Methodology 

Collection and Chemical Analysis of Sediment
   
Sediments core samples were collected from five locations along the lower Black River.  Four locations were downstream from the section dredged by USS/KOBE in 1990.  These were between river mile (RM) 0.2 and RM 2.5; the final station was taken in the mainstem upstream above RM 4.0 (Table 1).  We wanted to take three samples in the reach subject to remedial dredging in 1990 between RM 2.9 and RM 3.6.   However repeated attempts to collect samples with the vibracorer failed because of the destiny of the sediment in this region.  I assume that loose depositional material had not been redeposited in this stretch, and the prevailing substrate was a dense clay which the vibracorer could not penetrate.  

Sediments were collected as core samples following the USEPA draft protocols (1997) and ASTM protocols (1994).  Core sections were collected on the first day of October, 1997.  Cores adjacent to the primary sample were collected at each site for use in laboratory toxicity testing by Dr. Allen Burton. The primary sediment core at each site was divided into two to four sections depending upon depth of sediment.  Sections were frozen and sent to AXYS Analytical Service, Ltd., Vancouver, B.C. for PAH, PCB and metal analysis. AXYS reported contaminants as dry weight sample concentrations.  Samples were analyzed by gas chromatography and mass spectrometry according to USEPA approved analytical methodology including QA/QC procedures as described in a separate document.  

Fish Collecting and Processing  

Brown bullhead were collected from the Black River AOC near Lorain, Ohio.  Collections were made from 0.5K below to 1.0K above the 002 outfall from the former USX coking facility located approximately 3.5K upstream from the river's confluence at Lake Erie.  Within the collection reach, the Black River receives effluents from a USX steel plant and historically received effluent from the associated coking facility.  Bullhead were collected by overnight sets of fyke nets, with either 12mm or 25mm stretched mesh and 9-16m leads.  Forty-five fish of age 3 or older were collected from the Black River 1997, using a minimum total length of 250mm as a criteria for age selection (age 3 or older). 

Necropsy and Histology

Each specimen was anesthetized with MS-222, measured (total length), weighed, and then euthanized by cervical dislocation according to animal welfare protocol approved by Ohio State University. Specimens were  processed according to previously published methodology (Baumann et al., 1990).  Each individual was inspected for grossly observable lesions involving the skin, the gills, and tissues within the oral cavity.  An incision was made in the abdomen from the anus to the isthmus and the viscera were examined for lesions.  Any pathologies or unusual features that are grossly visible were noted.  Livers were excised, examined, weighed, and imme­diately preserved in 10% neutral buffered formalin for histopath­ological analysis.  Gross lesions on other tissues will also be excised along with surrounding tissue and preserved in formalin.

Tissue blocks were cut from 3 to 5 sections, depending on the size of the liver.  These blocks will be dehydrated with an alcohol series up to absolute, infiltrated with paraplast, embedded, and sectioned at 5µm.  Tissue sections were rou­tinely processed and stained with hematoxylin and eosin.  Tissue slides from all livers examined were archived in the Leetown Science Center, BRD, USGS or in the Registry of Tumors in Lower Animals, George Washington University.  Preparation of tissues for histopathology followed the protocol of the Registry of Tumors in Lower Animals as previously published (Baumann et al., 1990).

Aging  

Pectoral spines were collected in the field for later aging in the laboratory.  Spines were processed according to the methodology in Baumann et al. (1990).  Each spine will be placed into individual 20ml scintillation vials and decalcified with 5% aqueous hydrochloric acid.  When spines were flexible but not limp (14-18h), hydrochloric acid was replaced by 50-60% isopropyl alcohol for storage.  A 6-10mm sections of spine just anterior to the basal groove were cut, trimmed and sectioned at different thickness (25-150mm) to increase the possibility of easily readable annual rings.  Identical ages assigned by two different readers was the criteria for age determination.  A third reading is made wherever disagreement exited in the first two readings.  

Chemical Analyses of Fish 

An additional 15 brown bullhead of 250mm or greater were collected by fyke net as described above for chemical analysis. Whole fish were wrapped in cleaned foil and frozen on dry ice.  Fish were packaged as 5 composites of 3 fish each and sent to Axys Analytical Service Ltd., Vancouver, B.C.  Axys reported PCBs and PAHs as wet weight sample concentrations.  Each sample will be homogenized and analyzed by gas chromatography and mass spectrometry according to USEPA approved analytical methodology.  One fish sample will be submitted as a spiked QA/QC control for spike performance and recovery.  All data is reported as lipid-normalized values.

Sample Locations 

Five sites, numbered from upstream to downstream,  were sampled between River Mile (RM) 4.2 and RM 0.27 near the mouth of the river (Table 1).  This reach bracketed the coke plant outfall (002) at ~RM 3.5 which had been the primary point source for PAH entry into the river.  Samples were attempted at several additional locations near the coke plant outfall, between site 1(RM 4.11) and site 2 (RM 2.88).  All such attempts were unsuccessful, because the vibracorer on the RV Mudpuppy could not penetrate the sediment sufficiently to get a sample.  This point is important in interpreting the results of the survey, especially when extrapolating comparative areas of sediment containing PAH at elevated levels, and will be discussed later.  Cores obtained from other sites varied in depth from 9 inches (site 3) to 4½  feet (site 1).    Three additional samples were obtained by the Ohio EPA  who sampled with us on the same date.  They used an Eckman dredge which obtained essentially a surface sample.  While their sites are listed with ours using the same longitude and latitude (Table 1),  their samples, taken from a different boat, varied from our location by approximately 5 to 10 meters. 

Results and Discussion

PCB Concentrations in Sediment

All of the surface or top layer core sediment samples contained 100 ppb total PCBs or less (Figure 1).   There was surprisingly little variation in the amounts, indicating a lack of active point sources (including dump site drainage) within the reach.  Secondary core segments, below 4” in depth, showed a great deal more variation and much higher PCB concentrations (Figure 1).  Peak values at these depths occurred at site 1 and site 3.  However even at site 3 the concentration was less than ½ ppm.  A comparison of the average surface sample concentration from the Black River with other Lake Erie tributaries indicates that PCBs are not a problem in this system (Table 2).  While levels are not pristine (still five times those in the Chautauqua-Conneaut systems, for instance), they are about an order of magnitude less than PCB concentrations in such rivers as the Ottawa, Detroit, and the Ashtabula-Chagrin systems (Table 2). 

Even the levels in these more polluted systems are below the most notorious hot spots, such as New Bedford Harbor, where sediments in the upper harbor contained above 10ppm PCBs and sediments in the lower harbor contained above 50ppm PCBs (US EPA, 1996).  In the Great Lakes drainage, the Fox River system is the most infamous for PCB contamination, where Little Lake Butte des Morts had sediments with up to 250ppm PCBs (Wisconsin DNR, 1998).  The latter amount is almost 3 orders of magnitude greater than the highest concentrations found in the Black River.   

PCB Residues in Fish

PCB residues in mature Black River bullhead composites (whole carcass) ranged from 0.75 to 1.3 ppm with a mean of approximately 1ppm (Table 3).  Again this value is below the level of PCBs found in the most polluted Lake Erie tributaries such as the Ottawa River and the Buffalo -Eighteen Mile system, but above more pristine locations such as the Chautaugua-Conneaut system.  In the Ohio River, PCB residue levels in channel catfish generally decline from upstream (near industrial centers in PA and WVA) to downstream (Table 3).  While the “small” channel catfish are about the same size (and would have similar feeding habits with) large brown bullhead taken in our samples on the Black, the Ohio River PCB concentrations are determined on fillets only.  Since the fillet concentrations for the Ohio River fish are identical to those of whole Black River bullhead, we can safely assume that the Black River fish are less contaminated. 

The 1ppm level is about half of the traditional 2ppm “action level” used by the US FDA to prevent contaminated fish from entering the market place (US FDA, 1998).  However a better reference is the national survey of PCB residues in fish conducted by the US EPA at 362 locations across the country.  The mean total PCB value for all fish surveyed was 1.9 ppm; however the median value was only 0.2ppm (US EPA, 1992).  Twenty-six percent of the sites surveyed had fish tissue concentrations greater than 1ppm.  Levels of PCB in fish tissues in the Great Lakes have undergone a steep decline from the late 1960s and early 1970s due to restrictions on sale and use (Schmitt et al. 1990).  

PAH Concentrations in Sediment  

All locations recorded total PAH concentrations in sediment below 15ppm with one exception (Figure 2).  Site 3 at the downstream end of the turning basin had total PAH concentrations of 40 to 50 ppm in both the second level of the core sample (4-9”, Table 1) and in the Ohio EPA “surface” sample taken near by.  Other than site 3, differences were not great between the surface (0-4” core sections) and the second level samples taken at  each location, and no pattern was apparent.  The two deeper core samples taken at site 1 had even lower total PAH concentrations, 2.7 ppm for the 2-4 foot depth and 0.8 ppm for the 4-4½  foot depth.  This last sample was clearly lighter in color than the upper layers and composed of what appeared to be a heavy clay.  Judging from the material adhering to the corer at the locations where samples could not be taken due to penetration problems (near the coke plant outfall), this material might have been a major component of the sediment there.

To examine the results in terms of individual compounds, I selected one representative of each of the ring groups from 3 to 6 (Table 4).  Phenanthrene (PHE) is often the most common of the environmental PAH pollutants.  Benz(a)anthracene (BaA), Benzo(a)pyrene (BaP), and Indeno(1,2,3-cd)pyrene (INP) are all recognized carcinogens in fish or rodents or both (Black et al. 1985, Couch and Harshbarger 1985, Metcalfe et al. 1988, Rice et al. 1984).  Since PAHs of different molecular weights (ring numbers) might well have different half-lives in an environmental deposit, Table 4 should provide a good overview of historic trends.  Two values from the current survey (1997) are given in Table 4.   “Black 1997” represents the mean of all of the surface samples taken except that near the mouth, since the latter is beyond the area in which the historical composites were collected.  “OEPA 97” represents values from the “hot spot”.  The mean surface values in 1997 range from one to three orders of magnitude lower than the historically high sediment PAH values from the 1980 and 1982 collections.  They are also lower than values recorded in 1992, two years after the dredging.  Even the OEPA hot spot is approximately an order of magnitude (or more) lower than concentrations from the early 1980s, and about equivalent to the values from 1992.  

The mean 1997 PHE value is lower than that observed in  1994 and on a par with the value from 1984; values for the longer chain compounds are somewhat elevated over the 1984 and 1994 concentrations (Table 4).  The 1984 collection was a year after closure of the coking facility in 1983.  However coke production had already declined in 1992 due to a depressed market for United States steel in that year,  which is part of the reason that the 1980 sediment values are elevated over the 1982 values.  Another factor which influences such inter-year comparisons is the location at which samples were collected, and sampling methodology.  The 1980 collections may be biased toward higher concentrations than the 1982 collections because they were all taken in closer proximity  to the coke plant outfall. 

The exact locations at which sediment grabs were taken were not recorded for most of the early surveys.  Thus we cannot know, for instance, exactly where the 1984 samples were collected and how that might have influenced the resulting values.  However we did in this survey try to collect samples several times between RM 3 and RM 4 closer to the location of the coke plant outfall.  Core samples could not be collected at these locations because sediment characteristics would not permit it.  The sediment may have been too course (gravel) or bedrock may have been close to the surface.  However it is more likely judging by our sample at RM 4.11, that the sediment consisted of heavy clay.  These locations which could not be sampled were within the area subjected to remedial dredging in 1989-1990.   The most likely scenario is that little deposition of fine material has occurred since the dredging in areas we tried to sample.  PAH levels in the heavy clay at RM 4.11 were minimal (0.8 ppm), and thus if surface samples from this area were taken, they would have reduced the mean values of PAHs for 1997.   

In fact while Black River PAH values for longer chain compounds are two to three time higher that those from the less industrialized Cedar-Portage system(PHE values are actually lower in the Black),they are lower than mean values from the Detroit River (Table 4).  In fact the OEPA hot spot values are lower than the Detroit River mean values, except for INP which is approximately equivalent.  If we visualize the Black River’s bottom as a highway and feeding table for benthic fishes, and use mental GIS to envision this ribbon of sediment as color coded for PAH contamination at different orders of magnitude,  the post-point source, post-dredging improvement becomes apparent.  With the dredged section having dark blue low levels of PAH, some greens here and there around it, and a few yellow hot spots, it looks much different from the expanse of reds, oranges, and yellows of the early 1980s.  Based on this survey, the sediment PAH levels seem much reduced in the mid-lower river (RM 4.2 to RM 2.8), where the greatest concentrations historically occurred.   Since expense and time preclude comparative sediment work from being thorough enough to accurately describe risk to native fauna, we need to examine the fish to check the validity of our conclusions.    

PAH Concentrations in Fish  

The 5 composites of three fish each taken in the spring of 1998 averaged 1.66 mg/g total PAH and displayed relatively little variation, ranging from 1.3 to 1.9 mg/g total PAH.  Individual PAH compounds also had minimal variation among the composites, so the mean can be used as a measure of PAH contamination in brown bullhead (Table 5).  Individual PAH analyzed in early 1980s surveys were selected for comparison (Baumann et al, 1987).  Longer chain compounds, such as BaP, were seldom detected in early surveys both because metabolism keeps the level of such parent compounds low in fish even in polluted systems, and because detection limits in the early 1980s were relatively high.  PAH residues in bullhead  in 1998 were one to two orders of magnitude lower than those in fish collected in 1980 and 1981 (Table 5).  They are also substantially lower than residues in 1982, even though the reduction in steel production had an immediate effect on fish exposure.  Residues are still higher, however, than those in even industrialized rivers without PAH point sources, such as the Fox and Munuscong Bay of the St. Marys.  However values are only 30% to 50% of those found in bullhead from the Cuyahoga River in 1984. 

Four ring compounds (FLU-CHR) declined less than did PHE (3 ring), reflecting similar trends in sediment (Table 4).  BaA in sediment went from almost an order of magnitude less than PHE  in 1980 to approximate equivalence in 1997 (Table 4).  CHR in fish residues, which is close to BaA in size, went from two orders of magnitude less than PHE in 1980-1981 to under one order of magnitude less in 1998 (Table 5).  Such data indicates that the smaller PAH compounds may degrade more rapidly in the environment that the longer chain compounds.  Since most of the carcinogens are larger four, five, and six ring compounds, the carcinogenicity of remaining PAH in sediment may increase on a per weight basis of extracted residue through time.  Thus using total PAH of historically contaminated sediment to estimate risk to benthic fauna may underestimate risk of neoplasia.     

Tumor (Neoplasm) Prevalence in Brown Bullhead  

Liver tumor prevalence has proven to be one of the best indicators of carcinogen exposure for benthic fishes (Baumann 1992 and Baumann et al. 1996).  In 1982 60% of the fish examined had liver neoplasms, with almost two-thirds of those lesions being advanced enough to be classified as cancer (Figure 3) (Baumann et al. 1990).  Another nearly 20% had changes in liver cell clusters (hepatocellular alteration) which indicated that these areas might progress to neoplasms (and eventually to cancer).  Thus only 20% of the bullhead captured in 1982 had livers which were “normal” in the sense of being free from changes in liver tissue that progress to tumors.  The coke plant closed in 1983, and four years later the elimination of the point source was reflected in the health of the bullhead, even prior to remediation.  Only 10% of the population had cancer, with over 22% more having neoplasms (Figure 3)(Baumann and Harshbarger 1995).  However a fairly large percentage (25%) had areas of hepatocellular alteration.  Even so, the percentage of fish with normal livers had slightly more than doubled since the sample in 1982.

Remedial dredging occurred in 1989 and 1990 using an open clamshell type of dredge.  This dredging method combined with the length of time the coking plant had been in operation (depth of PAH contamination in sediment), allowed some redistribution of previously buried sediment with high PAH concentrations (Baumann and Harshbarger, 1998).  Brown bullhead living in the river in the summer of 1990 were exposed to this material.  Thus adult bullhead collected two and three years later in 1992 and 1993 again had a very high prevalence of neoplasms similar to that seen in 1982 (approximately 60%) (Figure 4).  The percentage of advanced neoplasms (cancers) was even higher than in 1982 (46-48%).  The percentage of normal livers was greater than 1982 than in 1992, but declined in 1993 to closer to the 1982 level.     

By 1998, now eight years after the dredging, fish pathology should reflect the post-dredging status of PAH contamination in Black River sediment.  Those fish present during the actual dredging in 1990 (and exposed to newly dredged, PAH-laden sediment) should be largely eliminated from the population based on a life span that seldom reaches eight years of age.  Only 6.7% of the fish surveyed by liver histopathology had cancer in 1998 (Figure 5).  Since these were the only neoplasms found, the total neoplasm rate was also 6.7%, which is only 20% of the neoplasm prevalence recorded in 1987, four years after the coke plant closure but prior to remediation, and about 11% of the high rates in the early 1980s and 1990s.  Even though a high percentage of the population still have areas of hepatocellular alteration (24.5%), the percentage of fish with normal livers is higher than it has ever been, at nearly 70% (Figure 5).  Not only  is this over 3 times the percentage of fish with normal livers found in 1982 and almost 3 times the number is 1993, but it represents a 60% increase in this subgroup compared to 1987. 

Brown Bullhead Age Structure 

The high cancer prevalence in the brown bullhead from the Black River caused a truncated age distribution historically (Baumann et al. 1990).  High mortality rates eliminated older age classes that might be expected in Great Lakes tributaries at this latitude.  Thus another check on cancer prevalence in this population is to compare age frequency distributions.  During 1980 and 1982 no fish were found older than age 5, despite a huge sample size (N=522) (Figure 6).  Five year olds made up only 5½ % of the population, with the rest of those age 3 and older (susceptible to fyke not capture) being ages 3 and 4 Baumann et al. 1994).  In 1987, after the coke plant closure but before remediation, a few six year olds were captured, and the percent of the population older than age 4 had increased to 11.5%.  Thus even though the outright cancer prevalence that year (1987) was only 10%, the population was still primarily composed of young fish. 

In 1992 and 1993 (N=166) the situation was similar to that in 1987 (Figure 7).  The population had a higher concentration of three year olds than in the early 1980s, but a slightly higher percentage of five year olds as well (along with one age 6 fish).  The situation in 1998 is dramatically different from all of the previous years sampled (Figure 7).  Over 60% of the fish in 1998 were age 5 or older with age 6 or older fish making up over 35% of the population.  This is in stark contrast to all of the previous years in which age 3 and 4 fish comprised approximately 90% of the brown bullhead captured.  The 1998 data are also more similar to that from Muskellunge Lake, NY, a more pristine site where over 45% of the bullhead captured were older than age 4 and 19% were age 6 or greater (Sinnot and Ringler, 1987).  

Cancer Prevalence-Age Interaction  

Cancer rates increase with age both because exposure to carcinogenic metabolites increases with time and because of the latent period between exposure and tumor development.  Thus fish populations (or human populations) more heavily exposed to carcinogens not only have a higher prevalence of cancer but also have cancer appearing in younger age groups.  Because of this phenomenon, tumor frequencies for heavily polluted systems often understate the actual risk difference when compared to less polluted sites.  For instance, the cancer prevalence in 1982 (38.7%) seems to be 5.8 times greater than that in 1998 (6.7%).  However age 3 fish in 1982 had a 31% cancer prevalence, while no age 3 fish in 1994 or 1998 had any cancer.  One of the two fish with cancer in 1998 was age 6, an age not reached by any fish captured in 1981 or 1982.  Thus the two cancer percentages are not directly comparable.

More age-specific tumor rates also provide greater resolution of differences in bullhead tumor prevalence during coke plant operation in 1982,  four years after elimination of the point source in 1987, and eight years after remediation in 1998. In the May of 1982, the 1977 year class (age 5) had a cancer frequency of 60%.  By September of 1982 that year class had disappeared, and the cancer prevalence in the 1978 year class (age 4) had increased over summer from 37% (in the May sample) to 54% (in the September sample) (Baumann et al. 1987). The cancer prevalence in age 5 and older fish had declined to 33% by 1987, or just over one-half the prevalence in the early 1980s.  However in 1998 fish of age 5 and older had a cancer prevalence of only 7%.  This further decline in cancer frequency by over 75% from 1987 can be attributed to the remedial dredging.  However even this statistic understates the real difference between these years, since the percentage of age 5 and older fish that were at least 6 years of age was 50% in 1998, only 22% in 1987, and zero in 1982.

Needed Research 

More research is needed on the results of remedial dredging on the health of native benthic species.  Comparative studies on the short and long term effects of differing dredging methodologies would be helpful in planning future undertakings.  Other types of remediation of contaminated sediment, such as the capping that has just taken place in the Ottawa River, also need documentation of actual effects on fish health.  Fish health measures also need to be refined, with determination of area, volume, or incidence of lesions in each fish.  More base studies are also needed to determine the background prevalence of lesions in fish of differing ages from river systems without major point sources of PAHs and other carcinogens.