Concentrations of polychlorinated biphenyls (PCBs) in water, sediment, and aquatic biota in the Houston Ship Channel, Texas

Available online at www.sciencedirect.com

Chemosphere 70 (2008) 593–606 www.elsevier.com/locate/chemosphere

Concentrations of polychlorinated biphenyls (PCBs) in water, sediment, and aquatic biota in the Houston Ship Channel, Texas Nathan L. Howell a, Monica P. Suarez b, Hanadi S. Rifai a

a,*

, Larry Koenig

c

Civil and Environmental Engineering Department, University of Houston, 4800 Calhoun Road, N107 Engineering Building 1, Houston, TX 77204-4003, USA b Parsons Water and Infrastructure, 3555 Timmons Lane, Suite 1100, Houston, TX 77027, USA c Texas Commission on Environmental Quality, P.O. Box 13087, Austin, TX 78711-3087, USA Received 19 March 2007; received in revised form 28 June 2007; accepted 5 July 2007 Available online 11 September 2007

Abstract Polychlorinated biphenyls (PCBs) were quantified in water, sediment, and catfish and crab tissue collected from the Houston Ship Channel (HSC) in Texas. The total concentrations of the 209 PCB congeners ranged from 0.49 to 12.49 ng l 1, 4.18 to 4601 ng g 1dry wt, 4.13 to 1596 ng g 1 wet wt, and 3.44 to 169 ng g 1 wet wt, in water, sediment, catfish and crab tissue, respectively. All media showed maximum concentrations greater than studies in other regions with the highest concentrations occurring in the most industrialized segments of the channel. Inter-media correlations suggested that sediment is a source to water. Galveston Bay sediment concentrations compared to a previous study showed a declining trend though the rate of the decline may be slowing. Detailed homolog profiles revealed that the industrialized part of the channel may be receiving PCB-laden sediment from its tributaries. An unusually high fraction of the deca-chlorinated congener (PCB-209) was found in all media. Seen in only a few other studies and in previous air concentrations in the channel, this may point to unusual Aroclor mixtures used in the history of the HSC or to contemporary sources from local industry. A comparison of PCB concentrations obtained using Aroclor, representative congener, and all congener methods, indicated that Aroclors are not an appropriate surrogate for total PCBs and that the NOAA NST method is more representative than the NOAA EPA method.  2007 Elsevier Ltd. All rights reserved. Keywords: Congener profile; Fingerprint; Aroclor

1. Introduction Polychlorinated biphenyls (PCBs) are a family of hydrophobic chlorinated compounds characterized by high persistence, toxicity, bioaccumulative properties and widespread distribution in the environment. These compounds were first synthesized in 1929 and used extensively for different industrial applications (e.g., transformers) up until the late 1970s when their manufacture, processing and distribution were banned in the US. However, despite the ban and evidence that shows that the concentrations of these toxics in the environment have declined (USEPA, 1999a), *

Corresponding author. Tel.: +1 713 743 4271; fax: +1 713 743 4260. E-mail address: [email protected] (H.S. Rifai).

0045-6535/$ - see front matter  2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.07.031

they continue to be catalogued as one of the major contaminants in several geographical areas. For example, there were 884 fish advisories in the US for PCBs in 2003 that translated into 2 079 985 lake acres and 133 876 river miles under advisory (USEPA, 2004). In particular, a fish advisory has been in place for the Houston Ship Channel (HSC) since 2001 due to the high levels of PCBs in catfish and blue crabs, and more recently in speckled trout (TDH, 2001; TDSHS, 2005). The HSC is home to a large variety of petrochemical complexes with approximately 40% of the US chemical production and oil refineries concentrated along its banks. An evaluation of USEPA’s toxic release inventory (TRI) data shows a declining trend of PCB releases in the study area of Harris County. The data show that while 13 600 pounds

594

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

of PCBs were disposed/released to the environment in 1989, only 383£ of PCBs were reported for the year 2003. Furthermore, less than 1% of the total releases reported in 2003 corresponded to air and water releases. (USEPA, 2006) However, as will be seen in this paper, PCB levels remain relatively high in fish and crab tissue, sediment and in the water column in the HSC presumably due to continuing and historical sources. This study was initiated to assess the levels, distribution, sources, and fate of PCBs in the HSC. PCB data were gathered in three media (water, sediment, and biota) at three different points in time. Additionally Aroclor and all congener PCB totals were quantified. Spatial analysis using total PCB concentrations and homolog distribution levels was undertaken as was a comparative analysis of Aroclor, representative congener, and total PCB data. 2. Materials and methods 2.1. Sampling sites and sample acquisition Water, sediment, and fish (catfish and crab tissue) samples were collected three times between summer 2002 and spring 2003 (sum total of 53, 98, 84, and 58 samples for water, sediment, catfish, and crab, respectively) from several locations along the channel as shown in Fig. 1. Water samples were collected at surface depths of 0.3 m using a high-volume sampler that yielded sample volumes between 268 and 731 l, thus ensuring adequate detection limits. One

micron Fisher-Robinson combustion-cleaned (450 C) glass fiber filters (GFFs) and stainless-steel columns packed with solvent-rinsed XAD-2 resin were used as collection substrates to retain both the particle-bound and dissolved phases of PCBs, respectively. Both the GFFs and the resin columns were wrapped in heavy-duty aluminum foil, labeled, placed in separate Ziploc bags, and refrigerated (<4 C) in the dark until analysis. Grab sediment samples were collected using either stainless steel Ponar, Eckman, or Peterson dredges. Prior to sample collection, the dredge, stainless steel spoon, and stainless steel bucket were rinsed with DI water, then ambient water. Between three and five grab samples were composited using only the top 5 cm of sediment. The samples were thoroughly homogenized, deposited into a precleaned amber glass jar with a Teflon seal, and kept at 4 C until analysis. Grab samples were not collected from the same area; instead, the dredge was used at various locations around the sampling boat. Fish tissue was collected based on the preferred order: hardhead catfish (Arius felis), blue catfish (Ictalurus furcatus), gafftopsail catfish (Bagre marinus), and channel catfish (Ictalurus punctatus). Tissue samples were obtained by compositing three to five same-species individuals from a given location without regard to gender. A target length of 30 cm was used for hardhead catfish. After collection, the fish were measured, weighed and filleted with a clean stainless steel knife, packed in aluminum foil, placed into bags, and frozen.

Fig. 1. Water, sediment, and fish sampling stations in the HSC.

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

Crab tissue was collected from blue crabs (Callinectes sapidus). A minimum width at the maximum point of 5 in. was targeted for blue crab harvesting. Once a minimum of three crabs were collected, they were measured, weighed and cleaned by removing the carapace and shaking out all internal organs. Then, the target tissue was wrapped in aluminum foil, placed in polyethylene bags, and frozen. 2.2. Analytical methods PCB congeners in water, sediment, fish, and crab were quantified by high-resolution gas chromatography/high resolution mass spectrometry (HRGC/HRMS) using USEPA method 1668A (USEPA, 1999b), which quantifies all 209 congeners. Sediment TOC was determined via USEPA Method 415.2. All of these analyses were completed by a commercial laboratory that met the data quality objectives established for the study. Briefly, GFF, XAD-2 resin, and sediment samples were spiked with a mixture of 13C-labeled PCB compounds (Cambridge Isotope Laboratories, Woburn, MA) and Soxhlet-extracted (after being dried with sodium sulphate) with dichloromethane for a minimum of 16 h and second with toluene. The extracts were subsequently combined. Depending on the appearance of the sample (i.e., color in the sample), back-extraction with potassium hydroxide and sulfuric acid was undertaken and the resulting extract was re-concentrated using rotary evaporation. The extract was subjected to a cleanup procedure using a silica column to make the extract suitable for injection into the GC. The column contained alternating layers of neutral, basic, and acidic silica gel. The column was pre-eluted with hexane, the sample applied, and then PCBs were eluted with hexane. Finally, the sample was concentrated to a final volume of 20 ll. In a similar manner, tissue samples, after homogenization, were spiked with 1 ml of a labeled PCB compound solution and treated with powdered anhydrous sodium sulfate for 12–24 h. The samples were then subjected to Soxhlet extraction with a mixture of methylene chloride:hexane (1:1) for 18–24 h and concentrated to near dryness using a macro-concentration device. Hexane and a labeled cleanup standard spiking solution were added to the solution; the solution was carefully dried, and the lipid content was determined. Subsequently, the lipid content in the sample was removed using a column packed bottom to to top with 2 g silica gel (100–200 mesh), 2 g potassium silicate, 2 g anhydrous sodium sulfate, 10 g acid silica gel (30% w/w), and a second 2 g anhydrous sodium sulfate. The column was pre-eluted with 100 ml of hexane before the sample extract was applied. Then, the extract was concentrated, and cleaned up in the same way as the water and sediment samples. In all of the collected samples, 2 ll of a labeled injection internal standard spiking solution was added to the final sample extract prior to injection. An aliquot of 1–2 l was injected l was injected into the GC using splitless injection.

595

Chromatographic separation of the PCB congeners was achieved using a 30 m · 0.25 mm i.d. (0.25-lm film thickness) Agilent DB-1 column. The GC conditions used were those recommended by the USEPA method: injector temperature, 270 C; initial temperature, 75 C; initial time, 2 min; temperature program, 75–150 C at a rate of 2 C min 1 and then 150–270 C at a rate of 2.5 C min 1; final time, 7 min. Helium was used as the carrier gas at 1.2 ml min 1 at 200 C. Quantitation was then performed on each chromatographic extract using the concentration of the labeled PCB (based on its response factor and the concentration of the internal standard), the volume of the extract, and the original sample weight or volume (depending on media type). 2.3. QA/QC Field blanks were collected with the samples at a frequency of 5% or higher and handled in the same manner as the actual samples. For the surface water samples, the field blank consisted of an uninstalled XAD-2 resin column and a GFF. For the sediment samples, the field blank consisted of clean playground sand (previously baked at 400 C to remove the organics, and then re-wetted). For tissue, the field blank consisted of an off shore, ocean fish species such as tuna purchased from a supermarket. In all of the cases, the blanks were transported to the field, packed and placed in coolers along with the actual samples for further analysis after the sampling. In the lab blanks, PCB congeners 8, 11, 28/20, and 44/47/65 were detected in four procedural water blanks, but these concentrations were below the lab maximum acceptance limit of 0.050 ng/sample and should not invalidate any results. Duplicates were collected during the three sampling events at frequencies >6% for water and sediment samples, and >10% for catfish and crab samples, respectively, to assess precision. The precision was evaluated using the relative percent difference (%RPD) as an indicator of the sum of 18 NOAA congeners listed by the USEPA (2000). The collected samples met the established QA/QC requirements for the study (an allowable %RPD < 50% and recovery range of 25–150% (USEPA, 1999b) for all labeled-PCB analytes). 3. Concentrations in water, sediment, and tissue 3.1. Water Water samples exhibited concentrations (dissolved + suspended phases) for each of the PCB congeners ranging from non-detect to 0.52 ng l 1 and the total concentrations (sum of all congeners) fluctuating between 0.49 and 12.49 ng l 1 with an average value (mean ± SD) of 2.47 ± 2.17 ng l 1 for the three sampling events. The range of PCB concentrations was quite broad compared to US bays, and the maximum concentration in the HSC was also the highest total PCB concentration out of all data

596

Location Lake Michigan, USA Baltimore Harbor, MD, USA Palos Verdes Peninsula, CA, USA New York Harbor, NY, USA San Diego Bay, CA, USA Delaware River, USA Chesapeake Bay, USA Salton Sea Lake, CA, USA Narragansett Bay, RI, USA Hanoi City, Vietnam Venice lagoon, Italy Busan Bay, Korea Bahrain coastal region Rio de la Plata estuary, Argentina Singapore, Southwestern coast Salton Sea Lake, CA, USA Hudson River, NY, USA Hanoi City, Vietnam Los Padres Lake, Argentina Eman River, Sweden Bahrain coastal region

Water (ng l 1)

Sediment (ng g 1 d. wt.)

Fish (ng g

1

w. wt)

Aroclors Sed (·103 ng kg 1) Fish (ng g 1)

0.34–1.74d 0.1–1.52e 0.06–1.14f 6.7–9.4d 0.024–0.419g 1.2–6.5d 8–2150h 116–304i 20.8–1760j 0.74–33.68k 2–2049l 5.71–199m 0.18–7.41n 0.04–98.5o 1.4–329.6p

0.79–45.8a 6–1590a 0.30–12.2a

10.1–114.5i 500–4000d 5.45–58.12b,k 3.5–730.6q 20.6–1,244d 2.39–40.1b,n

6.78–103.79b,c

2.25–45b,c

Year

Water sample collection

Reference

1991 1996–1997 1997 1998 1999

Peristaltic pump (GFF/GLSE) 18-l tanks (GFF/XAD-2) High-Vol (GFF/XAD-2) High-Vol (GFF/XAD-2) High-Vol (GFF/XAD-2) 20-l tanks (GFF/XAD-2)

Pearson et al. (1996) Bamford et al. (2002) Zeng et al. (1999) Totten et al. (2001) Zeng et al. (2002) Rowe et al. (2007)

1996 2000 1997–1998 1997 1996–1998 2000 2000–2001 2002–2003 2003

Ashley and Baker (1999) Sapozhnikova et al. (2004) Hartmann et al. (2004) Nhan et al. (2001) Frignani et al. (2001) Hong et al. (2005) De Mora et al. (2005) Colombo et al. (2005) Wurl and Obbard (2005)

2000 1994– 1995 1997 2000 1991 2000–2001

Sapozhnikova et al. (2004) Ashley et al. (2000) Nhan et al. (2001) Gonza´lez-Sagrario et al. (2002) Bremle et al., (1995) De Mora et al. (2005)

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

Table 1 P PCB concentrations in water, sediment, fish from other studies

Houston Ship Channel, USA NOAA 18 Congeners

0.12–2.61r

0.473–1418r

2.2–866r

ND-120 ND-480

2002–2003

High-Vol (GFF/XAD-2)

This study

Houston Ship Channel, USA 209 Congeners

0.49–12.49d

4.18–4601d

4.13–1596d

ND-120 ND-480

2002–2003

High-Vol (GFF/XAD-2)

This study

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

GFF, glass fiber filters; GLSE, liquid–liquid Goulden large-sample extractor. HSC NOAA 18 concentrations are included to aid comparison. a Aroclor 1254 + Aroclor 1260; ·103 ng kg 1 d. wt. b Muscle; ng g 1 dry weight. c Aroclor 1254 + Aroclor 1260; ND, non-detect. d 209 PCB congeners. e 24 individual PCB congeners + 21 chromatographically unresolved congener groups. f PCB-8, PCB-18, PCB-28, PCB-29, PCB-44, PCB-50, PCB-52, PCB-66, PCB-77, PCB-87, PCB-101, PCB-104, PCB-105, PCB-118, PCB-126, PCB-128, PCB-138, PCB-153, PCB-154, PCB-170, PCB180, PCB-187, PCB-188, PCB-195, PCB-200, PCB-206, PCB-209. g PCB-18, PCB-28, PCB-29, PCB-44, PCB-50, PCB-52, PCB-66, PCB-77, PCB-87, PCB-101, PCB-104, PCB-105, PCB-118, PCB-126, PCB-128, PCB-138, PCB-153, PCB-154, PCB-170, PCB-180, PCB-187, PCB-188, PCB-195, PCB-200, PCB-206, PCB-209. h 113 PCB congeners. i 55 PCB congeners. j PCB-8, PCB-18, PCB-28, PCB-29, PCB-44, PCB-50, PCB-66/95, PCB-87, PCB-101/90, PCB-105, PCB-118, PCB-126, PCB-128, PCB-138/163/164, PCB-170/190, PCB-180, PCB-187/182/159, PCB188, PCB-195, PCB-200, PCB-206, PCB-209. k PCB-44, PCB-49, PCB-52, PCB-101, PCB-105, PCB-118, PCB-128, PCB-138, PCB-149, PCB-153, PCB-170, PCB-180, PCB-200. l PCB-18, PCB-28, PCB-52, PCB-77, PCB-101, PCB-118, PCB-126, PCB-153, PCB-138, PCB-169, PCB-180, PCB-194. m PCB-8, PCB-18, PCB-28, PCB-29, PCB-44, PCB-52, PCB-66, PCB-87, PCB-101, PCB-105, PCB-110, PCB-118, PCB-128, PCB-138, PCB-153, PCB-170, PCB-180, PCB-187, PCB-195, PCB-200, PCB-206, PCB-209. n PCB-44, PCB-49, PCB-52, PCB-87, PCB-101, PCB-105, PCB-118, PCB-128, PCB-138, PCB-149, PCB-153, PCB-170, PCB-180, PCB-187, PCB-201. o 41 PCB congeners. p PCB-18, PCB-28, PCB-31, PCB-33, PCB-44, PCB-49, PCB-53, PCB-70, PCB-74, PCB-82, PCB-87, PCB-95, PCB-99, PCB-101, PCB-105, PCB-118, PCB-128, PCB-132, PCB-138, PCB-153, PCB156, PCB-169, PCB-170, PCB-171, PCB-177, PCB-180, PCB-183, PCB-187, PCB-190, PCB-194, PCB-195, PCB-199, PCB-205, PCB-206, PCB-208, PCB-209. q PCB-18, PCB-28, PCB-31, PCB-44, PCB-47, PCB-49, PCB-52, PCB-87, PCB-90, PCB-101, PCB-105, PCB-110, PCB-118, PCB-138, PCB-149, PCB-151, PCB-153, PCB-156, PCB-170, PCB-180, PCB-194, PCB-195, PCB-199. r 18 NOAA congeners.

597

598

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

surveyed (see Table 1, the listed areas are industrial and similar to HSC). This result stands when considering either the smaller set of NOAA 18 congeners or the complete 209 set. Since the HSC exhibits some of the highest water concentrations nationally, a particularly potent PCB source may exist here. The data indicated that all of the stations along the HSC exceeded the established USEPA water quality standard (using Aroclor analysis) for human health protection of 0.17 ng l 1 (USEPA, 1999a) by 65% or more. Seasonal congener profiles in Fig. 2 show that the tetra-chlorinated congeners dominated the congener profile for Summer and Fall 2002 while the di-chlorinated congeners dominated in Spring 2003. In both the Summer and Spring, there was a quantifiable fraction of deca-chlorinated congeners, which is represented only by PCB-209. The presence of PCB-209 in the profile in such a way as to defy the decreasing trend with increasing chlorination is unusual. The water PCB-congener profile in the HSC differs from that found by Park et al. (2001) in the Galveston Bay area for rain. Their data show the group of the tri-chlorinated congeners as the predominant group in both air and rain. This difference may suggest a low impact to the HSC channel system from atmospheric inputs, and may also suggest the presence of non-atmospheric sources of PCBs. The data exhibited seasonal variation; the average PCB concentrations through summer 2002, fall 2002, and spring 2003 were 1.69, 2.95, and 6.58 ng l 1, respectively (statistically different averages to 95% confidence). The spring 2003 sampling event only sampled four stations which may have skewed this average, but the nearly twofold increase in

water concentrations in the transition from summer to fall could be indicative of a general trend towards increased volatilization in warmer conditions. 3.2. Sediment Total PCB concentrations in sediment samples varied between 4.18 and 4601 ng g 1 dry wt with an average concentration of 168 ng g 1 dry wt and a standard deviation of 598 ng g 1 dry wt. The minimum sediment concentration was consistent with other bays, but the maximum concentration was in the higher grouping of other studies (Table 1). The NOAA 18 is not a perfect comparator to all of the studies shown, but when using it against similar numbers of congeners (12–41) the HSC concentrations are along the same concentration levels as Narragansett Bay and Venice lagoon Amongst all seasons sampled, the tetra, penta, and hexa-chlorinated homologs made almost equal contributions to the observed total PCB concentrations (average contributions of 20%, 21%, and 23%, respectively, see Fig. 2). The fall 2002 homolog profile showed increased levels of hexa and hepta congeners and almost zero deca congeners while Summer 2002 and Spring 2003 averaged out to 11% for PCB-209. Total organic carbon (TOC) ranged from 0.1% to 5.35% with an average of 1.32 ± 0.86%. A positive but weak correlation was found between TOC and total PCBs (p < 0.05; r2 = 0.13, 3 outliers removed), which is not unexpected because the organic compounds in the TOC have a similar hydrophobic nature to PCBs. Additionally a spatial comparison of sediment TOC and

35%

Sediment

Water

35%

Summer

30%

Percent Contribution

Percent Contribution

40%

Fall

25%

Spring

20% 15% 10% 5%

30%

Summer Fall Spring

25% 20% 15% 10% 5%

0%

0% 1

2

3

4

5

6

7

8

9

10

1

2

3

Congener Homolog Group 35%

5

6

7

8

9

10

35%

Fish Summer

30% 25%

Fall 20%

Spring

15% 10% 5% 0%

Crab

Percent Contribution

Percent Contribution

4

Congener Homolog Group

30%

Summer

25%

Fall

20% 15% 10% 5% 0%

1

2

3

4

5

6

7

Congener Homolog Group

8

9

10

1

2

3

4

5

6

7

8

9

10

Congener Homolog Group

Fig. 2. Seasonal variations in spatially averaged homolog profiles (using an all congener PCB addtion method) in water, channel sediment, catfish, and crab tissue samples. (Fall 2002 water PCB-209 erroneous results not included.)

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

PCB showed that TOC was higher in 1007–1006 just as was the case with PCBs. Thus, it is likely that PCB and TOC are both sourced within these segments, and co-locational sourcing of TOC–PCB increases their chance for correlation as shown by Ouyang et al. (2006). The dataset gathered suggested that total PCBs in the sediment in the Galveston Bay have declined from their historical values at some locations. Santschi et al. (2001) found sediment concentrations in 1995 at a site slightly east of Segment 2421 to be 6.8 ng g 1 dry wt while they estimated average concentrations in the 1960s from that area to be 14 ng g 1 dry wt. Segment 2421 (sampled eight years later) showed an average sediment concentration (using three stations closest to Santschi et al.’s sample 13 309, 14 560, and 16 213) of 6.1 ng g 1 dry wt using the 18 representative congeners which Santschi et al. used. This decrease yields an approximate first order rate constant of 0.013 yr 1 for the eight year period compared to the 0.045 yr 1 rate constant from Santschi et al. describing the decline between 1970 and 1990. The PCB sediment concentration dataset over time, however, is fairly limited and does not allow a more thorough analysis. 3.3. Tissue 3.3.1. Catfish Total PCB concentrations in catfish varied from 4.13 to 542 ng g 1 wet wt (176 ± 286 ng g 1 wet wt), and from 16.5 to 1596 ng g 1 wet wt (182 ± 128 ng g 1 wet wt) for inchannel and tributary tissue samples, respectively for the three sampling periods. These observed concentrations are much smaller than those reported in the Hudson River, but they are comparable to Los Padres Lake, Argentina and Eman River, Sweden (see Table 1). Over 95% of the total of collected samples during the three events were above the USEPA screening value of 20 ng g 1 wet wt (USEPA, 2000). Most of the total PCB concentration in catfish tissue was attributed to the hexachlorinated congeners followed by the penta-chlorinated ones, with average contributions of 29% and 24%, respectively, a finding that was similar to the sediment profiles. Fish gender was recorded at the time of collection, but a gender sample was not preserved since fish were composited irrespective of gender. This compositing could create unusual variance in concentration since gender was not considered, but it is not possible to separate out that effect with the data as it stands. 3.3.2. Crab Average total concentrations of PCBs of 29 ± 21 and 21 ± 18 ng g 1 wet wt were quantified for in-channel crab samples in summer 2002 and fall 2002, respectively. It was noted that, on average, the total PCB concentrations in catfish tissue exceeded those found in crab tissue by a factor of approximately six. It was also observed that 57% of the crab stations (Summer 2002 + Fall 2002) exceeded the USEPA screening value of 20 ng g 1. As with

599

catfish and sediment, hexa-chlorinated congeners followed by the penta-chlorinated congeners were the major contributors to the total PCB concentration, with average contributions of 32% and 27%, respectively. 3.3.3. Lipid correlations The total PCB concentrations in catfish and crab samples were correlated to lipid content. Percent lipids averaged 1.6 ± 1.0% and 0.7 ± 0.3% for fish and crab, respectively. No correlation was observed between concentration and lipid content in either species (p > 0.05). Correlation has been found in many other studies (Hope et al., 1998; Gonza´lez-Sagrario et al., 2002; Persson et al., 2007), and so the absence of correlation in this data set is unusual. 4. Analysis and discussion 4.1. Inter-media correlations It has been suggested that the recycling of PCB-contaminated sediment is one of the major potential sources of contamination to aquatic environments (Jeremiason et al., 1998; Zeng et al., 2002). Therefore, the observed PCB concentrations in water (dissolved and suspended phases) were correlated to the sediment concentrations. The analysis revealed a positive correlation between event-averaged water and sediment concentrations (p-value < 0.05, r2 = 0.48, n = 27), suggesting PCBcontaminated sediment as a potential source of PCBs to the water column. Five outliers were removed according to a three times interquartile range outlier test for each media, and this method of outlier detection is used for all regression analyses. As in the water–sediment data analysis, catfish and crab total PCB concentrations were correlated to the total PCB water concentrations. Although the fit in both cases was weak (r2 values of 0.18 and 0.16 for catfish (n = 44, 3 outliers) and crab tissue (n = 41, no outliers), respectively), positive significant correlations were found (p-value < 0.05). The analysis also correlated catfish and crab tissue data with sediment. While both were correlated with sediment (at a = 0.05, r2 = 0.26 with 18 outliers and r2 = 0.22 with 8 outliers for catfish and crab, respectively), catfish had the stronger correlation. 4.2. Dissolved phase–suspended phase partitioning The dissolved–suspended phase concentrations in water in the HSC and tributaries showed that 45 out of a total 53 station–season combinations were higher in the dissolved phase as compared with the suspended phase. Additionally, segments 1006 and 1007 had the greatest dissolved– suspended phase concentration differences (in favor of the dissolved phase). These results suggested that PCBs are sourced by the sediments via suspended particles (because hydrophobic PCBs would favor the suspended phase if

600

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

quantified in the suspended phase). The KD was averaged spatially but separated based on whether it was within the main channel or in one of the side bays on the edges of Segment 1005. Not much differentiation is seen in the average values per congener between the main channel and the side bays. The trend in partitioning increases substantially with increased congener chlorination as is expected. But beyond this mostly unsurprising trend is what one finds in the variability of KD between the side bays and the main channel. The relative standard deviation of KD is much higher within the side bays for the lighter

the system was in equilibrium rather than being kinetically controlled) and that this is occurring most heavily in segments 1006 and 1007. This finding is consistent with results reported in Gevao et al. (1997) and Zeng et al. (1999). Partitioning was further analyzed in this study using the non-equilibrium distribution coefficient (KD = Csusp/Cdiss, l/g dry) because equilibrium concentrations cannot be assumed. Fig. 3 presents a comparison of the KD for 17 select congeners from NOAA National Status and Trends (NST) program (NOAA, 1993) (PCB-187 excluded from the full 18 congeners in NOAA NST because it was not

40.00 In Channel

Distribution Coefficient (Kd, l/g dry)

35.00

Side Bays

30.00

25.00

20.00

15.00

10.00

5.00

12 8/ 16 6 12 9/ 13 8/ 16 PC 3 B 15 3/ 16 8 PC B 17 0 PC B 18 0/ 19 3 PC B 19 5 PC B 20 6 PC B 20 9 PC B

11 8

PC B

10 5

PC B

PC B

1/ 11 3

66

90 /10

PC B

PC B

52 PC B

PC B

44 /4 7/ 65

0

PC B

18 /3

8

PC B

PC B

20 /2 8

0.00

4.50 In Channel

Relative Standard Deviation of Kd

4.00

In Side Bays 3.50 3.00 2.50 2.00 1.50 1.00 0.50

9 20

20

5 19

19

6 PC B

PC B

3 PC B

0 17

18 0/

PC B

16 8 3/ 15

13

8/ PC B

PC B

16 3

6 12 9/

PC B

12

8/

16

11 8 PC B

PC B

5 10

13 /1 01 /1

90 PC B

PC B

66 PC B

52

7/ /4

PC B

65

8 /2

44 PC B

/3

20 PC B

18 PC B

PC B

8

0

0.00

Fig. 3. Suspended–dissolved phase distribution coefficient (Csusp/Cdiss) average values and relative standard deviations in the channel and the side bays for the NOAA NST 18 congener set.

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

4.3. Spatial trends

chlorinated congeners compared to the relative standard deviation found in the channel. At higher chlorination levels, the relative standard deviations between side bays and the main channel are nearly congruent. The increased variability within the side bays could simply result from compositional differences in the texture of the suspended particles found in the side bays as compared to the main channel, or it could be an indication of less equilibrium conditions in partitioning that occurs in the side bays due to different flow and mixing conditions from the side bays. Particle texture information was not assessed at the time of study, but something may at least be learned from the TOC analysis of sediment between the segment 1005 side bay and main channel segments. A comparison of TOC between the two reveals averages (mean ± 95% confidence) of 0.55 ± 0.1% and 1.15 ± 0.1% for the main channel and side bays, respectively. The higher TOC content may help explain the higher concentrations that are seen in the side bays over the main channel 1005, and this distinct TOC difference indicates differently textured particles if the suspended particles are sourced by benthic sediments. Information concerning equilibrium or disequilibrium comes from Park et al. (2001) who showed that PCB deposition as well as net PCB exchange out of Galveston Bay is occurring, and more research in comparing the main channel to the side bays may illustrate how these perturbations are handled in the two different flow regimes.

4.3.1. Total PCBs Fig. 4 depicts total PCB concentrations spatially in all media from the farthest upstream 1007 segment to the farthest downstream 2421 segment. Segments 1007 and 1006 exhibit the greatest total PCB concentrations. The maximum concentrations for all media are in Segment 1007 for all except water where the highest concentration is in Segment 1006. This suggests continuing sources into the water column in these segments or the presence of ‘‘pockets’’ of historical contamination that act as sources to the water column. The two ‘‘hot spot’’ segments (1006 and 1007) contain the greatest density of heavy industry along their lengths including refining, chemical production, petroleum distribution, catalyst processing, and waste treatment and disposal facilities. Hong et al. (2005) and Wurl and Obbard (2005) found higher PCB concentrations near industrialized areas, and thus it is not at all surprising to find that the more highly industrialized Segments 1007 and 1006 contain higher PCB concentrations. The analysis of the spatial distribution of PCBs in sediment and tissue was extended to the tributaries as shown in Fig. 4. The highest observed concentrations in the tributaries for sediment, catfish, and crab were in Segment 1006 at station 11273 (19.5 upstream from Morgan’s Point), the entry point of Patrick Bayou into the HSC. Particular to

350

2421 Side Bays

7

EPA standard

6 5 4 3 2

40

1

-1

Side Bays

250

EPA standard Tributaries

200 150 100

20 ng l

50

-1

0 35

30

25

20

15

10

5

0

-5

-10

-15

-20

0 45

40

35

30

25

1005

2421

3500

Main channel

1000

Side Bays

3000

Tributaries

2500 2000

800 600

1500

400

1000 200

500 0

0 45

40

35

30

25

20

15

10

5

0

-5

15

10

5

0

-5

-10

-15

-20

-10 -15 -20

Distance from Morgan's Point (Km)

140 1007

1005

1006

2421 Main channel Side Bays

120

Tributaries Crab Conc (ng g-1 w wt)

1006

Main Channel and Side Bays Sediment Conc (ng g-1 dry wt)

1200 1007

20

Distance from Morgan's Point (Km)

4000

Conc (ng g-1 dry wt)

300

Main channel

Distance from Morgan's Point (Km)

Tributaries Sediment

2421

1005

60 50

100 EPA standard Tributaries

80

40 30

60 20 ng l

-1

40

20

20

10

0

Main Channel and Side Bays Crab Concentration (ng g-1 w wt)

45

1006

8

Main channel

0.17 ng l

1007

Water Concentration (ng l -1)

1005

1006

Catfish Concentration (ng g-1 w wt)

9 1007

601

0 45

40

35

30

25

20

15

10

5

0

-5 -10 -15 -20

Distance from Morgan's Point (Km)

Fig. 4. Total PCB concentration spatial distribution in water, sediment, catfish, and crab tissue in the HSC main channel, side bays, and tributaries.

602

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

sediment, there were two tributary concentrations above the main channel concentrations (Vince Bayou 31.5 km upstream and Patrick Bayou 19.5 km upstream), which indicated the possibility of tributary PCB sediment transport to the channel. Catfish in general showed high tributary concentrations in Segment 1007 while both the crab and sediment samples did not. This difference in the catfish tributary spatial profile could suggest that the crabs’ pathway of uptake relates more to the tributary sediments as a source while the catfish population may be influenced more by other sources such as lower trophic organisms consumed by the catfish. The total PCB concentrations and their spatial trends all point to higher PCB impact in the upstream channel Segments 1007 and 1006. Thus, these segments were examined more closely at a congener homolog level to determine some linkages in the transport along the channel as well as to the sources that contribute to the PCB concentrations. 4.3.2. Homolog groups Fig. 5 shows the in channel spatial trend of all samples averaged in time moving from segment to segment. The crab and fish profiles both exhibit a trend that leans towards heavier congeners in segment 1007 (more centered around the hexa homolog group) and consistent focus on more of the penta and hexa homologs in the remaining downstream segments. The sediment spatial profile indi-

cates that the PCB heavy Segment 1007 and 1006 have profiles which are quite similar to one another and are more dissimilar from the homolog profiles in Segments 1005 and 2421. Lastly, the relative abundance of the deca-chlorinated congener (PCB-209) exhibits a distinct maximum in Segment 1005 flanked by much lower relative abundances in the neighboring upstream and downstream segments. Tissue does not exhibit the PCB-209 increase in 1005, which may be a result of less efficient transfer of superhydrophobic PCBs in the food web due to restricted membrane permeability (Kannan et al., 1998). Over 90% of the total PCB-209 water concentration was made up of the suspended phase (at all points excepts 3), and so it is likely that bioavailability of PCB-209 may also be hindered by its limited presence in the dissolved phase. Segments 1006 and 1007 were analyzed further because of the distinction in the sediment homolog fingerprints as compared to the other two segments. A homolog profile comparisons between main channel and tributary sediments in 1006 and 1007 was performed. While it cannot be denied that high concentrations exist in 1007, 1006 shows a much closer sediment fingerprint between the tributaries and the channel. Further focus on Segment 1006 from the various bayous that adjoin the channel there show that all three bayous show similar profiles to the overall in-channel 1006 profile. The summed squared difference of the relative percentages for homologs between a tributary and the main channel was calculated as 159, Fish

Water 2421 1006

0.300

1007

0.250 0.200 0.150 2421

0.100

1007

0.250 0.200

1006

0.150

2421

0.100 0.050 0.000

1005 1006 1

2

3

4

2

1007

5

6

7

8

9

Homolog Group

3

4

1007

5

6

7

8

9

Homolog Group

10

Sediment

10

Crab 2421

0.400

0.400

2421

1005

0.350

1007

0.300 0.250 0.200 0.150

2421

0.100

1005

0.050

1006

0.000 1

2

3

4

5

1007 6

Homolog Group

7

8

9

1005

0.350

1006

Relative Fraction

Relative Fraction

1006

0.300

0.050 0.000

1005

0.350

1005

1

2421

0.400

1005

Relative Fraction

Relative Fraction

0.350

1006 1007

0.300 0.250 0.200 0.150

2421

0.100

1005

0.050

1006

0.000 1

2

3

10

Fig. 5. PCB main channel homolog profiles organized by segment.

4

5

1007 6

7

Homolog Group

8

9

10

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

158, and 20 (squared percents) for Carpenter, Patrick, and Greens Bayous respectively. Therefore, Greens Bayou, which is the farthest upstream tributary within 1006 (25.5 km from Morgan’s Point), appears to most closely resemble the fingerprint of the sediment within the channel. The water spatial profile shows that the maximum water total PCB concentration occurs just after this point though the sediment profile actually declines before peaking again just following Greens Bayou. So Greens Bayou sediment could be a sediment transport source to the main channel of segment 1006, but it is possible that other sources may exist from other sediment or from other media entirely as well. The presence of PCB-209 in Segment 1005 as noted previously is an important observation since Segment 1005 is the only segment that contains side bays. A comparison between side bay and main channel profiles for sediment and for water indicated fairly similar profiles for most of the homolog groups, but both media showed a greater relative fraction of PCB-209 in the main channel segment over what is found in the side bays. A comparison of the sediment and the suspended particles profile in the main channel showed that PCB-209 dominated in the sediment phase over the suspended particles phase. It is important to remember that though Segment 1005 exhibits the highest relative fraction of PCB-209, Segment 1006 still contains the highest concentration of that individual congener by an order of magnitude in both water and sediment media. Segments upstream and downstream from 1006 all exhibit a relatively low magnitude of PCB-209 compared with 1006. It is unusual to find PCB-209 in the environment, and many research studies do not give much consideration to PCB-209 presumably because it was not present in the most common Aroclor mixtures (Aroclors 1016, 1242, 1248, 1254, and 1260) that were industrially produced in the US. (Frame et al., 1996) Yet PCB-209 has been reported in prior studies. Kannan et al. (1997) examined a site in coastal Georgia which was known to contain Aroclor 1268, a heavier technical mixture that contains PCB-209. Rowe et al. (2007) discovered PCB-209 in ‘‘unusually high concentrations in the suspended solids’’ in the Delaware river accompanied by significant concentrations of octaand nona-chlorinated congeners and attributed the PCB209 to a nearby titanium dioxide purification plant. Park et al. (2001) found that PCB-209 was a ‘‘dominant’’ homolog which was deposited into Galveston Bay by way of wet and dry particulates, and Ishikawa et al. (2007) confirmed that PCB-209 is generated in thermal refuse treaters. So, if air is assumed to be a source for the HSC, as the PCB-209 deposits in particulates to the surface, those particulates in water still retain much of the congener from that transfer. 4.4. PCB quantification methods There are three main methods used for PCB quantification in the literature and in practice. One method extrapo-

603

lates PCB signals back to Aroclor signatures, another uses representative PCB congeners within homolog groups, and a third sums all 209 PCB congeners. In general, these three methods, in the order they were just given, demonstrate a trend in increasing accuracy and difficulty of quantification. The progression from Aroclors to homologs to total 209 congeners occurred because the analytical resolution increased from the early 1980s up to the present (Shifrin and Toole, 1998). This study compared these different methods using the full 209 congener PCB totals as the standard of comparison. 4.4.1. Aroclor method evaluation Aroclor analysis was conducted concurrently with full congener analysis in the summer and fall of 2002. Results indicated that the majority of the samples collected in the three media (>90%) were non-detect. Aroclors were below the detection limit in all the water samples (n = 36), and Aroclor-1254 was the only species detected in one of the in-channel sediment samples collected (n = 44). For the water samples, it was clear that while the PCB-congener samples exhibited concentrations above the USEPA water standard of 0.17 ng l 1, all of the Aroclors in all of the samples collected were found below the detection limit. This essentially indicated no relationship or correspondence between Aroclor data and total PCB in water and sediment, a finding that has been reported by Frignani et al. (2001). Additionally, the USEPA assessment comparison value of 20 ng g 1 for catfish (n = 34) and crab (n = 40) was exceeded in 44% and 5% of the Aroclor collected samples, respectively, as opposed to the 95% and 43% exceedances obtained using the total PCB data, also indicating that Aroclor data in tissue are not adequate to assess PCB contamination. 4.4.2. Homolog method evaluation Two common homolog addition methods were compared with the complete 209 congener addition. The NOAA NST Program uses a list of 18 congeners that include PCB congeners 8, 18, 28, 44, 52, 66, 101, 105, 118, 128, 138, 153, 170, 180, 187, 195, 206, and 209 (NOAA, 1993). The 18 congeners are summed and subsequently multiplied by a standard factor of two, a practice that yields accurate results. USEPA (2000) uses 18 different congeners from the NOAA NST list that include PCBs 8, 18, 28, 44, 52, 66, 77, 101, 105, 118, 126, 128, 138, 153, 169, 170, 180, and 187. (USEPA cites this list as coming from a later NOAA document. Hence, it is referred to here as ‘‘NOAA EPA’’.) The main differences between the two lists are that the NOAA NST list provides at least one representative congener from every homolog group while the NOAA EPA list includes coplanar PCBs 77 and 126, which are considered more toxicologically potent and ‘‘PCDDlike’’. The NOAA EPA method is used heavily in industry by virtue of its endorsement by the USEPA while at least

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

Water NOAA NST

12

NOAA EPA

10

45 Degree Line

8 6 4 2

Concentration (ng g-dry wt-1)

5000

14

Represenative Congener Total

Represenative Congener Total -1 Concentration (ng l )

604

4000

NOAA NST

3500

NOAA EPA 45 Degree Line

3000 2500 2000 1500 1000 500 0

0 0

2

4

6

8

10

12

0

14

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

All Congener PCB Total (ng g-dry wt-1)

-1

All Congener PCB Total (ng l )

NOAA NST

1400

NOAA EPA

1200

45 Degree Line

Represenative Congener Total

Fish

1600

200 150 100

1000

50

800

0

600

0

50

100

150

200

400 Zoomed Inset

200 0 0

200

400

600

800

1000

1200

1400

1600

1800

Concentration(ng g-wet wt-1)

1800

(ng g-wet wt-1)

Represenative Congener Total Concentration

Sediment

4500

120

Crab NOAA NST

100

NOAA EPA 45 Degree Line

80 60 40 20 0 0

20

-1

All Congener PCB Total (ng g-wet wt )

40

60

80

100

120

All Congener PCB Total (ng g-wet wt -1)

Fig. 6. Comparison between representative congener and all congener methods in a total PCB calculation.

two other studies conducted by Lauenstein et al. (2002) and Fikslin and Santoro (2003) used the NOAA NST list. Fig. 6 compares the accuracy of these two homolog methods against the all congeners standard in all media. As can be seen the NOAA EPA method consistently under predicts the true total PCB concentration in all media. The NOAA NST method provided a much closer fit to the true total PCB concentration, with some under prediction in sediment and some over prediction in tissue (occurs more drastically in crab as compared to fish). Comparison of the sum of squared error (SSE) shows that in water, sediment, catfish, and crab, the NOAA EPA method has greater SSE over NOAA NST by factors of 3, 12, 27, and 2, respectively. Fikslin and Santoro (2003) used similar comparisons to evaluate the accuracy of the NOAA NST list against a sum of 81 congeners standard. They find significant inaccuracies in sediment and biota. A comparison was also made between the all congener, NOAA NST, and NOAA EPA approaches by way of the homolog profiles. These profiles generally showed agreement in the PCB signatures in all media with a few important exceptions. The mono-chlorinated group only appeared in the all congener profile for water and sediment. Moreover in the water samples, the di-chlorinated homolog group percentage was underrepresented by both the NOAA NST (3.1%) and NOAA EPA (3.5%) methods as compared with the all congener method (12.8%). The error in these lighter homologs could be important information to have if evaluating the degradation in higher chlorinated PCBs that occurs in the environment. The other exception particularly important to the HSC is the fact that the NOAA EPA method does not include the deca-chlorinated homolog group made up of PCB-209.

5. Conclusions Overall total PCB concentrations in the Houston Ship Channel are some of the highest or the highest seen in water, sediment, and fish with the greatest concentration ranges when compared to other bays similarly placed in industrial areas. While this may be in part due to the congener addition method used, the concentrations in water and tissue exceeded USEPA guidance levels in a high percentage of samples. Crab and catfish body burdens were more correlated to sediment than water when comparing regression values and homolog profiles. The Galveston Bay outflow portion of the channel had sediment concentrations that showed evidence of decline in time although the rate of decline has slowed from the early 1990s. The highest PCB concentrations were found in the two main channel upstream segments that have the greatest concentration of historical and current industrial facilities. Segment 1006 showed some correlation between its tributary and main channel sediments suggesting the movement of PCB source sediment from specific upstream tributaries into the channel. Seasonal trends were observed in water concentrations showing that volatilization is occurring and that it is occurring in greater measure in warmer seasons. The decachlorinated homolog (PCB-209) was detected in all media inciting further inquiries into its sources to the HSC since PCB-209 was present only in rarer Aroclor technical mixtures (e.g. Aroclor 1268). Rowe et al. (2007) found contemporary effluent sources for this congener in the Delaware River while Park et al. (2001) suggested atmospheric deposition as a possible source in nearby Galveston

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

Bay. Whether any or all of these sources are significant to the channel, the presence and potential use of PCB-209 as a marker in studying PCBs is an important finding yet to be seriously pursued in the literature. Finally, an Aroclor and full congener analysis comparison showed that Aroclor measurements are not effective for assessing total PCB levels in order to determine risk. Furthermore, in analyzing representative congener methods, it was shown that the NOAA EPA method was less accurate at total PCB concentrations over the NOAA NST method. This showed that the choice of which congeners to include is as vital to an environmental assessment as the number of congeners when determining how to choose a representative congener method. Thus far, the literature has not shown much concern for conclusions which could be incorrectly drawn from an incomplete consideration of all PCB congeners. Acknowledgments This research was made possible with funding from the Texas Commission on Environmental Quality (TCEQ), the Texas Advanced Technology Program, and the USEPA. Their support is gratefully acknowledged. References Ashley, J.T.F., Baker, J.E., 1999. Hydrophobic organic contaminants in surficial sediments of Baltimore Harbor: Inventories and sources. Environ. Toxicol. Chem. 18, 838. Ashley, J.T.F., Secor, D.H., Zlokovitz, E., Wales, S.Q., Baker, J.E., 2000. Linking habitat use of Hudson River striped bass to accumulation of polychlorinated biphenyl congeners. Environ. Sci. Technol. 34, 1023– 1029. Bamford, H.A., Ko, F.C., Baker, J.E., 2002. Seasonal and annual air– water exchange of polychlorinated biphenyls across Baltimore Harbor and the northern Chesapeake Bay. Environ. Sci. Technol. 36, 4245. Bremle, G., Okla, L., Larsson, P., 1995. Uptake of PCBs in fish in a contaminated river system. Bioconcentration factors measured in the field. Environ. Sci. Technol. 29, 2010. Colombo, J.C., Cappelletti, N., Barreda, A., Migoya, M.C., Skorupka, C.N., 2005. Vertical fluxes and accumulation of PCBs in coastal sediments of the Rio de la Plata estuary, Argentina. Chemosphere 61, 1345–1357. De Mora, S., Fowler, S.W., Tolosa, I., Villeneuve, J.P., Cattini, C., 2005. Chlorinated hydrocarbons in marine biota and coastal sediments from the Gulf and Gulf of Oman. Mar. Pollut. Bull. 50, 835–849. Fikslin, T.J., Santoro, E.D., 2003. PCB congener distribution in estuarine water, sediment and fish samples: Implications for monitoring programs. Environ. Monit. Assess. 87, 197–212. Frame, G.M., Cochran, J.W., Bowadt, S.S., 1996. Complete PCB congener distributions for 17 aroclor mixtures determined by 3 HRGC systems optimized for comprehensive, quantitative, congener-specific analysis. J. High Resolut. Chromatogr. 19, 657–668. Frignani, M., Bellucci, L.G., Carraro, C., Raccanelli, S., 2001. Polychlorinated biphenyls in sediments of the Venice Lagoon. Chemosphere 43, 567–575. Gevao, B., Hamilton-Taylor, J., Mubdoch, C., Jones, K.C., Kelly, M., Tabner, B.J., 1997. Depositional time trends and remobilization of PCBs in lake sediments. Environ. Sci. Technol. 31, 3274–3280. Gonza´lez-Sagrario, M.D., Miglioranza, K.S.B., de Moreno, J.E.A., Moreno, V.J., Escalante, A.H., 2002. Polychlorinated biphenyls in

605

different trophic levels from a shallow lake in Argentina. Chemosphere 48, 1113–1122. Hartmann, P.C., Quinn, J.G., Cairns, R.W., King, J.W., 2004. Polychlorinated biphenyls in Narragansett Bay surface sediments. Chemosphere 57, 9–20. Hong, S.H., Yim, U.H., Shim, W.J., Oh, J.R., 2005. Congener-specific survey for polychlorinated biphenyls in sediments of industrialized bays in Korea: Regional characteristics and pollution sources. Environ. Sci. Technol. 39, 7380. Hope, B., Scatolini, S., Titus, E., 1998. Bioconcentration of chlorinated biphenyls in biota from the North Pacific Ocean. Chemosphere 36, 1247. Ishikawa, Y., Noma, Y., Yamamoto, T., Mori, Y., Sakai, S.-i., 2007. PCB decomposition and formation in thermal treatment plant equipment. Chemosphere 67, 1383–1393. Jeremiason, J.D., Eisenreich, S.J., Baker, J.E., Eadie, B.J., 1998. PCB decline in settling particles and benthic recycling of PCBs and PAHs in Lake Superior. Environ. Sci. Technol. 32, 3249–3256. Kannan, K., Maruya, K.A., Tanabe, S., 1997. Distribution and characterization of polychlorinated biphenyl congeners in soil and sediments from a superfund site contaminated with aroclor 1268. Environ. Sci. Technol. 31, 1483–1488. Kannan, K., Nakata, H., Stafford, R., Masson, G.R., Tanabe, S., Giesy, J.P., 1998. Bioaccumulation and toxic potential of extremely hydrophobic polychlorinated biphenyl congeners in biota collected at a superfund site contaminated with aroclor 1268. Environ. Sci. Technol. 32, 1214–1221. Lauenstein, G.G., Cantillo, A.Y., O’Connor, T.P., 2002. The status and trends of trace element and organic contaminants in oysters, Crassostrea virginica, in the waters of the Carolinas, USA. Sci. Total Environ. 285, 79–87. Nhan, D.D., Carvalho, F.P., Am, N.M., Tuan, N.Q., Yen, N.T.H., Villeneuve, J.P., Cattini, C., 2001. Chlorinated pesticides and PCBs in sediments and molluscs from freshwater canals in the Hanoi region. Environ. Pollut. 112, 311–320. NOAA, 1993. Sampling and analytical methods of the national status and trends program national benthic surveillance and mussel watch projects 1984–1992 volume I: overview and summary of methods. In: Lauenstein, G.G., Cantillo, A.Y. (Eds.), National Oceanic and Atmospheric Administration (NOAA). Silver Spring, MD, p. 157. Ouyang, Y., Zhang, J.E., Ou, L.T., 2006. Temporal and spatial distributions of sediment total organic carbon in an estuary river. J. Environ. Qual. 35, 93–100. Park, J.S., Wade, T.L., Sweet, S., 2001. Atmospheric deposition of organochlorine contaminants to Galveston Bay, Texas. Atmos. Environ. 35, 3315–3324. Pearson, R.F., Hornbuckle, K.C., Eisenreich, S.J., Swackhamer, D.L., 1996. PCBs in Lake Michigan water revisited. Environ. Sci. Technol. 30, 1429–1436. Persson, M.E., Larsson, P., Holmqvist, N., Stenroth, P., 2007. Large variation in lipid content, Sigma; PCB and delta; 13C within individual Atlantic salmon (Salmo salar). Environ. Pollut. 145, 131–137. Rowe, A.A., Totten, L.A., Xie, M., Fikslin, T.J., Eisenreich, S.J., 2007. Air–water exchange of polychlorinated biphenyls in the delaware river. Environ. Sci. Technol. 41, 1152–1158. Santschi, P.H., Presley, B.J., Wade, T.L., Garcia-Romero, B., Baskaran, M., 2001. Historical contamination of PAHs, PCBs, DDTs, and heavy metals in Mississippi River Delta, Galveston Bay and Tampa Bay sediment cores. Mar. Environ. Res. 52, 51–79. Sapozhnikova, Y., Bawardi, O., Schlenk, D., 2004. Pesticides and PCBs in sediments and fish from the Salton Sea, California, USA. Chemosphere 55, 797–809. Shifrin, N.S., Toole, A.P., 1998. Historical perspective on PCBs. Environ. Eng. Sci. 15, 247. TDH, 2001. News release: TDH study shows fish, crabs from most of Galveston Bay safe to eat. Texas Department of Health (TDH), Austin, TX.

606

N.L. Howell et al. / Chemosphere 70 (2008) 593–606

TDSHS, 2005. News: Speckled Trout Consumption Limit Advised. Texas Department of State Health Services, Austin, TX. Totten, L.A., Brunciak, P.A., Gigliotti, C.L., Dachs, J., Glenn, T.R., Nelson, E.D., Eisenreich, S.J., 2001. Dynamic air–water exchange of polychlorinated biphenyls in the New York–New Jersey Harbor Estuary. Environ. Sci. Technol. 35, 3834–3840. USEPA, 1999a. Fact Sheet: Polychlorinated Biphenyls (PCBs) UpdateImpact on Fish Advisories. USEPA, Office of Water, p. 7. USEPA, 1999b. Method 1668, Revision A: Chlorinated Biphenyl Congeners in Water, Soil, Sediment, and Tissue by HRGC/HRMS. USEPA National Office, Washington, DC, p. 133. USEPA, 2000. Guidance for Assessing Chemical Contaminant Data for Use In Fish Advisories. USEPA.

USEPA, 2004. Fact Sheet: National Listing of Fish Advisories. Office of Water. USEPA, 2006. Envirofacts Data Warehouse. Washington, DC. Wurl, O., Obbard, J.P., 2005. Organochlorine pesticides, polychlorinated biphenyls and polybrominated diphenyl ethers in Singapore’s coastal marine sediments. Chemosphere 58, 925–933. Zeng, E.Y., Peng, J., Tsukada, D., Ku, T.L., 2002. In situ measurements of polychlorinated biphenyls in the waters of San Diego Bay, California. Environ. Sci. Technol. 36, 4975–4980. Zeng, E.Y., Yu, C.C., Tran, K., 1999. In situ measurements of chlorinated hydrocarbons in the water column off the Palos Verdes Peninsula, California. Environ. Sci. Technol. 33, 392–398.