bound PCB homologue with the deeper sediment pore water. Chemical gain, for example, by dechlorination of higher chlorinated homologues, is also possible.
For model calculations, flows through the model domain were specified on a seasonal basis based on aggregated results from the Blumberg-Mellor three-dimensional, intratidal hydrodynamic model for the 1989 water year (see Miller and St. John, Chapter 11). Tidal dispersion coefficients were taken from Thomann et al. (1989). Seasonal estimates of suspended solids concentrations, resuspension rates, burial rates and downstream transport of suspended solids were determined from seasonal solids balances using estimates of sediment loads, measured deposition rates from dated sediment cores, aerial estimates of deposition zones, dredging records, and a specified settling velocity of 3.2 m d-1 (Farley et al., 1999). Particulate (POC) and dissolved organic carbon (DOC) concentrations were specified on a seasonal basis based upon aggregated results from the System-Wide-Eutrophication-Model (SWEM) for the 1989 water year (see Chapter 11, this volume).
In addition, hydrodynamics, sediment transport, and organic carbon distributions, model descriptions for PCB partitioning between the freely-dissolved, DOC-bound, and particulate phases are essential in describing the various flux terms in the mass conservation equations. In our modeling studies, PCB partitioning is assumed to be fast compared to other environmental processes and is modeled as instantaneous (or equilibrium) reactions. PCB concentrations in the various phases can then be expressed in terms of the total PCB concentrations using the equilibrium partitioning relationships to solids and DOC (Kd = r./Cdis) and KDOC = (CDOC/DOC)/Cdis) and the total mass conservation equation (C = ^Cdis + ^CDOC + .mr).
For our current model applications, partitioning between freely-dissolved and DOC-bound phases is described as a direct function of Kqw (Kdoc = aDoc Kqw) where the factor aDOC is used to account for differences in PCB partitioning to lower molecular weight DOC compounds and octanol. Partitioning between freely-dissolved and sediment phases is also expressed as a direct function of Kow and the organic carbon fraction (foc) of the sediment (Kd = foc Kqw) (Karickhoff, 1981; DiToro et al., 1991). This simple "foc Kqw" relationship, however, may not provide an adequate description of PCB partitioning to suspended solids for cases like the Lower Hudson where phytoplankton comprise a large portion of the suspended organic material. PCB partitioning between the freely-dissolved and suspended solid phases is therefore expressed by a more general relationship that accounts for enhanced partitioning for lower chlorinated homologues (possibly due to the stronger binding of PCBs to cell membranes) and reduced partitioning of higher chlorinated homologues (due to occurrence of cell growth during the longer periods of PCB uptake for higher chlorinated homologues) (Skoglund and Swackhamer, 1994). This relationship is given as:
aphyto foc Kow
1 + V ku I aphyto JocKow where aphyto is the sorption enhancement factor for PCBs to phytoplankton cells andkg/ku is the ratio of phytoplankton growth to the PCB uptake rate. (Setting aphyto = 1 and kg/ku = 0reduces this expression to our previous formulation, K<j = focKow.)
Homologue-specific Kqw values for the model were determined from weighted averages of observed congener distributions in 1992 highresolution surface sediments and 1993 perch data (TAMS/Gradient, 1995) and Kow values for individual PCB congeners (as reported in Hawker and Con-nell, 1988). Resulting log Kow values are given as 5.0 (di-CB), 5.6 (tri-CB), 6.0 (tetra-CB), 6.45 (penta-CB), and 6.85 (hexa-CB). The value of aDOC is taken as 0.1. This represents an order of magnitude decrease in the PCB partitioning to DOC and is consistent with results of Burkhard (2000). The remaining two sorption coefficients, aphyto and kg/ku,were not specified a priori but were considered as adjustable parameters in calibrating the fate model results to observed PCB homologue concentrations in surface sediments.
In addition to PCB partitioning behavior, specification of kinetic rate coefficients for volatilization, pore water exchange, and chemical transformations are also required. The volatilization rate coefficient was calculated using the two-layer model of the air-water interface (Schwarzenbach, Gschwend, andlmboden, 1993), and assuming that PCB transfer through the water side of the interface is controlling the overall volatilization rate. A volatilization rate coefficient of 0.5 m d-1 was estimated using the O'Connor-Dobbins formula (O'Connor and Dobbins, 1958) with an average tidal velocity of 0.5 m s-1, an average depth of 6 m, and an average PCB molecular diffusivity coefficient of 0.4 x 10-5 cm2 s-1. This value is consistent with results of Clark et al. (1996) for gas exchange rates as determined from a sulfur hex-afluoride and helium-3 tracer study in the tidal freshwater Hudson.
Dissolved chemical exchange between pore water and the overlying water column has been shown in recent studies of the Upper Hudson to occur at fairly high rates (kf' = 1-15cmd-1) (Connolly et al.,
2000). These high rates of pore water exchange have been attributed to mixing of sediment particles by bioturbation coupled with diffusion of dissolved contaminant through the water side of the benthic boundary layer (Thibodeaux, Valsaraj, and Reible,
2001). For our current model application, kf' was specified as 5 cm d-1. Even with this higher rate of exchange, settling and resuspension of particle-bound PCBs still appeared to dominate the PCB transfer rates across the water-sediment interface in the tidal freshwater and mid-estuary regions.
Finally, a number of studies have shown that certain PCB congeners may be transformed in aquatic environments by degradation under aerobic conditions or microbial dechlorination under anaerobic conditions (Abramowicz, 1990). Although these processes were found to be active in altering PCB distributions in the Upper Hudson, aerobic degradation and anaerobic dechlorination do not appear to be significant in the Lower Hudson River. PCB transformation rates for the Lower Hudson and New York Harbor were therefore considered to be negligible in our model calculations.
pcb bioaccumulation modeling
Accumulation of PCBs in Hudson River white perch and striped bass is calculated using the food chain model of Thomann et al. (1992a,b). In this approach, PCB accumulation within a given organism is viewed as a dynamic process that depends on direct uptake from the water, ingestion of contaminated prey, depuration (from urine excretion and egestion of fecalmatter), andmetabolic transformation of PCBs within the organism.
Model equations for the uptake and release of PCBs into a given organism are typically written in terms of |g PCB g-1 (wet weight) of organism (vk) (Thomann et al., 1992a,b). The general form of this equation is given as:
- [ke + km + kg]vk where the first term represents the change in PCB homologue concentration in organism 'k' with time; the second term represents the direct uptake of PCB homologues from the water phase by diffusion across an external cell or gill membrane; the third term represents back diffusion of PCB homologues across the membrane; the fourth term represents PCB homologue uptake through the ingestion of contaminated food or prey and is dependent on the chemical assimilation efficiency (.aM) and the food consumption rate (I^) for organism 'k' feeding on organism 'l'; and the last term represents decreases in PCB homologue concentration in organism 'k' due to excretion (ke), metabolic transformation (km), and growth dilution (kg). In this equation, growth dilution is included as a loss term to account for the reduction in PCB homologue concentration due to the increase in size of the organism. A complete listing of terms in the equation is given in the appendix to this chapter.
A time-variable, age-dependent striped bass food chain model was previously developed for the Hudson River Estuary by Thomann et al. (1989; 1991). The model includes a five component, water-column food chain that consists of phytoplankton, zooplankton, small fish, seven age classes of perch, and seventeen age classes of striped bass. In applying the model to the Lower Hudson, PCB homologue concentrations in water and phytoplankton are taken directly from the transport and fate model calculations. Phytoplankton are preyed upon by a zooplankton compartment, the characteristics of which is considered to be represented by Gammarus. The small fish compartment, which feeds on zooplankton, is meant to reflect a mixed diet of fish of about 10 g in weight and includes age 0-1 tomcod and herring.
White perch is considered as a representative size-dependent prey of the striped bass and is assumed to feed exclusively on zooplankton. Based on feeding studies where stomach contents of striped bass were examined (Gardinier and Hoff, 1982; O'Connor, 1984; Setzler et al., 1980), the 0-2-year-old striped bass are assumed to feed on zooplankton; 2-5-year-old striped bass are assumed to feed on a mixture of small fish and 0-2-year-old perch; and 6-17-year-old striped bass are assumed to feed on 2-5-year-old perch.
Growth rates were determined from results of Poje, Riordan, and O'Connor (1988) for zooplankton; from a generalized growth-weight relationship for small fish (Thomann et al., 1989); from the age-weight data of Bath and O'Connor (1982) for white perch; and from the age-weight data of Setzler et al. (1980) and Young (1988) for striped bass. Details of age-dependent weights and growth rates are given in Thomann et al. (1989) and are summarized in Farley et al. (1999).
Respiration rates for zooplankton, small fish, white perch, and striped bass were estimated using formulations given in Thomann and Connolly (1984) and Connolly and Tonelli (1985). Details of respiration rates, along with lipid content, dry weight fractions, and food assimilation efficiency are given in Farley et al. (1999). These values are used with the gill transfer efficiency (p.), chemical assimilation efficiency from food (.a) and PCB homologue-specific parameters for Kow, to calculate gill uptake rates (ku = pRoxygen/Coxygen), backdiffusion rates (kb = ku/(flipid Kow)), and food ingestion rates (I = (R+kg)/a). Log Kqw values were previously given as 5.0, 5.6, 6.0, 6.45, and 6.85 for di- through hexa-CB. The chemical assimilation efficiency (a.) was set equal to the food assimilation efficiency (a) of 0.3 for zooplankton and 0.8 for fish. Gill transfer efficiency (p.) was the only remaining parameter and was adjusted in calibrating model results to observed PCB homologue concentrations in white perch. This value was then used for all fish species throughout the Lower Hudson, New York Harbor, Long Island Sound, and New York Bight.
In bioaccumulation calculations, migration of striped bass adds a further complication in specifying time-dependent exposure concentrations. Migration patterns used in the calculations were assigned based on Waldman (1988); Waldman et al.,
(1990) and are described in Thomann et al. (1989; 1991). These are summarized as follows: Striped bass are born on May 15 of each year and the yearlings are assumed to remain in the mid estuary (as defined by Km 30 to 126; RM 18.5 to 78.5). The 2-5-year-old stripedbass are considered to migrate from the mid estuary into New York Harbor in June and spend the summer months (July through September) in Long Island Sound and the New York Bight. Lastly, 6-17-year-old striped bass are assumed to spend most of their year in the open ocean, but migrate into Long Island Sound and the NewYork Bight around March 15 and return to the mid estuary aroundApril 15 to spawn. Theyremain in the mid estuary until the middle of July.
Was this article helpful?