Peter Landrum - NOAA GLERL (Emeritus)
This project continues GLERL’s efforts to predict the bioavailability of sediment-associated organic contaminants to benthic organisms. In this past year, specific measurements of desorption for three Lake Michigan sediments and measurements of concentrations in Diporeia and sediment were analyzed and showed that PAH were likely accumulated from sediment and followed the model developed from laboratory studies while PCB congeners were not accumulated from sediment and did not follow the model. An additional field study was undertaken to return to two of the three Lake Michigan stations and measure the bioaccumulation by oligochaetes along with measures of Tenax extraction and sediment concentration. It is expected that oligochaetes will be more strongly connected to sediment concentration even for the PCBs. The samples are extracted but the analysis remains to be completed. In addition to studies performed at GLERL, a collaborative study with Drs. Michael Lydy and Jing You compared the use of Tenax extraction with matrix-solid phase micro extraction (M-SPME) to determine which method better predicted the bioaccumulation of several classes of organic contaminants. The M-SPME was superior in predicting the bioaccumulation of organic contaminants to Lumbriculus variegatus for both laboratory and field studies.
Sediment-associated contaminants are considered responsible for maintaining contaminant levels in the food web and preventing rapid declines because of the long time course for contaminant burial below the bioactive zone in sediment. GLERL has long-term studies examining the sedimentation and reworking of particle-associated contaminants into sediment and the subsequent return of these contaminants back into the food web through the benthos. GLERL is one of very few laboratories in the world examining the factors that affect the bioavailability of sediment-associated contaminants. Because of the clear importance of these contaminants in nearshore environments, research at GLERL provides the fundamental science to support interpretation of hazard and development of regulation for sediment-associated contaminants. Efforts to date have been only partially successful in developing approaches to predict the bioavailability of sediment-associated organic contaminants to aquatic organisms. The main approach, normalization of concentrations to the organic carbon content of the sediment has worked well with some compounds in some sediments and failed with others. GLERL has shown that the factors controlling bioavailability of sediment-associated organic contaminants are controlled by more than just organic carbon and include the amount of plant pigment in sediments. In some cases, particularly for polycyclic aromatic hydrocarbons, the quality of the organic matter appears to be particularly important for determining bioavailability. However, this is not the whole answer. Currently, no method exists that allows a manager to characterize the expected bioavailability of sediment-associated organic contaminants with any certainty. This work has as a primary purpose to establish improved approaches for predicting the bioavailability of sediment-associated organic contaminants.
Assessing the environmental risks associated with contaminated sediments requires determining the conditions under which sediment-associated toxic organic compounds accumulate in benthic organisms (bioaccumulation), determining whether the toxins are transferred up the food chain (biotransfer; biomagnification), and developing models for describing bioaccumulation. This work will help define the relationship between sediment characteristics and the bioavailability of organic contaminants, which in turn should improve our ability to predict the effects of contaminants associated with sediments.
This project continues investigations into the fundamental processes that limit the bioavailability of sediment-associated contaminants to the food web with emphasis on the benthos. The specific experimental studies that are proposed for the upcoming year include:
Based on the laboratory work demonstrating the relationship between contaminant desorption from sediments and bioaccumulation of sediment associated contaminants, a field test to test the laboratory relationship was pursued. Sediments and Diporeia were collected from three sites in Lake Michigan and analyzed for concentrations of PCB and PAH. In addition, the desorption of these compound classes was determined using Tenax® resin. The data was analyzed and the expectation was a log-log relationship between the biota-sediment accumulation factor (BSAF) and the fraction of contaminant desorbed. For the PAH congeners, 80% of the data points fell within a factor of two of the laboratory relationship but for the PCB congeners there was no relationship between desorption and bioaccumulation (FY 05 Figure 1). The PCB appeared to behave more as would be expected from equilibrium partitioning theory.
The reason for the difference in behavior between the two classes of compounds is not known. The data do suggest that Diporeia obtain their PAH largely from sediment while the PCB likely are accumulated from another source. As a result of these findings, a second field collection this time of sediment and oligochaete worms was made to determine whether organisms that are more strongly associated with the sediment show a similar profile. The samples have been extracted and efforts to complete the analysis are continuing.
A second method to quantify the extent of bioavailability for sediment associated contaminants uses a solid phase micro extraction (SPME) technique to identify the amount of available compound. Studies in collaboration with Mike Lydy and Jing You were initiated to compare the utility of the Tenax® method with the SPME to see which method would better describe the bioavailability of contaminants from a range of compound classes. Further, the laboratory study comparing these methods was augmented by a field study. The work used the oligochaete, Lumbriculus variegatus, and focused on PAH, PCB, Pyrethroid, and Organophosphate compounds. The SPME approach appeared to provide a better description of the bioaccumulation than the Tenax® desorption method. Further, the SPME approach appeared to well describe the bioaccumulation for both field and laboratory studies (FY 05 Figure 2).
The laboratory work evaluating the bioavailability of sediment-associated contaminants measured the bioaccumulation of two PCB congeners 2,4,5,2’,4’,5’ - hexachlorobiphenyl (HCBP) and 3,4,3’,4’-tetrachlorobipenyl (TCBP) and two PAH congeners, benzo[a]pyrene (BaP) and pyrene(PY) to an oligochaete Lumbriculus variegatus and an amphipod Diporeia sp. from seven laboratory spiked sediments. Previous work demonstrated that the desorption rate for the rapidly desorbed fraction was correlated to the amount of plant-derived organic matter in the sediment as represented by chlorophyll-a, lignin, or lipid-like materials. The data for the bioaccumulation was compared to the rate of desorption and resulted in good correlations between the log BSAF and fraction rapidly desorbed with both organisms (FY 04 Figure 1)
For both organisms, the use of equilibrium partitioning theory was an adequate model to describe the difference in the bioavailability among the sediments for HCBP. This theory describes the partitioning based on the concentration in the lipids of organisms relative to the organic carbon of the sediment. For Lumbriculus variegatus, the average BSAF was 2.87 ± 0.6 and was only slightly larger than the 1.7 value expected theoretically. For Diporeia, the average value was 1.5 ± 0.9 and while smaller equilibrium partitioning theory was an adequate representation of the bioavailability. For the planar compounds, BaP, PY and TCBP, equilibrium partitioning was not adequate as all the values were much smaller than unity. Since planar compounds are known to bind more strongly to soot carbon, a correction was used to determine whether the additional sorption of the soot carbon (black carbon) would correct for the apparent failure of equilibrium partitioning to describe the BSAF values. After correcting for the amount of soot carbon, the BSAF values for L. variegatus were all greater than unity except for one sediment from West Bearskin Lake in Minnesota. Thus, equilibrium partitioning was an adequate description for the bioavailability across all the sediments. For Diporeia, the correction was still not adequate with half or more of the sediments exhibiting values lower than unity for BaP and TCBP. The reason for the difference between species was likely due to differences in feeding behavior as L. variegatus is a general feeder while Diporeia is a selective feeder.
Overall, the measure of the desorption was a more useful descriptor for both species in that the predictive ability (within about a factor of 2) was about equal. Therefore, a field program was designed to determine whether the laboratory result could be applied to Lake Michigan sediments with Diporeia. To examine this sediments and Diporeia were collected from three stations in Lake Michigan. The desorption was measured using Tenax® extraction of an aqueous slurry of the sediment. A 24 h extraction was selected to be representative of the amount desorbed by the rapidly desorbing compartment. The sediments, organisms and Tenax® resin were analyzed by gas chromatography with electron capture detection for the PCB congeners and gas chromatography/mass spectrometry for the PAH compounds. The chromatograms have been completed and are currently undergoing data analysis to extract the concentration values. It is expected that a figure similar to figure 1 will result relating the bioaccumulation potential of the sediment and the observed bioavailability.
The bioavailability of selected organic contaminants from sediment was evaluated based on sediment characteristics and measures of contaminant desorption versus measured bioaccumulation by the worm Lumbriculus variegatus and the amphipod Diporeia spp. The contaminant desorption from sediment, which likely dictates the bioavailability, was fit with a three compartment empirical model. The rapidly desorbed fraction for planar compounds was correlated to the fresh organic matter such as plant pigments, lipid like materials, and lignin. Similarly, the bioaccumulation was also strongly correlated to these same materials for the planar compounds. The one non-planar compound studied behaved substantially different and the bioaccumulation was inversely correlated to the amount of organic carbon. The empirical model used for these studies was compared to a radial diffusion model to describe the desorption characteristics. The empirical model allowed a better description of the data with no increase in error in the estimated terms.
In addition to the work on bioavailability in laboratory spiked sediment, the Tenax extraction technique for measuring desorption was applied to selected Lake Michigan sediments where the organism and sediment concentrations had previously been determined. In the literature, there is a strong correlation between the extent of Tenax extraction and the bioaccumulation of sediment-associated contaminants. However, there was no significant correlation with the Lake Michigan sediments. This may have been due to the long storage (approximately 3 years) between the collection and original analysis of the sediments and organisms and the determination of the PCB desorption with Tenax extraction. The work should be repeated with fresh samples
Substantial data analysis was performed on the data generated from seven laboratory-dosed sediments. Data analysis was completed on the role of sediment characteristics on desorption of contaminants. A manuscript was generated and the factors controlling desorption were different for the planar compounds, pyrene (PY), benzo[a]pyrene (BaP), and 3,4,3’,4’-tetrachlorobiphenyl (TCBP) compared to the non-planar 2,4,5,2’,4’,5’-hexachlorobiphenyl (HCBP). In the case of the planar compounds, the relatively fresh organic matter such as plant pigments, lipid like compounds, and lignin content correlated with the rapidly desorbing compound. There was no significant correlation for these compounds between total organic carbon, total nitrogen content, amount of soot carbon or the amino acid content. For the non-planar HCBP, the flux off the particles was negatively correlated with the organic carbon content of the sediment and the particle size characteristics of the sediment. Because these are laboratory dosed sediments, the extent of equilibrium was evaluated to assess the applicability of the findings to field sediments. The rapidly and slowly desorbed fractions were in equilibrium based on the desorption kinetics but the very slowly desorbed compound was not in equilibrium. This compartment would require about 1 to 3 years to reach equilibrium for the different compounds and sediments. Thus, true equilibrium may not occur within the biologically active zone for the very slow component even in the field because conditions change on a fairly frequent basis in this sediment layer. Overall, the sediment organic carbon characteristics were very important for controlling the rate of desorption particularly for the rapidly desorbing compound, which is understood to be the bioavailable compound in sediment.
Because it is so important to understand the desorption of contaminants
from sediments due to the strong connection to contaminant bioavailability,
a modeling investigation was undertaken to compare an empirical model
used on the work described above and a radial diffusion model. The empirical
model assigns the desorbed compound to one of several, both two and three
compartment models were used, compartments that have similar desorption
characteristics. This model follows the fraction of contaminant remaining
in the sediment as a series of first order decay processes
where St is the sediment-sorbed amount at time t, S0 is a sediment-sorbed amount at the start, F is a size fraction of a chemical in rapid, slow, and very slow sediment compartment, k is a desorption rate coefficient for respective compartments (h-1), and t is time (h). The radial diffusion model makes use of Fick’s Law of Diffusion and assumes that the compounds are initially evenly distributed in the particles and that there is a weighted average particle radius that will effectively describe the range of particles in the sediment.
The representation below described the contaminant as desorbing from two compartments and follows the amount of compound desorbed from the sediment. Dobs f = The diffusion rate from the fast desorbing compartment, T = Time (h), a = the mean particle diameter for the sediment, Dobs s = The diffusion rate from the slow desorbing compartment, and f = the fraction of the total contaminant attributed to the fast desorbing compartment.
The data for the comparison used the desorption data collected from the seven laboratory dosed sediments. While the radial diffusion model appears to be more mechanistic, in fact it is actually just as empirical at the model describing the desorption as a series of compartments with first order decay. This is the result of our lack of understanding of the true number of compartments and the actual distribution of the compound in the sediment to properly describe a2. When the data were fit with a two-compartment version of the empirical model and the radial diffusion model, the radical diffusion model could not describe the diffusion of benzo[a]pyrene from the sediments as two compartment processes. This occurs because of the relatively small rapidly desorbed fraction for benzo[a]pyrene in two of the sediments. Further, the extent of error and the quality of the fit to the data measured as coefficient of determination, model selection criteria and coefficient of variation of the estimates were all poorer with the radial diffusion model. When the three compartment empirical model is used to describe the desorption process, the fit of the model is generally better when comparing coefficient of determination and model selection criteria over the two compartment models. However, the coefficient of variation of the coefficient estimates again rises to be comparable to those of the radial diffusion model. Comparing the ability of the models to describe the amount of desorbed compound measured after about one day, the three-compartment model was more successful than the radial diffusion model. However, both the three compartment model and the radial diffusion model were about equally useful in describing the bioavailability of the compounds across the seven sediments (Figures FY03 F1 and FY03 F2).
The overall fit for the two models is not different in that they account for 40 to 50 % of the variability in the data.
The bioaccumulation data were also compared to the characteristics of the sediments to determine which parameters might best describe the bioaccumulation. There was a similar difference between the planar and non-planar compounds as described for the desorption. The bioaccumulation factor (BAF), showed a trend of a decline with the amount of soot carbon in sediment but once the BAF was normalized for organism lipid and total organic carbon there was no relationship between bioaccumulation and soot carbon. Because the amount of soot carbon correlated strongly with the amount of organic carbon in the sediment, the decline in BAF with soot carbon was really simply the impact of total organic carbon on bioaccumulation. The BAF also increased with the amount of chlorophyll A in the sediments. Similar correlations were observed with total plant pigments, the total amount of lignin and the amount of lipid like materials in the sediment. Although there was some trend in most of the data, these were not statistically significant because of the large variation in the amount of organic carbon. Once the bioaccumulation was normalized for the amount of organic carbon to form a BSAF, there was a strong correlation between BSAF and the amount of chlorophyll A (Figures 3 and 4)
Each of the compounds shown has a significant positive correlation with chlorophyll A but chlorophyll A does not describe the extent of variation between the compounds particularly for Diporeia spp.
The hexachlorobiphenyl performed differently and there was not a good correlation to describe the variation in the bioavailability across sediments. However, this is not a great difficulty as the variation in BSAF for Diporeia spp was from 2.5 to 3.5 across all the sediments and that for Lumbriculus variegatus was 0.5 to 2.5 with most of the points in the range of 1 to 2.5. Thus, the variation for this compound is small, and the total organic carbon content is the primary driving force for bioaccumulation of HCBP.
In a separate set of measurements, the desorption characteristics of PCB congeners from Northern Lake Michigan sediments were determined. The objective was to ascertain whether these measurements could be correlated to the observed bioaccumulation by the organisms. Of the four sediments originally examined, only two were available for Tenax extraction. In addition, a third sediment from the seasonal PCB study was also extracted. They were subjected to a 6 h extraction with Tenax in an aqueous suspension and the Tenax was analyzed for PCB congeners. Unlike published results with other sediments, there was no significant correlation between the fraction extracted with the Tenax and the accumulation by Diporeia spp. This negative result was surprising but may be due to the impact of storage of the sediments, approximately 3 years, between the time of analysis of the organisms and the Tenax extractions of the sediment. This experiment should likely be repeated with freshly collected organisms and sediments. The outcome did not lead to the expected improved description of the bioavailable fraction of PCB from sediments for the Northern Lake Michigan stations.
A set of seven sediments were spiked with pairs of contaminants and equilibrated for a minimum of 60 d to reach equilibrium. The sediment characteristics and the sediment-contaminant interactions were determined through a series of tests designed to be compared with the observed bioavailability of the contaminants to two aquatic invertebrates, Lumbriculus variegatus and Diporeia spp. The four compounds chosen for study were, benzo[a]pyrene (BaP), 2,4,5,2’,4’,5’-hexachlorobiphenyl (HCBP), 3,4,3’,4’-tetrachlorobiphenyl (TCBP), and pyrene (PY). The BaP and PY were 3H labeled while the two PCB congeners were 14C labeled. The compounds were spiked in equal molar concentrations to the sediments. The data was collected last year and this year’s efforts were devoted to analyzing the collected data and beginning the development of manuscripts. The desorption of contaminants from the sediments was found to follow a three compartment model with rapidly, slowly, and very slowly desorbing pools. The rapidly desorbing pool is thought to serve as the dominant bioavailable pool and the total flux from the sediment drives the bioaccumulation. In general, the desorption rate of the compounds from sediments was different for the planar compounds, BaP, PY and TCBP than for the HCBP and varied among the sediments. The factors that produced positive influences on the desorption flux and size of the rapidly desorbing pool for the planar compounds were the amount of plant pigments found in sediment and the amount of lipids (non-polar extractable residue). However, these parameters did not explain the desorption of HCBP, which is not a planar molecule. Other, measured parameters such as the distribution on specific particle size fractions and the aromaticity of the extracted humic matter were also important for determining the desorption of some of the planar molecules. For the HCBP, the desorption was controlled in part by the amount of organic carbon and the partitioning to fine particles. This desorption behavior is consistent with the importance of organic carbon for determining the bioavailability of HCBP as observed in this work and previously. The bioavailability of the contaminants from sediment was proportional to the flux of the contaminants from sediment. Thus, those parameters that affect desorption are generally the same parameters that are associated with the bioaccumulation of the contaminants. A manuscript on the desorption study is being written and a manuscript comparing the bioaccumulation to sediment characteristics will follow.
1. The continued investigations into the factors that control the bioavailability of sediment-associated contaminants were focused on a set of laboratory experiments that exposed Diporeia spp. and Lumbriculus variegatus to laboratory dosed sediments. Seven sediments from various locations in the United States were collected and dosed with a combination of 14-C 2,4,5,2’,4’,5’-hexachlorobiphenyl and 3-H benzo[a]pyrene or 14-C 3,4,3’,4’-tetrachlrorbiphenyl and 3-H pyrene. The sediments and the sorption of the contaminant with the sediments were characterized using several measurements: contaminant desorption rates, extraction efficiency by invertebrate gut juice, organic matter content including soot carbon content, the amount of extractable humic matter and the aromaticity of the humic matter, the amount of lignin, the amount of plan pigments, the amount of extractable amino acids and the particle size distribution of the sediments and the contaminants. All of the exposures and the characterization measurements were completed. The data for the hexachlorobiphenyl and the benzo[a]pyrene data are partially analyzed.
Table 1. Characteristics of the Sediments
|Sediment||TOC%||TON%||SOOT C%||Pigm. (µg/g)|
|Lake Huron #9||3.70||0.474||0.27||5.04|
|Lake Huron #54||3.18||0.394||0.32||6.34|
|West Bearskin Lake||1.25||0.125||0.05||5.11|
Table 2. Desorption characteristics of Benzo[a]pyrene from the sediment
|Lake Huron #9||6.8||0.030||15.3||0.0050|
|Lake Huron #54||18.8||0.026||19.9||0.0057|
|West Bearskin Lake||8.5||0.026||14.2||0.0029|
It has previously been thought that the amount of bioavailable compound and the amount of rapidly desorbable compound would correlate with the amount of organic matter in sediment. However, there was not a good correlation between the desorption rate or fraction of rapidly desorbable compound. There did appear to be a relationship between the uptake rate and the flux of contaminant off sediments calculated as the K(rapid) times the concentration in the rapidly desorbable pool. The data analysis is expected to be completed in the next fiscal year. Several outside investigators have collaborated in this work, Dr. Jussi Kukkonen (University of Joensuu, Finland) who was visiting GLERL and performed most of the work, and Drs. J. Gunnarsson and D. Weston (University of California, Berkeley), and Dr. S. Mitra (USGS Menlo Park, CA) who did specific analyses with the sediments.
2. PCB congeners are accumulated in different patterns in Diporeia between sites in the southern and northern basins of Lake Michigan. Samples were collected and the toxicokinetics of 4 selected 14-C labeled PCB congeners dosed into sediments from four stations along the eastern side of Lake Michigan were determined in the amphipod Diporeia spp. The sediment characteristics and the concentrations of environmentally resident PCB congeners were determined in sediment and synoptically in Diporeia samples. The bioaccumulation factor is relatively well determined by the toxicokinetics (e.g., Figure FY01-1). However, attempting to use sediment organic carbon as a normalizing factor did not improve the predictability between the laboratory predicted concentrations and the concentrations measured in the environmental samples (Figure FY01-2). The BSAF values ranged from 0.2 to 33 across a range of PCB congeners of 4.69 - 6.92. Additional, analysis of the data will be performed this upcoming fiscal year.
3. The seasonal field data for PCB concentrations in sediments, settling detritus and Diporeia spp. were previously collected and the Diporeia spp. data were compared to the results of toxicokinetic model (FY01-EQ1) parametrized with the seasonal sediment concentrations and the settling detritus concentrations. The model made use of parameters generated as part of several laboratory studies published in Landrum et al. 2001 below).
The model was generally able to predict the concentrations found in Diporeia at various points in the year e.g., Figure FY01-1.
The feeding rates used in the model were developed for Diporeia feeding on surficial sediments thus the slope for the prediction is near 1. However, when settling detritus is used as the food source, the predicted accumulation is much greater than the measured because the feeding rate on settling detritus is likely lower than that for surficial sediment as it has greater nutritional value. If the feeding rate were half that of the feeding rate on surficial sediment then the lines would essentially overlap. It is likely that the feeding rate on settling detritus would be lower since it has a higher nutritional quality.
1. A study on the role of contaminant desorption from sediment particles examined the kinetics processes for accumulation by Diporeia and n-pore filters from sediments and water to test the hypothesis that the desorption rate of organic non-polar contaminants from sediment particles limits the bioaccumulation of contaminants even in short-term bioassays. The rates of contaminant accumulation were determined on n-pore filter membranes, as a non-biological material that must receive contaminant only via passive processes, e.g. desorption and diffusion, compared to Diporeia spp. that can also obtain contaminant via ingestion and can migrate to new patches of dissolved contaminant. To sort out the kinetics for the filters, the uptake was measured in sediment (with and without Diporeia), water, and suspended sediment exposures. The relative rates for the n-pore filters were then compared to those for accumulation by Diporeia. From the preliminary data for TCBP, it is clear that Diporeia increase the rate of diffusion of contaminant to the n-pore filters by a factor of approximately five. Further, the accumulation rate for the filters is approximately seven times less than that for Diporeia, suggesting that routes other than diffusion are important for biological accumulation (Figure FY00-1).
This conclusion is also supported from the temperature response, which was much lower (approximately a factor of 20) for the non-biological material than for the Diporeia. Modeling the data, to fit a concentration based accumulation model, we observed that the implied minimum volume of interstitial water
1. to which the organisms would have to be exposed to demonstrate the measured rate of accumulation was directly proportional to kde, as kde decreases, the volume of I.W. required for exposure will increase for a given uptake rate. As accumulation from feeding becomes greater, the fraction resulting from I.W. exposure to maintain a fixed accumulation will decrease. The data from this study have been collected and the data analysis and modeling should be completed this fall.
2. To expand the predictability of the accumulation of PCB congeners by Diporeia as an out growth of the Lake Michigan Mass Balance Study, the toxicokinetics were examined for four sediments from Northern Lake Michigan. In these sediments, the profiles for accumulation appear to differ from the profiles observed for southern Lake Michigan. The sediments were collected and dosed with selected PCB congeners. The toxicokinetics were determined and the distribution of the contaminant within the sediment on a particle size basis was completed. The uptake clearance decreased with increasing sediment organic carbon concentration across all compounds tested. BSAFs values were statistically different among sediments and between compounds. Furthermore, the BSAF values increased with log Kow up to 5.92 (Figure FY00-2). In addition, sediments and organisms from each station were analyzed for a suite of PCB congeners. Data analysis is expected to be completed this next year.
3. This research continued the project begun under Measurements of Select Toxic Organic Contaminants in Surficial Sediments of Lake Michigan: Estimation of Lakewide Inventories and Seasonal Settling Fluxes, which was a part of the Lake Michigan Mass Balance Program. Briefly, Diporeia, surficial sediment (SS), settling detritus (SD), overlying water (OW) and suspended particles (SP) were collected from a Lake Michigan site for one year. Normalized PCB concentrations from Diporeia and the potential food sources were established so that log Kocs and BSAFs (lipid-normalized benthic organism contaminant concentration / organic carbon (OC)-normalized sediment contaminant concentration) of 14 congeners (log Kows 5.16 to 7.80) could be compared to theoretical values at equilibrium, which were predominantly determined at single time points. At equilibrium, log Koc should approximately equal log Kow. In addition, BSAFs should have a theoretical value of ~2, regardless of particle, chemical and animal properties. Seasonal factors that affect sorption and exposure were investigated. Models that included some of these factors were addressed in order to attempt improved predictions of PCB bioaccumulation.
Data highlights are as follows: SP partitioning slopes (of the relationship between log Koc and log Kow) increased from 1.01 in June to 1.25 in December, 1997. The December slope occurred during lake turnover, with year-high OC-normalized PCB concentrations due to increased particle-associated PCB concentrations and decreased OC content. SD partitioning slopes were ~ 0.5 from spring through summer, 1998. In contrast to 1997 SD partitioning, these slopes were due to greater partitioning of lower log Kow PCBs compared to partitioning for higher log Kow PCBs. The small spring slopes were coincident with a large, settling particle event that originated from coastal erosion.
For all exposure sources, the trend across seasons was increasing BSAF as log Kow increased. BSAFs bracketed the theoretical value for all contaminant sources, but animal:SS, :SP and :SD levels varied by factors of 6, 8 and 200, respectively. Animal:SD BSAFs were the highest among the contaminant sources in late summer, 1997. However, in May and June, 1997, animal:SD BSAFs were between 0.5 and 4, when Diporeia actively feed on SD. When we compared SD OC-normalized concentrations to SD BSAFs, data were well correlated (negatively). In other words, SD BSAF seasonal patterns were created by SD OC-normalized concentrations since lipid-normalized concentrations were relatively stable over time. (Figure FY00-3 compares lipid and SD OC-normalized PCB concentrations for 1997.)
Because SD BSAF patterns did not reflect more stable lipid-normalized concentrations, using BSAF as a predictor of PCB bioaccumulation from SD may not be appropriate. In order to attempt improvements in the predictability, a steady state model was investigated that incorporated toxicokinetic factors in order to predict Diporeia dry weight concentrations of contaminant on a seasonal basis. Such factors included contaminant assimilation efficiencies, Diporeia feeding rates, food source dry weight concentrations of contaminant, and Diporeia elimination rates. Data were well predicted. Animal feeding rates played a large role in the success of the model predictions. Currently, data are being analyzed and interpreted for presentation at the 21st annual Society of Environmental Toxicology and Chemistry meeting in Nashville, TN on November 12 -16, 2000.
1. One of the critical measures for estimating the accumulation of contaminants is to determine the amount accumulated from ingestion. To obtain an improved estimate of exposure via this route, feeding rates for Diporeia were measured for Lake Michigan sediment at 2, 4, 8 and 15 C. When placed in a modeling framework with assimilation information, previously determined, the amount of accumulation from ingestion can be estimated. An attempt was made to determine the feeding rate on diatoms, which are important food sources for Diporeia, but methodology problems precluded our ability to make accurate measurements at this time. However, feeding rate data from sediments were measured and used in combination with the desorption data below to build a contaminant accumulation model for Diporeia.
The experimental work to estimate the rate of contaminant desorption from sediment was completed for two non-polar organic contaminants, tetrachlorobiphenyl (TCBP) and benzo(a)pyrene (BaP). Diporeia accumulation of TCBP and BaP were also determined. All data were determined over a temperature range of 4 to 15 C. The initial data analysis from these studies was completed and an initial modeling framework established. The modeling exercise suggests that the desorption rate from sediment dominates the rate processes for the accumulation from interstitial water. The relative proportion of the contaminant accumulated from ingestion versus from interstitial water increases with increasing temperature. This is driven primarily by the increase in feeding rate with increasing temperature.
2. The sediment samples were collected to examine the differences in the toxicokinetics between stations from northern Lake Michigan and southern Lake Michigan. These sediments have been dosed and exposures should be started in December.
3. Feeding studies for Diporeia spp. Were completed for feeding on lake sediments at several temperatures. However, all attempts to obtain mass balance for Diporeia feeding on diatoms were unsuccessful. New experimental design will need to be investigated.
4. Experimental work was completed on the toxicokinetics in crayfish. The crayfish is an important critical food web organism for the transfer of contaminants up the food chain from zebra mussels. The data analysis is not yet complete and will be evaluated prior to committing additional resources to the problem.
5. This progress is reported here from a project begun under Measurements of Select Toxic Organic Contaminants in Surficial Sediments of Lake Michigan: Estimation of Lakewide Inventories and Seasonal Settling Fluxes which was part of the Lake Michigan Mass Balance Program. The goals of this project are to: 1) monitor the seasonal variation of PCB accumulation in Diporeia spp. in order to determine the dominant sources of these contaminants, and 2) determine the validity of the equilibrium partitioning (EqP) theory in predicting biota exposure to sediment-associated contaminants. Animals, surficial sediment, settling detritus, and dissolved/suspended particle samples were collected from the Grand Haven 45m site. To date, PCB concentrations from Diporeia and the potential food sources have been measured using a GC/ECD (gas chromatograph with electron capture detection). From initial data analysis, BSAFs generally increased seasonally throughout 1997 (Figure FY99-1). In addition, BSAFs generally increased with log Kow. However, it was clear that additional sediment samples, already collected would need to be analyzed for PCBs before seasonal changes could be evaluated with any certainty. These samples are currently undergoing analysis.
It is currently not possible to accurately predict the exposure of benthic (sediment dwelling) invertebrates to sediment-associated contaminants. Previous approaches for evaluating exposure have relied on equilibrium partitioning theory to predict the overall exposure. Equilibrium partitioning theory suggests that the maximum exposure would be equivalent to exposure to the contaminant concentration in the interstitial water (the water occupying the spaces between sediment particles) assuming that the system is at equilibrium. However, this approach has not been successful for predicting exposure for many cases. Thus, to improve our understanding of the processes that govern the exposure, work was initiated to determine the magnitude of exposure coming from each of the two major routes of exposure, sediment interstitial water and ingestion of sediment particles. Experimental work was designed to examine the abiotic exposure of sediment-associated contaminants and make a comparison to the accumulation by biota. The abiotic medium will only be exposed to interstitial water and material that can desorb from particles to the interstitial water. The biota will also be exposed through the ingestion route. With longterm experiments, the magnitude of the two routes of exposure will be determined. The experimental design was completed and initial experiments with abiotic media were performed. Exposures of the abiotic material were performed at 4, 10 and 15C. The initial experiments comparing the exposure of benthos to the abiotic material have begun. This work is expected to provide the focus for more detailed studies to improve prediction of bioaccumulation from sediment exposures.
The bioavailability of sediment-associated contaminants is known to decline with the length of time that contaminants are sorbed to sediments. The extent to which this occurs appears to be more extreme for compounds with lower log octanol-water partition coefficients (log Kow), such that the magnitude is greater for phenanthrene than for benzo[a]pyrene. Since most of the studies have been performed with relatively short-term experiments, the difference between the extent to which this occurs may be one of kinetics, rate, instead of extent. Bioavailability studies were performed with benzo[a]pyrene sorbed to sediments where the duration of contact ranged from 1 week to 13 months (Kukkonen and Landrum, 1998). The bioavailability was evaluated periodically. The bioavailability declined rapidly and by 6 months had reached a plateau of about 44% reduction compared to compound freshly sorbed to sediment. However, the chemical partitioning between sediment particles and interstitial water were consistent between freshly sorbed and long-term sorbed contaminant. Thus, chemical measures of partitioning do not lead to improved understanding of bioavailability.
Because lipids are the dominant force in determining the bioaccumulation of organic contaminants by aquatic organisms, a review of the role of lipids was completed (Landrum and Fisher, 1998). Normalization of accumulation of concentrations in organisms to the lipid content eliminates most of the variability between species in bioconcentration studies. Thus, predictions of bioconcentration can be made for lipid normalized bioconcentration with physical parameters such as the log of the octanol-water partition coefficient (log Kow). Studies of the relationships between lipids and bioconcentration have resulted in new insights in food web transfer of contaminants, the mechanism producing membrane narcosis, and the importance of lipids in the transfer of contaminants through the reproductive cycle.
Greenberg, M.S., Burton, G.A., Landrum, P.F., Leppänen, M.T., and Kukkonen, J.V.K. 2005. Desorption kinetics of fluoranthene and trifluralin from Lake Huron and Lake Erie sediments. Environ. Toxicol. Chem. 24:31-39.
Ingersoll, C.G., E.L. Brunson, S. Wang, F.J. Dwyer, G.T. Ankley, D.R. Mount, J. Huckins, J. Petty, and P.F. Landrum. 2003 Uptake and depuration of non-ionic organic contaminants from sediment by the oligochaete, Lumbriculus variegatus. Environ. Toxicol. Chem. 22:872-887.
Kukkonen, J.V.K. and P.F. Landrum. 1998. Effect of particle-xenobiotic contact time on bioavailability of sediment-associated benzo(a)pyrene by benthic amphipod, Diporeia spp. Aquat. Toxicol. 42:229-242.
Kukkonen, J.V. K., P.F. Landrum, S. Mitra, D.C. Gossiaux, J. Gunnarsson, and D. Weston. 2004. The role of desorption for describing the bioavailability of selected PAH and PCB congeners for seven laboratory spiked sediments. Environ. Toxicol. Chem. 23:1842-1851.
Kukkonen, J.V. K., S. Mitra, Landrum, P.F., D.C. Gossiaux, J. Gunnarsson, and D. Weston. 2005. The contrasting roles of sedimentary plant-derived aromatic carbon versus sedimentary black carbon on hydrophobic organic contaminant availability to Diporeia spp. and Lumbriculus variegatus. Environ. Toxicol. Chem. 24:877-885.
Landrum, P.F. and S.W. Fisher. 1998. Influence of lipids on the bioaccumulation and trophic transfer of organic contaminants in aquatic organisms. In M.T. Arts and B.C. Wainman Eds., Lipids in Freshwater Ecosystems, Springer-Verlag, NY, pp. 203-234.
Landrum, P.F., J. Kukkonen, M.J. Lydy and H. Lee II. 1999. Measuring absorption efficiencies: some additional considerations. Environ. Toxicol. Chem. 18:2403-2405
Landrum, P.F. 2000. A short summary of the sources and distributions of contaminated sediments in the Great Lakes. Toledo J. Great Lakes Law Sci. Policy 3:19-25.
Landrum, P.F., E. Tigue, S. Kane Driscoll, D. Gossiaux, P. Van Hoof, M. Gedeon, and M. Adler. 2001. Bioaccumulation of PCB congeners by Diporeia spp: Kinetics and factors affecting bioavailability. J. Great Lakes Res. 27:117-133
Leppänen, M.T., Landrum, P.F., Kukkonen, J.V.K., Greenberg, M.S., Burton Jr, G.A., Robinson, S. D., and Gossiaux, D.C., 2003 Investigating the role of desorption on the bioavailability of sediment-associated 3,4,3’, 4’-tetrachlorobiphenyl in benthic invertebrates. Environ. Toxicol. Chem. 22: (In Press).
Lozano, S. J., M. L. Gedeon, and P. F. Landrum. 2003. The effects of temperature and organism size on the feeding rate and modeled chemical accumulation in Diporeia spp. for Lake Michigan sediments. Journal of Great Lakes Research 29(1):79-88.
Lydy, M.J., J.L. Lasater and P.F. Landrum. 2000. Toxicokinetics of DDE and 2-chlorobiphenyl in Chironomus tentans. Arch. Environ. Contam. Toxicol. 38:163-168.
Ma, X, K.A. Bruner, S.W. Fisher and P.F. Landrum. 1999. Absorption of hydrophobic contaminants from ingested Chlamydomonas rheinhardtii and Chlorella vulgaris by zebra mussels, Dreissena polymorpha. J. Great Lakes Res. 25:305-317.
Nuutinen, S., L.J. Schuler, P.F. Landrum, J.V.K. Kukkonen, and M.J. Lydy. 2003. Toxicokinetics of organic contaminants in Hyalella azteca. Arch. Environ. Contam. Tox. 44:467-475.
Van Hoof, P.L., J.V.K. Kukkonen and P.F. Landrum. 2001. Impact of sediment manipulation on the bioaccumulation of field-contaminated and laboratory-dosed PAHs by an oligochaete (Lumbriculus variegates, Muller). Environ. Toxicol. Chem. 20:1752-1761.