Marine Pollution Bulletin 56 (2008) 1781–1787

Contents lists available at ScienceDirect

Marine Pollution Bulletin

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

Bayesian spatial modeling of Lavaca Bay pollutants

Wesley Bissett a,*, L. Garry Adams b, Robert Field a, William Moyer a, Tim Phillips c, H. Morgan Scott c,Terry Wade d, Steve Sweet d, James A. Thompson a

a Texas A&M University, College of Veterinary Medicine, Department of Large Animal Clinical Sciences, 4475 TAMU, College Station, TX 77843-4475, USAb Texas A&M University, College of Veterinary Medicine, Department of Veterinary Pathobiology, 4467 TAMU, College Station, TX 77843-4467, USAc Texas A&M University, College of Veterinary Medicine, Department of Veterinary Integrative Biosciences, 4458 TAMU, College Station, TX 77843-4458, USAd Texas A&M University, College of Geosciences, Geochemical and Environmental Research Group, 3149 TAMU, College Station, TX 77843-3149, USA

a r t i c l e i n f o a b s t r a c t

Keywords:Mercury

Polycyclic aromatic hydrocarbonsBayesian geo-statistical analysisSuperfundLocational risksLavaca BayTexas

0025-326X/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.marpolbul.2008.06.010

* Corresponding author. Tel.: +1 979 845 3541; faxE-mail address: wbissett@cvm.tamu.edu (W. Bisse

Locational risk of increased mercury and PAH concentrations in Lavaca Bay, Texas sediments and easternoysters (Crassostrea virginica) harvested from Lavaca Bay, Texas were analysed. Chemical analysis resultswere evaluated utilizing Bayesian geo-statistical methods for comparison of the model fit of a randomeffects model versus a convoluted model which included both random and spatial effects. For thoseresults fit best with the convoluted model, continuous surface maps of predicted parameter values werecreated. Sediment and oyster concentrations of mercury and the majority of measured PAHs were fit bestwith the convoluted model. The locational risks of encountering elevated concentrations of these pollu-tants in Lavaca Bay sediments and oysters were highest in close proximity to industrial facilities.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Contamination of Lavaca Bay, a secondary bay located along themid-Texas coast, has been well documented and has resulted inthe designation of a portion of the bay as a Superfund site due tomercury and polycyclic aromatic hydrocarbon (PAH) contamina-tion (USEPA, 2006). The Superfund program was established in1980 and allows the United States Environmental ProtectionAgency (USEPA) to address hazardous waste sites through eithergovernment or corporate remediation efforts (USEPA, 2008).Lavaca Bay’s Superfund designation occurred as a result of mercuryreleases by the Aluminum Company of America’s (ALCOA) chlor-al-kali unit located adjacent to Lavaca Bay. The chlor-alkali unit wasin operation between 1966 and 1979 with the bulk of mercuryreleased between 1966 and 1970. During this timeframe, waste-water was transported to a gypsum-lagoon located on an island lo-cated just off-shore of the industrial facility. After a settling period,water from the lagoon was discharged into Lavaca Bay (USEPA,2006). These releases led to elevated levels of methyl-mercury insediment, blue crabs and some species of finfish. As a result of thiscontamination the Texas Department of State Health Services(TDSHS) has banned consumption of fish and crab harvested fromthe affected portions of Lavaca Bay (Prosperie et al., 1999). It wasinitially thought that sediment deposition would gradually burythe mercury resulting in decreasing levels of bio-availability. This

ll rights reserved.

: +1 979 847 8863.tt).

has occurred in many but not all locations of the bay with deposi-tion of clean sediment leading to decreased levels of mercury in theupper, biologically active, layers of sediment. Mercury levels inseafood harvested from the area have also declined, but not as rap-idly nor to the extent predicted (Bloom and Lasorsa, 1999; Evanset al., 2000; Sager, 2002). Studies performed by ALCOA concludedthat multiple sources of mercury-release into Lavaca Bay were stillpresent. These included inputs from contaminated groundwater,run-off from the island disposal site, and re-suspension of contam-inated sediments by barge and ship-traffic (USEPA, 2006).

In addition to mercury, elevated levels of PAHs, a large family oftoxic chemical compounds have also been a concern in Lavaca Bay.Elevated PAH levels are the result of releases by a coal-tar facilityformerly in operation on the ALCOA property (USEPA, 2006). Astudy by Carr et al. (2001) documented that toxic levels of PAHswere still present in Lavaca Bay but the exact location of the high-est contamination varied among studies. Work cited in the USEPA(2006) Record of decision indicated that the highest levels of PAHswere detected in the closure area near the former Witco location.Marine sediment within the closure area has had consistent ele-vated PAH concentrations indicating a continued source of PAH re-lease into Lavaca Bay (USEPA, 2006).

2. Objectives

The objectives of this study were twofold. The first objectivewas to evaluate the extent and spatial distribution of current levels

1782 W. Bissett et al. / Marine Pollution Bulletin 56 (2008) 1781–1787

of mercury and polycyclic aromatic hydrocarbons found in LavacaBay sediment. The second was to determine the extent to whichthese contaminants have accumulated in pelagic fauna and theresulting spatial distribution using the Lavaca Bay oyster (Crassos-trea virginica) as the indicator organism.

3. Materials and methods

3.1. Sample collection

Sediment sampling was performed across Lavaca Bay withcollection locations selected to provide information at fresh-waterin-flow sources and industrial release points as well as locationsdistant from these two types of inputs into the bay. Sediment sam-ples were also collected adjacent to sampled oyster reefs. The low-er portions of Lavaca Bay were not sampled as extensively as theupper bay due to weather conditions at the time of sediment sam-ple collection in July, 2002.

The top 3–5 cm of sediment was retrieved with a Shipex grabsampler and placed in glass jars. Jars were filled and placed onice prior to delivery to the testing laboratory at Texas A&M Univer-sity’s Geochemical and Environmental Research Group (GERG)facilities. The samples were then frozen at �20 �C until chemicalanalysis was performed.

All identified viable oyster reefs in Lavaca Bay were sampled.Oyster collection locations are provided in Fig. 1. A minimum of25 oysters per reef were collected from eight different reefs. Theoysters were then separated, cleaned, and placed unopened inplastic zip-lock bags and stored on ice prior to delivery to the GERGlaboratory. After delivery to the laboratory, the oyster-shells wereopened with an oyster-knife, the oyster tissue frozen in glass jars,and the shells discarded.

3.2. Chemical analysis

Chemical analysis was performed following the standardsestablished by the National Oceanic and Atmospheric Administra-tion for the National Status and Trends Program (NOAA, 1998).Briefly, accelerated solvent extraction techniques were performedto extract sediment and tissue samples for surface prospecting ali-phatic and aromatic hydrocarbon analyses. Silica and aluminacolumnar chromatography was then utilized for purification of ex-tracts prior to analysis of aliphatics and PAHs. Gas chromatographyand mass spectrometry were utilized for quantitative determina-tion of PAH content mercury analysis was accomplished with astrong acid digestion of sediment samples followed by cold vaporatomic absorption spectrometry (CVAA).

3.3. Statistical analysis

Each reef and sediment-location was identified by its latitudeand longitude. These coordinates were used to plot the locationusing a commercial GIS software program.1 The map was then pro-jected into Universal Transverse Mercator 1983 (UTM83), zone14 units. The UTM83 coordinates were exported and used for all sta-tistical analyses. The spatial modeling of the contaminants was mod-eled using generalized linear kriging expanded to include a nugget, or‘‘random” effect at each location (Diggle and Ribeiro, 2007). The mod-el used a Bayesian method of inference, with vague prior beliefs and aMarkov Chain Monte Carlo (MCMC) implementation. The MCMCimplementation was performed by use of a readily available softwarepackage (Spiegelhalter et al., 2003). The prior beliefs included a non-

1 ArcGIS� Version 9.1, Environmental Systems Research Institute, Inc., Redlands,California.

informative normal distribution for the intercept with mean = 0 andprecision = 0.0001, and vague gamma priors (Gamma[0.01, 0.01]) forvariance components, including the range and nugget (spatially ran-dom location effect) and spatial effects (spatially dependent locationeffect). For all models, the distance-based variance function wasexponential with the covariance between locationi and locationj

modeled as a function of the distance between the two locations dij

and the rate of decline of covariance (/) as follows:

f ðdij;/Þ ¼ exp � /dij� �� �

Convergence was evaluated by visual examination of the his-tory plots of the two chains and visual examination of the Brooks,Gelman and Rubin statistics. For parameter estimation, the initial500 iterations were discarded to allow for convergence then every10th iteration was retained until 1000 iterations had been saved.For each contaminant, models with and without a spatial effectwere compared by use of the Deviance Information Criteria (DIC)(Spiegelhalter et al., 2002). An improvement of greater than 3.0in the DIC for the full model with the spatial effects was consideredto indicate an important spatial process.

For contaminants judged to have important spatial processes,Bayesian spatial prediction was performed for a grid of points witheach point representing the centroid of a 0.25-km � 0.25-km areaencompassing Lavaca Bay. One chain was utilized for predictioncalculations. A 1000-iteration burn-in was performed. An addi-tional 1000 iterations were performed and retained for the poster-ior distribution. Results of prediction modeling were imported intosatellite imagery of Lavaca Bay obtained from Google Earth.2 Thefont size at each prediction location was adjusted to provide asmooth prediction surface.

4. Results

4.1. Lavaca Bay sediment mercury concentrations and spatialdistribution

Locations within the closure area were found to have mercuryconcentrations considered to be harmful to exposed marine organ-isms. The Sediment Quality Guidelines developed for the NationalStatus and Trends Program were used to classify measured concen-trations as harmful or not. The Sediment Quality Guidelines weredeveloped for selected chemicals and trace metals that had exten-sive information available on what constitutes an exposure likelyto result in an adverse response in exposed populations. Specifi-cally, the effects range median values were used as research hasshown that the potential for adverse response in marine organismsincreases substantially when exposures above these levels occur(MacDonald et al., 1996). The highest concentration of mercurymeasured, 1.14 lg/g, dry weight, exceeded the effects range med-ian concentration of 0.71 lg/g, dry weight. Summary values formercury concentrations found in Lavaca Bay sediments are pro-vided in Table 1.

Comparison of the fit provided by the convoluted model versusthe random effects model demonstrated that mercury concentra-tions in Lavaca Bay sediments were fit best with the convolutedmodel indicating a spatial correlation in the data. Evaluation ofthe map of predicted mercury concentrations revealed a consistentspatial pattern across Lavaca Bay. The highest mercury concentra-tions were predicted in the vicinity of ALCOA and extended in anortherly direction. Elevated mercury levels were predicted to ex-tend beyond the current closure area to a point north of the High-way 35 causeway. The map of predicted sediment mercuryconcentrations is provided in Fig. 2a.

2 Google Earth�, Google, Inc., Mountain View, California.

Fig. 1. TDSHS closure area, Superfund Site, and oyster reef locations.

Table 1Summary data from Lavaca Bay sediments with ERL and ERM guideline values (MacDonald et al., 1996)

Chemical Guidelines Sediment summary data

Effects range low (ERL) Effects range median (ERM) Average Minimum Maximum Standard deviation Median

Mercury (lg/g, dry weight) 0.15 0.71 0.16 0.02 1.14 0.23 0.07

PAHs (ng/g, dry weight)Total PAHs 4022 44,792 2842.03 15.01 59,961.18 10,646.24 164.41Naphthalene 160 2100 6.05 0.20 96.31 17.02 2.21Acenaphthylene 44 640 21.43 0.04 458.04 82.29 0.59Acenaphthene 16 500 46.12 0.17 1122.26 195.95 0.81Fluorene 19 540 31.39 0.23 706.50 125.20 1.34Anthracene 85.3 1100 63.38 0.05 1484.73 259.00 2.30Pyrene 665 2600 355.87 0.33 7902.54 1395.29 10.37Benzo(a)anthracene 261 1600 184.21 0.15 3825.92 685.95 5.86Chrysene 384 2800 155.33 0.17 3138.34 563.87 6.13Benzo(a)pyrene 430 1600 259.08 0.25 5370.15 962.58 9.11Dibenzo(a,h)anthracene 63.4 260 21.61 0.04 436.00 78.71 0.752-Methylnaphthalene 70 670 105.68 0.17 2154.40 386.97 4.18

W. Bissett et al. / Marine Pollution Bulletin 56 (2008) 1781–1787 1783

4.2. Lavaca Bay tissue mercury concentrations and spatial distribution

Mercury concentrations within the closure area were below therecommended tissue residue criterion of 0.3 lg/g, fresh weight, for

the general public and above the recommended tissue criterion of0.049 lg/g, fresh weight, for subsistence fishermen. These recom-mendations were established to prevent harmful effects in humanpopulations associated with consumption of contaminated seafood

Fig. 2. Predicted mercury concentrations in (a) Lavaca Bay sediments and (b) Lavaca Bay oysters.

1784 W. Bissett et al. / Marine Pollution Bulletin 56 (2008) 1781–1787

(USEPA, 2000). The highest mercury concentration, 0.223 lg/g,fresh weight, was recorded within the closure area. Summary val-ues for mercury concentrations found in Lavaca Bay oysters areprovided in Table 2.

Based on the arbitrary criterion for improvement in model fit,tissue mercury concentrations were judged to have an importantspatial process within Lavaca Bay. Tissue concentrations of mer-cury were predicted to be elevated in the vicinity of ALCOA andDredge Island. The predicted area of highest concentrations formercury was limited to the area around Dredge Island. With theexception of the well-defined area near ALCOA, mercury concen-trations were predicted to be distributed across Lavaca Bay uni-

Table 2Summary data from Lavaca Bay oyster tissues with guideline values (USEPA, 2000)

Chemical USEPA guidelines(wet weight)

Average, dry weightbasis (wet weight)

Minimum, dbasis (wet w

Mercury (lg/g) (0.3a, 0.049b, 0.1c) 0.71 (0.07) 0.2 (0.02)

PAHs (ng/g)Total PAHs 1125.91 (98.65) 123.02 (10.7Acenaphthylene (5470a, 673b) 5.59 (0.49) 0.92 (0.08)Acenaphthene (5470a, 673b) 23.39 (2.05) 3.47 (0.30)Fluorene (5470a, 673b) 9.34 (0.82) 2.39 (0.21)Anthracene (547a, 67.3b) 16.86 (1.48) 2.39 (0.21)Pyrene (5470a, 673b) 49.52 (4.34) 3.55 (0.31)Benzo(a)anthracene 46.99 (4.12) 1.70 (0.15)Chrysene (547a, 67.3b) 54.86 (4.81) 2.86 (0.25)Benzo(a)pyrene (5.47a, 0.673b) 60.63 (5.31) 0.59 (0.05)Dibenzo(a,h)anthracene (1.094a, 0.1346b) 7.72 (0.68) 0.07 (0.01)Benzo(b)fluoranthene (54.7a, 6.73b) 92.76 (8.13) 3.08 (0.27)Benzo(k)fluoranthene (54.7a, 6.73b) 29.40 (2.58) 1.12 (0.10)Benzo(g,h,i)perylene (547a, 67.3b) 35.46 (3.11) 0.40 (0.04)Indeno(1,2,3-c,d)pyrene (54.7a, 6.73b) 44.62 (3.91) 0.83 (0.07)Fluoranthene (5470a, 673b) 96.32 (8.44) 4.38 (0.38)Phenanthrene (5470a, 673b) 38.24 (3.35) 5.31 (0.47)

a USEPA general public tissue residue criterion.b USEPA sustenance fishermen residue criterion.c USFWS criterion for protection of fish-eating birds and wildlife.

formly with no obvious spatial pattern but only site-specificrandom variation. The map of predicted mercury concentrationsin oyster tissues is provided in Fig. 2b.

4.3. Lavaca Bay sediment PAH levels and distribution

Concentrations of PAHs in Lavaca Bay sediments varied sub-stantially across Lavaca Bay with most having at least one locationwithin the Superfund site exceeding the effects range median con-centration. Total PAHs ranged from 59,961.2 ng/g, dry weight nearALCOA, to a low of 15 ng/g, dry weight. All PAH concentrations var-ied between locations with ranges between high and low values

ry weighteight)

Maximum, dry weightbasis (wet weight)

Standarddeviation

Median, dry weightbasis (wet weight)

2.39 (0.23) 0.72 0.34 (0.03)

8) 6616.31 (579.70) 1763.22 442.80 (38.80)37.68 (3.30) 9.81 2.15 (0.19)72.7 (6.37) 23.49 13.75 (1.20)51.27 (4.49) 11.97 5.12 (0.45)127.05 (11.13) 31.94 4.26 (0.37)449.94 (39.42) 104.47 10.62 (0.93)401.22 (35.15) 102.86 6.05 (0.53)399.47 (35.00) 105.93 9.45 (0.83)589.70 (51.67) 150.33 5.71 (0.50)66.94 (5.87) 17.22 0.98 (0.09)661.76 (57.98) 181.48 12.16 (1.06)238.20 (20.87) 63.39 3.53 (0.31)338.36 (29.65) 87.10 3.50 (0.31)398.88 (34.95) 101.61 4.91 (0.43)788.31 (69.07) 201.77 14.29 (1.25)314.15 (27.52) 74.24 9.99 (0.87)

W. Bissett et al. / Marine Pollution Bulletin 56 (2008) 1781–1787 1785

being approximately a 50-fold change to greater than a 7000-foldchange. Summary values for PAH concentrations found in LavacaBay are provided in Table 1.

The extreme variation in PAH concentrations between locationsnecessitated a logarithmic transformation for analysis. Based onthe arbitrary criterion for improvement in model fit, all of the sed-iment-PAHs were judged to be affected by an important spatialprocess within Lavaca Bay.

The predicted spatial distributions for all individual PAHs weresimilar with two areas identified as having the highest predictedconcentrations. As expected, one of these locations was locatedwithin the closure area near the north end of Dredge Island. Thesecond area with the highest predicted concentrations was locatedNorth of the Highway 35 causeway and outside of the current clo-sure area. Intermediate concentrations were predicted to occur be-tween these two locations. With the exception of the areasurrounding the two locations discussed above, predicted sedi-ment PAH concentrations were low throughout the remainder ofthe study area indicating that migration of the contaminants fromthe point of release was limited. The map of predicted sedimentbenzo(a)pyrene concentrations is provided as an example of thisfamily of chemicals in Fig. 3a.

4.4. Lavaca Bay tissue PAH levels and distribution

As in Lavaca Bay sediments, PAH concentrations in oyster tissuevaried substantially across Lavaca Bay. Total PAHs ranged from ahigh of 6616.3 ng/g, dry weight near ALCOA, to a low of 123 ng/g,dry weight. Summary values for PAH concentrations found in Lava-ca Bay oysters are provided in Table 2.

Based on the arbitrary criterion for improvement in model fit,35 of the 46 PAHs were judged to have an important spatial pro-cess within Lavaca Bay. The spatial distribution of predicted tissuePAH concentrations were similar to the distribution predicted forsediment concentrations with all PAHs having a similar distribu-tion in sediment and tissue. The highest concentrations were pres-ent along the eastern shore of Lavaca Bay and in close proximity toALCOA and the Superfund site. A spatial orientation of tissue PAH

Fig. 3. Predicted benzo(a)pyrene concentrations in (a)

concentration was apparent with examination of the maps.Predicted concentrations decreased as distance from these loca-tions increase however the predicted concentrations do not fallas quickly nor are they as consistent as that predicted for sedi-ment concentrations. A map of predicted benzo(a)pyrene concen-trations is provided as an example of this family of chemicals inFig. 3b.

5. Discussion

Results of sediment and oyster tissue analyses indicated thatportions of Lavaca Bay are still contaminated with mercury andPAHs. Evaluation of the predicted spatial distributions revealedthat the locational risks for exposure to these pollutants were attheir highest in close proximity to the industrial facility. Reviewof historical data showed that mercury concentrations in sedi-ments have been quite variable ranging from a low of 0.12 lg/g,dry weight in 1997 to a high of 39.3 lg/g, dry weight in 1987(NCCOS, 2006). Santschi et al. estimated that in the absence ofnew mercury releases into Lavaca Bay, surficial concentrations ofmercury should have decreased significantly with a recoveryhalf-life of approximately four years (Santschi et al., 1999). Thedramatic increase noted in 1987 and the increase in levels notedbetween 1997 (0.121 lg/g, dry weight) and 2002 (0.47 lg/g, dryweight) were indicative of the release of additional mercury intoLavaca Bay or re-suspension of buried mercury into the bio-activezone resulting from mechanical disruption associated with weath-er events such as storms and hurricanes and boating and shippingactivities.

ALCOA’s remediation efforts have included evaluation of contin-ued sources of mercury release into the bay system. There are threeground-water zones near the chlor-alkali processing unit that havebeen evaluated. Of these three, the second zone, located three toseven meters below sea-level, has a point of discharge directly intoLavaca Bay. A variety of different methods has been utilized toevaluate the extent of mercury loading in the Lavaca Bay systemresulting from discharge of contaminated ground-water from thesecond zone. The various methods utilized yielded a broad range

Lavaca Bay sediments and (b) Lavaca Bay oysters.

1786 W. Bissett et al. / Marine Pollution Bulletin 56 (2008) 1781–1787

of loading estimates ranging from 0.2 to 41 kg of mercury releasedinto Lavaca Bay annually. ALCOA has been operating a groundwa-ter extraction program since 1998 to prevent the flow of contami-nated groundwater into Lavaca Bay. The groundwater extractionsystem is part of the remedial measures approved in the USEPA’sRecord of Decision updated in January of 2006 and is assumed tohave resulted in a significant decrease in the release of mercuryinto Lavaca Bay (USEPA, 2006).

The role of the movement of mercury from the island disposalsite into Lavaca Bay has also been evaluated. The island’s soilsand surface waters have been found to contain elevated concentra-tions of mercury with continued contamination of Lavaca Baythrough leaching of mercury into the coastal waters. Estimates pre-sented in USEPA’s 2006 Record of Decision indicated that between3.6 and 5.9 kg of mercury have been released into Lavaca Bayannually with most of the release being at the northern side ofthe island (USEPA, 2006).

The increase in mercury levels noted between 1997 and thesampling performed during this study occurred after initiation ofgroundwater extraction efforts which were designed to reversethe flow of contaminated groundwater into Lavaca Bay. This sam-pling period was also performed after completion of remedial mea-sures designed to prevent the flow of contaminants from DredgeIsland into the bay system (USEPA, 2006). Results of this studyindicate that remediation efforts by ALCOA at the time of samplecollection had not successfully prevented the potential for marineorganisms to be exposed to harmful levels of mercury. Addition-ally, the area predicted to contain the highest concentrations ofmercury in sediments extended beyond the area historically con-sidered to present the greatest risk.

Tissue analysis also revealed the presence of elevated mercuryconcentrations. The highest mercury level at 2.39 lg/g, dry weight,was 20 times higher than the national median (NCCOS, 2006).While mercury levels have decreased to the point of not posing athreat to the health of the general public, they are still high enoughto pose a threat to subsistence fishermen. The sites with highestconcentrations of mercury were, like the highest sediment levels,in close proximity to ALCOA. The groundwater extraction systemand the Dredge Island fortifications were designed to prevent addi-tional movement of mercury into the Lavaca Bay ecosystem (USEP-A, 2006). The elevations in oyster mercury levels noted in thisstudy, particularly those just south of the Highway 35 causeway,indicated that at the time of sample collection either the remedialmeasures instituted were not completely effective and/or thatmechanical disruption of the sediment was re-suspending substan-tive amounts of mercury. The results of this study show that mer-cury was gaining access to the bio-active zone and was leading topotential adverse effects in marine species. These findings are inagreement with other studies (Sager, 2002).

Of the 45 different PAHs analysed in this study 11 including flu-oranthene, pyrene, benzo(a)anthracene, chrysene, benzo(a)pyrene,dibenzo(a,h)anthracene, naphthalene, acenapthalene, acenapth-ene, fluorene, phenanthrene, and 2-methylnapthelene were foundto exceed the probable-effects level at two different locations lo-cated adjacent to ALCOA. One of these locations had concentrationsof these same PAHs that exceeded the established effects rangemedian levels. Benzo(a)pyrene, the most potent animal-carcinogenamong the PAHs, provided the greatest reasons for concern overPAH contamination in Lavaca Bay. The highest level noted duringthis study, 5370.2 ng/g, dry weight was well above the effectsrange median concentration of 1600 ng/g, approximately 1.5 timeshigher than the apparent effects threshold-high level of 3600 ng/g,and seven times higher than the probable-effects level of 763 ng/g(MacDonald et al., 1996).

The elevated concentrations of PAHs found during this studyindicated that significant risks of adverse effects were likely to oc-

cur in marine life exposed to sediments in the vicinity of ALCOA.The source of PAHs is thought to be associated with pollution frompast industrial activities on the ALCOA property. The primarymechanism for current PAH release is thought to be movementof a dense, non-aqueous phase liquid from polluted groundwaterinto Lavaca Bay sediments. A dense, non-aqueous phase liquid isheavier than water and is not easily dissolved into water. It formsa liquid layer located below groundwater, and as noted earlier,there is a direct communication between the groundwater zone lo-cated 3–7 m below sea-level and Lavaca Bay (USEPA, 2006). Thepredicted spatial distributions of PAHs illustrated the potentialfor adverse effects over a larger geographical area than previouslythought. The area north of the State Highway 35 causeway pro-vides cause for concern. This area is outside of the closure areaestablished by the TDSHS and is an area frequented by recreationalfishermen. Potential explanations for predicted elevations in PAHconcentrations include an additional source of communication be-tween Lavaca Bay and polluted groundwater zones, release of addi-tional PAHs from another source, or a natural occurrenceassociated with under-ground oil seepage.

Oysters lack an efficient hepatic-detoxification-system and tendto bio-accumulate PAHs from the environment. For this reasonthey are often used to monitor industrial activities and accidentsassociated with higher environmental loads of PAHs (Orbea et al.,2002; Payne, 1977). The current study showed the highesttissue-levels from oysters located at the two reefs in closestproximity to ALCOA. The PAH levels at these reefs were high en-ough to constitute a threat to public-health. Benzo(a)anthracene,benzo(a)pyrene, benzo(b)fluoranthene, benzo(k)fluoranthene, anddibenzo(a,h)anthracene all exceeded the USEPA general public tis-sue residue criterion, with benzo(a)pyrene being almost ten timeshigher than the USEPA established level (USEPA, 2000).

Based on the oysters tendency to bio-accumulate PAHs, it wasexpected that spatial distributions of sediment and oyster tissuewould be similar. The predicted spatial distribution of tissue con-centrations of PAHs was strikingly similar to that predicted for sed-iments and was also cause for concern. The area with predictedelevations in close proximity to ALCOA was likely the result of pastactivities on the ALCOA property. The area north of the State High-way 35 causeway predicted as having elevated levels of PAHs inoysters was outside of the current closure area and provided thepotential for human exposures in fishermen harvesting and con-suming seafood from this area. Reasons for these elevations werethe same as those postulated for sediment PAH concentrations atthis same location.

Results of chemical analysis and predictive modeling performedduring this study showed that the Lavaca Bay ecosystem containeda complex mixture of chemicals and that the locational risks forexposure to the majority of chemicals evaluated were elevated incommon locations. This study provided convincing evidence thatmercury and PAHs, were present at high enough concentrationsto constitute a threat to environmental health and the health ofsubsistence fishermen without consideration of interactions be-tween multiple pollutants. With the highest measured and pre-dicted concentrations of multiple pollutants co-located atcommon locations, the potential for harmful effects and concernfor the health of the Lavaca Bay system is magnified.

Acknowledgments

We would like to express our gratitude to the Vivian L. SmithFoundation and the Point Comfort, Texas ALCOA facility for provid-ing the financial resources for completion of this project. Also, toour colleagues in Texas A&M University’s Food Animal Medicineand Surgery Service, your support and encouragement are greatlyappreciated.

W. Bissett et al. / Marine Pollution Bulletin 56 (2008) 1781–1787 1787

References

Bloom, N.S., Lasorsa, B.K., 1999. Changes in mercury speciation and the release ofmethyl mercury as a result of marine sediment dredging activities. The Scienceof the Total Environment 237–238, 379–385.

Carr, S., Biedenbach, J., Hooten, J., Russell, L., 2001. Sediment quality assessmentsurvey and toxicity identification evaluation studies in Lavaca Bay, Texas, amarine Superfund Site. Environmental Toxicology 16, 20–30.

Diggle, P.J., Ribeiro, P.J., 2007. Model-based Geostatistics. Springer Science, NewYork, NY.

Evans, D.W., Kathman, R.D., Walker, W.W., 2000. Trophic accumulation anddepuration of mercury by blue crabs (Callinectes sapidus) and pink shrimp(Penaeus duorarum). Marine Environmental Research 49, 419–434.

MacDonald, D.D., Carr, R.S., Caldet, F.D., Long, E.R., Ingersoll, C.G., 1996.Development and evaluation of sediment quality guidelines for FloridaCoastal Waters. Ecotoxicology 5, 253–278.

NCCOS (National Centers for Coastal Monitoring and Assessment), 2006. MonitoringData – Mussel Watch. Available from: <http://www8.nos.noaa.gov/cit/nsandt/download/mw_monitoring.aspx> [accessed 13 April 2006].

NOAA (National Oceanic and Atmospheric Administration), 1998. Sampling andAnalytical Methods of the National Status and Trends Program, MusselWatch Project: 1993–1996 Update. Lauenstein, G.G., Cantillo, A.Y. NOS ORCA130, 233.

Orbea, A., Ortiz-Zarragoitia, M., Sole, M., Porte, C., Cajaraville, M.P., 2002.Antioxidant enzymes and peroxisome proliferation in relation to contaminantbody burdens of PAHs and PCBs in bivalve mollusks, crabs and fish fromthe Urdaibai and Plentzia estuaries (Bay of Biscay). Aquatic Toxicology 58, 75–98.

Payne, J.F., 1977. Mixed function oxidases in marine organisms in relation topetroleum hydrocarbon metabolism and detection. Marine Pollution Bulletin 8,112–116.

Prosperie, S.L., Villanacci, J.F., Kathryn, A., Evans, Ingram, N.B., Wagner, L.B., et al.,1999. Preliminary public health assessment: ALCOA.

Sager, D.R., 2002. Long-term variation in mercury concentrations in estuarineorganisms with changes in releases into Lavaca Bay, Texas. Marine PollutionBulletin 44, 807–815.

Santschi, P.H., Allison, M.A., Asbill, S., Perlet, A.B., Cappellino, S., Dobs, C., et al., 1999.Sediment transport and Hg recovery in Lavaca Bay, as evaluated fromradionuclide and Hg distributions. Environmental Science and Technology 33,378–391.

Spiegelhalter, D.J., Best, N., Carlin, B.P., van der Linde, A., 2002. Bayesian measures ofmodel complexity and fit. Journal of the Royal Statistical Society Series B 64,583–639.

Spiegelhalter, D., Thomas, A., Best, N., Lunn, D., WinBUGS 1.4. [1.4]. 2003.Cambridge, UK, Imperial College of Science, Technology, and Medicine.

USEPA (United States Environmental Protection Agency), 2006. ALCOA/Lavaca Bay(Calhoun County) TEXAS EPA ID# TXD008123168 Site ID: 0601752.TXD008123168, 1-14-2. 2006h. United States Environmental ProtectionAgency.

USEPA (United States Environmental Protection Agency), 2000. Guidance forAssessing Chemical Contaminant Data for Use in Fish Advisories, vol. 1. FishSampling and Analysis, third ed., EPA823-B-00-007. Office of Water,Washington DC.

USEPA (United States Environmental Protection Agency), 2008. Superfund: BasicInformation. Available from: <http://www.epa.gov/superfund/about> [accessed02 June 2008].