Environ. Sci. Technol. 2003, 37, 667-672
Downloaded via AUSTRALIAN NATL UNIV on July 30, 2018 at 17:20:47 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Global Distribution and Budget of PCBs and HCB in Background Surface Soils: Implications for Sources and Environmental Processes S. N. MEIJER,† W. A. OCKENDEN,† A. SWEETMAN,† K. BREIVIK,‡ J . O . G R I M A L T , § A N D K . C . J O N E S * ,† Environmental Science Department, Institute of Environmental and Natural Sciences, Lancaster University, Lancaster LA1 4YQ, U.K., NILU, Norwegian Institute for Air Research, P.O. Box 100, N-2027 Kjeller, Norway, and Department of Environmental Chemistry, Institute of Chemical and Environmental Research (ICER-CSIC), Jordi Girona 18-26, 08034 Barcelona, Catalonia, Spain
This paper presents data from a survey of polychlorinated biphenyls (PCBs) and hexachlorobenzene (HCB) concentrations in 191 global background surface (0-5 cm) soils. Differences of up to 4 orders of magnitude were found between sites for PCBs. The lowest and highest PCB concentrations (26 and 97 000 pg/g dw) were found in samples from Greenland and mainland Europe (France, Germany, Poland), respectively. Background soil PCB concentrations were strongly influenced by proximity to source region and soil organic matter (SOM) content. Most (>80%) of the estimated soil PCB burden remains in the “global source region” of the Northern Hemisphere (NH) temperate latitudes (30-60° N) or in the OM-rich soils just north of that. %SOM correlated with PCB and HCB in the global data set, with the correlation coefficients being greater for HCB and the lighter PCBs than for heavier homologues. OM-rich soils in the NH consistently contained the highest burdens; such soils are a key global compartment for these compounds. Evidence for global fractionation of PCBs was found in the subset of soils from latitudes north of the global source region but was not discerned with the global data set. The full data set was used to estimate the burden for individual congeners/homologues in surface background soils and a global soil total PCB burden of 21 000 t. The significance of the inventory is briefly discussed in relation to the latest estimates of global production and atmospheric emission.
Introduction Soils are an important reservoir for many persistent organic pollutants (POPs) (1). Certain soils receive inputs of such chemicals at contaminated sites (from pesticide usage, sewage sludge disposal, and agricultural amendments), but on a global scale, very little of the soil surface receives such * Corresponding author e-mail:
[email protected]; phone: +44-1524-593972; fax: +44-1524-593985. † Lancaster University. ‡ Norwegian Institute for Air Research. § Institute of Chemical and Environmental Research. 10.1021/es025809l CCC: $25.00 Published on Web 01/16/2003
2003 American Chemical Society
direct inputs. In contrast, all soilsseven in remote areass receive inputs of POPs from atmospheric deposition. Only 0.04% of the global landmass is defined as “urban”; hence, “background soils” are a key environmental compartment to consider with respect to the global inventory of POPs. Various processes that can affect the global fate and budget of these compounds also occur in soils, namely, exchange with the atmosphere, biodegradation, formation of strongly bound (nonextractable) residues, and transfer to depth. POPs are very persistent in soils (1). Hence, as POPs are released or escape into the environment, the burden in soils globally becomes a complex function of the balance between inputs and losses. Similarly, the distribution of POPs in global background surface soils is a complex function of proximity to source regions, the long-range atmospheric transport (LRAT) potential of the POP in question, environmental/ climatic conditions and properties of the overlying vegetation, and soil type/system. Climate, soil, and canopy/vegetation type affect the processes and rates of air-surface exchange, degradation, and soil mixing/burial. Considerable information relating to POP sources, global distribution, budgets, environmental processes, and fates can therefore be gained from studying POPs distribution in background soils. In this paper, data are presented on polychlorinated biphenyls (PCBs) and hexachlorobenzene (HCB) in nearly 200 background surface soil samples collected from around the world. An estimate of the burden of these compounds currently residing in this environmental compartment is derived. The analysis of the data includes a subset of European background soils, from 41 sites in the U.K. and Norway that have recently been used to investigate the influences of SOM content, land cover, and global fractionation (2). Soil concentrations and distributions of these compounds are discussed with respect to sources and environmental processes.
Methods Soil Sampling. Surface soil samples (0-5 cm) were collected from 191 sites worldwide in 1998 (Figure 1). Sites were chosen to be remote from potential sources (i.e., away from populated/industrialized/agrochemical application areas) to ensure that they were representative of background levels in the areas from which they were collected. Soil sampling kits and instructions were sent to volunteers around the world, who were asked to choose sites that were far away from towns and cities, >2 km from busy roads, and >500 m from small dwellings and tracks. All samples were collected from the surface 0-5 cm in triplicate using a hand-held coring device. Overlying vegetation was removed prior to collection of the sample. Samples were double-wrapped in solvent-rinsed Al foil, sealed in plastic bags, and frozen prior to sending them to Lancaster University, where they were stored frozen until required for extraction. Separate samples were also collected at most sites to determine the bulk density. Analytical Methods. Samples were defrosted and mixed with sodium sulfate and then Soxhlet extracted for 12 h in dichloromethane (DCM) and 12 h in toluene. DCM and toluene extracts were combined after extraction. Extracts were cleaned using silica/alumina chromatography, followed by size-exclusion chromatography (S-X3 biobeads) (2). If necessary (i.e., if there was a precipitate formation on sample volume reduction), samples were treated with concentrated sulfuric acid before the size-exclusion chromatography step. They were analyzed by GC-MS for PCBs and HCB as described previously (2). Congeners included in the sum of the homologue groups and total PCBs are tri-PCB 18, 31, 28; VOL. 37, NO. 4, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
667
FIGURE 1. Map showing all sampling sites.
TABLE 1. Mean, Minimum, and Maximum Concentrations (pg/g dw) of Selected Compounds HCB PCB 28 PCB 52 PCB 90/101 PCB 118 PCB 153/132 PCB 138 PCB 180 ∑tri ∑tetra ∑penta ∑hexa ∑hepta ∑octa total PCB total PCB/unit SOM
mean
minimum
maximum
680 51 63 320 440 720 930 390 110 260 1 510 2 580 920 180 5 410 19 500
10 1.0 0.5 1.0 0.5 3.0 2.4 1.1 3.2 1.5 5.0 10 3.3 1.5 26 330
5 210 760 2 400 7 090 6 680 17 900 16 500 6 580 1 510 6 580 24 800 50 900 15 900 2 290 96 900 335 000
tetra-PCB 52, 49, 44, 70; penta-PCB 95, 90/101, 99, 87, 110, 123, 118; hexa-PCB 151, 149, 153/132, 141, 138, 158; heptaPCB 187, 183, 180, 170; and octa-PCB 199, 203, 194. %SOM was determined by loss-on-ignition. QA/QC. Blank samples were included at a rate of one for every five soils extracted. All results were blank corrected. The limit of detection (LOD) was calculated as three times the standard deviation (SD) of the mean blank. Concentrations below the LOD were replaced by half the LOD. All samples were spiked with a labeled recovery standard (containing 13C analogues of PCB 28, 52, 101, 153, 138, and 180) prior to extraction. Excellent recoveries were obtained, averaging 89 ( 21% for [13C]PCB 28, 92 ( 17% for [13C]PCB 52, 95 ( 16% for [13C]PCB 101, 103 ( 18% for [13C]PCB 153, 104 ( 16% for [13C]PCB 138, and 110 ( 23% for [13C]PCB 180, in all samples.
Results General Discussion of the Results. Table 1 presents a summary with the mean, minimum, and maximum concentrations (pg/g dw) for HCB, selected PCB congeners (PCB 28, 52, 90/101, 118, 153/132, 138, and 180), the PCB homologue groups, and the sum of all quantified PCBs. The full data set is available as Supporting Information. Differ668
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 4, 2003
FIGURE 2. Latitudinal distribution of total PCB in soil (panel a, pg/g dw; panel b, pg/g om) as related to PCB usage (from ref 3). ences of up to 4 orders of magnitude were found between sites; differences of over 3 orders of magnitude could occur within relatively small areas (within individual countries, e.g., France, Germany, and Norway). The lowest and highest PCB concentrations (26 and 97 000 pg/g dw) were found in samples from Greenland and mainland Europe (France, Germany, Poland), respectively. The lowest and highest HCB concentrations (10 and 5200 pg/g dw) were found in samples from Bear Island (Norway) and Europe (south Norway, Russia, U.K.), respectively. Clearly, there are substantial differences in background soil concentrations globally. It is appropriate to consider which factors cause this range. General Spatial Patterns. Figure 2 shows total PCB concentrations (expressed as both pg/g dw and pg/g SOM) ) as a function of latitude. The estimated cumulative historical usage of PCBs between 1930 and 2000 is also shown, compiled as described elsewhere (3) and spatially distributed on a 1°
FIGURE 3. Latitudinal distribution of HCB in soil (panel a, pg/g dw; panel b, pg/g om). Note: When expressed on an organic matter basis, there is one clear outlier, a very low organic matter forest soil (0.2% OM) from Brazil. × 1° grid (4), using population density as a surrogate to distribute the national usage data. About 86% of the total global usage of PCBs is estimated to have occurred between 30° N and 60° N (3). The latitudinal distribution of soil concentrations closely reflects this, with highest concentrations in Northern Hemisphere (NH) temperate locations (ca. 30-70° N). Most of the PCB burden emitted to the environment is in or close to the global source regions. Many of the highest PCB concentrations measured in the survey were to the north of the zone of maximum global usage (see Figure 2). Three factors are likely to have been influential in causing this observation. First, surface soils with the highest SOM contents globally occur around these latitudes, associated with forests, peat bogs, grassland systems, etc. (see Figure S1). These POPs have a strong affinity for SOM (2, 5), while vegetation canopies may enhance scavenging from the atmosphere and deposition to the soil (6). Second, persistence of PCBs in soils is controlled by climatic factors (1, 7). Enhanced deposition and greater persistence at higher latitudes, as compared to more limited depositionandpersistenceatlowerlatitudes,couldcontributes in partsto the trends in Figure 2. The third factor may be related to emissions. Although most of the historical emission of PCBs occurred between 30° and 60° N, measures to reduce their production and use were first introduced between these latitudes in western Europe, Japan, and America in the 1970s, while production of PCBs did not end in Russia until 1993 (3). The pattern of more recent emissions may therefore have shifted northwards relative to the historical cumulative usage (8). HCB concentrations (pg/g dw and pg/g SOM) are plotted against latitude in Figure 3. Differences in concentration between the various sites decrease markedly when results are expressed on a SOM basis. To our knowledge, no spatially resolved global usage or emission data are available for HCB. However, as for PCBs, the highest concentrations on a dry weight basis are clearly in the NH and were generally found in Europe. HCB environmental sources include pesticide usage, manufacturing, and combustion (9). Regressions. Regressions were calculated between soil concentration and the following variables: %SOM; estimated
PCB compound usage; estimated PCB emissions; population density; latitude. Organic Matter. The sum of the homologue groups, total PCB, and total HCB (pg/g dw) were regressed versus %SOM. All data were natural log-transformed to reduce scatter and allow linear fits. Table 2 shows the slope, 95% confidence interval (CI), r2, and significance of the regressions. All regressions are highly significant (at the p < 0.001 level). The r2 value decreased with increasing chlorination. Table 2 also presents an analysis when the soil samples were considered as subsets and split into three latitudinal bands, to coincide with the main PCB source area and the global regions north and south of that (i.e., south of the main source region: 90° S-30° N; source region: 30-60° N; north of the main source region: 60-90° N). All regressions are again highly significant (p < 0.001). Generally, the slopes and r2 of the regressions were highest in the northernmost region and lowest in the southernmost region. A further analysis with %SOM was conducted on the basis of soil land use/vegetative cover. This was to assess whether an effect on PCB concentrations because of high SOM content alone (e.g., peat samples) or a “canopy effect” could be observed. Samples were divided into the categories “F” for forest and “O” for open land. The O category included many different types of land with no or a relatively minor vegetative canopy, such as tundra, parkland, pasture, and “scrub”. Also included in this category were samples for which no field description was available. Figure 4 shows regressions of total PCB versus %SOM for the different categories (data log-transformed). The forest sites show slightly higher concentrationsson average by a factor of ∼2 (Figure 4)sthan the open sites for the same amount of SOM (statistically significant at the p < 0.01 level). This is in line with previous results (2). The %SOM is clearly an important variable that influences the concentrations of POPs in global background soils. The strength of this relationship is greater for HCB and the more volatile PCB congeners than for the heavier homologues (Table 2). This implies that the processes of atmospheric transport and air-surface exchange are important in influencing soil concentrations. Land use/vegetative cover presumably also influences the rate and processes of air-surface exchange and the dynamics of POPs within the soil of different ecosystems. %SOM has now been shown to influence soil PCB concentrations at the local (5), regional (2), and global scale (this study). With repeated air-surface exchanges, POPs would move toward an air-SOM equilibrium condition (2, 5). However, other competing processes (namely, localized emissions, degradation, burial, and strong retention by soils) will act against this state being achieved. Usage and Atmospheric Emissions. Figure 2 shows that the latitudinal distribution of global PCB concentrations in soil broadly follows estimated usage (3). This relationship was explored in more detail by investigating the correlation between soil concentrations and spatially distributed usage estimates derived on a 1° × 1° grid. Grid numbers were assigned to each soil sample on the basis of latitude and longitude of the sampling site, following the method used by Li et al. (10). Each sample was then paired with the usage calculated for that grid cell by Breivik et al. (3). If no usage data were available, the nearest grid cell was used for which usage information was available while checking that this would be representative for the site. Statistical information for the regressions, performed on log-transformed data, is given in Table 3. Significant (p < 0.001) correlations were found for measured total PCB soil concentrations with the estimated usage data when the data were expressed on dry weight or SOM basis. Expressing the data on a SOM basis slightly improved the correlation. The same exercise was performed for a selection of the PCB congeners to further investigate whether differences in VOL. 37, NO. 4, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
669
TABLE 2. Statistical Information from Regressions with % Organic Mattera all data slope (95% CI)
slope (95% CI)
r2 significance (p)
slope (95% CI)
r2 significance (p)
0.709
