Air−Boreal Forest Transfer and Processing of Polychlorinated

Jun 12, 2009 - Centre for Chemicals Management, Lancaster Environment Centre, Lancaster University, LA1 4YQ Lancaster, U.K., Department of Public ...
0 downloads 0 Views 1MB Size
Environ. Sci. Technol. 2009, 43, 5282–5289

Air-Boreal Forest Transfer and Processing of Polychlorinated Biphenyls CLAUDIA MOECKEL,† LUCA NIZZETTO,† BO STRANDBERG,‡ ANDERS LINDROTH,§ A N D K E V I N C . J O N E S * ,† Centre for Chemicals Management, Lancaster Environment Centre, Lancaster University, LA1 4YQ Lancaster, U.K., Department of Public Health and Community Medicine, Sahlgrenska Academy, University of Gothenburg, SE-40530 Gothenburg, Sweden, and Department of Physical Geography and Ecosystems Analysis, Lund University, So¨lvegatan 12, 22362 Lund, Sweden

Received December 16, 2008. Revised manuscript received May 20, 2009. Accepted May 21, 2009.

The exchange of persistent organic pollutants (POPs) between different compartments of a typical mature boreal forest was investigated. The study focused on fluxes of polychlorinated biphenyls (PCBs) between the atmosphere, vegetation and soil, and within the soil to assess whether this type of forest acts as a final sink or temporary repository for POPs. The study, at a Swedish site, suggested total PCB air-to-forest floor fluxes of 1.4 µg m-2 year-1. Much of this could be attributed to compounds bound to particles that may originate from needle surfaces. Degradation half-lives in soil between 6.4 and 30 years for tetra- to hepta-PCBs were obtained using a mass balance approach. This field data-based method derived degradation rates of POPs in background soils, although it may have underestimated the persistence of the heavy PCB congeners. Compounds reaching the forest soil appear to be stored efficiently and degraded slowly. As a first approximation, applying the findings from this study site to boreal forests on a global scale suggests that 2-21% (depending on the congener) of the estimated global atmospheric emission deposits to these ecosystems.

Introduction Forests are believed to play a major role in the global fate of persistent organic pollutants (POPs), acting as efficient transmitters for these compounds from the air to soils (1-4), which have been identified as a very important terrestrial storage compartment (5, 6). The so-called “forest filter effect” increases air-to-soil fluxes of POPs (1-3, 7) and could also reduce their long-range atmospheric transport (LRAT) (8) or enhance vapor pressure dependent fractionation (9). Even though the leaf area index (LAI) of forests in boreal regions is often lower than in temperate ones, they may still be important for removing POPs from the air because the cool climate supports the cold condensation effect (9), inducing these pollutants to deposit to available surfaces and re* Corresponding author phone: +44 (0)1524 593972; fax: +44 (0)1524 593985; e-mail: [email protected]. † Lancaster University. ‡ University of Gothenburg. § Lund University. 5282

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 14, 2009

straining revolatilization. Estimates suggest that boreal forest soils are able to store large amounts of POPs if they receive them (10). A number of studies have focused on characterizing air-leaf uptake pathways, kinetics, and storage locations within vegetation (11, 12). Other surveys investigated burdens in single compartments of forest systems, i.e., foliage (13, 14) or soil (15-17). So far, few studies have linked the ecosystem compartments by measuring/estimating fluxes between them (1-4, 7), and although the transfer of POPs to soil was considered to be an important factor in influencing the environmental fate of these chemicals, their behavior after depositing to the ground, including processes within the soil profile, has been little studied (17, 18). Soils could potentially act as ultimate sinks for POPs if they are able to retain or degrade these chemicals or as sources if they release them during organic matter decomposition (10). Such processes should, therefore, be considered. The present paper assesses the interaction of POPs between different compartments of a boreal forest ecosystem at a site well studied for other purposes mainly related to the determination of energy, water, and CO2 fluxes. Air (19), vegetation, litter, and soil samples were analyzed for POPs. However, the main focus here is to determine and discuss fluxes of polychlorinated biphenyls (PCBs) within this system in the medium-long term, to investigate whether this typical mature boreal forest acts as an actively exchanging capacitor or final sink for such pollutants, and to provide information on the potential role of boreal forests, in influencing the distribution and fate of POPs.

Data Collection All samples were collected in a boreal forest in the vicinity of the flux tower at Norunda Common, a well studied site 30 km north of Uppsala, Sweden (60°05′ N, 17°29′ E). The tower is located within a 100 year old stand of pine (Pinus sylvestris) and spruce (Picea abies) with a canopy height of 25 m. The site is surrounded by a 1-20 km wide belt of stands of these species of various age and height. Vegetation samples (Scots pine and Norway spruce needles and leaves of the most abundant herb level vegetation) were collected on the 10th of July 2007. In addition, using litter traps, freshly fallen litter was collected from the 25th of May to the 10th of July (i.e., covering the 6 week air sampling period) and separated with tweezers into needle and non-needle litter prior to analysis. Soil profile samples were taken from nine different layers including the litter layer (Oi) (some soil horizons were subdivided; see Table SI-1 in the Supporting Information) from a pit (50 cm × 50 cm), carefully avoiding cross-contamination between layers. Further details regarding the sampling site, collection of samples, processing, chemical analysis, and sample properties can be found in the Supporting Information (Text SI-1); details on air sampling are given in ref 19. An overview of PCB concentrations, found in the different sample matrices, is given in Table 1a. The results agree well with findings from other studies (14, 15, 17) and were used for the determination and discussion of fluxes in the following sections.

Results and Discussion Estimation of Air-to-Forest Fluxes. Given the lack of a standardized or direct method for the measurements of POP exchange between air and forest surfaces, air-to-forest ground fluxes have been estimated using three separate approaches: (i) estimating the amount of pollutants present 10.1021/es803505u CCC: $40.75

 2009 American Chemical Society

Published on Web 06/12/2009

VOL. 43, NO. 14, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5283

3.8 (19) 1.5 (7.6) 0.4 (2.0) 0.9 (4.3) 1.4 (6.9) 0.2 n.d. 26

52 90 + 101 118 138 153 + 132 180 203 ΣPCB

22 (0.6) 63 (1.7) 21 (0.6) 91 (2.5) 110 (3.1) 37 6.5 (0.2) 800

S-3b (pg g-1)

52 90 + 101 118 138 153 + 132 180 203 ΣPCB

PCB congener

22 (0.8) 57 (1.9) 14 (0.5) 80 (2.8) 88 (3.0) 29 n.d. 700

S-1b (pg g-1) 170 (3.7) 200 (4.3) 130 (2.9) 110 (2.4) 130 (2.8) 46 6.1 (0.1) 1500

P-3b (pg g-1)

2.0 5.8 1.9 8.3 10 3.4 0.60 73

spruce (ng m-2 year-1)

52 (1.2) 160 (3.6) 63 (1.4) 68 (1.5) 92 (2.1) 44 n.d. 1000

P-1b (pg g-1)

vegetation

41 (1.4) 85 (2.8) 38 (1.3) 94 (3.1) 110 (3.6) 30 n.d. 1000

18 (0.4) 41 (0.9) 44 (1.0) 98 (2.1) 96 (2.1) 46 6.7 (0.1) 650

L-nonNc (pg g-1) 120 (0.4) 400 (1.4) 380 (1.4) 770 (2.8) 900 (3.2) 280 45 (0.2) 5200

Oi (pg g-1)

30 34 23 19 23 8.0 1.1 260

8.2 23 10 27 32 8.5 1.0 230

Oe-1 (pg g-1)

fresh litter

93 (0.2) 510 (1.3) 460 (1.2) 1100 (2.8) 1100 (2.8) 390 50 (0.1) 7000

L-Nc (ng m-2 year-1)

(b) Air-to-Ground Fluxes

33 (1.0) 93 (2.7) 39 (1.1) 110 (3.2) 130 (3.8) 34 4.2 (0.1) 920

L-Nc (pg g-1)

fresh litter

pine (ng m-2 year-1)

understorey vegetation (pg g-1)

(a) Dry Weight-Based Concentrations

3.0 6.6 7.2 16 16 7.4 1.1 100

Oe-3 (pg g-1) 170 (0.1) 1800 (1.0) 1800 (1.0) 4600 (2.6) 4800 (2.7) 1800 340 (0.2) 27000

L-nonNc (ng m-2 year-1)

69 (0.1) 660 (1.0) 630 (1.0) 1700 (2.7) 1800 (2.8) 640 95 (0.1) 10000

Oe-2 (pg g-1)

Oa-1 (pg g-1)

33 110 100 200 240 74 12 1400

Oa-2 (pg g-1) 14 (0.2) 95 (1.1) 83 (0.9) 230 (2.6) 250 (2.8) 88 19 (0.2) 1500

soil Oi layer (ng m-2 year-1)

110 (0.1) 1100 (1.1) 1000 (1.0) 2800 (2.8) 2900 (2.9) 1000 220 (0.2) 17000

soil layers

75

9.0 (2.0) 11 (2.3) 14 (3.1) 4.5

A (pg g-1)

46

1.4 3.7 2.7 2.2 5.0

B (pg g-1)

(a) Dry weight-based concentrations in different forest compartments and ratios between concentrations of individual congeners and PCB-180 if applicable (values in brackets). (b) Air-to-ground fluxes, estimated from burdens found in different forest compartments. The results of all individual PCB congeners analyzed in all samples are given in the Supporting Information: Table SI-2 (dw-based concentrations) and Table SI-4 (fluxes). For better illustration, congener ratios are also shown in Figure SI-1. b Sample names: S-1, spruce needles, current year; S-3, spruce needles, third year; P-1, pine needles, current year; P-3, pine needles, third year. c Sample names: L-N, litter consisting of needles; L-nonN, litter consisting of material other than needles.

a

air (pg m-3)

PCB congener

trees

TABLE 1. Selected PCB Congeners in the Forest System Studieda

FIGURE 1. Processes contributing to air-to-forest ground fluxes of POPs and their half-lives in the soil. Italicized processes have been considered in the approach used to estimate degradation half-lives. in the conifer foliage annually deposited to the ground, (ii) estimating the amount of POPs that deposit bound to fresh litter, and (iii) measuring the amount of POPs present in the soil top layer and estimating the material’s residence time in this layer. Figure 1 provides a schematic showing processes that may have affected POP concentrations in the forest compartments analyzed to provide input data for the different approaches. This schematic illustrates that the three flux estimating approaches presented consider different mechanisms/pathways of POP accumulation. Approach i only accounts for gaseous exchange and wet and dry deposition of particle-bound POPs to conifer needles. Approach ii additionally considers POPs taken up by other above ground tree compartments. Approach iii also includes POPs deposited from understorey vegetation and in association with airborne particles not trapped by conifers and particles originating from needle surfaces (21), wet depositions, and net gaseous exchange between air and soils. Approach i. Vegetation has extensive hydrophobic surfaces which can sequester airborne PCBs, eventually transferring them to the soil when foliage is shed. The first approach, therefore, used concentrations in spruce and pine needles, the most abundant vegetation in this ecosystem. From concentrations in the older needles, yearly PCB airto-forest ground fluxes (Fveg, ng m-2 year-1) were calculated as Fveg ) cpinempine/year + csprucemspruce/year

(1)

where cpine and cspruce (ng g-1) are dry weight-based PCB concentrations in pine and spruce needles and mpine/year (170 g m-2 year-1) and mspruce/year (92 g m-2 year-1) are the forest ground area specific biomasses (dry weight) of the portions of needles that were assumed to be transferred to the ground annually (see Table SI-5 and Text SI-3 in the Supporting Information for more details). mpine/year and 5284

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 14, 2009

mspruce/year were based on the number of needle generations present on the trees (3 generations for pine, 8 for spruce), taking into account differences between the average weight of “new” needles (i.e., those that emerged in the year of sampling) and older ones. This resulted in Fveg for the sum of all PCB congeners analyzed (ΣPCB) of 330 ng m-2 year-1 (see Table 1b). This flux does not account for the contribution of understorey vegetation, tree parts other than needles, deposition of atmospheric particles, wax shedding, and direct air-soil exchange and will, therefore, underestimate air-to-forest system fluxes. Comparison of this approach with the others can provide insight into processes affecting POPs after they have accumulated in vegetation. Approach ii. This uses concentrations of PCBs found in fresh litter in the needle fraction and in other (non-needle) material (clitter-N and clitter-nonN, ng g-1) collected in litter traps. Air-to-forest ground fluxes Flitter (ng m-2 year-1) were calculated as Flitter ) clitter-Nmlitter/year flitter-N + clitter-nonNmlitter/year flitter-nonN (2) where mlitter/year (410 g m-2 year-1) is the forest area specific dry weight of annually produced litter (22) and flitter-N (0.61 g g-1) and flitter-nonN (0.39 g g-1) are the fractions of needles and other material that contribute to the total mass of litter, respectively. Unlike the first estimate, this also accounts for bark, twigs, and cones. The flux of PCBs associated with nonneedle material represents between ca. 21 and 50% of the total amount associated with the litter (see Table 1b), with the lowest contribution from non-needle material generally observed for relatively volatile compounds such as tetraCBs. For the sum of all PCBs analyzed, a flux of 330 ng m-2 year-1 was obtained (Table 1b), consistent with the estimated Fveg. However, Flitter was lower than Fveg for lighter PCBs (tetra-

and some pentahomologues) but higher for most of the heavier congeners. This shift in the congener pattern will be discussed in the next section. Approach iii. This estimates air-to-forest ground fluxes from the amounts stored in a layer of the soil with a known age, assuming no loss due to factors other than possible revolatilization. The layer under consideration needs to have been formed from vegetation that was exposed to known and relatively stable PCB concentrations. Therefore air-toforest ground fluxes FOi (ng m-2 year-1) were estimated from PCB concentrations found in the soil litter (Oi) layer (cOi, ng g-1), which is the only soil horizon that fulfills these criteria. Given an estimated residence time for material in the Oi layer (xOi) of 2.8 years (20) (see Supporting Information Text SI-2) and the forest ground area specific mass of the Oi layer mOi (750 g m-2), FOi was calculated as FOi )

cOimOi xOi

(3)

This approach gave a ΣPCB flux of 1400 ng m-2 year-1 (Tables 1b and SI-4 in the Supporting Infomation show fluxes of individual congeners). This greatly exceeds Fveg and Flitter, even when considering the uncertainty associated with the parameters (i.e., assuming a relative standard deviation of 30% for mOi particularly due to variations in the thickness of this layer and 10% for xOi according to ref 20). According to emission estimates (23), air concentrations might be expected to be up to 1.3 times higher when the vegetation that formed the Oi layer was growing. The contribution of depositing plant materials not accounted for by the vegetation or by the fresh litter samples (mainly herb level vegetation) could also enhance deposition fluxes to the ground. However, since dry weight (dw)-based concentrations in the Oi horizon are also higher than those in vegetation and fresh litter for nearly all compounds analyzed (see Tables 1b and SI-2 in the Supporting Information), there must be additional processes not accounted for by the other methods, affecting air-to-soil fluxes. Dry weight loss due to organic matter mineralization cannot explain the discrepancy, because soil litter is defined as the layer that contains poorly decomposed material (up to 10%). This suggests that the soil litter layer further accumulates vapor phase PCBs from the atmosphere or acts as a filter for depositing particles containing these chemicals (or both). Since direct air-soil exchange was assumed to be kinetically limited by fresh litter covering the surface (24) and, thus, to be less efficient than air-vegetation exchange, deposition of particle-bound PCBs is believed to contribute significantly to the load of pollutants in the soil. A process likely to be responsible for this is the continuing production of fresh waxes at the inside of the cuticle and shedding of old waxes from the surface of needles (25). Abrasion of wax particles may explain why the dw-specific wax content did not increase from first- to third-year spruce needles, although changes in the wax composition with needle age were observed for spruce in earlier studies (21, 25), showing that new waxes are produced. For pine, the ratio between the amounts of waxes produced and those shed in the same period was apparently higher than for spruce, allowing for rising levels of POPs in pine needles with increasing needle age, in contrast to spruce needles. These results indicate that air-to-soil fluxes of POPs, and, therefore, also the role of the boreal forest ecosystem as a sink for such chemicals in the environment would be significantly underestimated if calculations are based on concentrations in vegetation only. Approach i does not account for this process, simply because it regards needles as static compartments that do no release matter and pollutants during their lifespan, whereas the sampling device used to collect fresh litter for

approach ii was probably not able to retain submillimeter scale particles efficiently. PCB concentrations in the air and air-to-forest fluxes estimated from the concentrations in the Oi layer (method iii) agree within a factor of 2 with airborne levels and deposition fluxes measured by means of deposimeters by Brorstro¨m-Lunde´n and Lo¨fgren (3) in a different Swedish spruce forest between 1991 and 1995. Yearly deposition fluxes estimated in the current study were lower than those observed in two other flux studies: one conducted at both a coniferous and a deciduous forest site in Germany in 1995/ 96, where PCB concentrations in the air were higher by a factor of about 4 than those presented here (1), and one conducted in a Canadian deciduous forest with PCB concentrations in the air being similar to the ones reported now (4). This reflects the dependence of deposition fluxes of POPs on air concentrations but also implies a greater ability of very productive deciduous forests to remove PCBs from the atmosphere than coniferous ones. Observations on Air-Vegetation Transfers. An important issue for the fate of PCBs is whether vegetation reaches equilibrium with the air within the lifetime of needles or leaves. Approaching the vegetation’s capacity to accumulate PCBs would not only limit the removal of these compounds from the air but also may result in their release to the atmosphere soon after needles fall to the ground and decompose gradually. The spruce data may suggest equilibrium conditions being reached rapidly, because concentrations of PCBs remained constant on a dw basis from firstyear to third-year needles and decreased on a lipid weight (lw) basis (Tables SI-2 and SI-3 in the Supporting Information). Dry weight-based concentrations in pine needles increased with needle age as seen in other studies (e.g., refs 26 and 27). This may result from the increasing lipid content since lw-based PCB concentrations decreased with age (Table SI-3 in the Supporting Information). This was observed for the full range of PCBs investigated, even the heaviest ones. However, according to other studies (28, 29), heavy PCBs are very unlikely to have reached equilibrium between air and needles within the few weeks first-year needles had been exposed to the atmosphere between emerging and being sampled. Two processes, possibly occurring at the same time, may explain the observations. Uptake of PCBs might be plant side controlled at least for light congeners and, due to steady state conditions between the air and surface waxes of the cuticle (30), uptake may slow down long before internal parts have reached equilibrium. The results obtained for needle samples also support the wax-shedding hypothesis (21), particularly since comparison with high amounts of PCBs found in soil litter (Oi layer) suggests that needles take up these chemicals over their whole lifetime. A combination of these processes would also explain congener patterns of PCBs observed in air, vegetation, and litter (Table 1a and Figure SI-2 in the Supporting Information). The ratios between concentrations of selected PCB congeners and ΣPCB show that in both spruce and pine a distinctive shift toward heavier congeners occurred between air and needles, i.e., during the uptake of PCBs from the atmosphere by vegetation, indicating plant side limitation. Deposition of compounds bound to particulate matter may have contributed to this shift, but it is unlikely to explain its full extent as even at low temperatures only the heaviest congeners (hepta- and higher chlorinated PCBs) show significant particle bound fractions (19). However, no difference was observed between firstand third-year needles, suggesting that even the less hydrophobic congeners had not yet reached equilibrium between air and whole needles of either age class. The less pronounced congener pattern shift observed between air and pine needles compared to that between air and spruce needles may result from differences in the cuticle structure VOL. 43, NO. 14, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5285

of the two species, allowing for faster transfer of POPs from surface waxes deeper into the cuticle or whole needles of pine than can be achieved in spruce (27). Presumably the congener patterns changed only little during the transition from vegetation to litter, even though the analysis suggested a shift toward lighter congeners for spruce and toward heavier ones for pine. However, this is probably because the needle fraction of the fresh litter sample contained both spruce and pine needles. A slight shift toward heavier PCBs was observed between the fresh litter sample and the soil litter layer, possibly due to some loss of light congeners caused by revolatilization, leaching, or degradation. In addition to differences in the cuticle structure, as proposed above, higher concentrations of PCBs found in first-year pine needles compared to that in first-year spruce needles may also be due to stronger self-shielding of spruce needles against the wind as they grow at a higher density on the twigs than on pine needles. This would increase the thickness of the laminar layer of air surrounding the needles and, therefore, reduce the uptake rate for gas phase PCBs from the air. However, no differences were found between spruce needles collected from branches at five different heights between 5 and 25 m. This indicates that wind speed differences throughout the canopy were not enough to affect uptake rates of PCBs significantly. Similar observations were made with air samples collected passively within the canopy (19). On average, the specific needle area (cm2 g-1 dw) of Scots pine was slightly higher (34-91 cm2 g-1, depending on the sampling height) than that of Norway spruce (24-72 cm2 g-1, depending on the sampling height) in the area studied (31), but this small difference in the specific surface areas available for uptake of PCBs cannot fully explain the higher uptake by pine needles. PCB Deposition to Boreal Forests. Comments on Upscaling. The deposition flux estimates obtained by approach iii were used to assess, in first approximation, the potential role of boreal forests worldwide in removing atmospherically emitted POPs. It is stressed that this attempt, using the few experimental data available, required a number of important assumptions: (i) A directly proportional relationship exists between the air-to-forest floor flux and the above ground biomass. According to Hellstro¨m (32), the dry weight of needles as the compartment that exposes the largest surfaces to the air is well-correlated with their surface area. (ii) The mass ratio between different above ground vegetation compartments (needles, branches, stem, stump, roots) at Norunda reflects that at other boreal forest sites. (iii) PCB concentrations in the air of boreal forest areas worldwide are similar to those at Norunda at a given time and, therefore, (iv) the flux of PCBs to boreal forest floors is limited by the available vegetation surfaces but not by the efficiency of the transport of PCBs from source regions to boreal forests. It is appropriate to consider these assumptions further. Although boreal forests are not very diverse ecosystems regarding their major species, differences in forest properties are likely to account for the largest uncertainty associated with the assumptions made above. The validity of assumptions i and ii is likely to be affected by species composition, growth, and decomposition rates, which are influenced by temperature, availability of water, and soil properties. Particularly in eastern Siberian areas, dominated by deciduous larch trees, air-to-forest ground fluxes may be lower than assumed here as the needles are directly exposed to the air for only part of the year. From the literature available to date, it is not possible to quantify the significance of such differences for estimating global air-to-boreal forest ground fluxes of POPs. To validate or improve the approach and estimate and reduce the uncertainty caused by the limited number of environmental samples and the assumptions made, we recommend the analysis of litter of a known age 5286

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 14, 2009

TABLE 2. Estimated Global Air-to-Ground PCB Fluxes to Boreal Forests Compared with Global Emission Estimates

PCB congener

air-to-ground PCB flux to boreal forests (kg year-1)

global PCB emissionsa 2003-2005 (ref 23) (kg year-1)

air-to-ground PCB flux to boreal forests (% of yearly emissions)

52 90 + 101 105 110 118 123 138 149 153 + 132 158 180 194 199

350 1100 410 930 1100 180 2100 1300 2500 180 770 100 19

21000 11000 7300 11000 16000 3000 11000 6200 13000 1500 4000 900 820

1.7 10 5.6 8.1 6.9 6.1 19 21 19 12 19 11 2.3

a Breivik et al.’s upper emissions were used here (ref 23). For details refer to the text.

from different boreal sites, covering all major forest types found in this climate zone. Given the high uncertainties associated with the global emission estimates used (23), the simplifications made with assumptions iii and iv may be reasonable, because boreal regions form a relatively narrow belt around the globe and Norunda is located approximately in the middle of its north-south extension. Air concentrations measured at different background monitoring sites in boreal regions differed only by about a factor of 2 (33). However, so far, relatively few data have been reported on airborne PCBs in boreal regions. Thus, apart from considering revised emission estimates as they emerge (e.g., using the approach introduced by Gasic et al. (34)), additional measurements of POP concentrations in the air should be taken into account when they become available, to increase the accuracy of the approach used here. Regarding the latter, an even coverage of the north-south as well as the east-west extension of boreal forests is needed to address the possible effect of temperature gradients and distance from source areas. These caveats need to be borne in mind in the following section. Initial Attempt at Upscaling. In order to estimate the total above ground biomass of boreal forests worldwide, a global map of boreal forests was combined with a 0.5° by 0.5° grid of the global above ground forest biomass (35) using ESRI ArcMap 9.1. A global above ground boreal forest biomass of 1.3 × 1014 kg was obtained. The air-to-ground PCB fluxes at Norunda (Table 1b, last column, derived from PCB amounts in the Oi layer) were related to the above ground biomass BMa at this site (details on the estimation of BMa are given in the Supporting Information, Text SI-3). The resulting global air-to-ground PCB fluxes of boreal forests, based on concentrations in the Oi layer at Norunda, are shown in Table 2 together with estimates of current global atmospheric annual emissions (23). Note that Breivik et al.’s upper emissions were utilized here, because they are supported by estimates of the global burdens of PCBs in the environment (5, 34). Depending on the individual PCB congeners considered, results suggest that 1.7-21% of the PCBs emitted to the air are transferred to boreal forest floors every year. However, given that boreal forests represent ca. 4% of the earth’s surface, the present study suggests that a significant percentage of PCBs emitted into the environment reaches boreal forest soils. This highlights the importance of these ecosystems for POP storage on a global scale compared to

other stores/sinks, such as oceans (36), terrestrial systems closer to sources (8), urban areas (37), and atmospheric degradation reactions (38). However, it should be mentioned that the portion of PCBs boreal forests appear to take up according to this calculation is sensitive to the assumed atmospheric releases from sources. This highlights again the need for precise and validated emission estimates, as emphasized in refs 23 and 34. Lower values for relatively volatile congeners may be linked to some volatilization during the earliest stages of the vegetation-soil transition, as suggested by Moeckel et al. (17), possibly supporting the global fractionation theory (9). Estimating Degradation Half-Lives in the Soil. Here, an approach is presented to assess the fate of POPs after they have deposited to the forest ground. Is the studied forest an ultimate sink for POPs or just an actively exchanging capacitor? As the only process resulting in ultimate removal of PCBs, degradation may be very important for determining the environmental fate of these pollutants. A mass balance approach was used to investigate the importance of degradation of PCBs within the boreal forest soil compared to other removal processes. Here, the assumption was made that the concentrations in vegetation at a given time reflect those in the air, which in turn reflect emissions in source regions (i.e., doubling of emissions results in doubling of both concentrations in the air and in vegetation). As a result of this, the vegetation that was growing during the period of peak PCB emissions formed the soil layer with highest concentrations. The vertical distribution of PCBs in the soil profile studied combined with age estimates of soil layers (details in the Supporting Information Text SI-2) support this assumption showing that intralayer transfer of PCBs is limited. To estimate degradation half-lives (t1/2,deg, years) of PCBs, their mass balance in the Norunda soil profile was assessed by regarding the Oi layer as the interface compartment with the atmosphere, which delivers pollutants to the layers below (in this case Oe and Oa). Mathematically, the variation in the amount U (ng m-2) of PCBs present in the soil core per unit of ground surface can be expressed as follows: dU ) FIn - FVol - FDiss - FDOC - FDeg dt

(4)

where FIn is the flux entering the soil system, FVol is the revolatilization flux from Oe and Oa, FDiss is the leaching flux of dissolved PCBs, FDOC is the leaching flux of PCBs associated with dissolved organic matter, and FDeg is the loss flux due to degradation. Biosynthesis of PCBs within the soil is not believed to occur. Estimation of these fluxes (ng m-2 year-1) and associated assumptions are described in the Supporting Information. Using yearly time steps, eq 4 was integrated numerically for the period between 1930 and 2005 (the period covering the history of estimated PCB emissions (23) assuming U1930 ) 0). This was done to estimate the current amount (year 2005 because of the delay between uptake of PCBs into vegetation and their transfer to the ground) stored in the soil core. As FDeg was unknown, an algorithm was implemented which searches for the value of t1/2,deg that minimizes the difference between the estimated and measured U2005. The model accounted for compound specific yearly emission estimates from 1930 to 2005 (23) and leaching rates (rleach) between 0.006 and 0.11% per year (Table SI-6 in the Supporting Information). Bioturbation was considered to be negligible, an assumption that is believed to be reasonable for boreal forest systems ((39), see also age estimates for soil layers given in the Supporting Information Text-2). Since horizon age estimates (see the Supporting Information) suggest that the vegetation that provided the organic material present in the mineral

FIGURE 2. Half-life estimates of PCBs vs log KOW. Filled circles represent the median half-lives obtained from 1000 repeated simulations. The bottom and top of the gray bars show the 25 and 75% percentile, respectively, and the ends of the whiskers give the 5 and 95% percentile. For details on the uncertainty analysis refer to the Supporting Information (Text SI-4). horizons did not receive PCBs from the atmosphere but rather by a downward transport process, these layers were not taken into account for the calculation of Udeep. Any revolatilization was believed to be limited to the uppermost layer, because compounds volatilizing from deeper layers are likely to be trapped by the material covering them and gas phase migration of PCBs in soils is minimal (24). Degradation half-life estimates derived from this approach ranged from 6.4 years (PCB-52) to 30 years (PCB158) with relative standard deviations ranging from 9.4 to 28% (see Figure 2 which also shows uncertainties predicted by a Monte Carlo analysis as described in the Supporting Information Text SI-4). For most compounds analyzed, t1/2,deg increased with increasing chlorine content and also with increasing hydrophobicity as both properties are linked. Enzymes of most PCB-degrading bacteria require two adjacent unchlorinated carbon atoms in order to oxidize and break down PCB molecules (40) and, even if compounds contain such carbon atoms, a high chlorine content will generally increase their persistence. Therefore, light congeners are usually easier to metabolize than heavy ones (40). Moreover highly hydrophobic compounds show a strong affinity to organic materials such as humic substances in the soil and are, therefore, difficult for microorganisms to access. Surprisingly, hepta- and octaCBs seemed to degrade faster than several hexachlorinated congeners; although in PCB-180, PCB-183, and PCB-187, no unchlorinated carbon atoms were present at a 2,3- or 3,4-position, where the molecules could be attacked by enzymes of aerobic PCB-degrading bacteria and anaerobic dechlorination could be excluded under the conditions found in the soil profile studied. The observation may be due to an artifact, as discussed in the Supporting Information Text SI-4; most likely, the historical deposition peak was overestimated compared to currently depositing amounts and/or this peak was expected at an earlier time than it actually occurred. This is supported by the fact that the emission estimate approach underestimated PCB-180 concentrations measured in the air at another Swedish site between 1991 and 1994 (3) by 10% whereas concentrations of lighter congeners were slightly overestimated. Using the compound specific leaching rates given in Table SI-6 in the Supporting Information, the amount of PCBs leached into deeper layers (Uleach) from 1930 to 2005 was estimated. These estimates compare well with the amounts observed in the A- and B-horizon (UA+B) (Table SI-6 in the Supporting Information), even though they may rather reflect VOL. 43, NO. 14, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5287

a worst-case scenario for the more water-soluble congeners (see Supporting Information Text SI-4). Assuming that rleach reflects real leaching rates reasonably accurately, this means that no significant amounts of PCBs were washed out of the soil profile and leaching is a relatively minor process in this part of the forest in general. Similar observations were reported by Gocht et al. (41) for polycyclic aromatic hydrocarbons (PAHs) in Southern German spruce forest soils. This result agrees well with the relatively high pH of the mineral soil at Norunda, which precipitates organic matter that was dissolved in the organic horizon. Therefore, POPs that are bound to dissolved organic matter (DOM) are prevented from being transported further downward, but also truly dissolved POPs can be retained if they bind to the precipitated organic matter. In more acidic soils, such as podzols, UA+B (particularly UB) may account for a much higher proportion of the soil profile’s total POP burden (17). Implications for the Global Fate of POPs. Due to the forest filter effect, caused by the large lipophilic surfaces trees expose to the air, and the cool climate in boreal regions, which enhance partitioning of POPs to surfaces (“cold condensation”), the estimated amounts of PCBs depositing to boreal forest soils would represent a significant portion of the estimated global atmospheric emissions (1.7-21%). Meijer et al. (5) estimated the burden in global background soils to be quite a small fraction (ca. 1.6%) of the estimated total mass of PCBs produced but a substantial percentage (16-56%) of those that have been emitted to the atmosphere according to emission estimates (23). The calculations performed here indicate that, if such compounds reach the soil of boreal ecosystems, they are trapped there, leaching and/or degrading very slowly (17). In other words, these soils are likely to be not only long-term storage compartments but also final sinks for PCBs, retarding their re-entry to the atmosphere. Although validation of the results, using more field data on PCB concentrations in air and litter in boreal forests, is needed, these observations, taken together, suggest that the global supply of PCBs to boreal forests and their retention in the soils of such ecosystems are highly efficient. The method to estimate deposition fluxes used here, based on measurements made of POPs in the Oi soil horizon and the determination of the layer’s age, can potentially extend the interpretation of experimental observations at other sites and on a wider spatial scale, providing additional information to assess the efficiency of boreal forests in trapping and processing atmospherically transported POPs.

Acknowledgments We are grateful to the European Union Framework-VI project AQUATERRA for funding this research, to the Marie Curie Research Training Networks Programme under the European Commission for financially supporting L.N., to the technicians at the Norunda research site, and to Dr. Jackie Pates for the 137 Cs analysis.

Supporting Information Available Details about experimental methods including QA/QC information and the full data set of PCB concentrations in and fluxes to different forest compartments; soil layer age estimates including 137Cs data; forest biomass estimates; description of the model to estimate degradation half-lives of PCBs in the soil including input data used. This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Horstmann, M.; McLachlan, M. S. Atmospheric deposition of semivolatile organic compounds to two forest canopies. Atmos. Environ. 1998, 32, 1799–1809. 5288

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 14, 2009

(2) Nizzetto, L.; Cassani, C.; Di Guardo, A. Deposition of PCBs in mountains: The forest filter effect of different forest ecosystem types. Ecotoxicol. Environ. Saf. 2005, 63, 75–83. (3) Brorstro¨m-Lunde´n, E.; Lo¨fgren, C. Atmospheric fluxes of persistent semivolatile organic pollutants to a forest ecological system at the Swedish west coast and accumulation in spruce needles. Environ. Pollut. 1998, 102, 139–149. (4) Su, Y.; Wania, F.; Harner, T.; Lei, Y. D. Deposition of polybrominated diphenyl ethers, polychlorinated biphenyls, and polycyclic aromatic hydrocarbons to a boreal deciduous forest. Environ. Sci. Technol. 2007, 41, 534–540. (5) Meijer, S. N.; Ockenden, W. A.; Sweetman, A.; Breivik, K.; Grimalt, J. O.; Jones, K. C. Global distribution and budget of PCBs and HCH in background surface soils: implications for sources and environmental processes. Environ. Sci. Technol. 2003, 37, 667– 672. (6) Wania, F.; Su, Y. Quantifying the global fractionation of polychlorinated biphenyls. Ambio 2004, 33, 161–168. (7) Choi, S.-D.; Staebler, R. M.; Li, H.; Su, Y.; Gevao, B.; Harner, T.; Wania, F. Depletion of gaseous polycyclic aromatic hydrocarbons by a forest canopy. Atmos. Chem. Phys. 2008, 8, 4105– 4113. (8) Su, Y.; Wania, F. Does the forest filter effect prevent semivolatile organic compounds from reaching the Arctic. Environ. Sci. Technol. 2005, 39, 7185–7193. (9) Wania, F.; Mackay, D. Global fractionation and cold condensation of low volatility organochlorine compounds in polar regions. Ambio 1993, 22, 10–18. (10) Dalla Valle, M.; Jurado, E.; Dachs, J.; Sweetman, A. J.; Jones, K. C. The maximum reservoir capacity of soils for persistent organic pollutants and implications for their global cycling. Environ. Pollut. 2005, 134, 153–164. (11) McLachlan, M. S.; Welsch-Pausch, K.; Tolls, J. Field validation of a model of the uptake of gaseous SOC in Lolium multiflorum (rye grass). Environ. Sci. Technol. 1995, 29, 1998–2004. (12) Wild, E.; Dent, J.; Thomas, G. O.; Jones, K. C. The air-leaf transfer and within leaf movement and distribution of phenanthrene: further studies utilising two photon excitation microscopy. Environ. Sci. Technol. 2006, 40, 907–916. ¨ stman, C. Environmental monitoring (13) Kylin, H.; Grimvall, E.; O of polychlorinated biphenyls using pine needles as passive samplers. Environ. Sci. Technol. 1994, 28, 1320–1324. (14) Ockenden, W. A.; Steinnes, E.; Parker, C.; Jones, K. C. Observations on persistent organic pollutants in plants: implications for their use as passive air samplers and for POP cycling. Environ. Sci. Technol. 1998, 32, 2721–2726. (15) Meijer, S. N.; Steinnes, E.; Ockenden, W. A.; Jones, K. C. Influence of environmental variables on the spatial distribution of PCBs in Norwegian and U.K. soils: Implications for global cycling. Environ. Sci. Technol. 2002, 36, 2146–2153. (16) Krauss, M.; Wilcke, W.; Zech, W. Polycyclic aromatic hydrocarbons and polychlorinated biphenyls in forest soils: depth distribution as indicator of different fate. Environ. Pollut. 2000, 110, 79–88. (17) Moeckel, C.; Nizzetto, L.; Di Guardo, A.; Steinnes, E.; Freppaz, M.; Filipa, G.; Camporini, P.; Benner, J.; Jones, K. C. Persistent organic pollutants in boreal and montane soil profiles: distribution, evidence of processes and implications for global cycling. Environ. Sci. Technol. 2008, 42, 8374–8380. (18) Cousins, I. T.; Gevao, B.; Jones, K. C. Measuring and modelling the vertical distribution of semi-volatile organic compounds in soils. I: PCB and PAH soil core data. Chemosphere 1999, 39, 2507–2518. (19) Moeckel, C.; Harner, T.; Nizzetto, L.; Strandberg, B.; Lindroth, A.; Jones, K. C. The use of depuration compounds in passive air samplers: Results from field deployment, potential uses and recommendations. Environ. Sci. Technol. 2009, 43, 3227-3232. (20) Zhang, D.; Hui, D.; Luo, Y.; Zhou, G. Rates of litter decomposition in terrestrial ecosystems: global patterns and controlling factors. J. Plant Ecol. 2008, 1, 85–93. (21) Horstmann, M.; McLachlan, M. S. Evidence of a novel mechanism of semivolatile organic compound deposition in coniferous forests. Environ. Sci. Technol. 1996, 30, 1794–1796. (22) Lagergren, F. Annual leaf litter production. Data available on request at http://www.bgc-jena.mpg.de/public/carboeur/sites/ norunda.html, 2008. (23) Breivik, K.; Sweetman, A.; Pacyna, J. M.; Jones, K. C. Towards a global historical emission inventory for selected PCB congeners - A mass balance approach. 3. An update. Sci. Total Environ. 2007, 377, 296–307. (24) Harner, T.; Bidleman, T. F.; Jantunen, L. M. N.; Mackay, D. Soil-air exchange model of persistent pesticides in the United

States cotton belt. Environ. Toxicol. Chem. 2001, 20, 1612–1621. (25) Pru ¨ gel, B.; Loosveldt, P.; Garrec, J.-P. Changes in the content and constituents of the cuticular wax of Picea abies (L.) Karst. in the relation to needle ageing and tree decline in five European forest areas. Trees-Struct. Funct. 1994, 9, 80–87. (26) Strachan, W. M. J.; Eriksson, G.; Kylin, H.; Jensen, S. Organochlorine compounds in pine needles: methods and trends. Environ. Toxicol. Chem. 1994, 13, 443–451. (27) Di Guardo, A.; Zaccara, S.; Cerabolini, B.; Acciarri, M.; Terzaghi, G.; Calamari, D. Conifer needles as passive biomonitors of the spatial and temporal distribution of DDT from point sources. Chemosphere 2003, 52, 789–797. (28) Nizzetto, L.; Pastore, C.; Liu, X.; Camporini, P.; Jarvis, A.; Stroppiana, D.; Infantino, A.; Herbert, B.; Zhang, G.; Boschetti, M.; Brivio, P. A.; Jones, K. C.; Di Guardo, A. Accumulation parameters and trends for PCBs in temperate and boreal forest plant species. Environ. Sci. Technol. 2008, 42, 5911–5916. (29) Moeckel, C.; Thomas, G. O.; Barber, J. L.; Jones, K. C. Uptake and storage of PCBs by plant cuticles. Environ. Sci. Technol. 2008, 42, 100–105. (30) Barber, J. L.; Thomas, G. O.; Kerstiens, G.; Jones, K. C. Air-side and plant-side resistances influence the uptake of airborne PCBs by evergreen plants. Environ. Sci. Technol. 2002, 36, 3224–3229. (31) More´n, A.-S.; Lindroth, A.; Flower-Ellis, J.; Cienciala, E.; Mo¨lder, M. Branch transpiration of pine and spruce scaled to tree and canopy using needle biomass distributions. Trees-Struct. Funct. 2000, 14, 384–397. (32) Hellstro¨m A. Uptake of Airborne Organic Pollutants in Pine Needles, Geographical and Seasonal Variations. Doctoral thesis, Department of Environmental Assessment, Swedish University of Agricultural Sciences, Uppsala, Sweden, 2003. (33) EMEP. Co-operative programme for monitoring and evaluation of the long-range transmissions of air pollutants in Europe. EMEP POP data. Annual PCB average data is available for download at http://www.nilu.no/projects/ccc/emepdata.html, 2008.

(34) Gasic, B.; Moeckel, C.; MacLeod, M.; Brunner, J.; Scheringer, M.; Jones, K. C.; Hungerbu ¨ hler, K. Measuring and modeling short-term variability of atmospheric PCBs and characterization of urban source strength in Zurich, Switzerland. Environ. Sci. Technol. 2009, 43, 769–776. (35) Kindermann, G. E.; McCallum, I.; Fritz, S.; Obersteiner, M. A global forest growing stock, biomass and carbon map based on FAO statistics. Silva Fenn. 2008, 42, 387–396. (36) Dachs, J.; Lohmann, R.; Ockenden, W. A.; Mejanelle, L.; Eisenreich, S. J.; Jones, K. C. Oceanic biogeochemical controls on global dynamics of persistent organic pollutants. Environ. Sci. Technol. 2002, 36, 4229–4237. ˇ upr, P.; Lammel, G.; Holoubek, I. (37) Ru ˚ zˇicˇkova´, P.; Kla´nova´, J.; C An assessment of air-soil exchange of polychlorinated biphenyls and organochlorine pesticides across central and southern Europe. Environ. Sci. Technol. 2008, 42, 179–185. (38) Van Drooge, B. L.; Grimalt, J. O.; Garcia, C. J. T.; Cuevas, E. Semivolatile organochlorine compounds in the free troposphere of the northeastern Atlantic. Environ. Sci. Technol. 2002, 36, 1155–1161. (39) Lavelle, P.; Chauvel, A.; Fragoso, C. Faunal activity in acid soils. In Plant soil interactions at low pH; Date, R. A., Ed.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1995; pp 201-211. (40) Borja, J.; Taleon, D. M.; Auresenia, J.; Gallardo, S. Polychlorinated biphenyls and their biodegradation. Process Biochem. 2005, 40, 1999–2013. (41) Gocht, T.; Ligouis, B.; Hinderer, M.; Grathwohl, P. Accumulation of polycyclic aromatic hydrocarbons in rural soils based on mass balances at the catchment scale. Environ. Toxicol. Chem. 2007, 26, 591–600.

ES803505U

VOL. 43, NO. 14, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5289