Environ. Sci. Technol. 2009, 43, 3815–3823
Field Application of Activated Carbon Amendment for In-Situ Stabilization of Polychlorinated Biphenyls in Marine Sediment YEO-MYOUNG CHO,† UPAL GHOSH,‡ ALAN J. KENNEDY,§ ADAM GROSSMAN,‡ GARY RAY,§ JEANNE E. TOMASZEWSKI,| DENNIS W. SMITHENRY,⊥ TODD S. BRIDGES,§ AND R I C H A R D G . L U T H Y * ,† Department of Civil and Environmental Engineering, Stanford University, 473 Via Ortega, Room 313B Stanford, California 94305-4020, Department of Civil and Environmental Engineering, University of Maryland Baltimore County, Baltimore, Maryland 21250, Environmental Laboratory, U.S. Army Engineer Research and Development Center, 3909 Halls Ferry Rd, EP-R Vicksburg, Mississippi 39180, Swiss Federal Institute of Technology, Institute for Biogeochemistry and Pollutant Dynamics, ETH-Zentrum, CH-8092 Zu ¨rich, Switzerland, and Department of Education, Elmhurst College, 190 Prospect Avenue, Elmhurst, Illinois 60126
Received October 16, 2008. Revised manuscript received April 3, 2009. Accepted April 6, 2009.
We report results on the first field-scale application of activated carbon (AC) amendment to contaminated sediment for in-situ stabilization of polychlorinated biphenyls (PCBs). The test was performed on a tidal mud flat at South Basin, adjacent to the former Hunters Point Naval Shipyard, San Francisco Bay, CA. The major goals of the field study were to (1) assess scale up of the AC mixing technology using two available, large-scale devices, (2) validate the effectiveness of the AC amendment at the field scale, and (3) identify possible adverse effects of the remediation technology. Also, the test allowed comparison among monitoring tools, evaluation of longer-term effectiveness of AC amendment, and identification of field-related factors that confound the performance of in-situ biological assessments. Following background pretreatment measurements, we successfully incorporated AC into sediment to a nominal 30 cm depth during a single mixing event, as confirmed by total organic carbon and black carbon contents in the designated test plots. The measured AC dose averaged 2.0-3.2 wt% and varied depending on sampling locations and mixing equipment. AC amendment did not impact sediment resuspension or PCB release into the water column over the treatment plots, nor adversely impact the existing macro benthic community composition, richness, or diversity. The PCB bioaccumulation in marine clams was reduced when exposed to sediment treated with 2% AC in comparison to the control * Corresponding author phone: 650-723-3921; fax: 650-725-8662; e-mail:
[email protected]. † Stanford University. ‡ University of Maryland. § U.S. Army Engineer Research and Development Center. | Swiss Federal Institute of Technology. ⊥ Elmhurst College. 10.1021/es802931c CCC: $40.75
Published on Web 04/23/2009
2009 American Chemical Society
plot. Field-deployed semi permeable membrane devices and polyethylene devices showed about 50% reduction in PCB uptake in AC-treated sediment and similar reduction in estimated porewater PCB concentration. This reduction was evident even after 13-month post-treatment with then 7 months of continuous exposure, indicating AC treatment efficacy was retained for an extended period. Aqueous equilibrium PCB concentrations and PCB desorption showed an AC-dose response. Fieldexposed AC after 18 months retained a strong stabilization capability to reduce aqueous equilibrium PCB concentrations by about 90%, which also supports the long-term effectiveness of AC in the field. Additional mixing during or after AC deployment, increasing AC dose, reducing AC-particle size, and sequential deployment of AC dose will likely improve ACsediment contact and overall effectiveness. The reductions in PCB availability observed with slow mass transfer under field conditions calls for predictive models to assess the longterm trends in pore-water PCB concentrations and the benefits of alternative in-situ AC application and mixing strategies.
Introduction Sediments accumulate persistent, hydrophobic organic contaminants (HOCs), such as polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs), and dichloro-diphenyl-trichloroethane (DDT). These contaminants pose long-term risks to ecosystems and human health, for which PCBs are the most common chemicals of concern in contaminated sediments in the United States (1). Effective management of PCB-contaminated sediments remains a challenging problem (2). Sediment dredging and disposal are traditional remediation approaches for managing persistent contaminants in aquatic systems. Yet an assessment of data from 26 case studies of dredging projects indicates systematic difficulties in achieving cleanup levels (1). This is often a consequence of unfavorable site conditions and the inevitability of residual contamination and contaminant release during dredging (1, 3). Though dredging will remain an important component of contaminated sediment management, past experience and the complexity and heterogeneity of many sites suggest that a combination of approaches and new technologies are needed to develop economic and effective ways to treat sediment contamination. In-situ activated carbon (AC) amendment to the biologically active layer in sediment has been proposed as an alternative remediation strategy to manage persistent HOCs (3-5). In various laboratory studies (3, 6-8), we demonstrated that addition of AC to sediment reduced the availability of PCBs, PAHs, and DDT to water and uptake by organisms such as clams, amphipods, polychaetes, and mussels. Encouraged by those results, a field study was carried out with a battery of physicochemical tests and biological assessments. We report on the addition of AC to the upper sediment layer at South Basin in San Francisco Bay, CA using available large-scale mixing technologies for in-situ treatment of PCBs to sequester and reduce the availability of PCBs from sediment to biota and the aqueous phase. The sediment at South Basin is cohesive and sediment erosion tests showed that mixing AC into Hunters Point site sediment did not significantly affect the stability of surface sediments as measured by sediment erosion rate or critical shear stress for incipient motion (9). Hydrodynamic modeling showed that surface sediment would not erode under normal, nonstorm conditions and erosion would occur VOL. 43, NO. 10, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3815
FIGURE 1. Schematic of (A) San Francisco Bay; (B) Hunters Point Naval Shipyard and South Basin; and (C) four test plots (D-F) and mixing devices (AEI rotovator and CEI injector). The two plots indicated by shading (Plots D and F) were treated by mixing the sediment with AC to a nominal 30 cm depth. Plot C served as a control plot, and Plot E served as a reference plot. only for short periods of time under infrequent high-wind storm conditions (9). In 2004, we conducted a preliminary field study that showed deployment of mixing equipment for incorporation of AC into sediment and validated new field monitoring tools to measure AC distribution, PCBavailability and resuspension (10). For advancing the technique from the laboratory to field scale application, the current study had several major goals: (1) demonstrate the effectiveness of AC application for two large-scale mixing technologies, (2) further develop field monitoring tools, (3) evaluate benefits of AC treatment in the field, (4) assess possible adverse impacts of AC amendment and the long-term effectiveness of the AC amendment, and (5) identify field conditions that affect field-assessment performance. In this paper, we report results for those goals and considerations.
Materials and Methods Study Design, Site, and Test Plots. The study design comprised four test plots having a surface area of 34.4 m2 each, located within the inter tidal region at South Basin adjacent to the former Naval Shipyard at Hunters Point, San Francisco, CA (Figure 1). The plots were located in the southeast portion of the cove, 23 m away from the shoreline to avoid possible impacts from the former landfill area on the north side of the cove that had been a source of contamination (11). The plots were separated 4-6 m to minimize physical influence between adjacent plots and the shape of the plots were dependent on the type of mixing device employed (Figure 1 C). Plots designated D and F were treated with approximately three dry wt% of activated carbon (AC, Calgon TOG-NDS, 50 × 200). The field equipment was a barge-mounted rotovator system for Plot D for direct mixing of AC into sediment, and a crawler-mounted AC-slurry injector system for Plot F. A target mixing depth of nominally 30 cm was selected, as this comprised the biologically active zone and met regulatory concern to avoid deep mixing. Plot 3816
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 10, 2009
C was mixed by the rotovator system without AC and served as a control plot, whereas Plot E served as a reference plot. The AC was applied in January 2006. Major field assessments were conducted three times: pretreatment assessment (December 2005), 6 month posttreatment assessment (July 2006), and 18 month posttreatment assessment (July 2007) (Supporting Information (SI) Table S2). The assessments comprised (1) total organic carbon (TOC) contents of sediment, (2) sediment PCB levels, (3) aqueous equilibrium PCB concentrations, (4) 28-day in-situ semi permeable membrane device (SPMD) PCB uptake, (5) 28-day in-situ polyethylene device (PED) PCB uptake, (6) 28-day in-situ Macoma nasuta PCB bioaccumulation, (7) indigenous amphipod PCB bioaccumulation, (8) benthic community surveys, (9) PCB desorption, and (10) sediment black carbon (BC) contents. Additional sampling was performed at 24 month post-treatment (January 2008) for the characterization of surficial sediment material and ex-situ M. nasuta PCB bioaccumulation study. Field water samplings over the midsection of the test plots were conducted before, one-tidal cycle after, 6 months after, and 12 months after AC amendment. A time series SPMD uptake experiment was started 13 months after AC amendment (Feb 2007), and continued for 224 days. At each sampling event, different sampling locations were selected through stratified random sampling to avoid the altered area by previous sampling activities. Further information of the study design, field site, and test plots is given in the SI. Sediment Core Sampling and Analysis. Five sediment core samples from each test plot were taken at each sampling time point with a minimum depth of 30 cm. A clean cellulose acetate butyrate core linear (5 cm diameter, 90 cm long, Wildlife Supply Company, Buffalo, NY) was positioned perpendicular to the sediment surface and slowly tapped down into the sediment to minimize the disturbance of the sediment layer. The top of the core linear was capped to form a seal; the core was then slowly retracted and capped
and stored at 4 °C until further processing. The 30 cm long sediment core samples were divided into six cross sections of 5 cm lengths and processed and subsampled as shown in SI Figure S6. Carbon Analyses. To assess the application of AC, two independent analyses were conducted: total organic carbon (TOC) analysis and black carbon (BC) analysis. TOC analysis followed previously published methods (10) and is described in the SI. AC dose was calculated using a relationship between sample TOC and that for AC (TOC ) 86.1%) (10, 12): AC ) (TOC - TOC0) / (86.1 - TOC)
(1)
where AC is the amount of added AC (g/g), TOC is the measured TOC values after AC addition, TOC0 is the measured TOC values for Plot C (control). Black carbon (BC) measurement of sediment samples was performed by a wet chemical oxidation method using sulfuric acid and potassium dichromate (13) in which organic carbon derived from plant and biological matter is chemically oxidized followed by thermal oxidation of the residual BC (see SI). The relationship between BC and AC is AC ) (BC - 0.0022) / 0.7128
(2)
where AC is the amount of added AC (g/g dry sediment), BC is the measured BC (g/g dry sediment), and 0.0022 is the background BC for the control plot calibration curve. PCB Measurements. PCB aqueous equilibrium, desorption kinetics, sediment PCB extraction, and PCB analysis methods are described elsewhere (8, 14-16) and the SI. In-situ 28 Day Clam Bioaccumulation Test. To measure in-situ PCB bioaccumulation, 28 day clam exposure studies were conducted in test plots at three field assessment events: pretreatment, and 6 months and 18 months after AC amendment. The test followed previously developed field methods (10) (SI). After a 28 day exposure, caged clams (Macoma nasuta) were retrieved from the field and transferred to the laboratory for tissue analysis. Ex-situ 28 Day Clam Bioaccumulation Test. For ex-situ bioassays, sediment samples were collected 24 months after AC-treatment from each test plot at five randomly selected sampling locations. Sediment was collected using a stainless steel shovel to a depth of 15 cm. Five sediment samples from each plot were combined and sieved with 4 mm stainless steel mesh screen (OSH, Mountain View, CA) on site to remove large shell and coarse sand material, then the composite sediment was homogenized with an impeller mixer (Lightnin, ND-1A 115V, Rochester, NY) for five minutes to consistent texture. Homogenized sediments were layered into each of five replicate, 20 L aquaria (>4 cm depth) for each test plot (n ) 5) and overlying water (30‰) was gently added using a turbulence reducer and allowed to equilibrate overnight. The remaining sediment was used for chemical assessments. Ten Macoma nasuta were added to each test chamber; clams that failed to burrow after 24 h were replaced. The exposure was conducted for 28 days at 15 ( 1 °C with monitoring of water quality parameters (temperature, pH, D.O., salinity, and ammonia) and 70% water exchanges three times per week. Following the 28 day exposure, the clams were removed from the test sediments and allowed to purge their guts in reference sediment (from the site of clam collection) for a 48 h period, followed by transfer to clean seawater in aquaria for an additional 24 h period. Clams from each replicate were counted for overall survival, shucked, rinsed in deionized water, and frozen at -80 °C. Following homogenization, the tissue was analyzed for PCBs, lipid, and moisture content as described for in-situ bioassays. Clams failing to burrow during the gut-purging period were not included in the analysis.
In-situ SPMD/PED PCB Uptake Test. Five semi permeable membrane devices (SPMDs) were deployed in each plot within 0-15 cm depth for 28-day SPMD exposure following previously described procedures (10). For the time series deployment, six SPMDs were attached to a 10 × 30 cm rectangular frame (SI Figure S8) made of stainless steel tubing and deployed into Plots C and D within a 15 cm depth, 13 months after AC treatment. Two SPMDs from each sampling frame were retrieved 97, 140, and 224 days later. PED sampling using 51 µm low density polyethylene and performance reference compounds are described by Tomaszewski and Luthy (17). Surficial Sediment Assessments. To characterize surficial sediment deposition, 24 months after AC treatment, a top 5 mm surface sediment layer was carefully removed by a stainless steel sampling blade at five randomly selected sampling locations from each test plot, and transferred into precleaned glass jars. The sediment was dried and homogenized by mortar and pestle, and approximately 2 g portions of dry sediment were removed for determination of TOC, BC, and C13 isotope content. The five sediment samples were combined into one composite sample for triplicate PCB measurements and aqueous equilibrium measurements. Carbon-13 isotope signals were measured simultaneously with TOC using an element analyzer coupled with an isotope ratio mass spectrometer (Thermo Finnigan Delta Plus continuous flow stable isotope ratio mass spectrometer, Carlo Erba NA-1500 elemental analyzer). For comparison, the composite sediment for the ex-situ tests was analyzed together. From the homogenate, three analytical replicates for each analysis were sampled. Water Column Sampling and Analysis. To investigate possible PCB release into the overlying water column by either AC amendment or mechanical mixing, overlying water above the four plots was sampled simultaneously before treatment and soon after the high tide covered the plots after treatment with AC. The procedure is adapted from the surface water sampling method used in the U.S. Environmental Protection Agency Lake Michigan Mass Balance Study (18) as described by Cho et al. (10). The inlet of the sampling tube was anchored 15 cm above the sediment surface in the center of the test plot and submerged under water during high tide. The method involves pumping the water through a precombusted glass fiber filter paper with a nominal pore size of 0.7 µm to capture suspended particles, followed by passing the filtered water through a precleaned XAD-2 resin adsorbent column to trap dissolved PCBs. Up to 40 L of water was taken per sample from the field for duplicate analysis. The filters and resins were stored at 4 °C until analysis using previously published methods (10). Benthic Community Assessment. Benthic macroinvertebrates were collected in 0.15 m diameter by 0.15 m deep cores using PVC tubing. Five replicate cores were sampled from each plot at each sampling event and macroinvertebrates were enumerated from 0-5 cm depth to the lowest practical taxonomic designation according to available keys (19, 20). Samples were sieved over a 500 µm sieve using site water; the material retained on the sieve was fixed in 4% formalin containing Rose Bengal stain. After shipping, samples were transferred to 70% ethanol for longer-term storage. Samples from 18-month post-treatment were subsampled (quartered) due to the large number of macroinvertebrates present.
Results and Discussion Sediment Properties of Test Plots. Sediment PCB, TOC, and BC values in the top 15 cm sediment sections of the four test plots are shown in Table 1. Sediment concentrations were corrected by the percent recovery of PCB 65 surrogate to account for the limited extractability for AC-amended VOL. 43, NO. 10, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3817
TABLE 1. Sediment Properties of the Four Test Plots: Sediment PCB Levels (mg PCB/kg Dry Sed), Total Organic Content (% OC/Dry Sed), Black Carbon Content (g BC/g Dry Sed)a plot pretreatment (n ) 5) 6 mo. post-treat. (n ) 5) 18 mo. post-treat (n ) 5) pretreatment (n ) 5) 6 mo. post-treat. (n ) 5) 18 mo. post-treat (n ) 5) 6 mo. post-treat. (n ) 5) 18 mo. post-treat (n ) 5) pretreatment (n ) 1-2b) 6 mo. post-treat. (n ) 1-2b) 18 mo. post-treat (n ) 1-2b) 6 mo. post-treat. (n ) 1-2b) 18 mo. post-treat (n ) 1-2b)
C mix only, control
D rotovator, AC
E reference
sediment PCB level (mg PCB/kg dry sed) 1.35 ( 0.40 1.60 ( 0.72 1.62 ( 1.01 1.88 ( 0.34 1.78 ( 1.07 1.92 ( 0.83 2.32 ( 0.82 1.91 ( 1.10 2.04 ( 0.81 total organic content (% OC/dry sed) 0.44 ( 0.05 0.36 ( 0.06 0.47 ( 0.08 0.77 ( 0.16 3.38 ( 0.74 0.48 ( 0.05 0.63 ( 0.15 2.42 ( 0.59 0.57 ( 0.12 activated carbon (calculated from TOC) (g AC/g dry sed) 0.032 ( 0.010 0.021 ( 0.007 black carbon (g BC/g dry sed)content 0.0014 0.00075 ( 0.00001 0.00077 ( 0.0001 0.0015 0.025 ( 0.0002 0.0015 ( 0.0002 0.0019 ( 0.00002 0.017 ( 0.0001 0.0026 activated carbon (calculated from BC)(g AC/g dry sed) 0.033 ( 0.0003 0.020 ( 0.0002
F injector, AC 1.46 ( 0.37 3.43 ( 1.81 10.45 ( 16.94 0.45 ( 0.20 2.47 ( 1.20 3.31 ( 1.62 0.020 ( 0.011 0.032 ( 0.018 0.0019 ( 0.001 0.010 ( 0.0004 0.023 0.011 ( 0.0006 0.029
a Two sets of estimated activated carbon dose are presented using TOC and BC data. Each data entry represents the mean and one standard deviation. b Analytical replicates from composite sample.
sediments from Plots D and F (21). The total PCB levels among the plots were in the range of 1-2 mg/kg before and after treatment, except for Plot F after treatment. Plot F showed highly variable sediment PCB levels and may have been inherently more heterogeneous than the other three plots, and mixing may have dispersed a region of higher PCB concentration within the sediment of Plot F. After AC amendment, Plots D and F showed significant enhancement of TOC and BC values in the upper 15 cm sediment layer, which confirms successful AC incorporation into the designated plots in the range of 2-3.2% g AC/g dry sediment (Table 1 and SI). Because the two post-treatment assessments utilized different sampling locations, the ACtreated plots showed variable TOC values between the two assessments although the differences are not statistically significant for both plots (Student t test, p > 0.05). TOC data for 5 cm core sections down to 30 cm depth show that the mixing of AC into the plot resulted in variable distribution of the AC (SI). This uneven AC distribution was possibly induced by the unidirectional mixing motion of the mixing devices as well as by insufficient mixing time. In terms of variability, Plot F showed higher variability than Plot D (coefficient of variation), indicating that AC-mixing via the slurry injection device on Plot F was less homogeneous than the rotovator device employed at Plot D. As expected, the unmixed reference Plot E retained similar TOC values throughout the assessment events. The mixing-control Plot C showed an increase of TOC after mixing, which implies an effect of mixing on TOC redistribution within the top 15 cm sediment layer. The amount of deployed AC in the two plots was calculated from averaged post-treatment TOC values, giving 0.026 g/g dry sediment for both Plots D and F. The amount of AC dose estimated from BC values gave similar values as that estimated from TOC data (0.026 g/g and 0.020 g/g for Plots D and F respectively). The AC dose values for Plots D and F are close to but less than the target AC dose of 0.034 g/g dry sediment. The difference is likely due to some overmixing in the vertical (e.g., 30-40 cm) and horizontal dimensions with the large mechanical devices and the relatively small dimensions of the test plots. Also, operational loss during AC application and mixing could be a factor. Aqueous Equilibrium. Figure 2 shows PCB aqueous equilibrium concentrations and corresponding AC contents 3818
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 10, 2009
FIGURE 2. AC dose-response relationship for aqueous equilibrium PCB concentrations normalized by sediment concentration: (a) Plot D and (b) Plot F. Each column and error bar represents the mean and one standard deviation (n ) 5). in the sediment for Plots D and F. The data represent sampling at five different locations on each test plot for pre- and posttreatments. Aqueous equilibrium concentrations were normalized by PCB sediment concentrations to account for the variability of sediment levels. Adding activated carbon, whether by the rotovator in Plot D or the injector in Plot F,
FIGURE 4. Comparison of PCB bioaccumulation for field-deployed M. nasuta for 28 days. (n ) 3-5). The field bioassays at time 18-months post treatment are confounded by shallow burial depth and fresh surface sediment deposition. In each diagram, the bottom whisker is the smallest nonoutlier observation, the top whisker is the largest nonoutlier observation, the black bar represents the median, the lower bound of the box is lower quartile of data (25%), and the upper bound of the box is upper quartile of data (75%). FIGURE 3. PCB desorption for sediment samples collected from two AC-treated plots. The graphs show the fraction of PCBs desorbed for samples with different activated carbon (AC) contents. Each data point represents the mean and one standard deviation (n ) 1-2, analytical replicates). correlated with decreasing aqueous equilibrium PCB concentrations. In Plot D, more than 95% reduction in equilibrium aqueous PCB concentration was obtained with an approximate 3.2% dose of AC. The effect of reducing aqueous concentration was greater for tetra- and penta-chlorinated PCB homologues than for hexa-chlorinated PCBs, which is likely a consequence of slower mass transfer uptake of PCBs by AC for the higher homologues. Plot D showed a clear AC dose-response relationship, which correlates very well with the previous laboratory results with sediment from South Basin in which mixing with AC reduced aqueous equilibrium PCBs level from 44 to 87% as dose was increased from 0.34 to 3.4% (5). Meanwhile, this relationship was not as evident for Plot F, which is attributed to the more heterogeneous sediment properties of Plot F than Plot D in terms of both PCB and AC distributions. Field-deployed AC particles retained their potential to stabilize PCB contaminated sediment for at least up to 18 months after the initial field deployment event. Laboratory studies with sediment have shown similar long-term effectiveness with AC for 18 months of continuous mixing with sediment (22). PCB Desorption. Results of PCB desorption kinetics of sediment obtained before, and 6-18 months after, AC application are shown in Figure 3. The results are presented as fraction PCB desorbed at different times for different doses of AC achieved in the field as measured by black carbon (BC) content. The AC-treated plots (D and F, Figure 3a and b) show decreases in fraction of PCB desorbed with increasing dose of AC. Unlike aqueous equilibrium tests that used five field replicate sediment samples per each plot, the PCB desorption tests were conducted for one composite sample from each plot, so the heterogeneity of Plot F did not affect the results in this case (see SI). The desorption studies support the findings from the other measurements that PCB availability is reduced after AC application in the field.
In-Situ Macoma nasuta PCB Bioaccumulation. Data in Figure 4 show M. nasuta lipid normalized PCB concentrations for 28 day exposures conducted at different times during the study. Baseline lipid normalized PCB tissue residues in M. nasuta indicated no statistically significant differences among the field test plots prior to the addition of AC. Clams deployed six-month post-AC treatment in Plots D and F indicated a trend of lower PCB residues relative to the mixing control plot (Plot C). Clams deployed in the rotovator-mixed ACamended plot (Plot D) showed a 32% reduction (t test, p ) 0.02) at an average AC dose of 3.2% compared to the mixing control plot (Plot C). The injector-mixed AC-amended plot (Plot F) showed a 13% reduction compared to the mixing control plot (Plot C) at an average AC dose of 2.0%, although the p value is not less than the 0.05 alpha level (p ) 0.13). The smaller reduction for Plot F is probably due in part to the inherent heterogeneity of PCBs in Plot F with locally higher sediment PCB levels as discussed previously. PCB homologue-specific uptake values also showed similar reductions (data in SI) with greater benefit of AC amendment evident for less chlorinated PCB homologues, e.g., 50% less biouptake of tetra-PCBs in the rotovator-mixed plot compared to the mixing control. At the 18-month assessment, differences in tissue PCB residues for field-deployed clams were not observed among all plots (t test). To account for this result, we postulate that M. nasuta were exposed to deposited surficial sediment material rather than to underlying sediment that contained the AC treatment. This was a consequence of the clams surface-feeding and not burrowing beyond 5 cm, which was likely due to an observed firmer sediment texture that developed over time. We observed clear signs that sediment deposition had occurred by the 18 month post-treatment assessment, and as indicated also by the deployed clam cages that served as deposition markers for the time period 18-24 months. Prior radioisotope analysis of sediment cores from South Basin had suggested that new sediment might be deposited at a rate of 0.5-0.9 cm/yr (23). Although sediment deposition was not evident during the first six months of the study, by 18 month post-treatment, deposition was visually apparent. Given that M. nasuta can selectively surface VOL. 43, NO. 10, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3819
FIGURE 5. Activated carbon dose-response relationship for clam PCB bioaccumulation. ] McLeod et al. (6) laboratory studies that employed AC-sediment contact on a roller for 1 month (n ) 3-4). 9 Prior NAVFAC field study (10) and rotovator mixing for about half an hour total on the test plot (% difference of lipid normalized BAF compared to control plot) (n ) 3). 2This ESTCP field study based and rotovator mixing for about half an hour total on the test plot (% difference of lipid normalized tissue PCB residue compared to the mixing control plot) (n ) 3-5). *This ESTCP study based on rotovator mixing of field sediment and laboratory bioassay with additional mixing of collected field samples through sieving and 5 min homogenizing (% difference of lipid normalized PCB tissue residue compared to the mixing control plot) (n ) 5). Each point and error bar represents the mean and one standard deviation. deposit-feed by particle size at the mantle cavity and organic carbon enrichment (24-26), the freshly deposited material altered exposure and PCB accumulation. To confirm the surface deposition, we conducted further sampling and a set of additional laboratory tests, as presented in the SI. Collectively, carbon-13 isotope data showed that the origin and/or biological age of the surface sediment was clearly different from that of the 15 cm core composite samples. Also, we did not observe significant enhancement of TOC or BC contents within the surficial deposit for the AC-treated plots (Plots D and F) compared to the control plot. Lastly, there were no significant differences in aqueous equilibrium PCBs for the surficial deposits over the test plots. Through these sets of tests, we found that the reason for the masked biological signal was a combination of two field-related factors: a firmer and coarser sediment texture in the top sediment layer inhibiting clam burial, and in-coming sediment deposition that masked the underlying AC-treatment. These findings are illustrative of field-related factors that confound the performance and interpretation of in-situ biological assessments. Ex-Situ Macoma nasuta PCB Bioaccumulation. The fieldrelated issues were mitigated in the 24 month post-treatment ex-situ M. nasuta bioaccumulation tests because composite sediment samples were collected from the plots, coarse-sieved to achieve predominantly fine-grained test material, and homogenized briefly for five minutes prior to laboratory testing. When exposed in the laboratory for 28 days, M. nasuta survival was high (>82%) and further PCB tissue residues reductions were observed in comparison to the 6 month assessments. The PCB tissue uptake in Plot D was 36% lower than that found in Plot C (Figure 4) with 2% AC dose in the sample composite (t test, p < 0.001). In Plot F, 32% lower PCB tissue uptake was observed (t test, p < 0.001). PCB Bioaccumulation and Mixing Regime. In Figure 5, we compare various AC mixing regimes for efficacy of AC amendment. For laboratory tests with Hunters Point sediment, McLeod et al. (6) reported a significant dose-response relationship for clam bioaccumulation and AC doses ranging 3820
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 10, 2009
FIGURE 6. PCB uptake into field deployed SPMDs (n ) 3-5). In each diagram, the bottom whisker is the smallest nonoutlier observation, the top whisker is the largest nonoutlier observation, the black bar represents the median, the lower bound of the box is lower quartile of data (25%), and the upper bound of the box is upper quartile of data (75%). Open circles represent outliers. from 0.4 to 3.4%. This dose-response relationship is compared with our in-situ and ex-situ clam bioassays for the rotovator-mixed Plot D. Also shown in Figure 5 are field data from Cho et al. (10) for rotovator mixed AC in a test plot located about 8 m further in the mudflat from the current test plots. The data show that one-month mixing in the laboratory, or brief homogenizing (
