Ben K. Greenfield

Contaminant concentrations in sport fish from San Francisco Bay, 1997.

In 1997, seven sport fish species were sampled from seven popular fishing areas in San Francisco Bay. Mercury exceeded a human health screening value in 44 of 84 (52%) samples. All collected samples of leopard shark and striped bass exceeded the mercury screening value of 0.23 microg/g wet weight. PCBs exceeded the screening value in 51 of 72 (71%) samples. DDT, chlordane, and dieldrin, had lower numbers of samples above screening values: 16 of 72 (22%) for DDT, 11 of 72 (15%) for chlordanes, and 27 of 72 (37%) for dieldrin. Concentrations of PCBs and other trace organics were highest in white croaker and shiner surfperch, the two species with the highest fat content in their muscle tissue. Fish from one location, Oakland Harbor, had significantly elevated wet weight concentrations of mercury, PCBs, DDTs, and chlordanes compared to other locations. Removal of skin from white croaker fillets reduced lipid concentrations by 27-49% and concentrations of trace organics by 33-40%.

Nutrient supply and mercury dynamics in marine ecosystems: a conceptual model.

There is increasing interest and concern over the impacts of mercury (Hg) inputs to marine ecosystems. One of the challenges in assessing these effects is that the cycling and trophic transfer of Hg are strongly linked to other contaminants and disturbances. In addition to Hg, a major problem facing coastal waters is the impacts of elevated nutrient, particularly nitrogen (N), inputs. Increases in nutrient loading alter coastal ecosystems in ways that should change the transport, transformations and fate of Hg, including increases in fixation of organic carbon and deposition to sediments, decreases in the redox status of sediments and changes in fish habitat. In this paper we present a conceptual model which suggests that increases in loading of reactive N to marine ecosystems might alter Hg dynamics, decreasing bioavailabilty and trophic transfer. This conceptual model is most applicable to coastal waters, but may also be relevant to the pelagic ocean. We present information from case studies that both support and challenge this conceptual model, including marine observations across a nutrient gradient; results of a nutrient-trophic transfer Hg model for pelagic and coastal ecosystems; observations of Hg species, and nutrients from coastal sediments in the northeastern U.S.; and an analysis of fish Hg concentrations in estuaries under different nutrient loadings. These case studies suggest that changes in nutrient loading can impact Hg dynamics in coastal and open ocean ecosystems. Unfortunately none of the case studies is comprehensive; each only addresses a portion of the conceptual model and has limitations. Nevertheless, our conceptual model has important management implications. Many estuaries near developed areas are impaired due to elevated nutrient inputs. Widespread efforts are underway to control N loading and restore coastal ecosystem function. An unintended consequence of nutrient control measures could be to exacerbate problems associated with Hg contamination. Additional focused research and monitoring are needed to critically examine the link between nutrient supply and Hg contamination of marine waters.

Contaminant concentrations and histopathological effects in Sacramento splittail (Pogonichthys macrolepidotus).

Sacramento splittail (Pogonichthys macrolepidotus) is a species of special concern in California, due to multiple anthropogenic stressors. To better understand the potential impact of contaminant exposure, adult splittail were captured from the Sacramento-San Joaquin River Delta (California, USA) and analyzed for histopathology and contaminant exposure. Organochlorine contaminants (PCBs, DDTs, dieldrin, chlordanes, and PBDEs) and trace metals (Ag, As, Cd, Co, Cr, Cu, Fe, Mn, Mo, Ni, Pb, Se, Sn, V, and Zn) were detected in the tissues of all fish. In many samples, human health screening values were exceeded for PCBs (83 of 90 samples), DDTs (32 samples), and dieldrin (37 samples). In contrast, thresholds for fish effects were rarely exceeded. Histopathological analysis indicated the presence of macrophage aggregates in gonads, kidneys, and liver and a high incidence of liver abnormalities. In the liver, observed effects were often moderate to severe for glycogen depletion (55 of 95 fish), lipidosis (hepatocellular vacuolation; 51 fish), and cytoplasmic inclusion bodies (33 fish). Correlations between histopathology and tissue contaminant concentrations were weak and inconsistent. Significant correlations were observed between histopathology indicators and reductions in fish size, body condition, lipid content, and liver weight. These results suggest that splittail histopathology varies as a function of health and nutritional status, rather than exposure to legacy organic and metal pollutants.

Assessment of potential aquatic herbicide impacts to California aquatic ecosystems.

A series of legal decisions culminated in 2002 with the California State Water Resources Control Board funding the San Francisco Estuary Institute to develop and implement a 3-year monitoring program to determine the potential environmental impacts of aquatic herbicide applications. The monitoring program was intended to investigate the behavior of all aquatic pesticides in use in California, to determine potential impacts in a wide range of water-body types receiving applications, and to help regulators determine where to direct future resources. A tiered monitoring approach was developed to achieve a balance between program goals and what was practically achievable within the project time and budget constraints. Water, sediment, and biota were collected under “worst-case” scenarios in close association with herbicide applications. Applications of acrolein, copper sulfate, chelated copper, diquat dibromide, glyphosate, fluridone, triclopyr, and 2,4-D were monitored. A range of chemical analyses, toxicity tests, and bioassessments were conducted. At each site, risk quotients were calculated to determine potential impacts. For sediment-partitioning herbicides, sediment quality triad analysis was performed. Worst-case scenario monitoring and special studies showed limited short-term and no long-term toxicity directly attributable to aquatic herbicide applications. Risk quotient calculations called for additional risk characterizations; these included limited assessments for glyphosate and fluridone and more extensive risk assessments for diquat dibromide, chelated copper products, and copper sulfate. Use of surfactants in conjunction with aquatic herbicides was positively associated with greater ecosystem impacts. Results therefore warrant full risk characterization for all adjuvant compounds.

Polychlorinated biphenyl spatial patterns in San Francisco Bay forage fish

Industrialized waterways frequently contain nearshore hotspots of legacy polychlorinated biphenyl (PCB) contamination, with uncertain contribution to aquatic food web contamination. We evaluated the utility of estuarine forage fish as biosentinel indicators of local PCB contamination across multiple nearshore sites in San Francisco Bay. Topsmelt (Atherinops affinis) or Mississippi silverside (Menidia audens) contamination was compared between 12 targeted sites near historically polluted locations and 17 probabilistically chosen sites representative of ambient conditions. The average sum of 209 PCB congeners in fish from targeted stations (441±432 ng g(-1) wet weight, mean±SD) was significantly higher than probabilistic stations (138±94 ng g(-1)). Concentrations in both species were comparable to those of high lipid sport fish in the Bay, strongly correlated with spatial patterns in sediment contamination, and above selected literature thresholds for potential hazard to fish and wildlife. The highest concentrations were from targeted Central Bay locations, including Hunters Point Naval Shipyard (1347 ng g(-1); topsmelt) and Stege Marsh (1337 ng g(-1); silverside). Targeted sites exhibited increased abundance of lower chlorinated congeners, suggesting local source contributions, including Aroclor 1248. These findings indicate that current spatial patterns in PCB bioaccumulation correlate with historical sediment contamination due to industrial activity. They also demonstrate the utility of naturally occurring forage fish as biosentinels of localized PCB exposure.

Seasonal and annual trends in forage fish mercury concentrations, San Francisco Bay.

San Francisco Bay is contaminated by mercury (Hg) due to historic and ongoing sources, and has elevated Hg concentrations throughout the aquatic food web. We monitored Hg in forage fish to indicate seasonal and interannual variations and trends. Interannual variation and long-term trends were determined by monitoring Hg bioaccumulation during September-November, for topsmelt (Atherinops affinis) and Mississippi silverside (Menidia audens) at six sites, over six years (2005 to 2010). Seasonal variation was characterized for arrow goby (Clevelandia ios) at one site, topsmelt at six sites, and Mississippi silverside at nine sites. Arrow goby exhibited a consistent seasonal pattern from 2008 to 2010, with lowest concentrations observed in late spring, and highest concentrations in late summer or early fall. In contrast, topsmelt concentrations tended to peak in late winter or early spring and silverside seasonal fluctuations varied among sites. The seasonal patterns may relate to seasonal shifts in net MeHg production in the contrasting habitats of the species. Topsmelt exhibited an increase in Alviso Slough from 2005 to 2010, possibly related to recent hypoxia in that site. Otherwise, directional trends for Hg in forage fish were not observed. For topsmelt and silverside, the variability explained by year was relatively low compared to sampling station, suggesting that interannual variation is not a strong influence on Hg concentrations. Although fish Hg has shown long-term declines in some ecosystems around the world, San Francisco Bay forage fish did not decline over the six-year monitoring period examined.

Are Chinese consumers at risk due to exposure to metals in crayfish? A bioaccessibility-adjusted probabilistic risk assessment

Freshwater crayfish, the worlds third largest crustacean species, has been reported to accumulate high levels of metals, while the current knowledge of potential risk associated with crayfish consumption lags behind that of finfish. We provide the first estimate of human health risk associated with crayfish (Procambarus clarkii) consumption in China, the worlds largest producer and consumer of crayfish. We performed Monte Carlo Simulation on a standard risk model parameterized with local data on metal concentrations, bioaccessibility (φ), crayfish consumption rate, and consumer body mass. Bioaccessibility of metals in crayfish was found to be variable (68-95%) and metal-specific, suggesting a potential influence of metal bioaccessibility on effective metal intake. However, sensitivity analysis suggested risk of metals via crayfish consumption was predominantly explained by consumption rate (explaining >92% of total risk estimate variability), rather than metals concentration, bioaccessibility, or body mass. Mean metal concentrations (As, Cd, Cu, Ni, Pb, Se and Zn) in surveyed crayfish samples from 12 provinces in China conformed to national safety standards. However, risk calculation of φ-modified hazard quotient (HQ) and hazard index (HI) suggested that crayfish metals may pose a health risk for very high rate consumers, with a HI of over 24 for the highest rate consumers. Additionally, the φ-modified increased lifetime risk (ILTR) for carcinogenic effects due to the presence of As was above the acceptable level (10(-5)) for both the median (ILTR=2.5×10(-5)) and 90th percentile (ILTR=1.8×10(-4)), highlighting the relatively high risk of As in crayfish. Our results suggest a need to consider crayfish when assessing human dietary exposure to metals and associated health risks, especially for high crayfish-consuming populations, such as in China, USA and Sweden.

Soil geochemistry and digestive solubilization control mercury bioaccumulation in the earthworm Pheretima guillemi

Mercury presents a potential risk to soil organisms, yet our understanding of mercury bioaccumulation in soil dwelling organisms is limited. The influence of soil geochemistry and digestive processes on both methylmercury (MeHg) and total mercury (THg) bioavailability to earthworms (Pheretima guillemi) was evaluated in this study. Earthworms were exposed to six mercury-contaminated soils with geochemically contrasting properties for 36 days, and digestive fluid was concurrently collected to solubilize soil-associated mercury. Bioaccumulation factors were 7.5-31.0 and 0.2-0.6 for MeHg and THg, respectively, and MeHg accounted for 17-58% of THg in earthworm. THg and MeHg measured in soils and earthworms were negatively associated with soil total organic carbon (TOC). Earthworm THg and MeHg also increased with increasing soil pH. The proportion of MeHg and THg released into the digestive fluid (digestive solubilizable mercury, DSM) was 8.3-18.1% and 0.4-1.3%, respectively. The greater solubilization of MeHg by digestive fluid than CaCl2, together with a biokinetic model-based estimate of dietary MeHg uptake, indicated the importance of soil ingestion for MeHg bioaccumulation in earthworms.

A tiered assessment framework to evaluate human health risk of contaminated sediment

For sediment contaminated with bioaccumulative pollutants (e.g., PCBs and organochorine pesticides), human consumption of seafood that contain bioaccumulated sediment-derived contaminants is a well-established exposure pathway. Historically, regulation and management of this bioaccumulation pathway has focused on site-specific risk assessment. The state of California (United States) is supporting the development of a consistent and quantitative sediment assessment framework to aid in interpreting a narrative objective to protect human health. The conceptual basis of this framework focuses on 2 key questions: 1) do observed pollutant concentrations in seafood from a given site pose unacceptable health risks to human consumers? and 2) is sediment contamination at a site a significant contributor to seafood contamination? The first question is evaluated by interpreting seafood tissue concentrations at the site, based on health risk calculations. The second question is evaluated by interpreting site-specific sediment chemistry data using a food web bioaccumulation model. The assessment framework includes 3 tiers (screening assessment, site assessment, and refined site assessment), which enables the assessment to match variations in data availability, site complexity, and study objectives. The second and third tiers use a stochastic simulation approach, incorporating information on variability and uncertainty of key parameters, such as seafood contaminant concentration and consumption rate by humans. The framework incorporates site-specific values for sensitive parameters and statewide values for difficult to obtain or less sensitive parameters. The proposed approach advances risk assessment policy by incorporating local data into a consistent region-wide problem formulation, applying best available science in a streamlined fashion.

A multivariate analysis of CalEnviroScreen: comparing environmental and socioeconomic stressors versus chronic disease

BackgroundThe health-risk assessment paradigm is shifting from single stressor evaluation towards cumulative assessments of multiple stressors. Recent efforts to develop broad-scale public health hazard datasets provide an opportunity to develop and evaluate multiple exposure hazards in combination.MethodsWe performed a multivariate study of the spatial relationship between 12 indicators of environmental hazard, 5 indicators of socioeconomic hardship, and 3 health outcomes. Indicators were obtained from CalEnviroScreen (version 3.0), a publicly available environmental justice screening tool developed by the State of California Environmental Protection Agency. The indicators were compared to the total rate of hospitalization for 14 ICD-9 disease categories (a measure of disease burden) at the zip code tabulation area population level. We performed principal component analysis to visualize and reduce the CalEnviroScreen data and spatial autoregression to evaluate associations with disease burden.ResultsCalEnviroScreen was strongly associated with the first principal component (PC) from a principal component analysis (PCA) of all 20 variables (Spearman ρ = 0.95). In a PCA of the 12 environmental variables, two PC axes explained 43% of variance, with the first axis indicating industrial activity and air pollution, and the second associated with ground-level ozone, drinking water contamination and PM2.5. Mass of pesticides used in agriculture was poorly or negatively correlated with all other environmental indicators, and with the CalEnviroScreen calculation method, suggesting a limited ability of the method to capture agricultural exposures. In a PCA of the 5 socioeconomic variables, the first PC explained 66% of variance, representing overall socioeconomic hardship. In simultaneous autoregressive models, the first environmental and socioeconomic PCs were both significantly associated with the disease burden measure, but more model variation was explained by the socioeconomic PCs.ConclusionsThis study supports the use of CalEnviroScreen for its intended purpose of screening California regions for areas with high environmental exposure and population vulnerability. Study results further suggest a hypothesis that, compared to environmental pollutant exposure, socioeconomic status has greater impact on overall burden of disease.