Bioaccumulation of Contaminants by Transplanted Bivalves in the San Francisco Estuary: Status and Trends |
||
| Prepared for |
||
| Acknowledgments |
||
Contents |
SummarySan Francisco Bay Area dischargers, through the Bay Area Dischargers Association (BADA) and the South Bay Dischargers, have analyzed bioaccumulation in transplanted bivalves for their Local Effects Monitoring Program (LEMP) since 1989. Five separate surveys have been conducted. The objective of this report is to:
Beginning with the BADA surveys (Rounds 1-4), samples were located along gradients from three major estuary sewage outfalls belonging to the City and County of San Francisco (SFSE), the East Bay Municipal Utilities District, and the Contra Costa Central Sanitary District (CCCSD). Sites "near", "mid", "far" (from the outfall) and "out" (reference sites presumably outside outfall influences) were sampled. Bivalve deployments were made below the water surface and/or above the bottom in the case of SFSE and EBMUD, and just below the surface for CCCSD, and for 30 and/or 90 days. One to three replicate samples for analysis of trace metals and organics were collected at each station. Oysters, Crassostrea gigas, were used at CCCSD, and mussels, Mytilus californianus, were used at EBMUD and SFSE. These surveys collected samples during successive dry and wet seasons in the Estuary. Differences between the 30- and 90-day deployments and between the shallow and deep deployments during Round 1 and 2 were described in previous reports. Those reports made several conclusions that are important in considering the results of Rounds 3 and 4 presented in this document. In general, the 90-day deployments had higher concentrations than the 30-day deployments except for PAHs which were more often higher in the 30-day deployments. There were no clear patterns of greater accumulation at the surface or bottom (deep) deployments. However, pesticides in Round 1 often had higher concentrations in the deep deployments. Bioaccumulation of contaminants by transplanted bagged bivalves during LEMP Rounds 3 and 4 showed three general patterns:
In general, there was little evidence of definite temporal trends over the four LEMP Rounds. Copper and zinc increased over the four Rounds at EBMUD. PCB concentrations at EBMUD and copper at CCCSD suggested seasonal fluctuations at the "near" station, being higher during the dry season deployments than during the wet season. However, PCBs exhibited a decreasing trend over the four Rounds at the CCCSD locations. For mussels at EBMUD and SFSE, there were no temporal trends for mercury, nickel, lead, and pesticides. Arsenic was always lowest in Round 2, selenium and silver were always highest in Round 3, and DDTs were always lowest in Round 4. For oysters at CCCSD, there were no obvious trends over the four Rounds for cadmium, nickel, selenium, zinc, DDTs, PAHs, and pesticides. Only for PAHs (SFSE) , PCBs (EBMUD), and copper (SFSE) were "near-outfall" stations elevated relative to more distant stations. Most of the transect station comparisons suggest that, with the above exceptions, gradients attributable to these outfalls can not be detected. When comparing sites nearest the outfalls with the RMP "reference envelope", only silver, selenium, and PAHs exhibited higher concentrations nearest the outfalls, while other constituents could not be differentiated from Estuary "background". |
|
IntroductionSan Francisco Bay Area dischargers, through the Bay Area Dischargers Association (BADA) and the South Bay Dischargers, have analyzed bioaccumulation of contaminants in transplanted bivalves for their Local Effects Monitoring Program (LEMP) since 1989. Five separate surveys have been conducted (Table 1). Except for the first survey, the surveys have been termed "Rounds". The results of the first three surveys (South Bay, Round 1, Round 2) have been reported previously (references in Table 1). Information from Rounds 3 and 4 is included in this report. This report also contains an evaluation of trends in bioaccumulation through all of the surveys collected between 1989 and 1993 and a comparison of LEM data with other data from other studies employing transplanted bivalves in the Bay. The objectives of the LEMP were to evaluate the bioaccumulation of contaminants in transplanted bivalves in the vicinity of sewage outfalls compared to bioaccumulation in control or reference areas. The objectives of this report are to:
|
||
MethodsSampling MethodsIt is important to understand the sampling design (station location, number of replicates, etc.) used in each survey in order to interpret the results correctly, particularly if statistical analyses are to be used. The sampling designs used in each LEMP survey were somewhat different, particularly in the South Bay survey. The details of the sampling designs used in Rounds 1 and 2 are included in the respective reports listed on Table 1. Briefly, in the South Bay survey, three replicate samples were collected at each of five stations in the South Bay using the mussel Mytilus californianus. Only trace metals were analyzed, and the mussels were deployed for 200 days. No "near-outfall" or outfall gradient stations were sampled. The various rounds of BADA surveys were designed in an adaptive fashion, with each Round being adjusted based on evaluation of the results from the previous one. Beginning with the BADA surveys (Rounds 1-4), samples were located along transects from three major Estuary sewage outfalls belonging to the City and County of San Francisco (SFSE), the East Bay Municipal Utilities District, and the Contra Costa Central Sanitary District (CCCSD). Sites "near", "mid", "far" (from the outfall) and a reference site (labeled "out" in the figures), were sampled. Bivalve deployments were made below the water surface and/or above the bottom in the case of SFSE and EBMUD during Rounds 1 and 2, and just below the surface for CCCSD. Deployments were for 30 and/or 90 days. During Rounds 3 and 4, mussels along the SFSE and EBMUD transects were deployed 15 feet below the surface. One to three replicate samples for analysis of trace metals and organics were collected at each station. Oysters (Crassostrea gigas) were used at CCCSD, and mussels (Mytilus californianus) were used at EBMUD and at SFSE. These surveys collected samples during successive dry and wet seasons in the Estuary. It is important to clarify the use of the terms "location" and "site" in this report because of the different implications these terms have in explaining the statistical evaluation of results presented below. "Location" is defined as an area expected to show similar bioaccumulation patterns. A location is described by multiple "sites". At each site an array of transplanted bivalves was deployed containing one or more bags of bivalves, or "replicates." The results of Round 1 and 2 indicated that pollutants accumulated to a greater degree in samples deployed for 90 days than in those for 30 days. Using these data to further refine the sampling protocol, the extended deployment was used for all further transplants. Round 1 and 2 also indicated that no appreciably different accumulation patterns between deep versus shallow deployments could be observed, and therefore only mid-depth samples, at 15 feet below the surface in the case of SFSE and EBMUD, were deployed in Rounds 3 and 4. These refinements reduced the cost of the study significantly, without compromising its objectives. Due to the shallow outfall depth, the CCCSD sampling sites consistently had only one depth treatment just below the surface throughout all Rounds. A number of unforeseen events introduced considerable challenges to the sampling design that were beyond the control of the monitoring and surveying personnel. The Suisun Bay control location for Round 1 was attached to a mid-channel buoy which was later removed, presumably by personnel conducting buoy maintenance activities. In Round 3, the control array at the location shared by SFSE and EBMUD was removed from the Treasure Island site by the military who were conducting maintenance activities. For the same Round, the buoy used to deploy the array at the CCCSD near-field location was removed before the end of the deployment period. No mid-gradient samples for SFSE were collected because the piling used to deploy transplanted mussels was removed by the Port of San Francisco. The EBMUD mid-gradient array was also lost. The difficulty in deploying arrays in San Francisco and Suisun Bays where maritime and pleasure boating activities can interfere with the monitoring arrays became apparent. From Round 2 onward, this necessitated deployments to fixed structures and conducting routine checks on the arrays during each bivalve deployment period. All of these unexpected challenges contributed to a less than ideal sample retrieval. An important concept in environmental sampling design is that of replication within areas of interest. For example, if an objective of the LEMP is to estimate bioaccumulation "near-outfalls" (a location), then several sites at a "near-outfall" location must be sampled in order to estimate the concentrations at the "near-outfall" location and to statistically compare those concentrations with concentrations at other locations (e.g., reference locations). Measurements of concentrations (replicates) at single sites within the "near-outfall" location only provide estimates of variation at that site, not for the "near-outfall" location. Using within-site variation to draw conclusions on variation among locations is considered pseudoreplication (Hurlbert, 1984). In addition, a direct comparison of tissue concentrations at a presumed impact location with concentrations at only one or a few reference location(s) using a conventional ANOVA model may not be sufficient to differentiate human-induced impacts from background spatial variability. Contaminant concentrations in the Estuary are spatially variable. Using a single RMP location as a reference for comparison with an outfall location would require the tenuous assumption that the RMP location is representative of a broader portion of the Estuary. A better statistical model would incorporate information on broader-scale spatial variability rather than assuming that it is insignificant. Smith (1995) has developed an approach for comparing impacted and reference sites or locations that avoids pseudoreplication and accounts for broad-scale spatial variation among reference locations. SFEI (1996) applied this "reference envelope" approach to compare sites "near-outfalls" with multiple, rather than single, reference locations. Reference locations are used to define a "reference envelope." Non-reference locations are then considered impacted if they fall outside the "reference envelope". The "reference envelope" approach is used in this report primarily to meet the objective of comparing reference site bioaccumulation patterns with those nearest the CCCSD, SFSE, and EBMUD outfalls. The 1989 South Bay data, although not directly comparable with the Round 1-4 LEM data, are compared in very general terms with bioaccumulation patterns of Rounds 1-4, the State Mussel Watch and RMP results. The findings presented in SFEI (1996) are reiterated here to provide a context for interpretation of the LEMP outfall data. The use of two different bivalve species, Crassostrea gigas in Suisun Bay and Mytilus californianus at the other LEM sites, precludes comparisons between the CCCSD transect and those in the rest of the Bay. Analysis of trends over time, even within locations, is confounded by some of the sampling problems described above. The only consistent sampling among Rounds 1-4 is for 90-day deployments at the surface for CCCSD. However, given that the depth variable did not appear to influence bioaccumulation results in any significant way, the SFSE and EBMUD bottom/surface samples during Rounds 1 and 2 may reasonably be compared to the Round 3 and 4 samples deployed 15 feet below the surface. This is reflected in the consistent labeling of the graphs as "deep". Each survey was conducted during different months of the year which may have introduced another factor increasing variability. The South Bay LEM in 1989 used 200-day deployment periods, thus limiting comparability with the 90-day deployment periods for Central and Suisun Bays. All of these factors contribute to potentially high variability in the results, thus lowering the "signal to noise ratio" and limiting the level of certainty we may have in explaining some apparent patterns. Analytical MethodsAnalytical methods remained the same for all four sampling periods and are outlined in OíConnor et al. (1992). All trace metal concentrations are expressed in mg/g (ppm) dry weight; trace organics are expressed in ng/g (ppb) dry weight. PAHs are expressed as the sum of 24 individual compounds. Pesticides are expressed as the sum of 15 individual compounds, including DDTs. Twenty different PCB congeners were quantified and expressed as total PCBs during Rounds 1 and 2, while the PCB list for Rounds 3 and 4 only contained 19 congeners. Individual congeners measured in Rounds 1 and 2 and Rounds 3 and 4 overlapped to a large degree, but not completely. Therefore, comparisons of total PCBs between Rounds 1 and 2 and Rounds 3 and 4 should be exercised with some caution. Data AnalysisConsidering the above discussion, no statistical comparisons of differences in bioaccumulation among stations or over time are made in this report, with the exception of comparing "near-outfall" sites with reference sites using the "reference envelope" approach. Further, it is recommended that conclusions based on ANOVAs presented in the Round 1 and 2 reports be interpreted cautiously. Instead, more qualitative comparisons are made where possible, as well as comparisons with other bivalve bioaccumulation data. In addition to using the absolute tissue concentrations at the end of each deployment period and comparing them to initial tissue concentrations prior to transplanting the bivalves to the Estuary, this report uses "accumulation factors" (AFs) to illustrate depuration (loss of constituents from bivalve tissue) or accumulation. The AF is calculated by dividing the contaminant concentration in transplants by the initial bivalve concentration. For example, an AF of 1.0 indicates that the concentration of a specific contaminant remained the same during the deployment period compared to the initial contaminant level prior to transplanting the bivalve sample to the Estuary. An AF of 0.5 indicates that the bivalve sample lost 50% of the contaminant concentration during the deployment period, while an AF above 1 indicates accumulation. ResultsIn considering the results of LEMP Rounds 1-4, it is helpful to visualize what a "typical" pattern might look like, given the sampling regime used (Figure 1). Bivalves collected from their source (Bodega, Tomales, or Humbolt Bays) should have the lowest concentrations (T0). Travel blanks (TB) should not deviate significantly from those concentrations. If contamination from outfalls is a source of bioaccumulated material, then the "near" samples should have the highest concentrations, then concentrations should decrease successively at the "mid" and "far" stations. Samples "out" of the outfall's influence should have somewhat higher concentrations than the "time zero" or travel blanks, since the Estuary is known to have generally elevated contamination. If the bioaccumulated material is from other sources, then the gradient from "near" to "out" will not be as obvious as shown. Only a few of the LEMP samples exhibit this typical pattern. One good example is PAHs from SFSE during Round 4 (Figure 41). Differences between the 30- and 90-day deployments and between the shallow and deep deployments during Round 1 and 2 were described in previous reports. Those reports made several conclusions that are important in considering the results of Rounds 3 and 4 presented in this report. In general, the 90-day deployments had higher concentrations than the 30-day deployments except for PAHs which were more often higher in the 30-day deployments. There was no clear pattern of greater accumulation at the surface or bottom (deep) deployments. However, pesticides in Round 1 often had higher concentrations in the deep deployments. Since results of Rounds 3 and 4 have not been previously reported, concentrations and trends in those samples are described in some detail. For trace organics, both the dry weight and lipid-normalized dry weight results are presented in the text, while the figures show dry weight concentrations only. Data tables for all results, including the lipid-normalized data, are presented in Appendix 2. Bioaccumulation in Oysters Along the CCCSD Outfall TransectArsenic concentrations in the Rounds 3 and 4, 90-day, deep samples ranged between 7.3 and 11.6 ppm, similar to those measured in Round 2, but much lower than in Round 1 (Figure 2). Arsenic did not appear to accumulate in the oysters. There were no obvious differences in concentrations in the samples from along the outfall gradient (near, mid, far, out). The 30-day Round 2 samples generally had lower concentrations than the Round 1 samples. Cadmium concentrations measured during Rounds 3 and 4 ranged between 15 and 24 ppm and were similar to those measured in Rounds 1 and 2 (Figure 3). Cadmium appeared to accumulate in the oysters during the 90-day, but not during 30-day deployments. There were no apparent patterns in concentrations along the outfall gradient during either of the deployment periods. Chromium concentrations ranged between 1.81 to 10.17 ppm and were similar to those measured during Round 2, but not Round 1 (Figure 4). Chromium concentrations in oysters showed the same patterns as arsenic, with no consistent trends along outfall gradients and no consistent accumulation patterns. Samples deployed for 30 days showed very similar patterns to those deployed for 90 days, with Round 1 concentration ranges higher than those for Round 2. Copper concentrations in Rounds 3 and 4 ranged between 480 and 1,280 ppm (Figure 5). Copper appeared to accumulate in oyster tissues during Rounds 1 and 3, suggesting higher accumulation during the dry season than the wet season. There were no consistent trends in copper concentrations along the outfall gradient. As with chromium and arsenic, Round 1 showed the highest concentrations for both 30-day and 90-day deployment periods. Lead concentrations ranged between 0.3 and 1.9 ppm in Rounds 3 and 4 (Figure 6). Concentrations appeared to decrease over time at all stations. Lead accumulated in oyster tissues during Rounds 1 and 3 but not Round 4. There were no consistent trends along the outfall gradient sites. There were no Round 2 samples, thus comparisons between Rounds 1 and 2 cannot be made. Mercury concentrations in oysters ranged between 0.06 and 0.49 ppm in Rounds 3 and 4. For the 90-day exposure samples, all but Round 3 showed similar concentrations. Round 3 was much lower (Figure 7). For the 30-day deployment samples, Round 1 concentrations were higher than those for Round 2. Mercury accumulated in only a few samples. There were no consistent trends along the outfall gradient. For all Rounds for which control station data were available, the control stations had higher concentrations than the "near" outfall stations. Nickel concentrations measured in Rounds 3 and 4 ranged from 2.42 to 8.99 ppm, similar to the 90-day deployment concentrations measured in Rounds 1 and 2 (Figure 8). Nickel appeared to accumulate in most, but not all samples. There were no consistent patterns in concentrations along the outfall gradient; the "near" station had the highest concentrations during Round 4. An apparent artifact in "time zero" Tomales Bay oysters (mean concentration of almost 5 ppm) is revealed by travel blanks with a mean concentration of less than 1 ppm. All transect station concentrations were lower than the "time zero" blank sample. For the 30-day samples, there did not appear to be any major difference in concentrations between Rounds 1 and 2. Selenium concentrations in oyster tissues ranged between 0.26 and 20 ppm during Rounds 3 and 4 (Figure 9). Concentrations were highest during Round 3 (mid and far locations). No outfall-related trends were apparent in any of the samples. Selenium appeared to accumulate only during Rounds 1 and 3. The 30-day samples exhibited slightly higher concentrations during Round 1 than during Round 2. Round 3 and 4 silver concentrations ranged between 17.03 and 52.1 ppm (Figure 10). Concentrations were considerably higher during Round 1 than during any of the other times. There were no consistent outfall-related trends in silver concentrations, but accumulation was apparent during Round 1. Bioaccumulation occurred only during Round 1 and 2 and not during Round 4. There were no Round 2 samples, thus comparisons between Rounds 1 and 2 cannot be made. Zinc concentrations ranged between 1,200 and 2,600 ppm during Rounds 3 and 4 (Figure 11). Concentrations appeared to be highest in Round 1, but similar for all other Rounds. Zinc appeared to accumulate in oysters in all Rounds except Round 4. There were no obvious outfall-related trends in accumulations. Similar to other trace metals, the Round 1, 30-day exposure samples were higher than for Round 2. DDTs accumulated in oysters in all samples. Concentrations ranged between 68.7 and 125.7 ppb in Round 3 and 4 samples (Figure 12). Concentrations were similar for all Rounds. No outfall-related trends could be distinguished for any of the sampling periods. Lipid-normalized DDT concentrations indicated reduced concentrations in Rounds 3 and 4a trend that was not distinguishable in concentrations expressed in dry weights. The samples exposed for 30 days did not appear very different for Rounds 1 and 2 on either a dry weight or lipid-normalized weight basis. PAHs accumulated in oysters at all stations in all LEMP Rounds (Figure 13). As observed with DDTs, accumulation did not exhibit any outfall-related trends. Concentrations ranged between 913.5 and 2,175 ppb during Rounds 3 and 4. Round 1 showed slightly higher concentrations than any of the other Rounds, particularly at the station nearest the outfall. Again, lipid-normalized concentrations indicated a declining trend in PAH concentrations for the 90-day exposure samples, which was not apparent in the PAH concentrations measured on a dry weight basis. The samples exposed for 30 days showed the same trends as the 90-day samples expressed in dry weight, with Round 1 being slightly higher than Round 2. This difference disappeared when expressed in lipid-normalized weight. PCBs accumulated at all stations during all LEMP Rounds (Figure 14). Concentrations ranged between 43.7 and 108.1 ppb during Rounds 3 and 4. Again, no outfall-related trends could be distinguished, while a slight downward trend over time is apparent for all transect stations. This trend is more obvious when concentrations are expressed in lipid-normalized weight. The 30-day exposure, Round 1 samples appeared to have higher concentrations than the Round 2 samples when expressed in dry weights. However, this difference largely disappeared when expressed in lipid-normalized weights. Pesticides also accumulated at all stations during all LEM Rounds (Figure 15). Concentrations ranged between 89.9 and 176.2 ppb during Rounds 3 and 4. No apparent differences in pesticide concentrations among the four Rounds were obvious, and no outfall-related trends were apparent. The lipid-normalized data not only show a slight downward trend in concentrations over time, but also illuminated apparent differences between seasons and reproductive cycles, with dry season concentrations being higher than the wet season concentrations. The 30-day samples showed similar trends as other trace organic compounds. Bioaccumulation in Mussels Along the EBMUD Outfall TransectArsenic concentrations in the Rounds 3 and 4, 90-day, deep samples ranged between 9.8 and 12.6 ppm, similar to those measured in Rounds 1 and 2 (Figure 16). Arsenic did not appear to accumulate in the mussels. There were no apparent differences in concentrations among the 4 Rounds, although concentrations from Round 2 appeared to be lower. There were no obvious differences in concentrations in the samples from along the outfall gradient ("near", "mid", "far", "out"). The travel blank in Round 3 was unexplainably higher than in the other Rounds. For the 30-day samples, and the 90-day surface samples, Round 2 samples had generally lower concentrations than Round 1 samples. Cadmium concentrations measured during Rounds 3 and 4 ranged between 5.8 to 9.9 ppm, and were similar to those measured in Rounds 1 and 2 (Figure 17). However, concentrations measured in Round 4 appeared to be lower than the other Rounds, possibly because of the lower T0 values. Cadmium did not appear to accumulate in mussels. There were no apparent trends in concentrations along the outfall gradient. For the 30-day, and 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. Chromium concentrations ranged 2.5 to 11.3 ppm and were similar to those measured during Rounds 1 and 2 (Figure 18). There were no consistent trends of accumulation of chromium, and no consistent trends along the outfall gradient. Although the mean concentrations measured at the near and out locations in Round 4 were higher than other measurements, their 95% confidence intervals overlapped suggesting no significant differences. For the 30-day samples, and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. Copper concentrations in Rounds 3 and 4 ranged between 11.0 and 16.6 ppm (Figure 19). However, concentrations at the "near" station tended to be higher in Rounds 3 and 4 than in Rounds 1 and 2. Copper accumulated in mussel tissues during all LEM Rounds. There were no consistent trends in copper concentrations along the outfall gradient. For the 30-day samples, and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. Lead concentrations ranged between 2.0 and 4.2 ppm in Rounds 3 and 4 (Figure 20). Concentrations appeared to increase slightly over time at the EBMUD near station, but the 95% confidence interval suggests that it was not a significant trend. Lead accumulated in mussel tissues during Rounds 3 and 4 , but not during Round 1. There were no consistent trends along the outfall gradient sites. There were no Round 2 samples, thus comparisons between Rounds 1 and 2 cannot be made. Mercury concentrations in mussels ranged between 0.19 and 0.31 ppm in Rounds 3 and 4, and did not appear vary among the LEMP Rounds (Figure 21). Mercury accumulated in mussels in all samples. There did not appear to be any consistent trends over the outfall gradient. During Round 2, the "out" station had higher concentrations than the "near" outfall station. For the 30-day samples, and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. Nickel concentrations measured in Rounds 3 and 4 ranged between 3.2 and 6.6 ppm, similar to those measured in Rounds 1 and 2 (Figure 22). Nickel appeared to accumulate in mussels in all but Round 1. There were no consistent patterns in concentrations along the outfall gradient; the "out" station had the highest concentrations during Rounds 2 and 3. For the 30-day samples, and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. Selenium concentrations in mussel tissues ranged between 3.7 and 12.0 during Rounds 3 and 4 (Figure 23). Concentrations were highest during Round 3 ("near" and "far" locations). Only during this Round were decreasing concentrations along the outfall transect apparent. However, the travel blank was also higher than in other Rounds. Selenium appeared to accumulate during all LEMP Rounds, particularly Round 4. For the 30-day samples, and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. Silver concentrations ranged between 0.21 and 2.7 ppm (Figure 24). Concentrations were considerably higher during Round 3 than Rounds 4 or 1. Silver accumulated only in the Round 3 samples. There were no apparent outfall-related trends in silver concentrations. There were no Round 2 samples, thus comparisons between Rounds 1 and 2 cannot be made. Zinc concentrations ranged between 190 and 290 ppm during Rounds 3 and 4 (Figure 25). Concentrations appeared to increase slightly beginning with Round 3; Round 4 was obviously higher than Round 1. Zinc accumulated in mussels in all LEMP Rounds, but there were no indications of any outfall-related effects. For the 30-day samples, and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. DDTs accumulated in mussels in all samples. Concentrations ranged between 43.8 and 81.1 ppb in Round 3 and 4 samples (Figure 26). The Round 4 concentrations were lower than the others. Concentrations were always highest at the "near" outfall station suggesting outfall-related accumulation. Lipid-normalized DDT concentrations indicated much the same trend as those expressed on a dry-weight basis. For the 30-day samples and the 90-day surface samples there did not appear to be any difference in concentrations between Rounds 1 and 2. PAHs accumulated in mussels at all stations in all LEMP Rounds (Figure 27). Accumulation was an order of magnitude higher at the "far" location during Round 3 and 4 (not measured in Rounds 1 and 2), than in the other samples. Concentrations ranged between 193 and 7,142 ppb during Rounds 3 and 4. Concentrations at most ("far" excepted) stations were similar among the 4 Rounds. For the 30-day samples, and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2, except for two Round 2 samples (30-day surface, "far"; 90 day surface, "mid") which accumulated orders of magnitude more PAHs than the other samples. PCBs accumulated at all stations during all LEMP Rounds (Figure 28). Concentrations ranged between 74 and 214 ppb during Rounds 3 and 4. Concentrations at the "near" station showed some seasonal variation with levels being significantly higher during the dry seasons than the wet ones. Additionally, concentrations were always highest at the "near" outfall station than at the other gradient stations, suggesting a "near-outfall" source. For the 30-day samples and the 90-day surface samples, the Round 1 samples appeared to have higher concentrations than the Round 2 samples. However, as mentioned above, those results may be related to seasonal differences in accumulation. Examination of the lipid-normalized plots supports the idea that differences in lipid content in the mussels associated with their reproductive cycles may account for the apparent differences shown on the dry weight plots. Pesticides also accumulated at all stations during all LEM Rounds (Figure 29). Concentrations ranged between 76 and 113 ppb during Rounds 3 and 4. At the "near" outfall station there was no apparent difference in pesticide concentrations among the 4 Rounds, but the "near" station was always higher than the other gradient stations suggesting an outfall-related effect. For the 30-day samples and the 90-day surface samples, there did not appear to be any difference in concentrations between Rounds 1 and 2. Bioaccumulation in Mussels Along the SFSE Outfall TransectArsenic concentrations ranged between 9.9 and 12.3 ppm during Rounds 3 and 4 (Figure 30). Round 2 concentrations were lower than those from the other cruises; the dry season samples were higher than those from the wet season. Arsenic did not appear to accumulate in mussels from any of the LEMP Rounds. There were no consistent trends in accumulation at the stations along the outfall gradient ("near", "mid", "far", "out"). At the 30-day surface and deep stations and the 90-day surface stations, there were no obvious patterns of accumulation during Rounds 1 or 2. Concentrations during Round 2 were lower than Round 1. Cadmium concentrations ranged between 5.9 and 13.0 ppb in the mussels from Rounds 3 and 4 (Figure 31). There did not appear to be any difference in cadmium concentrations among the four LEMP Rounds. There were no consistent trends of accumulation of cadmium at the SFSE stations. There were no indications of outfall-related trends in cadmium accumulation. At the 30-day surface and deep stations and the 90-day surface stations, there were no obvious patterns of accumulation during Rounds 1 and 2. Concentrations were not obviously different in Rounds 1 and 2. Chromium concentrations in mussels during Rounds 3 and 4 ranged between 2.9 and 11.34 ppm (Figure 32). Round 1 chromium concentrations were higher than the other Rounds, but the T0 samples were also much higher. Chromium did not appear to accumulate in the mussels. There was no decreasing bioaccumulation indication of a gradient away from the outfall. At the 30-day surface and deep stations and the 90-day surface stations, there were no obvious patterns of accumulation during Rounds 1 and 2. Concentrations were not obviously different in Rounds 1 and 2. Copper concentrations ranged between 11 and 19 ppb in mussels during Rounds 3 and 4 (Figure 33). There did not appear to be any obvious trends in accumulation over the 4 LEMP Rounds. Copper accumulated in the mussels at all stations in all Rounds. Only during Round 4 did accumulation data suggest a possible outfall effect. None of the other Rounds showed any similar pattern. At the 30-day surface and deep stations and the 90-day surface stations, copper accumulated at almost all stations. Concentrations at the Round 1 30-day surface stations were lower than in Round 2, but otherwise there was no obvious difference in Rounds 1 and 2 copper concentrations. Lead concentrations ranged between 0.5 and 4.0 ppm during Rounds 3 and 4 (Figure 34). Concentrations were similar among the four LEMP Rounds, and were generally higher at the "near" outfall stations than at the other stations. Lead accumulated in most of the LEMP samples. Analysis on Round 2 was not conducted, thus no comparisons between Rounds 1 and 2 for the 30-day and 90-day shallow samples are possible. Mercury concentrations in mussels ranged between 0.039 and 0.86 ppb in Rounds 3 and 4 (Figure 35). Concentrations were similar among all LEMP Rounds. However, the "far" station in Round 4 had appreciably higher accumulation than at the other stations or Rounds. Mercury accumulated at most stations in most Rounds, except for Round 1. There were no trends in accumulation along the outfall gradient. At the 30-day surface and deep stations and the 90-day surface stations, there were no clear patterns of accumulation. There were no obvious differences in Rounds 1 and 2 mercury concentrations. Nickel concentrations ranged between 3.2 and 7.1 ppm at the SFSE stations during Rounds 3 and 4 (Figure 36). There were no apparent differences between the four LEMP Rounds. Nickel accumulated in mussels during all Rounds except Round 1. There were no consistent relationships to the SFSE outfall gradient. At the 30-day surface and deep stations and the 90-day surface stations, there were no clear patterns of accumulation in the 30-day samples. However, accumulation was present in the 90-day surface samples. There were no obvious differences in Rounds 1 and 2 nickel concentrations. Selenium concentrations ranged between 2.4 and 10 ppb in Rounds 3 and 4 (Figure 37). Concentrations during Round 3 were higher than in the other Rounds including the travel blank (TB). However, examination of the 95% confidence interval suggests that there were no real differences among the Rounds. Selenium did not accumulate consistently in all LEMP Rounds. Only Round 4 showed consistent accumulation of selenium. There were no indications of outfall-related accumulation. At the 30-day surface and deep stations and the 90-day surface stations, selenium appeared to accumulate in the Round 2 samples, but not during Round 1. There were no obvious differences in Rounds 1 and 2 selenium concentrations. Silver concentrations ranged between 0.21 and 2.79 ppm during Rounds 3 and 4 (Figure 38). Round 3 concentrations were considerably higher than during Rounds 1 and 4. Silver appeared to accumulate in the few samples collected, but there were no indications of accumulation gradients along the transect. Analysis on Round 2 was not conducted, thus no comparisons between Rounds 1 and 2 for the 30-day and 90-day shallow samples are possible. Zinc concentrations in mussels ranged between 190 and 370 ppm during Rounds 3 and 4 (Figure 39). Round 4 concentrations were slightly higher than the other Rounds, but not significantly. Zinc accumulated in mussels at all SFSE stations during all Rounds. There were no clear trends of outfall-related accumulation of zinc. At the 30-day surface and deep stations and the 90-day surface stations, there was some evidence of accumulation at some of the stations, particularly the 90-day surface stations. There were no obvious differences in Rounds 1 and 2 concentrations. DDT concentrations ranged between 43.8 and 78.0 ppb in mussels during Rounds 3 and 4 (Figure 40). Round 4 concentrations were lower than the other Rounds. At the "near" station concentrations were highest during Round 2. DDT accumulated in all samples during all Rounds, but there was no clear elevation at the "near-outfall" station. DDTs appeared to accumulated at the 30-day surface and deep stations and the 90-day surface stations. There were no obvious differences in the Round 1 and 2 concentrations. PAH concentrations ranged between 335 and 1,756 ppb during Rounds 3 and 4 (Figure 41). Concentrations were highest during Round 3. PAHs accumulated in all mussels during all Rounds. Accumulation of PAHs appeared to be related to a "near-outfall" source, particularly during Round 4 where the pattern of bioaccumulation was typical of an outfall gradient. PAHs appeared to accumulate at the 30-day surface and deep stations and the 90-day surface stations, and exhibited higher concentrations at the "near-outfall" station. Round 2 concentrations at the "near" outfall station were higher than the Round 1 concentrations only during the 30-day exposures. PCB concentrations ranged between 77 and 169 ppb in mussel tissues during Rounds 3 and 4 (Figure 42). Concentrations during Round 4 were considerably lower than the previous three Rounds. PCBs accumulated in all samples during all Rounds, but there was no indication that the accumulation was outfall-related. PCBs accumulated at the 30-day surface and deep stations and the 90-day surface stations. Because there were no variance estimates at the outfall gradient stations, it cannot be determined whether there were any real differences between the Round 1 and 2 concentrations. Pesticide concentrations ranged between 81 and 102 ppb during Rounds 3 and 4 (Figure 43). Round 2 concentrations were the highest, but there did not appear to be any consistent trends among the four LEMP Rounds. Pesticides accumulated in all samples during all Rounds, but did not appear to be outfall-related. At the 30-day surface and deep stations and the 90-day surface stations, pesticides bioaccumulated at all stations. There were no obvious differences in Rounds 1 and 2 concentrations except for the 90-day surface deployments where Round 2 had higher concentrations than Round 1.
Discussion and ConclusionsContaminant Accumulation by TransplantsAccumulation factors (AFs) for trace elements in bivalves were generally low (Figure 44). Oysters accumulated significant quantities of nickel and chromium (AFs >10), but little or none of the other elements. Selenium concentrations decreased in transplanted oysters. Transplanted mussels approximately doubled their concentrations of most trace elements. Mussels accumulated little or no arsenic or cadmium. In contrast to the trace elements, trace organics were strongly accumulated in both mussels and oysters (Figure 45). In many cases concentrations were readily detected in transplants but not detected in controls, making calculation of an AF impossible, but nevertheless indicative of strong accumulation. In oysters, AFs for all of the major organics measured (a -chlordane, dieldrin, p,p'-DDD, p,p'-DDE, PCB 153, S PCB, S PAH) were all greater than 7. In mussels, AFs for the organochlorine pesticides and PAHs (but not PCBs) were lower than in oysters, but were still appreciable, ranging between 2 and 4. The only exception was S PAH at EBMUD, which did not accumulate much above the initial concentration. The observed variation in AFs is a function of differences in background concentrations in the Estuary relative to those at the Tomales Bay and Humboldt Bay control locations, inter-specific differences in accumulation in oysters and mussels, and local variation in contaminant concentrations. For example, oysters consistently accumulate higher concentrations of PAHs than mussels, as seen in side-by-side deployments in the RMP (SFEI, 1996). This is at least a partial explanation for the differences in AFs at CCCSD and EBMUD. Local contamination is also important, however, and is the major reason that AFs for mussels at SFSE are almost as high as for the CCCSD oysters. Spatial Patterns of AccumulationAs discussed in the Results section, if outfalls were sources of biologically available contaminants, then concentrations in bivalves would be expected to decrease with distance from the outfalls. In general, the LEMP data do not fit this pattern. The only instance where this pattern was consistently observed was for PAHs at the SFSE outfall. "Near-outfall" concentrations in all sampling rounds were consistently higher at SFSE than concentrations at more distant stations. Other contaminants were frequently elevated at the "near-outfall" locations but in an inconsistent or indistinct manner; these included nickel at CCCSD, pesticides and PCBs at EBMUD, and copper at SFSE. No other contaminants were distinctly elevated at the "near-outfall" locations. One non-outfall location, the "far" station for EBMUD, consistently had very high concentrations of PAHs (up to 7,000 ng/g dry-weight in Round 3). Trends Over TimeThis discussion focuses on trends at the 90-day samples for SFSE, EBMUD, and CCCSD deployments at the "near" outfall stations, since the data set for "near" stations is the most complete for all Rounds. Evaluation of any trend was based on observations on the report figures, and not on any statistical analysis, as explained above. In general, there was little evidence of definite trends over the four LEMP Rounds or discernible seasonal variation of contaminant accumulation, with a few exceptions. Copper and zinc increased over the four Rounds at EBMUD. PCB concentrations at EBMUD and copper at CCCSD suggested seasonal fluctuations at the "near" station, being higher during the dry season deployments than those deployed during the wet season. However, PCBs exhibited a decreasing trend over the four Rounds at the CCCSD locations. For mussels at EBMUD and SFSE, there were no temporal trends for mercury, nickel, lead, and pesticides. Arsenic was always lowest in Round 2, selenium and silver were always highest in Round 3, and DDTs were always lowest in Round 4. For oysters at CCCSD, there were no obvious trends over the four Rounds for cadmium, nickel, selenium, zinc, DDTs, PAHs, and pesticides. Comparisons with Other DataThe ranges of concentrations measured during LEMP Rounds 3 and 4 are tabulated along with the results from the 1993 RMP, the Western States Petroleum Association (WSPA) LEM from 1993, and the South Bay LEM from 1989-1990 (Table 2). Figures 46-58 display the distributions of measured contaminant concentrations from the major monitoring programs employing transplanted bivalves in San Francisco Bay from 1980 to the present, including LEMP data from "near-outfall" locations. A large database exists for contaminant concentrations in transplanted mussels (Mytilus californianus) in the Bay due to the long-term monitoring of the State Mussel Watch program and the more recent inception of the RMP. The RMP provides the only frame of reference in the Bay for bioaccumulation in transplanted oysters. Table 2 presents a summary evaluation of the relationships among these datasets for each contaminant. Two contaminants, silver in mussels and oysters and selenium in mussels, reached distinctly elevated concentrations at the "near-outfall" locations, accumulating to concentrations that were high relative to both the RMP and SMW data. Silver in mussels and oysters and selenium in mussels in 1994 were also shown to exceed the upper boundary of the "reference envelope" based on 1994 RMP data (SFEI, 1996). S PAHs at SFSE were clearly and consistently elevated relative to RMP data and exceeded the upper boundary of the "reference envelope" based on 1994 RMP data (SFEI, 1996). Relative to the SMW data, however, SFSE S PAH concentrations were not exceptionally high. SFSE S PAHs concentrations were in the upper half of the SMW distribution, but the highest SFSE concentration approximated the 75th percentile of the SMW distribution. While a PAH signal has clearly been detected near the SFSE outfall, the concentrations measured are not exceptionally high relative to data for the Bay as a whole. For two contaminants, cadmium and copper, LEMP concentrations were high relative to RMP data, but RMP data were generally lower than SMW data. Even SMW stations that should have been comparable to RMP stations had higher concentrations. These results raise the possibility that methodological differences might be a factor contributing to the patterns observed among these datasets, particularly in cases where the RMP and SMW distributions are clearly offset from each other. The South Bay data were generally within the range of the LEMP data, with some notable exceptions: arsenic and selenium concentrations were much lower than concentrations in any of the other datasets, suggesting methodological artifacts. The South Bay data as a whole were slightly lower than the "near-outfall" data points at SSFE and EBMUD for silver, copper, and cadmium. Compared to the 1993/94 RMP data, copper and cadmium concentrations in the South Bay were relatively high. Efficacy of Bivalve MonitoringAs has been shown, some difficulties exist with regard to interpretation of transplanted bivalve data. Some of the difficulty is due to sampling design problems, and the use of several species. Another difficulty is a low degree of accumulation of some contaminants in transplants relative to initial concentrations. Just what the bioaccumulation of contaminants may mean in an ecological context is also difficult to determine. Very few studies of ecological impacts (growth, mortality, reproduction, etc.) due to bioaccumulation have yet been undertaken, and those that have were conducted on a limited number of trace substances only. However, the human health implications from eating contaminated shellfish are better known, and a number of trace contaminants are regulated by the United States Food and Drug Administration. Monitoring with transplanted bivalves does indicate that contaminants are capable of entering the food web and accumulating. Evaluation of bioaccumulation along a waste discharge gradient, as has been done by the LEMP, also provides an indication of whether or not the outfalls may be directly contributing to the contaminant load available for bioaccumulation. Only for PAHs (SFSE) , PCBs (EBMUD), and copper (SFSE) were outfalls shown to be potential direct contributors. Silver (all outfalls) and selenium (SFSE and EBMUD) did not show a gradient with distance from the outfall, but concentrations measured near the outfalls were high relative to both RMP and SMW data, suggesting that concentrations in the areas surrounding the outfalls are elevated. Most of the transect station comparisons suggest that, with the above exceptions, gradients attributable to individual waste discharges cannot be detected. When comparing sites nearest the outfalls with the RMP "reference envelope" of all stations, only silver, selenium, and PAHs exhibited higher concentrations in bivalves nearest the outfalls, while other constituents could not be differentiated from Estuary "background." It is clear, however, that the Estuary as a whole exhibits elevated concentrations for most trace elements and substantially elevated concentrations of trace organics compared to "clean" areas. Another question to consider is whether other species or methods of sampling should be used. C. gigas no longer lives in the Estuary (although it once did). Transplanting organisms into bags placed on the bottom of the Estuary which is not their natural habitat may cause unnatural bioaccumulation patterns. Perhaps resident/native bivalves, such as Ostrea lurida should be used, since they would have integrated their lifetime exposure and do not exhibit the large seasonal fluctuation in body burdens due to their different reproductive strategy as brooders. If transplanted bivalves near waste water outfalls are considered as indicators of the success of pollution prevention programs, the consistent use of one species at all locations should be considered in order to compare locations over time. The RMP is interested in these questions as well, and resolving them will no doubt require some pilot studies to determine the best design option. In 1996, the RMP incorporated into its annual work-plan a special study reviewing in detail the long-term bioaccumulation database. The Program Review in 1997 is expected to provide expert advice on the value of continued bivalve monitoring. |
||
|
|
||
References CitedDavis, J.A., and T.H. Daum. 1993. Bioaccumulation of contaminants by bivalve mollusks in the vicinity of municipal wastewater discharges to San Francisco Bay: Wet season results, 1991/1992. SFEI Report to BADA, 38 pp. Hurlbert, S.H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr., 54:187-211. Jenkins, Sanders, and Associates. 1993. A project pilot study for local monitoring of petroleum refinery effluent. Final report to Western States Petroleum Association, Glendale, CA. 113 pp. Larry Walker Associates and Kinetics Laboratories. 1991. Bioaccumulation Monitoring of trace elements in south San Francisco Bay using transplanted marine mussels (Mytilus californianus). Report for EOA, Inc. for the City of San Jose and Sunnyvale NPDES permit monitoring. 22 pp. O'Connor, J.M., J.A. Davis, and T.H. Daum. 1992. Bioaccumulation of contaminants by bivalve mollusks in the vicinity of Municipal wastewater discharges to San Francisco Bay. SFEI Report to BADA, 73 pp. San Francisco Estuary Institute. 1996. Annual Report of the 1994 Regional Monitoring Program. SFEI, Richmond, CA. 339 pp. San Francisco Estuary Institute. 1994. Annual Report of the 1993 Regional Monitoring Program. SFEI, Richmond, CA. 214 pp. San Francisco Estuary Institute. 1994b. 1994 RMP Implementation Plan. SFEI, Richmond, CA. Smith, Robert W., EcoAnalysis, Inc. 1995. Southern California Regional Monitoring Project: The Reference Envelope Approach to Impact Monitoring. Southern California Coastal Water Research Project. |
||
San Francisco Estuary Institute
7770 Pardee Lane
Oakland, CA 94621-1424
(510) 746-SFEI (7334)
Fax (510) 746-7300
www.sfei.org