Environ. Sci. Technol. 2007, 41, 7723-7729
Persistent Chlordane Concentrations in Long Island Sound Sediment: Implications from Chlordane, 210Pb, and 137Cs Profiles L I J I A Y A N G , †,‡ X I Q I N G L I , §,| J O H N C R U S I U S , ⊥ U R S J A N S , †,‡ MICHAEL E. MELCER,# AND P E N G F E I Z H A N G * ,†,§,@ Department of Earth and Atmospheric Sciences and Department of Chemistry, City College of New York, CUNY, New York, New York 10031, Ph.D. Program in Chemistry and Ph.D. Program in Earth and Environmental Sciences, Graduate School and University Center, CUNY, New York, New York 10016, United States Geological Survey, Woods Hole, Massachusetts 02543, and Department of Math and Science, United States Merchant Marine Academy, Kings Point, New York 11024
Concentrations of chlordane, a banned termiticide and pesticide, were examined in recently collected surficial sediment (10 sites) and sediment cores (4 sites) in Long Island Sound (LIS).The highest chlordane concentrations were observed in western LIS, near highly urbanized areas. Chlordane concentrations did not decrease significantly in the past decade when compared to the data collected in 1996, consistent with the observation of nearconstant chlordane levels in blue mussel tissues collected during the same time period. Chlordane concentrations in many of the sites exceeded levels above which harmful effects on sediment-dwelling organisms are expected to frequently occur. Chlordane concentrations in two of the four sediment cores showed a peak below the sediment surface, suggesting reduced chlordane inputs in recent years. The lack of a chlordane concentration maximum below the sediment surface in the other two cores, coupled with the lack of a well-defined 137Cs peak, indicated significant sediment mixing. Simulations of 137Cs and 210Pb profiles in sediment cores with a simple sediment-mixing model were used to constrain both the deposition rate and the bioturbation rate of the sediment. Simulations of the chlordane profiles indicated continued chlordane input to LIS long after chlordane was phased out in the U.S. Continued chlordane input and significant sediment mixing may have contributed to the persistent chlordane
* Corresponding author phone: (212)650-5609; fax: (212)650-6482; e-mail:
[email protected]. † Ph.D. Program in Chemistry, Graduate School and University Center, CUNY. ‡ Department of Chemistry, City College of New York, CUNY. § Department of Earth and Atmospheric Sciences, City College of New York, CUNY. | Present address: College of Environmental Sciences, Peking University, Beijing, 100871, People’s Republic of China. ⊥ United States Geological Survey. # United States Merchant Marine Academy. @ Ph.D. Program in Earth and Environmental Sciences, Graduate School and University Center, CUNY. 10.1021/es070749a CCC: $37.00 Published on Web 10/16/2007
2007 American Chemical Society
concentrations in surficial sediment, which poses longterm threats to benthic organisms in LIS.
Introduction Long Island Sound (LIS) is an estuary surrounded by the coast of Connecticut to the north and Long Island (New York) to the south (Figure 1). New York City is located at the west end of the Sound. The Sound, about 150 km long and about 30 km across at its widest point, is one of the largest estuarine systems on the Atlantic coast of the U.S. (1). The Sound provides vital transportation and rich fishing and shellfishing grounds for commercial interests. It is also the home to numerous beaches and other recreational facilities. The Sound, however, has been contaminated with various pollutants, including organochlorine pesticides such as chlordane (2). Technical chlordane is a mixture of over 140 compounds and was used worldwide as a pesticide for farmlands, home lawns, and gardens and as a termiticide for house foundations (3). Many chlordane components and metabolites are toxic (4), are suspected to be carcinogenic (5), and may have estrogenic effects (6). Chlordane was first used in the U.S. around 1948, and its application for the control of insects on agricultural crops and vegetation was banned in 1983 (3). All commercial chlordane use in the U.S. was cancelled in 1988, and the sole chlordane use between 1983 and 1988 was to control subterranean termites (3). The concentrations of chlordane and other contaminants in LIS sediment were previously surveyed by the National Status and Trends (NS&T) Program of the National Oceanic and Atmospheric Administration (NOAA) (2, 7). Between 1985 and 1996, NOAA sampled sediment and benthic-dwelling organisms at over 10 sites in the Sound, all of which were carefully selected to avoid point sources and to represent general contamination conditions in the Sound (8). The surveys showed a clear decline in chlordane concentrations in the surficial sediment of two sites associated with the major rivers (LICR and LIHR, Figure S1a in the Supporting Information; see Figure 1 for locations) and two western LIS sites (LIHH and LITN, Figure S1b) between the late 1980s (when chlordane was phased out in the U.S.) and 1996 (the date of the last NOAA sediment survey). Monitoring of chlordane concentrations in blue mussel (Mytilus edulis) tissues by NOAA has continued to the present and has shown near-constant levels in the past decade (Figure S2), indicating near-constant chlordane concentrations in the top several centimeters of sediment where the mussels dwell. This observation is intriguing, given that chlordane has been banned for nearly 20 years. The fate of chlordane in sediment may be influenced by microbial processes such as biodegradation and physical processes such as sedimentation and sediment mixing. On the one hand, in depositional environments, highly contaminated sediment may be buried by increasingly clean sediment over time, leading to a decline of contaminant concentration in surficial sediment (i.e., natural capping) (9). On the other hand, sediment mixing (caused by bioturbation, bottom currents, etc.) may redistribute the contaminant deposited earlier (deeper in the sediment bed) to the sediment surface, leading to persistent concentrations in surficial sediment (10). The objectives of this work were to determine chlordane concentration changes in the surficial sediment since the last NOAA survey in 1996, to examine the influence of VOL. 41, NO. 22, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7723
FIGURE 1. Sampling sites in Long Island Sound. The Sound is about 150 km long and 30 km across at its widest point. sedimentation and sediment mixing on the observed concentration changes, and to infer the history of chlordane deposition from the sedimentary record. Surficial sediments were collected at the sites previously surveyed by the NS&T Program. Sediment cores were also collected at selected sites to examine the profiles of chlordane concentrations and 137Cs and 210Pb activities. To our knowledge, very few studies have examined chlordane concentration profiles in sediment cores (11-13), and no such study has been performed in LIS. The radionuclide activity and chlordane concentration profiles were simulated using a sediment-mixing model to infer the history of chlordane releases into the Sound and the relative importance of sedimentation and sediment mixing on the vertical transport of chlordane in sediment following deposition. The enantiomeric composition of chlordane in LIS sediment has been examined in a previous paper (14), which concludes that enantioselective biodegradation of chlordane was inhibited in LIS sediment.
Materials and Methods Sample Collection and Acquisition. Surficial sediments were collected in 2005 and 2006 at eight LIS sites (LICR, LIPJ, LIHR, LISI, LIHU, LIHH, LIMR, and LITN, Figure 1) that were previously surveyed (1986-1996) by the NS&T Program’s Mussel Watch Project (8) and two additional sites (LIMB and LILN, Figure 1) where very high chlordane concentrations (>21 ng/g dry weight) were observed in a separate sediment toxicity survey (2). Sediment cores were collected in 2005 at four sites that showed the highest chlordane concentrations in past surveys (LIHH, LILN, LIMB, and LITN, Figure 1). Aliquots of archived surficial sediments at five sites (LICR, LIPJ, LIHR, LIHH, and LITN) collected by the Mussel Watch Project were obtained from NS&T Specimen Bank (7, 8) to determine chlordane concentrations in the 1980s. Analysis of the archived sediments also allowed comparisons between the concentrations obtained using our method and NOAA’s method (7). More details about the sample collection and acquisition can be found in our previous paper (14). Sample Extraction, Chlordane Analysis, and Quality Control. The protocol to extract archived and recently collected sediment samples was described in detail in our previous paper (14). Chlordane in sediment was analyzed using a Fisons GC 8000 gas chromatograph equipped with 7724
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 22, 2007
an AS 800 autosampler and a Fisons ECD-800 Nickel-63 electron capture detector (ECD) (Fisons Instruments SpA, Milan, Italy). A SLB-5ms capillary column of 60 m × 0.25 mm × 0.25 µm size (Supelco, Bellefonte, PA) was employed. The carrier gas was helium (99.999%), and the pressure was set to 240 kPa. Samples (2 µL) were injected splitless (split opened 1 min after injection) at an injector temperature of 280 °C. The oven temperature was held at 100 °C for 2 min, ramped to 160 °C at 10 °C/min, then ramped from 160 °C to 290 °C at 5 °C/min, and held at 290 °C for 5 min. The detector temperature was set to 300 °C. The makeup gas for the ECD detector was 5.2% methane in argon, and the pressure was set at 130 kPa. Chlordane concentrations were quantified using external standards (ranging from 1 to 25 ng/mL) and were corrected by the recoveries (ranging from 41 to 57%) of the surrogate, PCB 166. Each sample was injected 3 times, and the relative standard deviation of the three injections was typically within 2%. Analytical blanks, processed using an identical procedure for real samples, were included in every batch of five sediment samples. Chlordane concentrations were below the detection limit in the blanks. To check the reproducibility of extraction, four replicates of the LIMB surficial sediment were extracted, and the relative standard deviation of the four replicates was 7%. To determine the completeness of extraction, the LIMB surficial sediment was extracted a second time immediately after the first extraction. Both cis-chlordane (CC) and transchlordane (TC) were below the detection limit in the second extraction. Recoveries of chlordane were also determined by spiking randomly selected sediment samples with about 2-5 times the native amounts of chlordane. The spiked sediments were extracted, cleaned up, and analyzed in the same way as real samples. The chlordane concentrations corrected by the surrogate (PCB 166) recoveries were typically within (15% of the concentrations corrected by the recoveries of spiked chlordane, validating the concentration correction using surrogate recoveries for the rest of the samples. 137Cs and 210Pb Analysis and Modeling. To infer the history of chlordane releases to LIS sediment, the sediments from the cores collected at LILN, LITN, and LIMB were analyzed for 137Cs and 210Pb by γ counting. Detailed background on the use of 137Cs and 210Pb as chronological tools to date recent sediment can be found elsewhere (e.g., ref 15). Between 5
FIGURE 2. Comparison of total chlordane concentrations in surficial sediments to threshold and probable effect concentrations. ERL, ERM, TEL, and PEL stand for effect range-low, effect range-medium, threshold effect level, and probable effect level, respectively. Sampling sites were sorted from east to west (see Figure 1 for site locations). and 10 g of dried, disaggregated sediment was sealed in vials and stored for at least 21 days to allow for the in-growth of 222Ra and 214Pb to approximate equilibrium values. Samples were counted for 2-5 days using a Princeton Gamma-Tech Ge well detector (Princeton, NJ), detecting the 46.3 keV 210Pb peak and the 661.6 keV 137Cs peak. Detector efficiencies were determined by counting standards filled to the same vial height as the samples. EPA standard pitchblend ore was used for 210Pb, and SLOSH III standard (16) and NIST SRM 4350b were used for 137Cs. Supported 210Pb activities were determined from the total 210Pb activity at the base of the core, whereas excess 210Pb was calculated by subtracting the supported 210Pb from the total 210Pb activity. A simple sediment-mixing model was used to simulate 137Cs, 210Pb, and chlordane profiles. In this model, the sedimentation was treated as a process of advection (second term on RHS of eq 1), while sediment mixing was treated as a process of diffusion (first term on RHS of eq 1), as described in the following equation:
[
]
∂FA ∂ Db∂FA ∂FA -s ) - λFA ∂t ∂z ∂z ∂z
(1)
where F is the density of the solid phase (assumed to be 2.5 cm3 g-1), A is the nuclide activity (disintegrations per min (dpm) g-1) or chlordane concentration (ng/g), t is the time (year), z is the depth (cm), Db is the sediment mixing rate coefficient (cm2 year-1), s is the sedimentation rate (cm year-1), and λ is the decay constant (0.0311 year-1 for 210Pb, 0.0230 year-1 for 137Cs, and 0 for chlordane).
Results and Discussion Chlordane Concentrations in Surficial Sediment. The concentrations in the recently collected surficial sediments range from 0.5 ng/g dry wt (LIHR) to 13.0 ng/g dry wt (LIMB) for TC and from 0.4 ng/g dry wt (LIHR) to 12.0 ng/g dry wt (LIMB) for CC (Figure 2 and Table S1). Chlordane concentrations were highest at western LIS sites (LITN, LIHH, LILN, and LIMB, Figure 2), as was observed in previous surveys (e.g., ref 2). Higher organochlorine contamination in estuaries near highly urbanized areas in this region (e.g., New York City and Newark) has also been observed by others (17, 18). One possible cause for higher chlordane concentrations in western LIS is that chlordane input into western LIS was
FIGURE 3. Concentration profiles of trans-chlordane (TC) and cischlordane (CC) in sediment cores collected at LIMB, LIHH, LILN, and LITN. greater than to the rest of the Sound. For example, runoff from house foundation soils at western LIS is expected to carry more chlordane due to the greater population (house) density. Another possible reason is that chlordane input into eastern LIS (from riverine/agricultural sources through the Connecticut River and the Housatonic River) was transported to the western LIS by bottom water that flows west and southwest (1). The chlordane may have then been deposited in the western LIS as a result of quiescent and eutrophic conditions in this sheltered part of the Sound that allows fine-grained, organic-rich material to accumulate. Indeed, the total organic carbon (TOC) content was highest in the western LIS (Table S1). Plotting total chlordane (TC + CC) concentrations against TOC content yielded a coefficient of determination (R2) value of 0.68 (Figure S3), indicating a fairly strong correlation between chlordane concentration and TOC content. TOC content in sediment samples of LIMB and LITN cores showed a gradual decrease with depth (Figure S4,) and there was no clear relationship between chlordane concentrations and TOC content in the sediment core (cf. Figures 3 and S4). The observed chlordane concentrations have decreased appreciably at 2 out of the 10 sites since the last survey in the 1990s (by a factor of about 5 at LIHR and a factor of 2 at LIMB), decreased slightly at LIMR and LILN, and increased somewhat at the remaining sites (Figure S1 and Table S1). These observations indicate that overall chlordane concentrations in surficial sediments did not decrease significantly in the past decade, consistent with the near-constant chlordane concentrations in mussel tissues in the past decade (Figure S2). Since chlordane was banned in late 1980s in the U.S. and its use decreased afterward, chlordane concentrations in sediment might be expected to decrease in the past decade. Chlordane concentration increases at some sites in the past decade may be due to spatial variations in chlordane concentrations (e.g., sampling locations were a few meters to a few hundred meters apart between this and previous surveys, Table S1) or due to the differences in sample extraction and analysis. Ratios of CC concentrations VOL. 41, NO. 22, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7725
FIGURE 4. 210Pb (triangles) (a), 137Cs (circles) (b), and total chlordane (TC + CC, squares) (c) profiles from the LIMB core and sediment-mixing model simulations (lines), assuming a constant flux of 210Pb and a time-varying 137Cs flux to surface waters (assuming input from fallout only in panel b). Complete list of model inputs is provided in Table S2. Brief model simulation descriptions include (a and b) mass accumulation rate ) 0.27 g cm-2 year-1 (sedimentation rate ≈ 0.7 cm year-1) with bioturbation to 19 cm averaging 28 cm2 year-1 (solid line) and 3 cm2 year-1 (dashed line) and (c-e) mass accumulation rates of 0.23 g cm-2 year-1 (dotted line), 0.27 g cm-2 year-1 (solid line) and 0.30 cm-2 year-1 (dashed line) (sedimentation rates ) 0.57, 0.68, and 0.74 cm year-1), with bioturbation to 19 cm averaging 5 cm2 year-1 (dotted line), 3.4 cm2 year-1 (solid line), and 5 cm2 year-1 (dashed line). 137Cs input includes a flux over and above that expected from fallout in panel d, as discussed in the text. (e) Mass accumulation ) 0.27 g cm-2 year-1 and bioturbation to 19 cm averaging 3.4 cm2 year-1, with a peak in chlordane delivery to surface waters in 1971 (line), based on usage records (see text). of archived sediments obtained in this study to the corresponding concentrations from NOAA ranged from about 0.6 (LITN) to 3.0 (LIPJ), indicating that different analytical procedures may indeed yield significantly different concentration values for the same sediment (Table S1). Threshold effect concentrations (below which harmful effects on sediment-dwelling organisms are not expected) such as effect range-low (ERL) and threshold effect level (TEL) and probable effect concentrations (above which harmful effects on sediment-dwelling organisms are expected to frequently occur) such as probable effect level (PEL) and effect range-median (ERM) are often used to assess the ecotoxicological impacts of organic contaminants in surficial sediment (19, 20). The ERL, TEL, PEL, and ERM for total chlordane in marine sediment were 0.5, 2.26, 4.79, and 6.0 ng/g d.w., respectively (21-23). Chlordane concentrations exceeded the ERL value at all 10 sites and exceeded the TEL value at all sites except LIHR (Figure 2). Chlordane concentrations at 4 of the 5 most western LIS sites (LIHH, LIMB, LILN, and LITN) also exceeded the PEL and the ERM values. Therefore, Figure 2 clearly indicates that sediment toxicity by chlordane exists at all sites sampled and that chlordane still poses significant threats to benthic organisms in western LIS, even after the use of chlordane has been completely banned for nearly two decades in the U.S. Chlordane, 137Cs, and 210Pb in Sediment Cores. TC and CC concentrations followed similar trends with depth at all sites where sediment cores were collected (Figure 3). Total chlordane profiles were very similar to 137Cs profiles (Figures 7726
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 22, 2007
4 and 5). In the LIMB core, the excess 210Pb decreased from the maximum value of 7 dpm g-1 at the sediment surface to values below 1 dpm g-1 at a depth of ∼33 cm (Figure 4, supported 210Pb activity was 2.2 dpm g-1). The 137Cs activity and chlordane concentration maintained intermediate levels at the sediment surface, reached a maximum at a depth near 28 cm, and decreased to near-zero values at a depth close to ∼50 cm (Figures 3a and 4). Chlordane profiles in the LIHH core were similar to those in the LIMB core, albeit the concentrations were lower (Figure 3a). In the LITN core, the excess 210Pb decreased from a maximum of ∼8 dpm g-1 at the sediment surface to less than 1 dpm g-1 at a depth of ∼40 cm (Figure 5). The 137Cs activity and chlordane concentration maintained relatively constant levels from the sediment surface down to ∼30 cm and then dropped to near-zero values at a depth of ∼40 cm (Figures 3b and 5). In the LILN core, the 210Pb excess activities generally decreased from ∼5 dpm g-1 at the sediment surface to ∼2 dpm g-1 at a depth of ∼50 cm (Figure 4 of ref 14). The 137Cs activity and chlordane concentration were variable in the core and showed no clear peaks, although there was a hint of a possible 137Cs maximum at a depth of ∼25-30 cm (Figure 3b and Figure 4 of ref 14). Note that there was no detectable 60Co in any of the cores, which is consistent with an atmospheric fallout origin for the 137Cs rather than releases from nuclear power plants in the LIS vicinity (e.g., ref 24). Because the LIMB core has both a well-defined onset of 137Cs activity in the core and a well-defined 137Cs peak, some simple sediment accumulation rate estimates can be made. In the other cores, the radionuclide data are more complex
FIGURE 5. Excess 210Pb (triangles) (a), 137Cs (circles) (b), and total chlordane data (squares) (c) from the LITN core, as well as sedimentmixing model simulations (lines), assuming a constant flux of 210Pb and a time-varying 137Cs flux to surface waters (assuming input from fallout only). The chlordane input in panel c peaks in 1971 based on usage records (see Table S3). Complete list of model inputs is provided in Table S2. Brief model simulation descriptions include mass accumulation rates of 0.24 g cm-2 year-1 (solid line) and 0.30 (dashed line) g cm-2 year-1 and bioturbation to 20 cm at an average rate of 110 cm-2 year-1. and will be interpreted later with the aid of a sedimentmixing model. On the basis of a simple interpretation that the depth of maximum 137Cs activity (∼30 cm) at LIMB corresponds to 1963, a sedimentation rate of 0.7 cm year-1 was estimated. This rate is very similar to the estimate of 0.8 cm year-1, assuming that the initial detection of 137Cs (at ∼44 cm) corresponds to 1952 and that the first detection of significant chlordane (∼49 cm) corresponds to its first use in the U.S. around 1948 (3). These sedimentation rates translate to mass accumulation rates of ∼0.3 g cm-2 year-1, assuming an average determined porosity of 0.84. 137Cs, 210Pb, and chlordane can be transported downward by sedimentation, bioturbation, and possibly diffusion (e.g., desorption and/or diagenetic mobility (the process by which certain particle-bound substances are released into solution by a variety of reactions in the sediment)). Desorption and/ or diagenetic mobility of 137Cs have been observed in various locations (25, 26). The 137Cs inventories in the cores range from 30 to 60% of that expected from atmospheric fallout, while the 210Pb inventories range from 170 to 210% of the value expected from fallout (Table 1). The lower 137Cs inventories and higher 210Pb inventories than those expected from atmospheric fallout are consistent with similar low 137Cs inventories in other estuarine environments attributed to the mobility of 137Cs under high-salinity conditions (27) and enhancement of the 210Pb inventory due to sediment focusing (the resuspension and movement of sediment from shallow to deep parts of the basin by water currents) (28). As noted previously, however, the depths at which chlordane concentrations and 137Cs activities first increase significantly were very similar (Figures 4 and 5). Since chlordane was first introduced into the environment around 1948 (3) and the first significant weapons fallout began in 1952, it is concluded that diagenetic mobility has had little effect on the penetration of 137Cs into the sediment. The alternative possibility that 137Cs and chlordane have diffused downward in the sediment by comparable distances is unlikely, given the different conditions required for mobility of the two contaminants (25, 29). Modeling Results. Sediment mixing in estuaries can be caused by different processes, including bioturbation, bottom currents, and storm events. Bioturbation refers to the enhanced transport of particles, solutes, and sorbed species in bed sediment by the activities of benthic organisms, such as feeding, burrowing, excavation, tube construction, and irrigation (30). Bioturbation can redistribute sediment (and any associated contaminants) from its original stratigraphic position and may penetrate to depths of tens of centimeters beneath the sediment surface (31). The treatment of sediment mixing as a diffusive process in this work (first term on the
TABLE 1. Parameter Values Derived from Best-Fit Simulations of Radionuclide Data and Inventories of 210Pb, 137Cs, and Chlordane parameter mixed layer depth (cm) sediment mixing rate coefficient (bioturbation rate) (cm2 year-1)b porosity rangec mass accumulation rate (g cm-2 year-1) sedimentation rate (cm year-1)d Peclet number 210Pb (dpm cm-2) 137Cs (dpm cm-2) chlordane (ng cm-2)
LIMB (Figure 4c-e)
LITN (Figure 5)
LILNa (no figure)
19
20
50
3.4 (solid)
110
80
0.80-0.89 0.27 (solid)
0.73-0.90 0.85 0.24 (solid) 0.26
0.68 (solid)
0.5 (solid)
0.7
3.8 54.8 6.13 417
0.09 67.8 3.58 449
0.4e 54.5 2.89 159
a Not estimated with confidence. b Av for entire mixed layer (decreases with decreasing porosity). c Porosity data were fit with an exponential, with highest porosities at the sediment-water interface. d Mass accumulation rate is assumed constant in this model, but the sedimentation rate is dependent on the porosity, which decreases with depth in most sediment cores. For this estimate of sedimentation rate, the average porosity for the excess 210Pb- and 137Cs-containing sediments has been used. e Highest of all estimates (could be as low as 0.1).
right-hand side of eq 1) provided a reasonable description of the available radionuclide data in our study area and allowed quantifying the time-averaged impact of mixing processes using just two parameters, the sediment-mixing coefficient (Db, cm2 year-1) (32, 33) and the sediment mixed layer depth. The model uses a finite difference approximation of the previous equation (eq 1) as described by Santschi et al. (34) and subsequently modified by Crusius et al. (31). The sediment-mixing model was used to evaluate whether the 137Cs, 210Pb, and chlordane data at each site could be explained by a common set of sediment accumulation and sediment mixing (mainly bioturbation) rate estimates and whether the chlordane concentrations observed are consistent with the known usage history. The nuclide fluxes from atmospheric fallout were estimated from the 90Sr fallout record measured at New York City, assuming a 137Cs/90Sr ratio of 1.5 (35), and the 210Pb fallout record measured at New Haven, CT (36). While the 210Pb profile at LIMB could be simulated well assuming bioturbation with no sedimentation (Figure S5a), the 137Cs profile could not (Figure S5b), indicating that downward transport of 210Pb and 137Cs was not solely due to bioturbation. Likewise, the radionuclide VOL. 41, NO. 22, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7727
data were inconsistent with downward transport by sedimentation only because the 137Cs profile from such a model run maintained a maximum that is far sharper than is apparent in the data (Figure S5c,d). For the LIMB core, the 210Pb profile could be reproduced by a combination of sedimentation (s ) 0.8 cm year-1) and mixing to 19 cm (Figure 4a). A sediment mixing rate coefficient (Db) of 28 cm2 year-1 yielded a fair fit of the 137Cs profile (Figure 4b). This fit is not perfect in the sense that the constant 137 Cs activity in the top ∼19 cm of core and the slope immediately above the peak (∼19 to ∼25 cm) were not reproduced. The constant 137Cs activities in the top ∼19 cm of the core, despite a clear slope in the 210Pb profile over this depth range, could only be reproduced if there was continued 137Cs input to the sediment over and above that predicted from fallout. Possible sources include sediment (e.g., via sediment focusing) or material from the terrestrial watershed (37). Hence, a 137Cs flux over and above that expected from fallout was included. The best fit to both the 210Pb and the 137 Cs profiles, assuming such 137Cs input over and above fallout, resulted from a sedimentation rate of 0.8 cm year-1 (mass accumulation rate of 0.27 g cm-2 year-1) and a Db value of 3.4 cm2 year-1 (Figure 4d,e). For the LITN core, the best fit to both the 210Pb and the 137Cs profiles was achieved with a sedimentation rate of 0.6 cm year-1 and a Db value of 110 cm2 year-1 (Figure 5a,b). Because the 137Cs data from the LILN core do not show a clear maximum, nor do they decrease to near-zero values at the bottom of the core (Figure 4 in ref 14), sedimentation and bioturbation rates could not be estimated with confidence. The sedimentation rates and sediment mixing rate coefficients determined in this study are near or within the ranges previously determined in LIS sediment (∼0.06 to ∼0.6 cm year-1 and ∼6 to ∼110 cm2 year-1, respectively; Figure 1 of ref 38). Total chlordane (TC + CC) profiles were also simulated with the numerical model using the sedimentation rate and mixing rate coefficient derived from the best fit of radionuclide data, to examine the possibility of continued input after the 1980s. Annual chlordane input to surface waters was assumed to be proportional to records of chlordane usage in the U.S. (Table S3), adjusted to match the observed total sediment inventory. On the basis of these data, the chlordane input increased gradually from 1948 to 1964, peaked in 1971, decreased to low values by 1985, and ceased by 1995. For the LIMB core, using the model parameters obtained from the best fits of the 210Pb and 137Cs data (Table 1), the onset of chlordane in the sediment and the depth of the chlordane maximum were fit reasonably well (Figure 4e). However, the modeled surface sediment contained much less chlordane than observed, suggesting that there was significant chlordane input to LIS after the 1980s. The data from the other cores do not help to constrain post-1980 chlordane inputs due to very high sediment mixing, as discussed next. Significant chlordane input to LIS is supported by observations of unusually high chlordane concentrations in pond sediment from western Long Island (39). Modeling of the radionuclide profiles indicated that both sedimentation and sediment mixing were occurring at all three sites where such data were generated (Table 1). The lack of a well-defined peak in both chlordane and 137Cs in the LITN and LILN cores can readily be explained in light of the inferred rates of sedimentation and sediment mixing. The relative influence of sedimentation and sediment mixing in transport of particulate materials in the sediment-mixed layer can be assessed using the dimensionless Peclet number (Pe ) szb/Db, where zb is the mixed layer depth) (38, 40). Large values of Pe imply that sedimentation dominates the vertical transport of sediment, whereas small values (
