Global Biogeochemical Cycles

Isotopic trends and background fluxes of atmospheric lead in northern Europe: Analyses of three ombrotrophic bogs from south Sweden

Authors


Abstract

[1] Knowledge about the natural atmospheric background deposition rate of lead (Pb) prior to anthropogenic pollution is critical in the understanding of present-day pollution and for establishing realistic goals for the reduction of atmospheric Pb. We utilize stable Pb isotopes (206Pb and 207Pb) in radiocarbon-dated peat cores from three ombrotrophic bogs from south Sweden, to calculate fluxes and to survey atmospheric Pb trends prior 3500 BP (the so far known onset of large-scale anthropogenic pollution). The estimated atmospheric Pb deposition rate was between 1 and 10 μg Pb m2 yr−1 between 5900 and 3700 calendar years BP, which is 100 to 1000 times lower than present-day deposition rates. The majority of the samples older than 3500 calendar years BP had 206Pb/207Pb ratios ≤1.20, which is significantly lower than unpolluted Swedish mineral soils (206Pb/207Pb > 1.30), suggesting that even the natural atmospheric deposition of Pb was dominated by long-range transport, rather than local inputs from soil dust. Low 206Pb/207Pb ratios (1.16–1.18) of several samples indicate that this distant transport originated at least partly from early pollution sources. A possible climatic connection with the observed Pb deposition trends is suggested.

1. Introduction

[2] Atmospheric lead (Pb) pollution has a long history in Europe [Settle and Patterson, 1980; Nriagu, 1998]. From antiquity to the present, anthropogenic emissions have contributed significant inputs of Pb to remote boreal and arctic ecosystems [Boutron et al., 1994; Renberg et al., 1994]. This extensive pollution in time and space makes it impossible today to directly measure natural fluxes, i.e., deposition not affected by pollution, of atmospheric Pb in the Northern Hemisphere. Natural atmospheric Pb sources and deposition rates are key parameters to consider when understanding present-day pollution and for establishing realistic goals for the reduction of atmospheric Pb deposition in Europe. Estimates of natural atmospheric fluxes can only be obtained from paleoecological records older than 3500 years, when the atmosphere was not significantly polluted by anthropogenic emissions [Brännvall et al., 1999; Renberg et al., 2000; Shotyk et al., 2001].

[3] When estimating atmospheric deposition rates, ombrotrophic bogs and polar ice cores have an advantage in comparison to lake and marine sediments, since they only have the atmosphere as a source. Glacial ice cores, such as those from the Greenland summit, can provide long-term records of atmospheric Pb deposition with a good chronological precision based on lamination counting and age-depth modeling. However, fracture zones in the Greenland ice dating to the early and mid-Holocene and the very low Pb concentrations are fundamental problems when analyzing the atmospheric Pb record for this period [Hong et al., 1994, 1996]. Fluxes of Pb during the period between 7000 and 3000 BP have therefore not been documented from work from glacial ice. Ombrotrophic bogs, on the other hand, have by comparison much higher concentrations, but do not provide the advantage of annual lamination counting for dating. Rather, peat chronologies must rely on 14C dating.

[4] Important natural sources of atmospheric Pb are considered to be soil dust, volcanic emissions, forest fires, biogenic sources and sea salt spray; where soil dust and volcanic-derived atmospheric Pb are estimated to be the most important geological sources on a global scale [Nriagu, 1989]. Still, the contribution from each source has been debated [Nriagu, 1989; Hinkley et al., 1999] and studies concerning natural sources of atmospheric Pb to Northern Europe are lacking. In studies of the Pb pollution history in Sweden using ombrotrophic bogs [Brännvall et al., 1997], it was recognized that peat samples older than 3500 years, i.e., peat from pre-pollution times, had a stable Pb isotopic composition that strongly differed from the isotopic composition of Swedish mineral soils. This was interpreted to be caused by long-range transport of soil dust from continental Europe [Bindler et al., 1999].

[5] Analysis of stable Pb isotopes has proved to be a powerful tool for surveying sources of atmospheric Pb both during industrial time and during glacial and interglacial periods [Weiss et al., 1999; Jones et al., 2000]. Due to the extremely large isotopic differences between Swedish soils and other possible sources of atmospheric Pb, it is easy to distinguish local sources from external sources, thus creating favorable conditions for Pb provenance determinations [Renberg et al., 2002]. The objectives of this study are to further evaluate estimated background atmospheric Pb deposition rates [Bindler et al., 1999] and present the main prepollution trends in atmospheric Pb deposition for the period before 3500 BP, by analyzing Pb concentrations and stable Pb isotopes in three ombrotrophic bogs in southern Sweden, and thus contribute to the understanding of natural atmospheric Pb cycling.

2. Material and Methods

[6] Peat cores were collected from three ombrotrophic bogs, Dumme Mosse (DM), Årshultsmyren (ÅM) and Traneröd Mosse (TM) in southern Sweden (Table 1), using a Wardenaar corer [Wardenaar, 1987] for the top meter and a Russian peat corer [Aaby and Digerfeldt, 1986] (length 75 cm and diameter 5 cm) for deeper layers, in September 2001. A complete peat profile was collected using two parallel drives, where the cores overlapped by 25 cm in each drive. Cores were wrapped in plastic for preservation and transport. The study sites are classic Swedish domed ombrotrophic bogs studied for almost a century, [von Post and Granlund, 1926]. Detailed descriptions of macrofossil and pollen stratigraphy, and bog development for ÅM, TM and Store Mosse (57°15′N 13°55′E), an adjacent bog to DM, are previously published [Ringberg, 1984; Svensson, 1988; Thelaus, 1989; Malmer et al., 1997]. Ash content, assumed to equal the amount of mineral matter, was determined from peat first dried at 105°C and then heated at 550°C for 4 hours. The relative standard deviation (RSD) of the method is 0.15% as estimated from analyses of 32 duplicates or triplicates. Ash content together with peat characteristics (degree of decomposition and botanical components) were used for correlation of overlapping cores.

Table 1. Geographic Position, Annual Precipitation, and Bog-Complex Sizea
 Lat-LongAltitude, m.a.s.lSize, km2Precipitation, mm yr−1
Dumme Mosse57° 30′ N 14° 02′E21520700
Årshultsmyren56° 45′N 13° 25′E160101200
Traneröd Mosse56° 03′N 13° 11′E1650.85900

[7] To determine the Pb isotopic composition of nearby and regional soils, samples were collected from the C-horizon of Haplic podzols and Cambic regosols [Food and Agriculture Organization, 1990] in the vicinity of each bog and from other sites in south Sweden and in Denmark.

[8] Pb, Sc and Zr concentrations and stable Pb isotopes, 206Pb and 207Pb, of freeze-dried peat and soil samples were determined using ICP-MS (Perkin-Elmer model ELAN 6100) after a strong acid digestion (HNO3 + HClO4, 10:1) in open Teflon vessels. This method does not completely recover all Pb in samples rich in mineral grains. Consequently, in mineral soils this method might yield slightly different results than a total sample dissolution with a strong acid digestion that includes HF, while in organic-rich materials (i.e., organic-rich sediments, peat and soil O-horizon) the difference between the digestion methods is minor [Bindler et al., 1999; Renberg et al., 2002]. For deeper peat layers, where Pb concentrations were low, 0.5–1 g dry peat was used and for near-surface peat, where concentrations were higher, 0.1–0.3 g was used.

[9] Concentrations of Pb, Sc and Zr were verified against the certified multielement standard, SPEX ICPMS-2 (SPEX CertiPrep Certified Reference materials). A ten-point calibration within the range of 0.25–50 ng mL−1 was used. Acid blanks were in the range of 0.04 to 0.1 ng mL−1. Method detection limits (MDL) for elemental concentrations of Pb, Sc and Zr, determined as the reagent blank value plus three times the mean standard deviation of the 10 lowest measured concentrations, were 0.13, 30 and 4 ng mL−1, respectively. Analyzed samples were always at least three times higher than the MDL for Pb and Zr, while several Sc measurements were below. Precision for Pb concentration measurements, determined as the relative standard error of 50 analyses of four internal reference materials (lake sediments) over an 8-year period, was <10%. Accuracy for the elements, determined from the certified values for the SPEX ICPMS-2, was <10%.

[10] Analyses of Pb masses 206 and 207 were made using dwell times of 50 ms and were corrected empirically for fractionation by repeated analysis of the NIST SRM 981 reference material (0.000 ± 0.006% per a.m.u.). Precision for 206Pb/207Pb was <0.5% as determined by 50 measurements of an internal reference material (lake sediment) over an 8-year period (analyzed using two different instruments: Perkin-Elmer model Elan 5000 prior to 1999 and Perkin-Elmer model Elan 6100 since). Accuracy of the 206Pb/207Pb ratio measurements during the course of the study, as determined by replicate analysis of the NIST SRM 981 common standard, was 2.0% (2σ level, 95% confidence).

[11] AMS radiocarbon dating (Ångström Laboratory, Uppsala University, Sweden) was performed on Sphagnumspp. leaves and stems, which were first washed with distilled water in order to avoid interference by roots and humic substances. Two complementing dates were performed on bulk peat where the NaOH insoluble fraction was used for dating. The 14C dates were transformed to calendar years using Calib 4.3 [Stuiver and Reimer, 1993], and BP values in this work refer to calibrated years before AD 1950. Supporting dates for calculating peat accumulation rates were provided from Pb analyses (the Roman Peak ∼AD 0 [Renberg et al., 2001]). A Boltzmann equation was used for peat growth modeling (σ2 = 0.98, σ2 = 1.0 and σ2 = 1.0, for DM, ÅM and TM, respectively).

[12] Atmospheric Pb deposition rates were calculated by multiplying Pb concentrations with peat accumulation rates (g m−2 yr−1). Maximum and minimum values for the Pb deposition rates were calculated from the highest and lowest possible peat accumulation rates, respectively, according to a 95.4% confidence interval (2σ) of the calibrated calendar years.

[13] In order to determine the contribution of Pb not derived from local soils (Pb concdistant) to the total Pb concentration a simple binary mixing model was used,

equation image

where 206Pb/207Pbsample is the isotope ratio of a given sample, 206Pb/207Pblocal is the isotope ratio of the local soil, 206Pb/207Pbdistant is the ratio of an assumed nonlocal source and Pb concsample is the Pb concentration of the sample.

[14] Pb concentrations, for several subsamples within each peat profile, were normalized to either Sc or Zr to account for variations in mineral matter in the peat profiles. These are conservative metals with no significant anthropogenic source [Shotyk et al., 2000]. The enrichment of Pb in relation to Sc and Zr (PbEFSc or Zr) can be calculated using the equation

equation image

where Pb concfen is the concentration of Pb in the deepest sample from the fen peat and Sc or Zr concsample and Sc or Zr concfen is the concentration of Sc or Zr in a peat sample and in the deepest fen peat sample, respectively.

3. Results

[15] The peat from each of the three bogs shows no obvious signs of local disturbance (fire and land-use impact) throughout the cores. The peat has low average mineral matter content (1–2%), typical for ombrotrophic bogs (Figure 1). Surface peat, fen peat and sections with a high degree of decomposition have higher ash content (3–5%). The botanical transition between the fen peat and the overlying bog peat coincides well with decreasing ash content.

Figure 1.

Pb concentrations, 206Pb/207Pb ratios, PbEFSc and PbEFZr mineral matter and age (calendar year) versus depth in Dumme Mosse (DM), Årshultmyren (ÅM) and Traneröd Mosse (TM). Peat growth model are developed from a Boltzmann equation and horizontal bars mark 2σ age range. Age dates are derived from 14C dating of Sphagnum (Sph) or bulk peat (Bulk) and chronological Pb marker (Pb).

[16] The lowest measured Pb concentrations for the ombrotrophic peat are 0.02, 0.03 and 0.04 μg g−1 while mean concentrations for the period between ombrotrophic bog initiation and 3500 BP are 0.07, 0.09 and 0.23, in DM, ÅM and TM, respectively. By comparison, the highest Pb concentration in each bog is found between 4 and 20 cm depths and ranged between 86 and 120 μg g−1 (Figure 1). Higher concentrations are strongly related to increases in the proportion of Pb with low 206Pb/207Pb ratios. 206Pb/207Pb ratios below 1.16 are mainly found in the peat above 20 cm depth in all three bogs, but low 206Pb/207Pb ratios (1.16–1.18) comparable to those in the surface peat are also found for deeper layers of ombrotrophic peat (i.e., at depths of 245 and 330 cm in ÅM and 540 and 600 cm in TM). These isotope excursions in the deeper peat coincide with minor increases in Pb concentration.

[17] Pb found in the ombrotrophic stages of the bogs has an isotopic composition that strongly differs from the local mineral soils (Figure 1), as well as soil from other sites throughout Sweden [Bindler et al., 1999]. High 206Pb/207Pb ratios close to or above 1.3, similar to local mineral soils, were only found in the fen stage.

[18] The Pb/Sc ratio and Pb/Zr ratio in the Carex peat (fen stage) of the bogs are between 0.3–0.7 and 0.1–0.3, respectively. Higher ratios between 1.0–2.3 and 1.1–3.1, respectively, are found in local soils. The significantly lower PbEFSc and PbEFZr values (Figure 1) in the Carex peat as compared to local soils suggests that already the fen peat in each bog is depleted in Pb compared to local soils. Recent accumulated peat (0–20 cm) is enriched in Pb and has PbEFSc and PbEFZr values between 90–170 and 70–285 and 140–570 and 260–950 times higher values than the values found in the deep layers of ombrotrophic peat older than 3,500 years (Figure 1). Several samples from the ombrotrophic peat have Sc concentrations below the MDL, which makes the calculated PbEFSc less reliable. Despite this, PbEFSc and PbEFZr indicate comparable trends.

4. Discussion

4.1. Reliability of the Records

[19] The immobility of Pb in peat bogs has been questioned [Damman, 1978; Urban et al., 1990; MacKenzie et al., 1998a]. One could therefore argue that the decreasing concentrations and increasing 206Pb/207Pb ratios with depth are the result of a mixing of downward moving surface pollution with Pb transported upwards from the underlying mineral soil with groundwater. There are two main arguments that significant post-depositional downward migration of surface pollution is not occurring in the studied bogs. First, the main temporal 206Pb/207Pb and Pb concentration trends from 3000 years ago to the present are consistent among the study sites and with the known pollution history of Pb outlined in several papers, based on both estimations of past Pb production from historical data [Settle and Patterson, 1980; Nriagu, 1983] and from different paleoarchives, such as lake sediments, ice cores and peat bogs [Hong et al., 1994; Renberg et al., 1994; Martínez-Cortizas et al., 1997; Brännvall et al., 1999; Renberg et al., 2001; Shotyk et al., 2001]. The strong pollution influence is seen by low 207Pb/207Pb ratios similar to reported values for European ore deposits [Yener et al., 1991; Stos-Gale et al., 1995; Rohl, 1996; Sayre et al., 2000] and modern aerosols in south–east Sweden [Rosman et al., 1998]. The isotope ratio trends and values in the bogs match the ratios in the Greenland ice (Figure 2).

Figure 2.

206Pb/207Pb ratios and Pb concentrations since 3000 BP from the ice core at Summit, Greenland [Rosman et al., 1993] representing the known Pb pollution history of Europe in comparison with 206Pb/207Pb ratios and Pb concentrations from the ombrotrophic peat in DM, ÅM and TM.

[20] Another argument against the concept that vertical movement of surface pollution is causing the low 206Pb/207Pb ratios seen in the older peat can be established using mass balance calculations (equation (1)). Given a ratio for surface pollution (Pbdistant) of 1.16 and ratios for the underlying mineral soil of 1.32, 1.46, and 1.33 for DM, ÅM and TM, respectively, the amount of surface pollution Pb that is required to explain the low 206Pb/207Pb ratios can be calculated. The results of these calculations are unrealistic from a mass balance perspective; for example, the peat overlying the low 206Pb/207Pb excursions of the periods labeled E1 and E2 (Figures 3 and 4) would on average receive 2 to 3 times less Pb pollution per gram dry weight (individual samples receiving up to 58 times less pollution).

Figure 3.

Standardized 206Pb/207Pb ratios and standardized mineral matter content (see text for explanation) versus time for DM, ÅM and TM. Original data are from Figure 1. Shown are event 1 (E1), event 2 (E2), period 1 (P1), period 2 (P2) and periods with rapid increase of relative mineral matter content in the ombrotrophic peat.

Figure 4.

Box-plot of 206Pb/207Pb ratios found for potential local and distant sources of atmospheric Pb in comparison with 206Pb/207Pb ratios of the analyzed peat. Local soil sources are represented by Swedish and Danish C-horizons (15 and 4 sites, respectively) and lake sediments older than 3000 years reflecting soils within the catchments (30 sites). Distant natural sources are represented by global soil dust-sources (n = 28), including the Sahara desert, used for provenance determination of aeolian dust in the Greenland ice [Sun, 1980; Biscaye et al., 1997], volcanic rocks from Italy (n = 73) [Fitton and Upton, 1987; Ayuso et al., 1998], and volcanic rocks from Iceland (n = 36) [Sun, 1980]. Anthropogenic data are represented by reported 206Pb/207Pb ratios of ores and artifacts from early active mining districts in Anatolia, the Balkans, the British Isles, Greece and the Iberian peninsula (n = 522) [Yener et al., 1991; Sayre et al., 1992; Arribas and Tosdal, 1994; Stos-Gale et al., 1995; Rohl, 1996; Stos-Gale et al., 1998; Sayre et al., 2000] and modern aerosol data from south–east Sweden (n = 35) [Rosman et al., 1998].

[21] It can be argued that use of botanical indicators to separate ombrotrophic peat from peat affected by inputs from groundwater is not reliable, and studies from a Swiss bog have suggested that the use of calcium/magnesium molar ratios is a useful approach to make a geochemical distinction between ombrotrophic and minerotrophic peat sections [Shotyk, 1996]. This method is favored by the presence of Ca and Mg rich minerals, such as calcite and dolomite, which are very sparse in the gneissic and granitic bedrock in south Sweden, making this technique less suitable for the work presented here. Even though the botanical, physical and chemical properties of Swedish ombrotrophic bogs have been studied for more than a century, studies focusing on groundwater transport of Pb in peat are lacking. However, investigations in other botanically classified ombrotrophic Sphagnum bogs in Europe have shown that Pb is bound strongly to organic matter and is not transported upwards from the underlying mineral soils [Shotyk, 1996; MacKenzie et al., 1997, 1998b; Vile et al., 1999].

4.2. Atmospheric Deposition Rates Prior to 3500 BP

[22] The mean background Pb atmospheric deposition rate calculated from TM and ÅM was in the range of 1–10 μg Pb m−2 yr−1 (Table 2). The calculated value for TM is 4–10 μg Pb m−2 yr−1 during a period ranging from ∼5200 to 3700 years BP, while the lowest fluxes, 1–3 μg Pb m−2 yr−1, are calculated from ÅM during the period between 5900 and 4900 years BP. Due to plateaus in the radiocarbon calibration curve [Stuiver and Reimer, 1993], the calibrated age ranges of the overlying and underlying peat layers from DM overlap, which makes calculation of peat accumulation rates for this period less precise, and thus also Pb deposition rates. The maximum deposition rate for DM was therefore calculated using the highest reported regional accumulation rate for Sphagnum fuscum peat of 95 g m−2 yr−1 during the Holocene for the nearby bog Store Mosse, ca 45 km south of DM [Malmer et al., 1997]. This accumulation rate is probably an overestimate, which is indicated by the age-depth model used for the peat younger than 4100 BP in DM. This gave a calculated deposition range of 3–8 μg Pb m−2 yr−1 in DM.

Table 2. Radiocarbon Dates and Calculated Annual Pb Deposition Rates Prior to 3500 BP, From DM, ÅM and TMa
Sample No.BogDepth, cm14C yr BPCalibrated 14C yr BP (range 2σ)Bulk density, g dm−3Mean Pb concentration, μg g−1Pb deposition, μg m−2 yr−1
  • a

    Maximum and minimum depositions rates are calculated from the range of 2σ.

  • b

    Calculated from maximum reported peat growth from the adjacent bog Store Mosse [Malmer et al., 1997].

Ua- 14791DM2803640 ± 754217–3721   
Ua- 14792DM3503805 ± 754412–3983710.0893–8b
Ua- 14795ÅM3304420 ± 505285–4859   
Ua- 14796ÅM3704985 ± 605890–5604800.0691–3
Ua- 14793TM5303545 ± 553978–3650   
Ua- 14794TM6004290 ± 755208–4554730.2674–10

[23] These atmospheric deposition values of Pb are in good agreement with previous estimates from three other bog sites in Sweden and one in Switzerland of 3–11 μg m−2 yr−1 [Shotyk et al., 1998; Bindler et al., 1999]. Our estimated background values between 1 and 10 μg m−2 yr−1 are between 1000 and 100 times lower than the reported deposition rates of ∼1 mg Pb m−2 yr−1 for south Sweden determined by biomonitoring studies in 1995 using terrestrial mosses [Rühling and Tyler, 2001]. The enrichment of Pb (PbEFZr) by 260–950 times in surface peat (0–20 cm), compared to values found in the ombrotrophic peat older than 3500 years, further supports that the background deposition was magnitudes lower than the modern atmospheric deposition. It is important to note that the estimated deposition rate for AD 1995 is already a magnitude lower than 2 decades before, following the harder restrictions upon metal emissions from industries and the restrictions on of alkyl Pb in fuel [Rühling and Tyler, 2001].

4.3. Atmospheric Pb Deposition Trends and Sources Prior to 3500 BP

[24] To facilitate comparison between the temporal isotopic trends and the changes in mineral matter, the 206Pb/207Pb ratio and the mineral matter content are standardized to unit variance and zero mean (Figure 3). This was done by subtracting each value with the mean value for background ombrotrophic peat, and by dividing by the standard deviation.

[25] The temporal changes in the Pb isotopic composition of the bogs are similar (Figure 3), despite that the bogs are situated up to 200 km apart, which strongly suggest that the Pb is derived from distant sources by atmospheric transport. A further argument for this is that the 206Pb/207Pb ratios of the ombrotrophic peat are significantly lower than the ratios in local mineral soils (C-horizon) in Sweden (Figure 4). Possible origin of this distant Pb could either be from natural atmospheric Pb sources like soil dust, volcanic emissions, forest fires, biogenic sources and sea salt spray [Nriagu, 1989] or from mines exploited by early metal-using civilizations [Nriagu, 1983].

[26] Several samples from two periods, named event 1 (E1) and event 2 (E2) in Figure 3, at 5400–5000 BP and 4100–3800 BP, respectively, have an isotopic composition that can only be explained by inputs from a source with 206Pb/207Pb ratios as low as 1.16 (Figure 4). Even though data on the isotopic composition of different natural sources of atmospheric Pb are sparse and show large heterogeneity, none of the likely natural sources shown in Figure 4 have a 206Pb/207Pb ratio below 1.17. The similarity in 206Pb/207Pb ratios between these samples and Pb derived from ore deposits from early European mining districts [Yener et al., 1991; Stos-Gale et al., 1995; Rohl, 1996; Sayre et al., 2000] suggests that samples from E1 and E2 consist, at least partly, of anthropogenic derived Pb. Processing of Pb ores has a history older than 6,000 years and Greek mines operated well before 5000 BP [Elliot et al., 1937], which makes pollution from early metal-using civilizations fully possible [Nriagu, 1983]. The timing of E2 also coincides with the rise and demise of several early metal-using civilizations around the Mediterranean sea [Warren, 1989]. At that time, exploitation of Pb ores also occurred on the British Isles [Nriagu, 1983]. However, as E1 is only detected in two bogs by a limited number of samples at a very early date in the history of metallurgy, a complete separation between anthropogenic source(s) and natural source(s) with low 206Pb/207Pb ratios would need a complimenting multielement study, to fully ascertain the source of this Pb. Such a study should also include analyses of volcanic ash (tephra) to assess the possible role of volcanic eruptions. A similar approach would be needed to separate whether the periods in-between E1 and E2 (P1 and P2), having 206Pb/207Pb ratios similar to different natural atmospheric sources (Figure 4), really are a result of changing natural atmospheric Pb sources or a result of mixture between a local source and the distant source with low 206Pb/207Pb ratio detected during E1 and E2.

[27] The studied bogs are and were surrounded by forest with thin organic surface layers, and the underlying mineral soil constitutes a large potential source of atmospheric Pb through local dust emissions. Therefore, it may seem surprising that Pb from nearby mineral soils did not act as a dominant atmospheric Pb source, as is evidenced from the low 206Pb/207Pb ratios in the peat. The maximum potential contribution from local Pb mineral sources to the ombrotrophic peat can be calculated using the mixing model (equation (1)), applying a 206Pb/207Pb ratio of 1.30 (the lowest analyzed value for a local mineral soil) for the local source and a ratio of 1.16 (the lowest individual value from E1and E2) for the distant source. This calculation shows that local mineral soils, at most, contributed 15–70% of the Pb found in P1 and P1, while only 0–40% in E1 and E2. An estimation using more typical values of 1.45 for Swedish mineral soils [Bindler et al., 1999] and 1.17 (lower percentile for E1 and E2), reduces the potential contribution from local sources to less than 30% in the ombrotrophic peat. As it is reasonable to think of several natural sources of Pb occurring at the same time [Nriagu, 1989], a binary mixing model as used here most likely overestimates the contribution from local soils. Clearly, local mineral soils did not act as a dominant Pb source to the bogs during this time. This is not really a surprising result when considering the insignificant upward transport of Pb from the mineral soil to the overlying organic horizon in boreal soils, suggested by Bindler et al. [1999]. In such a case, the Pb composition of the soil organic horizon would over time achieve an “equilibrium” with atmospheric sources, rather than be significantly influenced by the large Pb inventory in the underlying mineral soil horizons. Hence, remobilization or redistribution of topsoil by winds and forest fires, for example, would only recirculate Pb that had originally been atmospherically transported from distant sources.

[28] The atmospheric deposition of dust is thought to be strongly affected by climate, and the occurrence of Pb derived from atmospheric dust in Greenland ice and Atlantic marine sediments has been used as a proxy for climate change [Biscaye et al., 1997; Jones et al., 2000]. Interestingly, E1 and E2 occur during periods characterized by abrupt climatic shifts indicated by a variety of climatic proxies in Europe [Digerfeldt, 1988; Svensson, 1988; Bond et al., 1997; Hughes et al., 2000]. The isotopic trends also show several similarities with the 1100-year pacing of wet shifts reconstructed from two English mires by Hughes and coworkers [Hughes et al., 2000]. In that work, wet-shifts, suggested to be associated with wide-scale climatic change, occur during E1 and E2, but also around 3500, 3000 and 2300, which is further supported by other peat records on the British Isles and in Scandinavia [Aaby, 1976; Svensson, 1988; Chambers et al., 1997]. During these periods, rapid changes in the 206Pb/207Pb ratio occurs in the peat in DM, ÅM and TM and, in some of the bogs, also significant changes in mineral matter (Figure 3), which suggest that the observed atmospheric Pb deposition trends are connected to climate variability. The way in which climatic variables affect the occurrence of atmospheric Pb is complex [Biscaye et al., 1997], and at present there is not a solid basis for identifying precisely in which way the recorded Pb isotopic trends are affected by climatic factors. It may be that under certain climate regimes, there are climatic conditions that favor atmospheric transport of Pb from specific regions, which might account for the brief potential anthropogenic signal at E1 and E2. Thus, Pb isotopes may serve as a proxy for air mass circulation. However, these indications of a climatic impact on the atmospheric deposition in Sweden stress the importance of combining further analysis of Pb isotopes with climatic proxies in peat. It is also important to include a multisite approach to cope with local variation in the atmospheric Pb deposition and influences of local sources.

5. Conclusions

[29] The atmospheric deposition rate of Pb in south Sweden was in the range of 1–10 μg Pb m−2 yr−1 between 5900 and 3700 BP. At that time, Pb was a very rare element in the atmosphere and the deposition rate was 100 to 1000 times lower than modern rates. Even if it is not realistic to strive for an atmosphere totally free from Pb pollution at the present time, it is essential to be aware of the low atmospheric deposition rates of Pb prior to extensive pollution. Low 206Pb/207Pb ratios in the bogs even during “pristine time” (prior to known large-scale pollution), suggest that the atmospheric input of Pb to the Swedish forest landscape was mainly dependent upon long-range transport, since unpolluted Swedish mineral soils (C-horizons) have much higher isotopic ratios and the Pb in the bogs can thus not have been derived from Swedish mineral soils. The very low 206Pb/207Pb ratios of the Pb deposited in the ombrotrophic peat during two events around 5200 BP and 4000 BP suggest that anthropogenic sources, such as metal production, rather than soil dust were active sources of long-range transported atmospheric Pb during these times. An additional link between atmospheric Pb deposition and climate during natural condition seems reasonable; however, since an unambiguous chronology of short-term climate changes does not exist for south Sweden, a complete discussion about this relationship cannot yet be made.

Acknowledgments

[30] Financial support was provided by the Swedish Environmental Protection Agency and the Swedish Research Council for Environment, Agricultural Science and Spatial Planning (FORMAS). We thank Ole K. Borggaard, Jan-Erik Wallin, Elisabeth Bohlin and Tom Korsman for providing help with sampling of Danish soils, pollen analysis, peat classification and statistical inspiration, respectively.

Ancillary