Environ. Sci. Technol. 2006, 40, 4639-4645
Flux Rates of Atmospheric Lead Pollution within Soils of a Small Catchment in Northern Sweden and Their Implications for Future Stream Water Quality J O N A T A N K L A M I N D E R , * ,† RICHARD BINDLER,† HJALMAR LAUDON,† KEVIN BISHOP,‡ OVE EMTERYD,§ AND INGEMAR RENBERG† Department of Ecology and Environmental Science, Umeå University, SE-901 87 Umeå, Sweden, Department of Environmental Assessment, Swedish University of Agricultural Sciences, SE-750 07 Uppsala, Sweden, and Department of Forest Ecology, Swedish University of Agricultural Sciences, SE-901 83 Umeå, Sweden
It is not well-known how the accumulated pool of atmospheric lead pollution in the boreal forest soil will affect the groundwater and surface water chemistry in the future as this lead migrates through the soil profile. This study uses stable lead isotopes (206Pb/207Pb and 208Pb/ 207Pb ratios) to trace the transport of atmospheric lead pollution within the soil of a small catchment and predict future lead level changes in a stream draining the catchment. Low 206Pb/207Pb and 208Pb/207Pb ratios for the lead in the soil water (1.16 ( 0.02; 2.43 ( 0.03) and streamwater (1.18 ( 0.03; 2.42 ( 0.03) in comparison to that of the mineral soil (>1.4; >2.5) suggest that atmospheric pollution contributes by about 90% (65-100%) to the lead pool found in these matrixes. Calculated transport rates of atmospheric lead along a soil transect indicate that the mean residence time of lead in organic and mineral soil layers is at a centennial to millennial time scale. A maximum release of the present pool of lead pollution in the soil to the stream is predicted to occur within 200-800 years. Even though the uncertainty of the prediction is large, it emphasizes the magnitude of the time lag between the accumulation of atmospheric lead pollution in soils and the subsequent response in streamwater quality.
Introduction Atmospheric lead (Pb) pollution derived from metal production and other human activities during the last four millennia in Europe has resulted in extensive contamination of the northern hemisphere (1, 2), including the boreal forest where about half the pollution was deposited prior to AD 1800 (3). For the boreal forest most of the airborne lead is found in the soil (4, 5). To what extent this accumulated pool of * Corresponding author phone: + 46 90 7869784; fax: + 46 90 7866705; e-mail:
[email protected]. † Umeå University. ‡ Department of Environmental Assessment, Swedish University of Agricultural Sciences. § Department of Forest Ecology, Swedish University of Agricultural Sciences. 10.1021/es0520666 CCC: $33.50 Published on Web 06/30/2006
2006 American Chemical Society
pollution Pb contributes to the content of Pb in surface waters draining forest soils is not well-known. Recently, the European Union adopted a new directive for water quality management that should be implemented before 2015, which requires that the levels of prioritized toxic elements, such as Pb, should only show minor differences compared to that of the natural state (6). The implementation of this policy is, however, hampered by limited knowledge about both natural surface water Pb concentrations and the time scales required for surface waters to respond to decreasing emissions. The flux of Pb leaving forested catchments through stream runoff is small in comparison to that of atmospheric deposition, which has been interpreted as evidence for an almost complete retention of atmospheric Pb in forest soils (7). This argument has two limitations: (i) it does not indicate whether the Pb in the water is coming from pollution or mineral weathering; (ii) the steady-state assumption underlying it cannot exclude the possibility of a pollution front moving slowly through the soil resulting in a future release. Lead is continuously transported downward in soils through cotransport with particulate and colloidal organic matter moving with infiltrating precipitation (8). Several studies have suggested that the loss of Pb from the organic layer to the underlying mineral soil in podzolssthe most common soil type in the boreal zonesis a slow process that proceeds over a centennial time scale (8-11). As a result, Pb is redistributed very slowly to deeper, water-saturated mineral soil layers from which a lateral transport of soil water and metals to surface waters occurs (12). Thus, it is not likely that the present Pb levels in streams are in steady state with atmospheric inputs. Several studies have stressed that it is more likely to expect increasing Pb levels as the modern pollution front continues to migrate downward (13, 14). To predict how the redistribution of pollution Pb in the boreal forest will affect future water quality, it is necessary to asses the migration rates of Pb through both podzolic and organic-rich riparian soil types, the latter considered to be the proximal source of metals to streams (12, 15). Stable Pb isotope analyses have increased our knowledge about the impact of pollution in different ecosystem compartments of the boreal forest, such as the atmosphere (16, 17), soil (4, 10, 18), and plants (5, 19). Pb isotope analyses of stream and soil water samples offer a possibility to separate and estimate the contribution from pollution and local geogenic catchment sources, respectively (20-22), because Pb derived from pollution has 206Pb/207Pb ratios mainly e1.17 while geogenic Pb in Sweden typically has ratios above 1.30 (23). In this study we use Pb concentration and stable isotope analyses (206Pb/207Pb ratio) of water draining from a boreal forest catchment as well as of soil water and solid-phase soil samples to asses the fate of pollution Pb in a forested catchment in northern Sweden. The objectives are to (1) determine the extent of Pb pollution in the stream; (2) estimate the transport rate of pollution Pb within a podzolic soil and a riparian soil; (3) use the estimated transport rates to make a first approximation of the future release of Pb stored in the soil to the stream by using a steady-state massbalance model.
Materials and Methods Field Sampling. The study was carried out on a 13 ha forested catchment (64°,14′ N, 10°,46′ E, 235-310 m above sea level) at the Svartberget research station in northern Sweden (Figure 1). Annual precipitation is about 600 mm year-1, where approximately one-third falls as snow. The studied stream, VOL. 40, NO. 15, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4639
was measured manually. The total vertical and lateral soil water flux was assumed to equal the yearly runoff, with water infiltrating vertically downward through the unsaturated zone to the water table. Assuming that all lateral flow occurs in the saturated zone below the water table in accordance with Darcy’s law (27), the daily lateral soil water fluxes were calculated using the relationship between groundwater level and stream discharge established during 1993-2001. More details of the hydrology are provided as Supporting Information.
FIGURE 1. Map of Sweden and the studied catchment with the soil lysimeter transect. Va¨straba¨cken, drains a 10-15 m thick till that overlays gneissic bedrock. The stream was straightened and deepened (1 m) in the 1920s. Soils are predominantly well-developed iron podzols with an organic layer 5-10 cm thick, an E-horizon around 10 cm thick, and a spodic B-horizon that gradually changes toward a C-horizon around 50-70 cm depth. The organic-rich riparian soil is found within ∼10 m of the stream and has an histic organic horizon with an average depth of 45 cm that overlays a mineral soil rich in organic matter (24). Detailed description of the physiographical, hydrological and hydrochemical characteristics of the catchment can be found elsewhere (12, 24-27). Streamwater sampling was carried out weekly (more often during spring flood) during 1993 and 2002 by collecting grab samples in HDPE acid-washed bottles soaked in heated successive baths of 33% and 10% HCl for 48 h. Bottles were rinsed, stored, and transported with ultrapure water (>18 MΩ). Streamwater samples were transported within 4 h to the laboratory where an aliquot for metal analyses was acidified with HNO3 and stored at 4 °C. In 2002, soil water samples were collected from tension lysimeters into acidwashed borosilicate flasks under a suction of 0.8 bar, then transferred to HDPE flasks. The lysimeters are made of HDPE plastic and porous ceramic and have equilibrated for 7 years in the soils prior to sampling. Detailed descriptions of their construction and installation are previously published (27). The lysimeters were placed 58 and 76 m downslope from the water divide in two different soil types characterizing the catchment: an organic riparian soil located 4 m from the stream and a well-developed iron podzol located 22 m uphill from the riparian site. Soil profiles were collected in 2002 and 2004 with a steel corer (diameter 4.2 cm) containing a removable plastic liner or from pits dug with a spade. Soil samples were placed in polypropylene bags and stored at 4 °C. Soil density was determined through vertical sampling in the pits at ∼10 cm intervals. High-resolution soil density estimates (1.13). In the uphill soil, the Pb concentration decreases from values around 5 to 12 µg L-1 in the top 6 cm of the soil to values as low as 0.2 µg L-1 in the deeper parts of the mineral soil (Figure 3a). The concentration in the soil water from the riparian soil is somewhat higher and decreases from 30 µg VOL. 40, NO. 15, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4641
FIGURE 4. (a) Inventory of pollution Pb, vertical and lateral fluxes of Pb, as well as the estimated MRT of Pb in the soil horizons of the uphill soil. The error bars indicate the upper and lower quartiles (p < 0.25) of the uncertainty of single estimates. (b) Similar variables as above for the riparian soil. L-1 in the peaty surface to values also around 0.2 µg L-1 at 35 to 65 cm depth (Figure 3b). High Pb concentrations in the soil water from the two profiles are associated with high TOC concentrations (Figure 3, parts a and b). The amount of Pb pollution stored in the soil can be calculated by multiplying the solution to eq 1 with the soil layer mass (kg m-2). Including an uncertainty in the calculation, where all variables vary around their average value following a normal distribution (( 3.5 σ), the Pb inventory for each soil profile is about 1.5 ( 0.5 and 1.0 ( 0.8 g m-2 in the uphill and riparian soil, respectively (Figure 4, parts a and b). The largest vertical flux of about 2 to 4 mg m-2 year-1 occurs from the O-horizon of both soil types (Figure 4, parts a and b). In the uphill soil, the vertical flux of Pb dominates over the lateral flux in every horizon except in the C-horizon. In contrast, the lateral flux of Pb in the riparian soil dominates in soil layers below 20 cm due to the large accumulated water inputs from upslope that traverse the riparian soil. The calculated MRT of Pb in the horizon of the uphill soil ranges from about ∼150 years in the organic layer to several thousand years in the mineral soil (Figure 4a). The MRT of Pb in the soil layers from the riparian soil decrease from ∼150 years in the organic layer down to a few decades in the deepest layer, indicating that the riparian soil acts as a short-term buffer of Pb transported laterally from the uphill soil (Figure 4b). Lead in the Stream. The streamwater flux and streamwater chemistry in 1993 and 2002 are shown in Figure 5. The flow rates are higher in 1993, as are Pb concentrations and TOC levels. Samples from 1993 have Pb concentrations typically around 0.5 ( 0.2 µg L-1 and TOC levels of about 25 ( 6 mg L-1, while the samples from 2002 have values around 0.2 ( 0.1 µg L-1 and 11 ( 3 mg L-1. The 206Pb/207Pb ratio of the streamwater did not show any systematic change over the seasons or between the years of 1993 and 2002. The samples from 1993 and 2002 have a mean 206Pb/207Pb ratio of 1.18 ( 0.03 and 1.18 ( 0.02, respectively. A 206Pb/207Pb ratio of 1.18 suggests that more than 86% of the Pb in the streamwater is derived from pollution (eq 1, 206Pb/207Pbgeogenic >1.5; 206Pb/207Pbpollution >1.13). The yearly loss of Pb from the catchment, calculated as the integrated sum of the daily runoff multiplied with the stream concentration measurements, is estimated to be 0.06 ( 0.03 mg Pb m-2 year-1 (∼4 g year-1) during the low-runoff year of 2002 and 0.2 ( 0.1 mg Pb m-2 year-1 (∼16 g year-1) during the high-runoff year of 1993. The contribution of wet atmospheric deposition of Pb to the stream is estimated to be
