Free Appendix - District Court of Delaware - Delaware

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 1 of 40

concerning PAH output from the Motiva outfalls and various other sites indicated that PAH contamination was higher in the particulate phase, with higher concentrations at the 601 outfall (treatment plant) than the 001 outfall or the cooling water intake canal. Most of the analyses were based on concentrations rather than PAH loadings. In fact, loading data from the Motiva refinery were generally ignored. This represented a basic misunderstanding of how PAH contamination of the receiving river was affected by the refinery. Station concentrations of PAHs were mainly in the particulate phase (68%). This, together with high PAH concentrations in the sediments in depositional areas, indicated that deposition of PAH-laden particulate matter was the mechanism for PAH distribution in sediments of the study area, and that the refinery could have been a source of such PAHs. Statements concerning bioavailability were not supported by experimental data. NOAA ERLs and ERMs were exceeded for various PAH indicators in depositional areas where experimental data indicated toxicity. These data should have prompted a detailed study of such areas. They did not. The bivalve studies and long core work were carried out at stations distant from the depositional areas. According to the report, the Triad results indicated multiple stressors at various stations that included PCBs, Hg, Zn, pesticides, and different combinations of PAHs. Instead of using loading of individual PAHs, the PAH data were re-organized into various groups which compounded problems associated with identification of actual PAH sources to the river. Triad data were grouped in ways that masked identification of causative effects. The experimental results were then analyzed in only the most restrictive of ways (ANOVA analyses). As was true of most of the analyses, the actual distribution of the Triad results in space and time was ignored in favor of artificial grouping of the data that then formed the basis for a series of statistical applications. According to the chemical fingerprinting of PAHs, refinery effluents were "dominated by petrogenic four-ring PAHs (Fluoranthenes/Pyrenes with two [FP2] and three [FP3] alkyl groups and Benz(a) anthracene/Chrysenes with two [BC2], three [BC3] and four [BC4] alkyl groups). In the authors' view, this pattern "exhibited little variability over time relative to the background patterns in the Delaware River." This conclusion was

A00039

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 2 of 40

not backed up by the loading data or the sediment PAH information. According to the Report, water and sediments were dominated by urban background signatures (two-three ring PAHs [fuel and bilge tank discharges]), pyrogenic four- to six-ring PAHS (combusted sources), and Perylene (diagenic product of plant decomposition). Based on the PAH signature analyses, refinery sediment PAHs were only found at moderate to low levels at a limited number of stations around the refinery outfall, and PAH concentrations in river stations "consistently showed Delaware River urban background signature, and exhibited little or no contribution from the refinery." The authors used the results from the longcore work to support their conclusions, but the long-core work was based on only two sample cores in areas of low PAH sediment concentrations. According to the Motiva Report, the refinery removes most of the PAHs, especially the two-three-ring PAHs, and the unimportant contribution of the four-ring PAHs was attributed to dilution. that dilution reduces the loading of PAHs to the river by the refinery, which is false. The bivalve studies in no way followed the experimental design for the only bioavailability part of the study. The work done with clams was limited to the sampling of existing bivalves in areas other than the Triad stations at times other than the Triad work. Sampling was restricted to 2001. The clam results were supposed to show that this population was able "to survive, grow, and reproduce in areas where the animals have been exposed to PAHs and other chemical stressors." The observational data indicated PAHs associated with background sources and not the refinery. However, since no bioavailability data were taken in areas sampled for the Triad analyses, the attribution of cause-and-effect to different contaminants was not possible. This was a major concern when the authors of the study dropped any chemical screening of the animals taken in the Triad analyses. The elimination of the clam bioavailability study thus compromised any cause-and-effect attribution, and represented a major problem with the study. As projected by our agreement during the negotiations concerning the study design, the lack of bioavailability data would preclude attribution of effects due to contaminants other than PAHs, an agreement that was violated by the study authors. This repeats a basic fallacy associated with the emphasis on concentrations rather than loading:

A00040

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 3 of 40

The "weight of evidence" analyses, based on the authors' flawed approach, found that there was no evidence that the refinery's historical excess pollutant discharges contributed to contamination of river sediments, and that there was no evidence of geographic contaminant patterns related to Motiva exceedances even though there were significant excessive discharges noted during the early part of the four-year study. Chronic sediment toxicity was attributed to PAHs, PCBs, and metals (Cu, Zn, As, Pb, Hg) and not to Motiva effluent based on statistical analyses of the data. conclusions did not follow from the data generated in the study. The authors'

IV.
A. Introduction

INDEPENDENT ANALYSIS OF MOTIVA DATA

Due to problems noted in the previous section (III), I conducted an independent analysis of the data generated by the Motiva study. Instead of using PAH concentrations as the principal data for analyses, I used loading data for individual PAHs, which is more representative of the potential influences of the refinery on the river. pollutants such as PAHs. It should be emphasized that mere dilution does not mitigate or even influence the actual loading of Attachment of PAHs to particulates would invalidate the identification of refinery sediment PAHs in places proximate to outfall areas, since it is likely that these particulates would be transported to depositional areas of the river rather than to areas of low deposition around the outfalls. The relationship of loading to sediment PAH concentrations and associated biological impacts was explored with actual sediment PAH compounds rather than the groupings of such compounds that were used by the Motiva researchers. The lack of any realistic determination of spatial and temporal aspects of such contamination by grouping the data prior to the statistical tests was addressed in this evaluation of the data. B. Methods and Materials The first step in the analysis of the Motiva data was to organize spreadsheets of the individual study components, and determine the spatial and temporal distributions of the data. The 601 and 001 loadings by PAH type were calculated by month over the study period. Since the data taken did not allow a review of time-based deposition rates of the

A00041

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 4 of 40

PAHs, the PAH loadings were averaged over various periods so that the temporal aspects of loadings vs. sediment concentrations could be evaluated more realistically. These data were compared to sediment PAH analyses made in 8/00 (loadings averaged from 3/998/00), 4/01 (loadings averaged from 1/01-4/01), 8/01 (loadings averaged from 5/01-8/01), 4/02 (loadings averaged from 1/02-4/02), and 8/02 (loadings averaged from 5/02-8/02). This re-organization was based, in part, on the assumption that sediment PAH concentrations were influenced to a higher degree by more recent loadings. This approach did not preclude effects due to loading from the more distant past, which included the pollutant discharges during the early months of the study. A series of statistical methods was used to analyze the data. Scatter-grams of the long-term field data were examined and either logarithmic or square root transformations were made, where necessary, to approximate the best fit for a normalized distribution. These transformations were used in all statistical tests of significance. A cross-correlation analysis was then carried out using the sediment PAH concentrations and the results of the Triad analyses. This analysis was backed up by regressions of various combinations of the database. The data were also organized for a series of PCA and associated PCAregression determinations using biological factors as the dependent variables. The high correlations disallowed any real determinations of the PCA-regression analysis beyond what was noted using the correlation and regression analyses. This process was not considered to be complete in terms of more advanced statistical determinations that are often used for publication purposes. It was designed to utilize the more realistic inclusion of primary data rather than the arbitrarily grouped analyses in the Motiva Report. This analysis was devised to determine if there was any relationship among PAHs derived from refinery loadings, observed sediment PAHs, and the Triad results. The approach used thus compensated for significant weaknesses in the Motiva Study data analyses as outlined above. C. Results and Discussion 1. River Background: Relationship to Refinery Effluent Loading

A00042

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 5 of 40

According to the Motiva Report, the expected distribution of the refinery effluent plume was located on the Delaware side of the river, with maximum coverage of about three miles upstream and seven miles downstream depending on the tidal stage (Figure 1). The distribution of simulated time-and-depth averaged current speeds is shown in Figure 2. The accumulation of PAHs was most likely in areas of weak circulation (Hamburg Cove, Goose Island Flats, Salem Cove, Reedy Island Bar). Based on the dye results, Hamburg Cove and the Reedy Island Bar were thus the most likely places to find settlement of particulate-bound PAHs that were loaded from the refinery into the river. Preliminary studies were used to determine the distribution of the main stations for the Triad analysis (Figure 3). The distribution of % TOC (Figure 4) indicated the most consistent area of accumulation in the Reedy Island Bar part of the river. The distribution of sand in the area (Figure 5) supported the above determinations, with less sand and higher silt-clay fractions in depositional areas associated with the Reedy Island Bar area. The highest sand deposition was in the immediate receiving area from the effluent canal, indicating that areas around the outfall were not necessarily representative of potential depositional areas where PAH concentrations would most likely be high. It has been well established in the scientific literature that specific components of total PAHs (TPAHs) are attached to particulates and usually concentrate in areas of deposition of the particulate matter. This finding was also made in the Motiva Study. Water concentrations of PAHs are rarely indicative of PAH distributions because of the ephemeral nature of PAH concentrations due to various factors such as loading variability, complex changes of currents, dynamic processes associated with resuspension of sediment PAH concentrations, and biological factors that include microbial activity. The basis for the Triad approach rests on the nature of PAH distribution as described above. The distribution of TPAH and Environmental Protection Agency (EPA) Priority Pollutant PAH (EPAH) are shown in Figure 6. The deposition of PAHs was generally consistent with the results of the dye and water current studies and the distributions of % TOC and % sand. There was some temporal variation in the PAH loading. The highest sediment PAH concentrations were found at station 56 off the effluent canal, behind Pea Patch Island, and around the Reedy Island Bar. The other depositional area, Hamburg

A00043

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 6 of 40

Cove, was not characterized by high concentrations of PAHs. The simplest explanation for this lack of PAHs in a depositional area upstream of the discharge canal is that there is, in big rivers such as the Delaware, a general downstream drift of particulates due to water currents. PAH concentrations in water across all stations were 57% particulate, 14% colloidal, and 37% dissolved phase. High molecular weight PAHs were 96% particulate and low molecular weight PAHs (two- and three-ring) were 47% particulate. Since total PAHs occurred largely in the particulate phase, this form of PAH transport should be considered when determining the distribution of refinery-based PAHs. Sediment concentrations of PAHs located in immediate receiving areas of the effluent canal were generally low compared to the sediment burdens of PAHs in areas south of the refinery effluent discharge point. The data indicated that PAHs attached to particulates were blown out of the effluent canal, and, due to the current structure and the nature of the downward drift of river-borne particulates, were concentrated in depositional areas south of the discharge canal. The major emphasis of the Motiva Study on stations offshore of the effluent canal for PAH fingerprinting based on water concentrations and sediment distributions was thus erroneous, and was not consistent with the early findings of the study. The sediment PAH concentrations were generally low at upriver stations (DR45, DR10, DR9B). If, as claimed by the authors of the Motiva study, the Delaware River (not the refinery) was responsible for the deposition of PAHs in sediments of the study area, why were there low concentrations of PAHs in areas proximal to the influence of the refinery? The significance of this finding would invalidate the primary conclusions of the Motiva Study. The accumulation, over time, of PAHs in areas of deposition is more consistent with the data than the hypothesized importance of river-associated PAHs. Such accumulation is also more consistent with what is known about PAH distribution in receiving areas such as the Delaware River. The bottom line is that the authors simply ignored their own data in the design and completion of the study. 2. Temporal PAH Loading Trends One important factor in the determination of the potential impact of refinery effluents on the Delaware River-estuary is related to the loading rates of PAHs. PAH

A00044

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 7 of 40

concentrations in water are not relevant since dilution of the effluent (as noted in the 001 discharges) has no effect on the loading rates per se. This differentiation was not made in the Motiva Report, which generally ignored loadings in favor of concentrations in many of the analyses such as fingerprinting. The loading of individual PAH compounds was not even calculated by the authors of the Report. Loading trends of PAHs from the 001 and 601 outfalls are shown in Figure 7. It should be noted that the loading rates represented samples taken only one day per month, and, since we had no daily flow rates, the actual monthly loadings were, on average, based on the single-day samples. With the exception of the March 1999 loading, the 001 TPAH discharge was usually greater than that of the 601 discharge. High total TPAH loadings (i.e., exceedances) were noted during March 1999, June 1999, September 1999, January 2000, and March 2000. With the advent of the 2001-2002 Triad study, refinery loadings were lower than the previous two years with peaks noted during spring 2001 and spring 2002. A comparison of the some of the dominant PAH compounds in the effluent loading from 001 and 601 (Figure 8) indicated that different combinations of PAHs were represented during peak PAH loading events. This contradicted Report statements of the temporal stability of the PAH contributions from the refinery. With the exception of Naphthalene, there was a general downward trend of the individual PAHs, with a peak noted for several PAH types during March 2002. However, these data indicate that there were individual loading trends for the various PAH types, and the averaging of the data, as carried out in the Motiva Report, did not address the qualitative and quantitative changes of PAH loading during the specific Triad sampling periods. A comparison of loading trends of TPAHs and EPAHs from outfall 001 with concentrations of TPAHs and EPAHs averaged over the Triad stations (Figure 3) during the individual Triad samplings from 8/00-8/02 is given in Figure 9. The trends for both EPAHs and TPAHs in sediments were similar in that there was a general downward trend from August 2000 through April 2002. Despite the unexplained increase in most of the sediment PAHs during the August 2002 Triad sampling, the trend of the sediment PAHs in the study area followed the downward trend of the 001 PAH loadings. This temporal trend of loadings from the refinery and sediment PAH concentrations indicated that there

A00045

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 8 of 40

was a connection between refinery loading and sediment PAH concentrations. These trends were largely ignored in the Motiva Report. Thus, contrary to the conclusions of the Report, sediment PAH concentrations in areas expected to be affected by the refinery followed PAH discharges from the refinery in time. A comparison of average PAH loadings over the study period by compound in outfalls 001 and 601 is shown in Table 1. Eleven of the top 15 PAHs loaded to the river by the refinery were common to both outfalls (i.e., were refinery derived). Dominant (top 15) compounds that were common to both the 001 and 601 outfalls (i.e., refinery effluents) were equivalent and were close to the totals of the average loading of the two outfalls. The top PAHs that were loaded by both outfalls were refinery-associated C3- and C2 Fluoranthenes/Pyrenes compounds, and included a number of PAHs that were not attributed to the refinery by the Motiva Report. Contrary to the Report, average loading from 001 was comparable to that of 601, and such discharges were dominated by compounds known to be related to refinery loadings (not river contaminants). According to the Motiva Report, refinery signatures had high levels of Fluoranthenes/Pyrenes (2 (FP2) and 3 (FP3) alkyl groups and Benz(a)anthracene/Chrysenes with 2 (BC2) and 3 (BC3) and 4 (BC4) alkyl groups). Fluoranthene, Benz(a)anthracene and Perylene were not in the wastewater treatment plant effluent and were used to differentiate river background water. However, this finding was an over-simplification in that various other compounds such as Phenanthenes/Pyrenes, Dibenzothiophenes, Naphthalenes, and Benzo(e)pyrene were commonly loaded by the refinery. When temporal periodicity was taken into account (Tables 2 and 3), compounds such as Benzo(a)pyrene, Benzo(a)anthracene, and Benzo(b)fluoranthene were also loaded into the river from 601. During exceedances, many of the most important PAH compounds were common to both outfalls (Table 2). During these periods, the loadings from 601 were considerably higher than those from 001. There was a major reduction of loading during the period of Triad analyses (Table 3). Data were averaged over fourmonth periods during the Triad analyses. When viewed as a function of time, there were major differences in the types of PAH compounds that dominated the loading from both outfalls. Dominance varied from the Fluoranthenes/Pyrenes (1999-spring 2000) to

A00046

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 9 of 40

Naphthalenes in 2001-2002. Once again, the differences between outfalls 001 and 601 in terms of the PAH compounds that dominated loading from the refinery were minimal. These changes in loading patterns were not addressed in the Motiva Report, even though such changes could have affected the biological response of the system. There was no resolution of the time factor in the loading of PAHs from the refinery relative to the sediment concentrations of such compounds. This problem, together with the simplistic way of dealing with the individual PAH compounds, contributed to the erroneous conclusions concerning the relationship of riverine PAH distributions relative to those associated with the refinery. The authors of the Motiva Report made the error of equating bioconcentration with bioavailability. Bioavailability has to do with which compounds in the sediments eventually get into associated food webs, and this can only be found through experiments and tissue concentrations of caged biota in depositional areas that are high in sediment PAHs. The importance of bioavailability has to do with whether or not contaminated sediments should be removed during restoration activities. Bioconcentration refers to concentrations found in animals in the study area. A comparison of bivalve tissue PAH concentrations and mean PAH loading at outfalls 001 and 601 (Table 4) indicates that various compounds loaded from the refinery were in the bivalve tissues. The C1, C2, and C3 Fluoranthenes/Pyrenes and C4 Phenanthrenes/Anthracenes that were loaded at outfall 601 were found at relatively high concentrations in the sediments and in bivalve tissues. The C4 Dibenzothiophenes and C4 Naphthalenes that were loaded at outfall 601 were not found in the sediments but were found in bivalve tissues. This may indicate different modes of bioconcentration (sediments and water), but this possibility can only be determined through experimental bioavailability experiments. The C2-C3 Phenanthrenes/Anthracenes and Pyrene were not high in refinery loadings and were concentrated in sediments and bivalve tissues. These compounds were from sources other than the refinery. Bivalve tissue PAHs were thus represented by a combination of refinery PAHs and PAHs from other sources. However, these numbers may not be representative of PAH tissue concentrations, because caged bivalve experiments were not carried out in

A00047

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 10 of 40

depositional areas characterized by high sediment PAHs. This uncertainty derives from the faulty choice of stations for the bivalve work. 3. Triad Results a. Sediment Quality Guidelines (NOAA) The National Oceanic and Atmospheric Administration (NOAA) established a series of Sediment Quality Guidelines that were designed to predict the concentrations of toxic agents that were of potential ecological concern (Long and Morgan, 1990; Long and Macdonald, 1998; Long et al., 1995) These guidelines were based on correlations between toxicity data and measured concentrations of toxic compounds from extensive data sets. Long et al. (1995) found that cumulative effects of mixtures of toxicants in sediments may drive the apparent effective concentrations of individual toxicants downward toward lower concentrations (i.e., possible synergistic effects). This conclusion does not preclude adverse effects from a single toxicant, especially one with the potential to cause adverse chronic toxic effects. Sediment concentrations of PAHs were compared to the 10% threshold of effects (Effects Range Low: ERL) and the 50% probability (Effects Range Median: ERM). The lack of bioavailability data for the various toxic agents in the Triad sediments precluded exact associations of sediment concentrations with toxicity. Results of the application of the NOAA guidelines to the PAH concentrations found in the sediments during the Triad samplings are shown in Table 5 and Figure 10. The low and high molecular weight PAHs almost uniformly exceeded the ERL limits for sediment toxicity. The ERM limits were exceeded mainly for the low molecular weight PAHs at stations DR53, DR56, DR67, DR68, and DR83. These data indicate that most of the stations were affected by the PAH levels in the sediments, regardless of the season or year of sampling. These effects included the so-called reference stations (DR10, DR9B). The low molecular weight PAHs were more closely associated with ERM-level effects than the high molecular weight PAHs. The primary PAH effects projected by the NOAA guidelines corresponded closely to depositional areas as denoted by % TOC and sand distributions.

A00048

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 11 of 40

b. Sediment Toxicity Tests Instead of graphing the statistical results, I used the actual experimental data for mapping and statistical analyses. The results of the Hyalella experiments are shown in Figure 11. During spring 2001, summer 2001, and summer 2002, there were relatively lower survival rates at stations noted to be depositional areas having relatively high sediment concentrations of PAHs. The Hyalella growth results were more complicated, with relatively lower growth at the reference station DR10 during spring 2001. This would limit the use of this station as the reference site for the statistical analyses. Growth was also reduced at various stations characterized by high PAHs during the summer of 2002. The clearest results of the series of experiments with Hyalella concerned reductions in reproductive effectiveness that were most apparent at various combinations of depositional stations noted for high sediment PAH concentrations. The highest percentages of gravid Hyalella were found in the refinery discharge areas. This trend was evident to a lesser degree in the growth and mortality results. The results of the Leptocheirus experiments are shown in Figure 12. The most distinct adverse effects on the mortality of this species were noted at the high depositional stations during spring 2001. There were lesser degrees of this trend during the other three experimental periods. Once again, in all four experimental periods, mortality was generally low with sediments taken from areas associated with the refinery outfall. The lowest growth indices (measured as weight) for Leptocheirus were noted at various combinations of depositional stations during all four experimental periods. During these periods, the greatest growth was noted at stations in the vicinity of the refinery outfall. The clearest patterns of adverse effects were noted with the reproductive capacity of Leptocheirus, where the least numbers of offspring were noted in depositional areas characterized by high PAH concentrations. High numbers of offspring of this species were noted in areas associated with the refinery outfall. The relatively higher rate of adverse effects on the infaunal species (Leptocheirus) relative to the epifaunal species (Hyalella) is not unusual when dealing with sediment contaminants such as PAHs.

A00049

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 12 of 40

Overall, there were general trends noted with the experiments carried out with Hyalella and Leptocheirus. Areas that were notably depositional with high concentrations of PAHs in the sediments were often associated with adverse impacts on the experimental subjects. These data represent further evidence that the sediment PAHs in refinery outfall receiving areas were not high and were not as toxic as sediments farther removed from the discharge area. A statistical analysis of these data will be presented below. c. Infaunal Community Analyses Results of the invertebrate community analyses for the Motiva Study are shown in Table 6. The relatively uniform results indicated high dominance by a few species and relatively low numbers of cumulative species richness. There was no real trend in the various community indices of this group of animals (Figure 13). Some stations The periodically showed decreases in the indices used for the community analysis.

dominant Rangia cuneata (Figure 14) was located mainly in areas associated with the refinery outfall and areas north of the outfall. There were virtually no Rangia in the depositional areas south of the outfall. This result does not coincide with the results of the bivalve studies carried out at different times and in different areas than the Triad analyses. Rangia is considered to be a species that is sensitive to pollution, and the Triad data were consistent with a possible adverse effect on this species in areas characterized by high sediment PAHs. The lack of bioavailability data preclude a more exact determination of the role of the high concentrations of PAHs in the depositional areas relative to the distribution of Rangia in the study area. Rangia distribution trends were followed generally by the distribution of Tubificoides heterochaetus, with reduced numbers in areas high in sediment PAHs. However, other dominant species such as Boccardia ligerica and Corophium lacustre were found in increased numbers in areas denoted by high PAHs. The community data generally did not reveal a definitive trend with respect to the distribution of PAHs in the study area. This could be related to the general contamination of sediments in the area. However, the distribution of dominant infaunal populations showed both positive and negative relationships to the distribution of sediment PAHs, which indicated that such distribution had an effect on such populations.

A00050

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 13 of 40

In various parts of the Triad studies, there were indications that depositional areas characterized by high concentrations of PAHs were associated with exceedances of NOAA ERL and ERM guidelines, toxic effects in laboratory experiments, and the absence of sensitive species. These results differ from the conclusions presented in the Motiva Report. 4. PAH Fingerprinting One key to the conclusion in the Report that the refinery effluents were not implicated in adverse impacts in the study area is the fingerprinting analyses. As pointed out in the review of the Report, the lack of attention to PAH loading in favor of concentrations represented a basic mistake due to the relative importance of loading to the distribution of refinery-released PAHs in sediments of the receiving system. In addition, identifying fingerprints of sediments close to the outfall as characteristic of refinery effluents ignores the fact that particulates carrying PAHs could actually distribute such refinery products to depositional areas distant from the outfall. The influence of the major exceedances during the early sampling were not accounted for in the various analyses by the authors. The data shown in the loading section of this report negated their various conclusions based on fingerprinting. The sediment PAH concentrations were organized by station groupings (according to the five time periods noted above). Stations in the Triad analyses were used for these groupings. The stations noted as "all" included the following: (DR1, DR2, DR23, DR26, DR10, DR9B, DR45, DR56, DR51, DR52, DR53, DR55, DR67, DR68, DR83). The use of stations noted as contaminated (DR53, DR55, DR67, DR68, DR83) was based on the biological and sediment PAH data of the Motiva Study. An analysis was carried out to determine if there was any correspondence between refinery PAH loadings and sediment PAH concentrations (total and by type). The loadings were averaged at four-month intervals prior to each Triad sampling period (spring and summer 2001; spring and summer 2002). The loading and sediment PAH data were organized in descending order, and the top 16 PAHs (in terms of loading from outfalls 601

A00051

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 14 of 40

and 001) were compared to the top 16 sediment PAH concentrations for the above dates. The 16 top PAHs in the sediment analyses represented an average of 61.77% of the total of all PAH sediment concentrations. These results were then summarized in a comparison of the loading and sediment concentrations of the occurrence of PAHs. loadings (Table 7). The highest sediment PAH concentrations often coincided with the refinery loadings from outfalls 110 and 601. According to the Motiva Report, petrogenic PAH source signatures were comprised of lower molecular weight (two- and three-ring) PAHs. It was this group of PAHs that was most closely associated with exceedances of NOAA ERM guidelines. According to the Report, the refinery effluent (outfall 601) was dominated by petrogenic four-ring PAHs (Fluoranthenes/Pyrenes with 2 [FP2] and 3 [FP3] alkyl groups and Benz- (a)anthracene/Chrysenes with 2 [BC3], 3 [BC3] and 4 [BC4] alkyl groups). Delaware water was characterized by Fluoranthene, Benz(a)anthracene and Perylene. Outfall 001 was mixed with (95%) cooling water. The petrogenic group was diluted rather rapidly by river water, according to the authors. However, the PAH loadings remained the same, regardless of such dilution, and these loadings were dominated by petrogenic types as outlined above. The origin of the Naphthalenes was unknown, according to the authors. This group of PAHs was excluded from the characterization of the refinery discharges by the authors, even though the Naphthalenes were well represented in the 601 effluent, which represented the undiluted refinery waste product. To simplify the analysis, the results shown in Table 7 were summarized in Tables 8 and 9. The distribution of individual PAHs in the 601 and 001 outfall loadings by date are shown in Table 8. The loading of the refinery effluent from outfall 601 to the river was characterized by the Naphthalene series, with increased loadings of individual compounds noted during summer 2001 and spring 2002. The Fluoranthene/Pyrene group (C1, C2, C3) comprised an important component of the 601 PAH loading, with high initial loadings (summer 2000) that decreased from spring 2001 through spring 2002. There was Highlights indicated PAH sediment concentrations of compounds common to both 601 and 001

A00052

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 15 of 40

an increase in the loading of these compounds during summer 2002. The Chrysenes (C1, C2, C3, C4) were also a consistent component of the 601 PAH loading; there was no definitive temporal trend of the loading of this PAH group. Peak loading of the Naphthalenes from outfall 001 occurred during the same increased loading of this group from outfall 601. This comparison would indicate that the Naphthalenes were part of the refinery PAH loading regardless of the fact that the exact origin of these PAHs remains unknown. The loading of the Chrysenes at outfall 001 was less definitive than that of the 601 outfall. during the 2002 period. The data indicate that there were relatively consistent characterizations of the refinery loadings that were primarily characterized by the Naphthalenes, the Fluoranthenes/Pyrenes, and the Chrysenes. There were also differential temporal changes in the qualitative and quantitative loadings of these and other PAHs. Periodic loading of other compounds was also noted in the temporal analyses of loading from the refinery. These loadings were ignored in the Motiva Report based on the fingerprinting results. The sediment PAH concentrations, with differentiation of those sediment PAHs that were associated with refinery loadings, are shown in Table 10. Specific Naphthalenes (Naphthalene, C1, C2) in both station groupings were associated with refinery loadings. This was also noted for specific Chrysenes (Chrysene, C1, C2, C3) and the Fluoranthenes/Pyrenes (C1, C2). The Phenanthrenes/Anthracnenes in the sediments were not associated with the 601 PAH loadings. The periodic association of Perylene in the sediments with the 601 loadings remains unexplained. These data show clear associations of higher sediment concentrations at the depositional stations for all three major groups (Naphthalenes, Chrysenes, Fluoranthenes/Pyrenes). The sediment PAH concentrations were somewhat different when compared to the 001 PAH loadings (Table 10). There were concentrations of Perylene, the Naphthalenes, and the Fluoranthenes/Pyrenes (C1, C2), but the Chrysenes were not represented in the 001 loadings. The associations of the There was a substantial loading of the Fluoranthenes/Pyrenes (C1, C2, C3) from the 001 outfall, although such loading decreased

A00053

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 16 of 40

001 loadings of the Phenanthrenes/Anthracnenes reinforced the observation that this group of PAHs was not associated with refinery effluents. However, the relatively high concentrations of the Naphthalenes, Chrysenes, and Fluoranthenes/Pyrenes in sediments of the noted depositional areas were closely associated with loadings from outfall 601, and, with the exception of the Chrysenes, outfall 001. This analysis indicates a mixture of refinery-derived and river-derived PAHs in the sediments, but the high concentrations of the refinery-derived PAHs in sediments of depositional areas runs counter to the overall observations and conclusions by the authors of the Motiva Report. A review of the distribution of PAH/TOC ratios in the study area (Table 10, Figure 15) indicated relatively high ratios at stations DR53, DR55, DR56, DR67, DR68, and DR83. These areas have already been identified as depositional areas with relatively high PAH concentrations. It is noteworthy that station DR1 (the effluent canal) had high sediment PAHs but relatively low PAH/TOC ratios. When sediments are high in TOC, there is a tendency for higher PAH concentrations. By standardizing the results by dividing the PAH concentrations by TOC concentrations, the accumulation factor in PAH distribution can be examined more closely. In this case, the high PAH/TOC ratios indicate that PAHs were better represented in sediments of areas characterized as depositional. This conclusion is not unexpected, considering the ecological characteristics of various PAH compounds. The low molecular weight PAHs, both as a group and individually (Figure 16), were concentrated in areas denoted as depositional during all four Triad samplings. These concentrations were directly related to refinery loadings, and were not associated with river-derived PAHs. This contradicts the findings in the Motiva Report. 5. Statistical Analyses The results of a cross-correlation analysis of the various sediment characteristics and the biological results of the Triad analysis is shown in Table 11. The analysis was backed up with a series of regression analyses. During spring 2001, Rangia biomass was negatively correlated (P < 0.05) with various refinery-associated PAHs. Hyalella growth

A00054

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 17 of 40

and survival were negatively associated with a mixture of river- and refinery associated PAHs. All three indices of the Leptocheirus experiments were negatively associated with a mixture of river- and refinery associated PAHs. During summer 2001, Rangia biomass was again negatively correlated (P < 0.05) with various refinery-associated PAHs. Tubificoides heterochaetus numbers were negatively associated with a mixture of riverand refinery associated PAHs, whereas Boccardiella ligerica numbers were positively associated with a mixture of river- and refinery associated PAHs. Hyalella % gravid and survival were negatively associated with a mixture of river- and refinery associated PAHs. Leptocheirus survival, growth and number of young were negatively associated with a mixture of river- and refinery associated PAHs. The significant association of adverse biological effects with individual refinery-loaded PAHs cannot be substituted for the experimental determination of bioavailability, but the results indicate that the conclusions of the Motiva Report concerning the lack of influence of refinery-loaded PAHs were incorrect. During spring 2002, noted decreases of sediment PAHs were associated with a lessening of correlations of adverse effects with sediment PAH concentrations. Only Leptocheirus survival, growth and number of young were negatively associated with a mixture of river- and refinery associated PAHs. However, by summer 2002, Rangia biomass was negatively correlated (P < 0.05) with various refinery-associated and riverassociated PAHs. All three indices of the Hyalella and Leptocheirus experiments were negatively correlated (P < 0.05) with various refinery-associated and river-associated PAHs. The emphasis throughout the Motiva Report was on the relative biological unimportance of refinery-based PAHs in the sediments of the receiving system. However, when a calculation was made concerning the relative contribution of Motiva-loaded PAHs relative to the total PAH concentrations in the sediments (expressed as percentages for all stations and for the depositional stations), around half of the PAHs noted in the sediments were derived from refinery PAHs (as defined by the authors and as noted in the loading from outfalls 001 and 601: see above). Motiva-loaded PAHs were thus significantly represented in sediments of depositional areas. It is not understandable how

A00055

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 18 of 40

the authors of the Report came to the conclusion that refinery-based PAHs were so unimportant in the overall effects on the receiving area. The premise of such unimportance was largely based on the fingerprinting data. However, the Report's fingerprinting analysis ignored the actual loading rates of PAHs from the refinery, and the contribution of the refinery to stations that were identified as depositional and strongly contaminated with PAHs. Table 12: Percentages of sediment PAHs derived from the Motiva effluent during the period of analysis (summer 2000, spring 2001, summer 2001, spring 2002, summer 2002).

Date Aug-00 Apr-01 Aug-01 Apr-02 Aug-02

%sed average all stations 48.2 46.5 55.7 46.8 50.4

%sed average contam stations 47 46.2 51.4 49.4 50.5

6. Application of Results to Motiva Report Conclusions An outline of the results of my review and analysis of the Motiva data is given in Table 13. Data organization for the Motiva Study was largely carried out without attention to spatial and temporal aspects of the PAH distributions, and with unsubstantiated assumptions that undermined PAH fingerprinting results. By odd grouping of the data, the statistics used by the authors added to this obfuscation of results. The loading data were ignored and mistakes were made concerning the substitution of PAH concentrations for PAH loading. Major omissions of the study included the lack of bioavailability data that added to poorly supported claims of causation based on statistical analyses in the Report. Bioavailability was mistakenly associated with bioconcentration. Triad organisms were not tested for toxic agents. The so-called bioavailability data were taken with only two samplings of bivalves at stations distant from the depositional areas of the system. There were no data to determine bioavailability at Triad stations. Without bioavailability information, it was not possible to separate the impacts of the various toxic agents in the

A00056

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 19 of 40

sediments for Triad organisms. The Triad results that indicated sediment toxicity in depositional areas were largely ignored by the study authors, who substituted statistical obfuscation for actual analysis of the primary data. The results of the long-core work were based on two stations that were not depositional. The results of the Motiva Study were skewed by mistaken assumptions made by the authors concerning how PAHs are carried into the river system and deposited in areas distant from the refinery outfall. The fingerprinting analyses assumed that areas around the effluent discharge area were indicative of refinery effluent signatures, when it is likely that specific refinery-associated PAHs were transported to depositional areas distant from the refinery outfall where they then contributed to high PAH concentrations and patterns of biological deterioration. The results of an evaluation of the primary Motiva data indicated that there was a relatively clear association of refinery-derived PAHs with high PAH concentrations in areas of deposition distant from the effluent canal. Combinations of river-derived and refinery-derived PAHs were significantly correlated with adverse biological effects. Highly significant negative correlations of sediment PAH concentrations (refinery-loaded) with various indices of biological functions would cast doubt on the main conclusions of the Motiva Report that chronic sediment toxicity was attributed to river-derived PAHs, PCBs, and metals (Cu, Zn, As, Pb, Hg) and not to the PAHs in the Motiva effluent. Further statistical analyses of the database are warranted.

A00057

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 20 of 40

Literature Cited
Bopp, R. F., M. L. Gross, H. Tong, H. J. Simpson, S. J. Monson, B. L. Deck, and F. C. Moser. 1991. Major incident of dioxin contamination: Sediments of New Jersey estuaries. Environ. Sci. Toxicol. 25: 951-956. Chaky, D. A. 2003. Polychlorinated biphenyls, polychlorinated dibenzo-P-dioxins, and furans in the New York metropolitan area: Interpreting atmospheric deposition and sediment chronologies. Ph. D. Dissertation, Rensselaer Polytechnic Institute, Troy, New York. Ehrlich, R., R. J. Wening, G. W. Johnson, S. H. Su, and D. J. Paustenbach. 1994. A mixing model for polychlorinated dibenzo-p-dioxins and dibenzofurans in surface sediments from Newark Bay, New Jersey using polytopic vector analysis. Arch. Environ. Contam. Toxicol. 27: 486-500. Hall, L. W. Jr. and D. T. Burton. 2003. A baseline study for assessing the potential aquatic ecological effects from Motiva Enterprises LLC Delaware City Refinery effluent using the sediment triad approach. Unpublished report. Long, E. R and L. G. Morgan, 1990. The potential for biological effects of sedimentsorbed contaminants tested in the national status and trends program. National Ocean Service, Rockville, MD. NOAA/TM/NOS/OMA-52. Long, E. R., D. D. MacDonald, S. L. Smith, and F. D. Calder. 1995. Incidence of adverse biological effect within ranges of chemical concentrations in marine and estuarine sediments. Env. Man. 19, 81-91. Long, E. R. and D. D. Macdonald. 1998. Perspective: recommended uses of empirically derived, sediment quality guidelines for marine and estuarine ecosystems. Human and Ecological Risk Assessment. Vo. 4, 1019-1039. New Jersey Department of Environmental Protection. 2002. Estimate of cancer risk to consumers of crabs caught in the area of the Diamond Alkali Site and other areas of the Newark Bay Complex from 2,3,7,8-TCDD and 2,3,7,8-TCDD equivalents. Unpublished Report. Wening, R. J., M. A. Harris, B. Finley, D. J. Paustenbach, and H. Bedbury. 1993. Application of pattern recognition techniques to evaluate polychlorinated dibenzo-pdioxin and dibenzofuran distributions in surficial sediments from the lower Passaic River and Newark Bay. Ecotoxicol. Environ. Saf. 25: 103-125. Zongwei C., V. M. S. Ramanujam, and M. L. Gross. 1994. Levels of polychlorodibenzop-dioxins and dibenzofurans in crab tissues from the Newark/Raritan Bay System. Environ. Sci. Technol 28: 1528-1534.

A00058

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 21 of 40

List of Tables
Table 1: Comparison of mean PAH loadings from outfalls 001 and 601. Monthly data were averaged over the study period. Table 2: Comparison of PAH loadings (outfalls 001, 601) during exceedances. Table 3: Comparison of mean PAH loadings at outfalls 001 and 601 during various periods. Table 4: Comparison of mean bivalve tissue PAH concentrations with mean PAH loading at outfalls 001 and 601 and mean sediment PAH concentrations. Table 5: Low (LMW) and high (HMW) molecular weight PAHs of sediments taken in the study area during spring 2001 (SP01), summer 2001 (SU01, spring 2002 (SP02), and summer 2002 (SU02) and expressed as Effects Range Low: (ERL) and Effects Range Median (ERM). Table 6: Biomass of infaunal macroinvertebrates taken during the four Triad samplings in the study area. Table 7: Comparison of PAH loading (gday-1) from outfalls 001 and 601 with sediment concentrations (ng g-1 DW). Sediment loading was averaged from the 4 months prior to the determination of sediment concentrations (August 2000, April 2001, August 2001, April 2002, August 2002). The loading data and sediment PAH concentration data were reorganized by order of loading rate and sediment concentration. The top 16 PAH loadings were compared to the highest PAH concentrations (16). Wherever there was a correspondence of PAH loading with a high PAH sediment concentration, the concentration notation was highlighted. This allowed a comparison of the PAHs that were loaded into the system with PAHs that were found in high concentrations in the sediments. The sediment PAH concentrations were averaged over all stations and over the depositional stations characterized by high overall sediment PAH concentrations (contam). Table 8: Distribution of individual PAHs in the 601 and 001 outfalls by date. Sediment loading was averaged over the four months prior to the determination of sediment concentrations (summer 2000, spring 2001, summer 2001, spring 2002, summer 2002). Table 9: Comparison of sediment PAH concentrations that coincided with loadings from the 601 and 001 outfalls. Where there was a high concentration of an individual PAH with a high loading, the sediment concentration figure was high- lighted and underlined. Table 10: Mean sediment PAHs by station (all dates averaged).

A00059

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 22 of 40

Table 11: Results of a cross-correlation analysis of sediment characteristics (% total organic carbon, % sand, individual PAHs, total PAHs, and EPA-designated PAHs) with biological data (Rangia cuneata biomass, Corophium lacustre numbers, Tubificoides heterochaetus numbers, Boccardiella ligerica numbers, numbers of infaunal numbers, species richness, Shannon diversity and Shannon evenness, Hyalella survival, Hyalella growth, Hyalella % gravid, Leptocheirus survival, Leptocheirus growth, and Leptocheirus % gravid). Data analysis was carried out by date (April 2001, August 2001, April 2002, August 2002). Significant correlations (P < 0.05) are highlighted. Table 12: Percentages of sediment PAHs derived from the Motiva effluent during the period of analysis (summer 2000, spring 2001, summer 2001, spring 2002, summer 2002). Table 13: Outline of the results of the review and re-evaluation of the Motiva Report.

List of Figures
Figure 1; Dye study showing distribution of Motiva effluent during maximum ebb tide and high slack tide (from, Hall and Burton, 2003). Figure 2: Simulated time-and-depth averaged current speeds (from, Hall and Burton, 2003). Figure 3: Stations used for the Triad analyses. Figure 4: Distribution of % total organic carbon in the sediments of the study area. Figure 5: Distribution of % sand in the sediments of study area. Figure 6: Distribution of total Polynucleated Aromatic Hydrocarbons (TPAH) and Environmental Protection Agency (EPA) Priority Pollutant PAH (EPAH) in sediments of the study area. Figure 7: Loading trends of TPAHs from the 001 and 601 discharge areas (3/99-8/02). Figure 8: Loading trends of individual PAHs from the 001 discharge area (3/99-9/02).

N0 N1 N2 N3 FP2 FP3 BC3

Naphthalene* C1-Naphthalenes C2-Naphthalenes C3-Naphthalenes C2-Fluoranthenes/Pyrenes C3-Fluoranthenes/Pyrenes C3-Chrysenes

A00060

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 23 of 40

Figure 9: Comparison of loading trends of TPAHs and EPAHs from outfall 001 with concentrations of TPAHs and EPAHs averaged over the Triad stations (Figure 3) during the individual Triad samplings from 8/00-8/02. Data were averaged for all stations (ALLSED) and for sediments taken in depositional areas having high concentrations of PAHs (Figure 6; CONTSED). Figure 10: NOAA ERL and ERM toxicity benchmarks. Figure 11: Results of Hyalella experiments (% survival, length [growth], % gravid). A. Graphical presentations B. Maps Figure 12: Results of Leptocheirus experiments (% survival, length [growth], % gravid). A. Graphical presentations B. Maps Figure 13: Maps of numbers, species richness and Shannon diversity of infauna taken during the four Triad samplings in the study area. Figure 14: Maps of biomass to dominant infaunal macroinvertebrates taken during the four Triad samplings in the study area. Figure 15: Distribution of PAH/TOC ratios by station during the Triad sampling periods (April 2001, August 2001, April 2002, August 2002). Figure 16: Distribution of the low molecular weight PAHs (µgkg-1 DW) in the study area by sampling period.

A00061

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 24 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 25 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 26 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 27 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 28 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 29 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 30 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 31 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 32 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 33 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 34 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 35 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 36 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 37 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 38 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 39 of 40

Case 1:88-cv-00263-SLR

Document 304-2

Filed 07/18/2005

Page 40 of 40