Sediment Toxicity Identification Evaluation (TIE) Studies at Marine SitesSuspected of Ordnance Contamination
R. S. Carr,1 M. Nipper,2 J. M. Biedenbach,1 R. L. Hooten,1 K. Miller,3 S. Saepoff4
1 U.S. Geological Survey, BRD, CERC, Marine Ecotoxicology Research Station, Texas A&M University—Corpus Christi, Center for Coastal Studies,6300 Ocean Drive, NRC, Suite 3200, Corpus Christi, Texas 78412, USA2 Texas A&M University—Corpus Christi, Center for Coastal Studies, 6300 Ocean Drive, NRC, Suite 3200, Corpus Christi, Texas 78412, USA3 Naval Facilities Engineering Service Center (NFESC), Restoration Development Branch, Port Hueneme, California 93043, USA4 US. Army Corps of Engineers, Seattle District, 4735 East Marginal Way South, Seattle, Washington 98134-2385, USA
Received: 12 December 2000/Accepted: 11 May 2001
Abstract. A sediment quality assessment survey and subse-quent toxicity identification evaluation (TIE) study was con-ducted at several sites in Puget Sound, Washington. The siteswere previously suspected of contamination with ordnancecompounds. The initial survey employed sea urchin porewatertoxicity tests to locate the most toxic stations. Sediments fromthe most toxic stations were selected for comprehensive chem-ical analyses. Based on the combined information from thetoxicity and chemical data, three adjacent stations in OstrichBay were selected for the TIE study. The results of the phaseI TIE suggested that organics and metals were primarily re-sponsible for the observed toxicity in the sea urchin fertiliza-tion test. In addition to these contaminants, ammonia was alsocontributing to the toxicity for the sea urchin embryologicaldevelopment test. The phase II TIE study isolated the majorityof the toxicity in the fraction containing nonpolar organics withhigh log Kow, but chemical analyses failed to identify a com-pound present at a concentration high enough to be responsiblefor the observed toxicity. The data suggest that some organic ororganometallic contaminant(s) that were not included in thecomprehensive suite of chemical analyses caused the observedtoxicological responses.
Previous surveys have shown that sediments in the vicinity ofnaval facilities in Puget Sound, Washington, were contami-nated with ordnance compounds, originating from past use,storage, improper disposal, and incineration of these com-pounds (URS Consultants 1995; EA Engineering, Science, andTechnology, Inc. 1996). It is not possible, however, to predictif sediment samples will be toxic on the basis of analyticalchemistry information alone (Longet al. 1995; MacDonaldetal. 1996). Toxicity tests are recognized as effective tools fordetermining the biological significance of contamination found
in coastal sediments (Carret al. 1996a, 1996c, 2000). Thecontaminants responsible for any observed toxicity can bedetermined using toxicity identification evaluation (TIE) pro-cedures. TIEs have been shown to be an effective tool for theidentification of causative agents of toxicity in pore watersfrom marine sediments (Hoet al. 1997; Carret al. 2001).
A sediment toxicity survey was used to select sites forcomprehensive toxicological and chemical analyses and for theselection of sites for TIE studies. For the sediment toxicitysurvey, surficial sediments were collected in the vicinity ofJackson Park and of Port Hadlock Naval Facilities, PugetSound, WA, and two stations in Sequim Bay, WA, werepreselected as reference sites. Sediments were analyzed forporewater toxicity using sea urchin (Arbacia punctulata) fer-tilization and embryological development tests. The most toxicsediments were characterized chemically. Based on the com-bined results of the chemical analyses and toxicity tests, threestations adjacent to each other at the Jackson Park NavalReservation site were selected for a TIE study.
Materials and Methods
Sediment Collection
Sampling sites in Puget Sound were selected based on naval activitiesin the area and on previous information on ordnance contamination.Sampling was concentrated in two main areas: the Jackson Park site,with 25 stations selected in Ostrich Bay (OB1-25, Figure 1); and thePort Hadlock site, with 25 stations selected adjacent to the Northernend of Indian Island (PT1-25, Figure 2). Two preselected referencestations (SQ1 and SQ2) were located in Sequim Bay. Our laboratoryreference station was located in Redfish Bay, Texas.
Sediment samples in Puget Sound were collected with the assistanceof the Washington State Department of Ecology, which provided theresearch vessel and sampling gear. Sampling was performed using amodified double Van Veen grab. Sediment samples were shipped tothe U.S. Geological Survey (USGS) Marine Ecotoxicology ResearchStation (MERS) in Corpus Christi, Texas, where the sediment wasprocessed and tests were performed. Surficial sediment samples wereCorrespondence to:R. S. Carr;email: [email protected]
collected from the 52 selected stations in Puget Sound in May 1998.Samples were placed in precleaned 1-gal high-density polyethylenecontainers, chilled, and shipped in insulated coolers with blue ice.Separate sediment subsamples were collected at stations OB1, OB8,and PT13 for replicate testing as part of the quality assurance programfor the toxicity testing and chemical analyses. These were placed inseparate containers and labeled as OB26, OB27, and PT26, respec-tively. Samples were received by the USGS in Corpus Christi the dayfollowing shipment. All porewater samples were extracted within 3days from the time of field collection of sediment.
Sediment Porewater Extraction Procedure
Pore water was extracted from the sediments using a pneumaticextraction apparatus (Carr and Chapman 1995). This extractor is madeof polyvinyl chloride and uses a 5mm polyester filter. It is the samedevice used in previous sediment quality assessment surveys (Carr andChapman 1992, 1995; Carret al.1996a, 1996b, 1996c, 2000; Carr andNipper 1999). After extraction, the porewater samples were centri-fuged in polycarbonate bottles at 1,2003 g for 20 min to remove any
suspended particulate material; the supernatant was placed in amberglass jars and frozen at220°C. The pore water was stored frozen untiljust prior to testing, when water quality parameters were measured andadjusted, if necessary. This extraction procedure has been shown insubsequent sediment-spiking studies with ordnance compounds toresult in minimal loss of ordnance compounds (Nipperet al. 2002a,2002b).
Two days before conducting a toxicity test, the samples were movedfrom the freezer to a refrigerator at 4°C. One day prior to testing,samples were thawed in a tepid (20°C) water bath. Temperature of thesamples was maintained at 206 1°C after thawing. Sample salinitywas measured and adjusted to 306 1‰, if necessary, using purifieddeionized water or concentrated brine. Other water quality measure-ments (dissolved oxygen, pH, sulfide, and ammonia concentrations)were made. Temperature and dissolved oxygen (DO) were measuredwith YSI meters; salinity was measured with a Reichert refractometer;and pH, sulfide (as S22), and total ammonia (expressed as nitrogen;NH4) were measured with Orion meters and their respective probes.Unionized ammonia concentrations (expressed as nitrogen; NH3) werecalculated for each sample using the respective salinity, temperature,pH, and NH4 values (Whitfield 1974). Following water quality mea-
Fig. 1. Sampling stations at Jackson Park site, Ostrich Bay, Puget Sound, Washington. Color differentiation of symbol indicates those stationsthat were significantly toxic relative to the reference site in the sea urchin (A. punctulata) fertilization and embryological development test(Dunnett’st test,a # 0.05 and detectable significance criteria applied)
Marine Sediment TIE Studies for Ordnance Compounds 299
surements and adjustments, the samples were stored overnight at 4°Cbut returned to 206 1°C before the start of the toxicity tests.
Toxicity Tests
Toxicity of the sediment pore water was determined using the seaurchin fertilization and embryological development tests withA.punctulata, following the procedures described previously (Carretal. 1996a, 1996b).A. punctulataurchins were obtained from GulfSpecimen Company, Inc. (Panacea, FL). Each of the 55 porewatersamples (including the three replicate subsamples) was tested in adilution series design at 100%, 50%, and 25% of the water qualityadjusted sample with five replicates per treatment. Dilutions weremade with 0.45mm Millipore filtered sea water (MFS). A referenceporewater sample collected from Redfish Bay, Texas, which hadbeen handled identically to the test samples, was included with eachtoxicity test as a negative control. This site is far removed from anyknown sources of contamination and has been used for more thana decade as a reference site. In addition, dilution blanks of MFS andbrine controls (purified deionized water with brine added to reach
a 30% salinity) were included. The brine control had the objectiveof identifying any possible adverse effects caused by the brine. Adilution series test with sodium dodecyl sulfate (SDS) was includedas a positive control and results were compared to the respectivecontrol charts.
Chemical Analyses
Thirteen sediment samples and the replicate sample of station OB8were selected for chemical analyses. The samples selected were fromstations OB6, OB8, OB12, OB15, OB16, OB18, OB22, and OB23,from the Jackson Park site, and PT7, PT11, PT12, PT16, and PT19from the Port Hadlock site. These sediments were shipped on dry iceto Columbia Analytical Services Inc. (CAS), Kelso, WA. Chemicalanalyses included a suite of ordnance compounds, trace metals, poly-cyclic aromatic hydrocarbons (PAHs), organochlorinated pesticides,polychlorinated biphenyls (PCBs), and butyltins, as well as particlesize distribution (Carr and Nipper 1999). Ordnance compounds weremeasured by HPLC using Method 8330 (US EPA 1994), and tracemetals were measured by ICP/MS using EPA Methods 200.8 (US EPA
Fig. 2. Sampling stations near the Port Hadlock naval facilities adjacent to Indian Island, Puget Sound, Washington. Color differentiation ofsymbol indicates those stations that were significantly toxic relative to the reference site in the sea urchin (A. punctulata) fertilization andembryological development test (Dunnett’st test,a # 0.05 and detectable significance criteria applied)
300 R. S. Carret al.
1996), except for mercury, which was measured by CVAA using EPAMethod 7471 (US EPA 1996). PAHs were measured using GC/MSselected ion monitoring using EPA method 8270 (US EPA 1996).PCBs were measured by EPA Method 8082 (US EPA 1996) usingGC/ECD. Organochlorinated pesticides were measured by GC/ECDusing method 8081A (US EPA 1996). Butyltins were measured byGC/FPD using the method of Kroneet al. (1989). Particle size distri-bution was analyzed by method PS-PSEP modified (PSEP 1986).
Data Analysis
The EC50 values for dilution series toxicity tests were calculated by thetrimmed Spearman-Karber method (Hamiltonet al. 1977) with Ab-bott’s correction (Morgan 1992). Statistical comparisons between eachtreatment and the reference pore water were made using analysis ofvariance (ANOVA) and Dunnett’s one-tailedt test (which controls theexperimentwise error rate) on the arcsine square root transformed datawith the aid of Statistical Analysis Systems software (SAS 1989). Priorto statistical analysis, the transformed data sets were screened foroutliers (SAS 1992). Outliers were detected by comparing the studen-tized residuals to a critical value from at distribution chosen using aBonferroni-type adjustment. The adjustment is based on the number ofobservations, n, so that the overall probability of a type I error is atmost 5%. The critical value, cv, is given by the following equation:cv 5 t(dfError,.05/(23 n)). After omitting outliers but prior to furtheranalysis, the transformed data sets were tested for normality and forhomogeneity of variance using SAS/LAB software (SAS 1992).
A second criterion was also used to compare test means to referencemeans. Detectable significance criteria (DSC) were developed to de-termine the 95% confidence value based on power analysis of allsimilar tests performed by our lab (Carr and Biedenbach 1999), tominimize statistical artifacts. This value is the percent minimum sig-nificant difference from the reference that is necessary to detectsignificant effects while minimizing type I errors. The DSC value forthe sea urchin fertilization assay is 15.5% ata # 0.05, and 19% ata #0.01. For the embryological development test the DSC values ata #0.05 anda # 0.01 are 16.4 and 20.6%, respectively.
TIE Procedures
Phase I: Based on the results of the toxicity tests and chemicalanalyses, three of the most toxic and contaminated stations, whichwere located in Ostrich Bay and adjacent to each other (Figure 1,inside circle), were selected for the TIE procedure. Twenty-eight litersof sediment from each station were collected in August 1998, com-posited into one 84-L sample, and shipped to the USGS MarineEcotoxicology Research Station (MERS) in Corpus Christi, Texas.Pore water was extracted from this sediment composite on arrival atthe laboratory, and processed as described above.
The sea urchin fertilization and embryological development testswere used with the TIE procedure. Initially, the toxicity of a frozen anda fresh porewater sample was compared. Because no significant dif-ference in toxicity was detected between the samples (Carr and Nipper1999), frozen pore water was used for the TIE procedure and treatedas described in the previous sections.
Baseline toxicity of the sample was assessed. Phase I TIE treatmentswere applied following the US EPA protocol (Burgesset al.1996) andconsisted of:
1. Aeration, for the assessment of the contribution of volatile chem-icals to the toxicity;
2. Filtration, for the assessment of the contribution of particulatematerial and particle-associated contaminants to the toxicity;
3. C18 column, for the assessment of the contribution of slightly polarand nonpolar organic chemicals to the toxicity;
4. EDTA addition, for the assessment of the contribution of metals tothe toxicity;
5. Sodium thiosulfate addition, for the assessment of the contributionof oxidants to the toxicity;
6. Graduated pH adjustments for the assessment of the contribution ofammonia to the toxicity.
Several contaminants were identified in the initial phase IIchemical characterization of the composite porewater sam-ple for which no toxicity information was available for thesea urchin tests. Therefore, toxicity tests were conducted fortributyltin (TBT, bis (tributyltin oxide, 96% purity, Sigma-Aldrich Chemical Co.), dibutyltin (DBT, dibutyltin oxide,98% purity, Sigma-Aldrich), and arsenic (arsenic (III) ox-ide, AS2O3, 99% purity, Sigma-Aldrich). The low solubilityof these chemicals in sea water required the use of carriersto prepare the stock solutions for the toxicity tests. Methanolwas used as the carrier for TBT, acetic acid for DBT, andHCl for arsenic. Appropriate carrier controls were includedwith each test series which were performed in a 50% serialdilution test design.
Phase II: Phase II TIE procedures were conducted for the identi-fication of organic contaminants responsible for toxicity, using theprocedures described in Ankleyet al. (1991). Organic extractionand fractionation was performed on the Ostrich Bay sample utiliz-ing J. T. Baker, Bakerbond Speedisks C18, a Speedisk remotesample adapter and a vacuum pump attached to a 2-L Erlenmeyervacuum flask.
Extraction. Two Speedisks were utilized to extract 5 L of porewater (2.5 L each). Each Speedisk was attached to the vacuum flaskusing a silicone rubber stopper adapter. The Speedisk was precon-ditioned with two 10-ml aliquots of HPLC-grade methanol. Thedisk was pulled under vacuum at a flow rate of 100 –150 ml/minuntil 3–5 mm of solvent was left covering the disk. Before theporewater sample was added, the disk was rinsed with two 10-mlaliquots of MilliQ purified water drawn down under vacuum until3–5 mm of water remained on the disk. The disk was not allowedto dry out between conditioning and sample addition. Sample wasadded to the disk to bring the level to the top of the disk reservoir(' 35 ml) and a (methanol/DI rinsed) remote sample adaptorattached to the top of the reservoir. The Teflont tubing from theadaptor was clipped to the side of a 4-L beaker containing 2.5 L ofthe sample pore water. The sample was drawn up the tubing into thereservoir under vacuum through the Speedisk at a flow rate of100 –150 ml/min. After the entire volume had been passed throughthe Speedisk, the disk was dried under full vacuum for 5 min.
Fractionation. Fractions of organics were eluted off theSpeedisks using the same vacuum system modified to collect thefractions into precleaned 20-ml glass vials. Eight methanol/waterfractions were eluted off each Speedisk. Percent methanol fractionswere as follows: 25, 50, 75, 80, 85, 90, 95, and 100%. Thenonmethanol component of the fractions consisted of 0.45-mmfiltered sea water. Elution was performed by adding a 3.5-mlvolume of the elution fraction to the disk and allowing a 1-min soakbefore collecting the fraction under vacuum into a precleaned glassvial until the Speedisk was dry. The procedure was repeated twicewith each fraction and collected in the same vial for a total volumeof 7 ml for each fraction. This procedure was repeated for eachmethanol/water fraction beginning with 25% methanol and continu-ing through all the fractions in increasing methanol concentrations.
Marine Sediment TIE Studies for Ordnance Compounds 301
Both Speedisks were eluted in the same manner from each disk andfractions of the same methanol/water concentration were combinedin a single vial (20-ml glass) and capped with a Teflon-lined capand refrigerated until toxicity testing. Elution blanks were preparedin the same manner as described above using a separate precondi-tioned and rinsed Speedisk.
Phase II toxicity testing. Blank and sample fractions were testedfor toxicity using the sea urchin fertilization test. Five replicates, eachcontaining 5 ml of dilution water (0.45mm filtered sea water) injectedwith 75 ml of each blank or sample fraction were tested. Comparisonswere made between sample fractions and blanks of the same methanol/water concentrations using pairedt tests.
Chemical Analyses: A subsample (4 L) of the frozen porewatersamples used for the TIE procedure was shipped on dry ice to Colum-bia Analytical Services. A subsample (4 L) of the fresh porewatersample used for the initial comparison of fresh and frozen porewatertoxicity was also sent for chemical analyses. The same methodsdescribed for the sediment analyses were applied for chemical mea-surements conducted with the pore water. Aliquots of the exposuremedia for each dilution for the tests with TBT, DBT, and arsenic werealso analyzed by CAS by the methods described previously for thesesubstances.
Data Analysis: Statistical comparisons between each TIE treatmentapplied to the porewater sample from Ostrich Bay and the baselinetoxicity of that same sample were made using ANOVA and Dunnett’stwo-tailed t test on the arcsine square root transformed data with theaid of SAS (SAS 1989). Dunnett’s one-tailedt test was used todetermine lowest observed effect concentrations (LOECs) and noobserved effect concentrations (NOECs) for the TBT, DBT, andarsenic toxicity tests. The trimmed Spearman-Karber method (Hamil-ton et al. 1977) with Abbott’s correction (Morgan 1992) was used tocalculate EC50 (50% effective concentration) values for the dilutionseries tests.
Results and Discussion
Sediment Quality Survey
Sea Urchin Fertilization Toxicity Test:Ostrich Bay, at theJackson Park site, had more stations that were toxic in theurchin fertilization test than did the Port Hadlock site (Figures1 and 2). Samples were considered toxic when their effect wassignificantly higher (a # 0.05) than that of the referencesample and the percent fertilization was below the detectablesignificance criteria (DSC) ata # 0.05 (Carr and Biedenbach1999). The most toxic samples were from stations OB12,OB16, and OB23, with significant toxicity at a 25% porewaterconcentration. The samples from stations OB8, OB11, OB15,OB17, and OB18 were significantly toxic at a 50% porewaterconcentration. In the 100% sample, stations OB5, OB7, OB14,OB20, OB22, and OB24 at Jackson Park site; stations PT15and PT23 at the Port Hadlock site; and station SQ2 in SequimBay were significantly toxic.
EC50 values were calculable for 22% (12 of 55) of thesamples from all sites. The remaining 78% of the samples werenot toxic enough for the calculation of an EC50, which wouldbe. 100%. Only one of the samples, from station OB12, hadan EC50 , 25%, indicating that it would still be toxic at aporewater concentration, 25%. The toxicity results for the
quality assurance replicate samples were practically identical.Stations OB1/OB27 and PT13/PT26 were nontoxic, withEC50 . 100%. Replicates OB8/OB26 were significantly toxicat a 50% concentration and EC50 values were 54.4 and 57.3,respectively, with overlapping 95% confidence limits.
Sea Urchin Embryological Development Toxicity Test:Sim-ilar to the fertilization test, samples used in the sea urchinembryological development tests were considered toxic whentheir effect was significantly higher (a # 0.05) than that of thereference sample and the percent normal embryos was belowthe DSC ata # 0.05. The urchin embryological developmenttest was more sensitive than the fertilization test. Only threestations, all at the Port Hadlock site (PT1, PT2, and PT4), didnot exhibit toxicity in the undiluted pore water (100% strength)(Figures 1 and 2). At the Port Hadlock site, the samples fromstations PT7, PT8, PT11, PT19, and PT24 were the most toxic,causing an adverse effect at the 25% concentration. Samplesfrom stations PT12, PT16, and PT21 were toxic at the 50%concentration and the remaining 14 samples were toxic whenundiluted (100%). Higher levels of toxicity occurred in thesamples from the Jackson Park site, with stations OB6, OB15,and OB22 toxic at a 25% concentration, stations OB4, OB5,OB8, OB10, OB11, OB12, OB13, OB14, OB18, OB19, andOB23 toxic at a 50% concentration, and the remaining 11stations toxic in the undiluted porewater sample.
The toxicity results for the quality assurance field replicatesamples were more variable with the embryological develop-ment test than with the fertilization test, but still close to eachother. Samples OB1/OB27 were toxic at 100% sample and hadEC50 values of 71.3 and 65.2, respectively. Samples OB8/OB26 were significantly toxic at a 50% concentration and EC50
values of 35.2 and 35.7, respectively. Samples PT13/PT26were toxic at 100%, but EC50 values. 100 and 68.0% weresignificantly different. The difference only occurred due to adecreased development rate in the 100% sample of PT26,whereas mortality rates in 50% and 25% concentrations werevery similar.
Reference Toxicant Test Results:The EC50 values for thereference toxicant, SDS, were 4.8 (4.4–5.1) in the fertilizationtest, and 3.8 (3.6–4.0) in the embryological development test.These values are within the acceptable limits for SDS toxicitytests, according to our laboratory’s control chart. This indicatesthat the organisms used for these tests were within the usualand acceptable range of sensitivity.
Porewater Quality Measurements:The salinity of the initialporewater samples ranged from 14 to 32%. Only 15 samplesrequired salinity adjustment to satisfy the test salinity require-ment of 306 1%. The original salinity of 10 samples was 32%and was adjusted to 306 1% by addition of purified deionizedwater. The salinity of five samples was between 14 and 28%and was adjusted using a 122% brine made from evaporatingsea water.
The initial DO of only one sample (SQ2) was, 80%saturation (75.1%) but was still considered sufficiently high tobe used in the tests without prior aeration, which might removesome of the toxicants of concern. The pH of all samples rangedfrom 7.37 to 7.78. Total ammonia (NH4) ranged from 0.006(PT1) to 3.5 (SQ1) mg/L, and unionized ammonia (NH3)
302 R. S. Carret al.
ranged from 0.00 to 45.9 (OB15)mg/L. These values are belowthe toxic range for fertilization and embryological developmenttests withA. punctulata(Carr et al. 1996b).
Sulfide concentrations were below detection limit (, 0.005mg/L) in all of the samples except SQ1 and SQ2, both of whichhad sulfide concentrations of 0.011 mg/L. The higher sulfidelevel and low DO in the samples from Sequim Bay suggest ahigh organic load, which could be associated with higher levelsof a variety of contaminants, therefore contributing to the toxiceffects of those samples to the sea urchin embryos. No chem-ical measurements were made with these samples to corrobo-rate this hypothesis.
Sediment Chemistry and Grain Size Distribution:Particlesize distribution varied among samples, from fine sedimentswith only 10–12% sand (station OB16, OB23) to coarse sed-iments with . 80% sand (stations OB6, PT7, PT12, PT16,
PT19). The only ordnance compound in measurable amountswas picric acid, found in concentrations varying from 0.1 to 1.0mg/kg in sediment (dry weight) from stations OB6, OB8,OB12, OB15, OB16, OB18, OB22, OB23, OB26, PT11, PT12,PT16, and PT19. Total butyltins occurred in these same sam-ples, except for PT19, in levels varying between 0.9 and 8mg/kg sediment dry weight.
Concentrations of total PCBs, and selected organochlori-nated compounds, PAHs, and metals were compared to theavailable Sediment Quality Guidelines (SQGs) (Longet al.1995; MacDonaldet al. 1996). Stations exceeding the effectsrange–low (ERL) and threshold effects level (TEL) are pre-sented in Table 1. The stations with the most SQG exceedanceswere OB18 and OB8, with 20 and 18 exceedances, respec-tively. Stations OB22 and OB23 had 17 exceedances each.Station OB22 was the only one with a chemical (phenanthrene)exceeding the probable effects level (PEL). Stations OB15 and
Table 1. Sediment threshold-effects level (TEL), probable effects level (PEL), and the effects range low and median (ERL and ERM, respec-tively) values for key contaminants and stations exceeding those values
Contaminant TEL PEL ERL ERM Stations Exceeding TEL or ER-L
a Above PEL.b Sum of the following low molecular weight PAHs; acenaphthene, acenaphthylene, anthracene, fluorene, 2-methylnaphthalene, naphthalene, andphenanthrenec Sum of the following high molecular weight PAHs; benz(a)anthracene, benzo(a)pyrene, chrysene, dibenzo(a,h,)anthracene, fluoranthene, andpyrened Sum of high and low molecular weight PAHs described above.
Marine Sediment TIE Studies for Ordnance Compounds 303
OB16 had nine SQG exceedances each, and OB12 had four.The stations from Port Hadlock site were less contaminated,with only four SQG exceedances at station PT11, two at PT16,and one at PT19. The high number of SQG exceedances atstations OB8, OB18 and OB22, including an exceedance of thePEL for phenanthrene at station OB22, agrees with the hightoxicity observed for those samples (Figure 1).
TIE Results
Phase I: A summary of chemical measurements made in thefresh and frozen porewater samples for the TIE study, includ-ing the sum of ordnance compounds, butyltins and PCBs, aswell as PAHs and metals that were in measurable amounts, ispresented in Table 2. Concentrations of organic chemicalsvaried only slightly between the frozen and fresh sample, withhigher levels of some PAHs in the frozen pore water, suggest-ing some loss in the fresh sample between the extraction andanalysis times. The concentrations of metals, on the other hand,dropped to approximately half of the initial concentration afterfreezing. The toxicity test results, however, showed that, forthese samples, there were no significant change in toxicity afterfreezing and thawing the sample as has been observed previ-ously (Carr and Chapman 1995). Therefore, frozen sampleswere used for the application of the TIE treatments.
The TIE procedure was applied to full-strength pore wateras well as to samples diluted to 50% and 25%. Results fromthe urchin fertilization test showed that the C18 columntreatment significantly reduced toxicity from the three dilu-tions, providing strong indication that organic compoundswere responsible for some of the observed toxic effects(Table 3). At the 100% concentration, the addition of EDTAalso caused significant reduction of toxicity, suggestingsome contribution of metals as causative agents of toxicity.The increase of pH to 9.0 significantly increased toxicity atthe three concentrations, but no significant reduction oftoxicity was observed with pH reduction. Therefore, ammo-nia was not considered as a toxicant of concern in thosesamples for the fertilization endpoint.
The toxicity of the 100% and 50% concentration samples forthe TIE procedures employing the embryological developmenttest was still very high after all TIE treatments were applied.The 25% concentration samples indicated reduction of toxicitywith several treatments: EDTA, C18 column, sodium thiosul-fate, pH reduction (Table 4). This suggests that several factorswere contributing to the toxicity of that sample, includingmetals, organic compounds, and ammonia. The initial union-ized ammonia concentration in the 100% fresh and frozenporewater samples was 104 and 109mg/L, respectively. TheLOECs for the sea urchin embryological development andfertilization tests for unionized ammonia are 90 and 800mg/L,respectively (Carret al. 1996a). The results of the chemicalanalysis of the sample used for the TIE indicated that parentordnance compounds could not have been responsible for thetoxicity observed in that sample, although previous surveys haddetected low levels of a variety of ordnance compounds insediments from Ostrich Bay and in shellfish tissue samplesfrom the vicinity of the Port Hadlock Naval Facility, WA (URS
Consultants 1995; EA Engineering, Science, and Technology,Inc. 1996).
Several metals and organotin compounds were detected inthe original TIE porewater sample from phase I for which notoxicity data was available for the sea urchin assays from theliterature. A summary of the results of tests performed in ourlaboratory with these contaminants is provided in Table 5. Acomparison of these data with the concentration of these chem-icals in the porewater sample (Table 2) indicates that theycould not have contributed significantly to the observed toxic-ity. None of the other metals or PAHs measured in the pore-water sample were present at concentrations near the knowntoxic concentrations for the sea urchin assays (Carret al.1996b). As the toxicity in phase I tests was reduced by theEDTA and C18 column treatments, it appears that the observedtoxicity in the sea urchin fertilization test, particularly, is dueprimarily to some unmeasured organic or organometallic com-pounds. In addition to these contaminants, ammonia was also acontributing factor to the observed toxicity for the embryolog-ical development test.
The reference toxicant (SDS) test conducted concurrentlywith the fresh and frozen sample comparison, resulted inEC50 values of 4.5 mg/L (4.1–5.0) and 7.1 mg/L (95%confidence limit not reliable) for the fertilization and em-bryological development tests, respectively. This was withinthe acceptable sensitivity range for the fertilization test, butindicated slightly lower sensitivity than usual for the em-bryological development test, for which the upper accept-
Table 2. Summary of chemical measurements in fresh and frozenpore water
a Frozen sample used in the TIE studies.b Not detectable.
304 R. S. Carret al.
able limit for the EC50 would be 6.8 mg/L. However, it wasconsidered acceptable for the purpose of this particular test,which was comparing two samples treated in different ways(fresh versus freezing). The reference toxicant tests con-ducted with SDS concurrently with the TIE proceduresproduced EC50 values of 4.5 mg/L (4.2– 4.8) and 3.3 mg/L(3.2–3.4) for the fertilization and embryological develop-ment tests, respectively. These values indicate that the testorganisms were within the acceptable range of sensitivitybased on the control charts.
Phase II: Because the Phase I TIE provided strong indica-tion that organic compounds were at least partially respon-sible for the observed toxic effects, column fractions be-tween 25% and 100% methanol were eluted from theSpeedDisk for the Ostrich Bay sample and aliquots of eluantwere redissolved in sea water for testing. The 80%, 85%,90%, and 95% methanol fractions, which would containnonpolar organic contaminants with increasing logKow val-ues, were toxic with the 85% and 90% fractions exhibitingthe highest toxicity in the sea urchin fertilization test (Table6). These toxic fractions were analyzed by CAS for the suiteof contaminants measured previously in the original pore-water sample. Apart from a few PAHs, which were observedin the low mg/L range in the concentrated eluant, the onlyother substances above the detection limits were phthalates
Table 3. Toxicity data for sea urchin,A. punctulata, fertilizationtest following phase I TIE procedures
Baseline OBa 100 22.4Baseline OB 50 40.4Baseline OB 25 74.8Baseline OB 12.5 94Baseline OB 6.25 95.2Baseline REFb 100 87.8Baseline REF 50 91.6Baseline REF 25 93Baseline REF 12.5 95.8Baseline REF 6.25 93.8Baseline MFSc 100 89.7Aeration OB 100 16.8Aeration OB 50 27.6Aeration OB 25 73.2Aeration MFS 100 83.8Filtration OB 100 30Filtration OB 50 42.6Filtration OB 25 81Filtration MFS 100 83.4C18 OB 100 56 **C18 OB 50 73.8 **C18 OB 25 89.2 **C18 MFS 100 84.2EDTA OB 100 44.6 *EDTA OB 50 42.6EDTA OB 25 81EDTA MFS 100 93.8Na thiosulfate OB 100 32.8Na thiosulfate OB 50 32Na thiosulfate OB 25 73.8Na thiosulfate MFS 100 91.6pH 7.2 OB 100 25pH 7.2 OB 50 26pH 7.2 OB 25 39.2 **pH 7.2 REF 100 60.2pH 7.2 REF 50 84pH 7.2 REF 25 82.4pH 7.2 MFS 100 85.8pH 8.0 OB 100 23pH 8.0 OB 50 29.2pH 8.0 OB 25 67.2pH 8.0 REF 100 38pH 8.0 REF 50 84.6pH 8.0 REF 25 90.4pH 8.0 MFS 100 92.8pH 9.0 OB 100 0 **pH 9.0 OB 50 0 **pH 9.0 OB 25 13.2 **pH 9.0 REF 100 11.8pH 9.0 REF 50 59.4pH 9.0 REF 25 75.6pH 9.0 MFS 100 89.4
a Pore water from site selected for TIE, from Ostrich Bay.b Reference pore water, from Redfish Bay, Texas.c Millipore filtered sea water.d Significantly different from Ostrich Bay baseline toxicity, *indicatessignificant difference at alpha# 0.05 and **indicates significantdifference at alpha# 0.01.
Table 4. Toxicity data for sea urchin,A. punctulata, embryologicaldevelopment test following phase I TIE procedures
Treatment Sample % Dilution
Mean %NormalPluteus Sig. Diff.d
Baseline OBa 25 7.00Baseline REFb 25 89.40Baseline MFSc 100 89.40Filtration OB 25 13.00Filtration MFS 100 91.80Aeration OB 25 3.80Aeration MFS 100 87.20EDTA OB 25 80.60 **EDTA MFS 100 85.80C18 OB 25 54.40 **C18 MFS 100 90.80Na thiosulfate OB 25 46.00 **Na thiosulfate MFS 100 86.20pH 7.2 OB 25 90.00 **pH 7.2 REF 25 91.80pH 7.2 MFS 100 87.20pH 8.0 OB 25 34.80 **pH 8.0 REF 25 89.00pH 8.0 MFS 100 89.60pH 9.0 OB 25 0.00pH 9.0 REF 25 88.20pH 9.0 MFS 100 91.60
a Pore water from site selected for TIE, from Ostrich Bay.b Reference pore water, from Redfish Bay, Texas.c Millipore filtered sea water.d Significantly different from Ostrich Bay baseline toxicity, *indicatessignificant difference at alpha# 0.05 and **indicates significantdifference at alpha# 0.01.
Marine Sediment TIE Studies for Ordnance Compounds 305
(primarily di-n-butyl phthalate), which most likely leachedfrom the prewashed plastic syringes used in the TIE proce-dure. The highest concentration of phthalates was observedin the 75% methanol fraction (2,000mg/L) which was nottoxic, thereby demonstrating that phthalates were not re-sponsible for the observed toxicity.
Conclusions
Although toxicity, as measured by the sea urchin embryolog-ical development test, which exhibited relatively high sensitiv-ity to several ordnance compounds (Nipperet al. 2001), wasobserved at the majority of stations at both Jackson Park andthe Port Hadlock sites, it is highly unlikely to be caused by theparent ordnance compounds. The only ordnance compounddetected in sediments from the most toxic stations was picricacid, in low concentrations of# 1 mg/kg sediment dry weight,and it was not detected at one of the most toxic stations to seaurchin embryos (PT7). No parent ordnance compounds weredetected in the porewater sample used for the TIE study, which
indicates that explosives of concern in this study were notresponsible for the toxicity observed in this composite samplefrom the most toxic stations. The Phase I TIE proceduresindicated that organic chemicals (e.g., PAHs, PCBs, pesti-cides), and metals to a smaller extent, were the likely causativeagents of the toxicity observed in the sea urchin fertilizationtest. The Phase I TIE procedures indicated that a mixture ofseveral classes of chemicals, including organic chemicals, met-als, and ammonia, were the likely causative agents of the toxiceffect in the sea urchin embryological development test.
Phase II TIE studies indicated that fractions eluted with80–95% methanol were toxic, but no contaminants of concernwere identified in these samples. It is likely that unmeasuredsubstances such as degradation products or metabolites of themeasured analytes were responsible for the toxicity identifiedin the TIE study. The specific contaminants responsible for theobserved toxicity were not identified in this study, but thecompounds in the comprehensive list of analytes which weredetected in the porewater sample were not present at highenough concentrations to be implicated.
Acknowledgments.This study was funded by the Naval FacilitiesEngineering Command through an Economy Act order between theEngineering Field Activity, Northwest (EFA NW), the Naval FacilitiesEngineering Service Center (NFESC), and the U.S. Geological Sur-vey. We are grateful for the expert technical assistance provided byLinda Price-May and for the initial review of the manuscript by Drs.Guilherme Lotufo and Jon Lebo. Sediment chemical measurementswere performed by Columbia Analytical Services under the directionof Richard Craven. We are grateful to Dale Norton of the WashingtonState Dept. of Ecology for arranging for the use of the R/VKittiwakeand to Dale Norton, Cindy O’Hare, EFA NW, and the crew of theKittiwake who assisted with the sediment sampling in Puget Sound.
References
Ankley GT, Schubauer-Berigan MK, Dierkes JR (1991) Sedimenttoxicity identification evaluation Phase I (characterization), PhaseII (identification) and Phase III (confirmation) modifications ofeffluent procedures. National Effluent Toxicity Assessment Cen-ter, Technical Report 08-91, US Environmental ProtectionAgency, Duluth, MN
Burgess RM, Ho KT, Morrison GE, Chapman G, Denton DL (1996)Marine toxicity identification evaluation (TIE). Phase I guidancedocument, EPA/600/R-96/054, US Environmental ProtectionAgency, Washington, DC.
Carr RS, Biedenbach JM (1999) Use of power analysis to developdetectable significance criteria for sea urchin toxicity tests. AquatEcosys Hlth Mangmt 2:413–418
Carr RS, Chapman DC (1992) Comparison of solid-phase and pore-water approaches for assessing the quality of marine and estuarinesediments. Chem Ecol 7:19–30
Carr RS, Chapman DC (1995) Comparison of methods for conductingmarine and estuarine sediment porewater toxicity tests—Extrac-tion, storage, and handling techniques. Arch Environ ContamToxicol 28:69–77
Carr RS, Biedenbach JM, Hooten R (2001) Sediment porewater tox-icity test survey and phase I sediment toxicity identification eval-uation studies in Lavaca Bay, Texas—an estuarine Superfund site.Environ Toxicol 16:20–30
Carr RS, Nipper MG (1999) Development of marine sediment toxicity
Table 6. Toxicity test results for sea urchin (A. punctulata) fertili-zation and embryological development tests with tributyltin, dibutyl-tin, and arsenic
a Results oft tests between column blanks and Ostrich Bay extractions(**indicates significance at 0.01).
306 R. S. Carret al.
data for ordnance compounds and toxicity identification evalua-tion studies at select naval facilities. Final report prepared forNaval Facilities Engineering Service Center, Port Hueneme, CA
Carr RS, Chapman DC, Howard CL, Biedenbach JM (1996a) Sedi-ment quality triad assessment survey of the Galveston Bay, Texassystem. Ecotoxicology 5:341–364
Carr RS, Chapman DC, Presley BJ, Biedenbach JM, Robertson L,Boothe P, Kilada R, Wade T, Montagna P (1996b) Sedimentporewater toxicity assessment studies in the vicinity of offshoreoil and gas production platforms in the Gulf of Mexico. Can J FishAquatic Sci 53:2618–2628
Carr RS, Long ER, Chapman DC, Thursby G, Biedenbach JM, Win-dom HL, Sloane GM, Wolfe DA (1996c) Sediment quality as-sessment studies of Tampa Bay. Florida. Environ Toxicol Chem15:1218–1231
Carr RS, Montagna PA, Biedenbach JM, Kalke R, Kennicutt MC,Hooten R, Cripe G (2000) Impact of storm water outfalls onsediment quality in Corpus Christi Bay, Texas. Environ ToxicolChem 19:561–574
EA Engineering, Science, and Technology, Inc. (1996) Remedialinvestigation— Jackson Park housing complex/Naval HospitalBremerton. Engineering Field Activity Northwest, Contract no.N44255-94-D-7309
Hamilton MA, Russo RC, Thurston RV (1977) Trimmed Spearman-Karber method for estimating median lethal concentrations intoxicity bioassays. Environ Sci Technol 11:714–719; correction12:417 (1978).
Ho KT, McKinney RA, Kuhn A, Pelletier MC, Burgess RM (1997)Identification of acute toxicants in New Bedford harbor sediments.Environ Toxicol Chem 16:551–558
Krone CA, Brown DW, Burrows DG, Bogar RG, Chan S-L, VaranasiU (1989) A method for analysis of butyltin species and measure-ment of butyltins in sediment and English sole livers from PugetSound. Mar Environ Res 27:1–18
Long ER, MacDonald DD, Smith SL, Calder FD (1995) Incidence ofadverse biological effects within ranges of chemical concentra-tions in marine and estuarine sediments. Environ Man 19:81–97
MacDonald DD, Carr RS, Calder FD, Long ER, Ingersoll CG (1996)Development and evaluation of sediment quality guidelines forFlorida coastal waters. Ecotoxicology 5:253–278
Morgan BJT (1992) Analysis of quantal response data. Chapman andHall, London, 511 pp
Nipper M, Carr RS, Biedenbach J, Hooten R, Miller K, Saepoff S(2001) Development of marine toxicity data for ordnance com-pounds. Arch Environ Contam Toxicol (in press).
Nipper M, Carr RS, Biedenbach JM, Hooten RL, Miller K (2002a)Toxicological and chemical assessment of marine sedi-ments spiked with ordnance compounds. Mar Ecol Prog Ser (inpress)
Nipper M, Carr RS, Biedenbach JM, Hooten RL, Miller K (2002b)Marine sediment porewater contamination with explosives: tox-icity, chemistry and influence of degradation products. Mar EcolProg Ser (in press)
PSEP (1986) Recommended protocols for measuring conventionalsediment variables in Puget Sound. Prepared for US EPA, Region10, Office of Puget Sound, Seattle, WA, by Tetra Tech, Inc.,Bellevue, WA
SAS Institute Inc. (1989) SAS/STAT user’s guide, version 6, 4th ed.,vol. 2. SAS Institute Inc., Cary, NC, 846 pp
SAS Institute Inc. (1992) SAS/LAB Software: user’s guide, version 6,1st ed. SAS Institute Inc., Cary, NC, 291 pp
URS Consultants (1995) Final remedial investigation/feasibility studyreport. US Navy CLEAN Contract, Engineering Field ActivityNorthwest, Contract no. N62474-89-D-9295
US EPA (1994) SW-846 Method 8330. Nitroaromatics and nitraminesby high performance liquid chromatography (HPLC), Revision 0.US Environmental Protection Agency, Office of Solid Waste andEmergency Response, Washington, DC
US EPA (1996) Test methods for evaluating solid waste. Physical andchemical methods. US EPA SW-846, 3rd ed., including finalupdate I. US Environmental Protection Agency, Washington, DC
Whitfield M (1974) The hydrolysis of ammonium ions in sea water—atheoretical study. J Mar Biol Assoc UK 54:565–580
Marine Sediment TIE Studies for Ordnance Compounds 307
