Historical accumulation rates of mercury in four Scottish ombrotrophic peat bogs over the past 2000 years

Historical accumulation rates of mercury in four Scottish ombrotrophic peat bogs over the past 2000 years

Science of the Total Environment 407 (2009) 5578–5588 Contents lists available at ScienceDirect Science of the Total Environment j o u r n a l h o m...

849KB Sizes 0 Downloads 5 Views

Science of the Total Environment 407 (2009) 5578–5588

Contents lists available at ScienceDirect

Science of the Total Environment j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v

Historical accumulation rates of mercury in four Scottish ombrotrophic peat bogs over the past 2000 years John G. Farmer a,⁎, Peter Anderson b, Joanna M. Cloy a, Margaret C. Graham a, Angus B. MacKenzie c, Gordon T. Cook c a b c

School of GeoSciences, University of Edinburgh, Edinburgh, EH9 3JN, Scotland, UK Contaminated Land Assessment and Remediation Research Centre, University of Edinburgh, Edinburgh, EH9 3JL, Scotland, UK Scottish Universities Environmental Research Centre, East Kilbride, G75 0QF, Scotland, UK

a r t i c l e

i n f o

Article history: Received 7 April 2009 Received in revised form 5 June 2009 Accepted 16 June 2009 Available online 30 July 2009 Keywords: Mercury Peat bogs Accumulation rates 210 Pb dating

a b s t r a c t The historical accumulation rates of mercury resulting from atmospheric deposition to four Scottish ombrotrophic peat bogs, Turclossie Moss (northeast Scotland), Flanders Moss (west-central), Red Moss of Balerno (east-central) and Carsegowan Moss (southwest), were determined via analysis of 210Pb- and 14C-dated cores up to 2000 years old. Average pre-industrial rates of mercury accumulation of 4.5 and 3.7 μg m− 2 y− 1 were obtained for Flanders Moss (A.D. 1–1800) and Red Moss of Balerno (A.D. 800–1800), respectively. Thereafter, mercury accumulation rates increased to typical maximum values of 51, 61, 77 and 85 μg m− 2 y− 1, recorded at different times possibly reflecting local/regional influences during the first 70 years of the 20th century, at the four sites (TM, FM, RM, CM), before declining to a mean value of 27 ± 15 μg m− 2 y− 1 during the late 1990s/early 2000s. Comparison of such trends for mercury with those for lead and arsenic in the cores and also with direct data for the declining UK emissions of these three elements since 1970 suggested that a substantial proportion of the mercury deposited at these sites over the past few decades originated from outwith the UK, with contributions to wet and dry deposition arising from long-range transport of mercury released by sources such as combustion of coal. Confidence in the chronological reliability of these core-derived trends in absolute and relative accumulation of mercury, at least since the 19th century, was provided by the excellent agreement between the corresponding detailed and characteristic temporal trends in the 206Pb/207Pb isotopic ratio of lead in the 210Pb-dated Turclossie Moss core and those in archival Scottish Sphagnum moss samples of known date of collection. The possibility of some longer-term loss of volatile mercury released from diagenetically altered older peat cannot, however, be excluded by the findings of this study. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Dated cores from ombrotrophic peat bogs are being used increasingly as archives of metal deposition from the atmosphere since the end of the last Ice Age and in particular since Roman times. This practice has probably been most successful and widely applied for lead, especially lead of anthropogenic origin released to the atmosphere since the onset of the Industrial Revolution (e.g. Shotyk et al., 1998; Klaminder et al., 2003; Cloy et al., 2008). There are, however, a growing number of such reports for other trace elements, including mercury (e.g. Benoit et al., 1994, 1998; Norton et al., 1997; Martinez-Cortizas et al., 1999; Bindler, 2003; Givelet et al., 2003; Roos-Barraclough and Shotyk, 2003; Shotyk et al., 2003, 2005; Steinnes and Sjøbakk, 2005; Coggins et al., 2006) although the extent to which mercury accumulation rates deduced from mercury concentration profiles in dated peat

⁎ Corresponding author. Tel.: +44 131 6504757; fax: +44 131 6504757. E-mail address: [email protected] (J.G. Farmer). 0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2009.06.014

cores accurately reflect a record of deposition from the atmosphere is currently a matter of debate (Biester et al., 2007). In the context of peat bog records of atmospheric deposition, mercury can be considered somewhat different from other trace elements for several reasons. Firstly, with respect to emissions to the atmosphere, there is a significant natural component of a magnitude comparable to that of current anthropogenic emissions. In a recent review of the literature, Gustin et al. (2008) estimated natural mercury emissions (e.g. from land, oceans, volcanoes) to be in the range 800–3000 t y− 1 (Nriagu (1989), for example, quoting 2500 t y− 1), compared with a range for current global anthropogenic emissions of 2000–2400 t y− 1. Secondly, anthropogenic emissions of mercury to the atmosphere comprise several different chemical forms: Hg(0) (unreactive gaseous elemental mercury), Hg(II) (reactive gaseous mercury, RGM) and Hgpart (mercury attached to particulate material), with relative proportions of 60:30:10 globally (Lohman et al., 2008), although N95% of global atmospheric mercury is believed to exist as gaseous-phase mercury, with a residence time in the atmosphere of 0.5–1 year (Schroeder and Munthe, 1998; Fitzgerald and Lamborg, 2005). Thirdly, these

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

species may be subject to interconversion, variation in degree of atmospheric transport, and deposition by different mechanisms, e.g. wet and dry (Benoit et al., 1998; Fitzgerald et al., 1998; Schroeder and Munthe, 1998; Fitzgerald and Lamborg, 2005; Gustin et al., 2008). Fourthly, as a result of its volatility, it is possible that some deposited mercury may be re-emitted to the atmosphere, a phenomenon that may be relevant with respect to long-term release of mercury from peat bogs (Biester et al., 2003). Direct estimates of anthropogenic mercury emissions to the atmosphere are available for only the past few decades. In the UK, there was a decline of 89% from 63.3 t in 1970 to 6.9 t in 2004 (NAEI, 2006), and, in Europe, a decline of 72% from 860 t in 1980 to 239 t in 2000 (Pacyna et al., 2006a). Globally, anthropogenic mercury emissions may have been as high as 3560 t in 1983 (Nriagu and Pacyna, 1988) before falling to 1881 t in 1990, but have since grown to 2190 t by 2000, with the increasing contributions (amounting to ∼55% of the total in 2000) from the developing countries of Asia at least matching the decline in those from the developed western world (Pacyna et al., 2006b). In 2000, ca. 48% and 65% of the European and global anthropogenic mercury emissions, respectively, were from stationary combustion sources, especially coal (Pacyna et al., 2006a,b). Similarly, direct measurements of atmospheric mercury concentrations and deposition are available for only recent decades. Measured wet deposition rates of mercury in remote areas are reported to have been as high as 20 μg m− 2 y− 1 in the 1980s but somewhat lower since then (Biester et al., 2007). Such values may not equate to the total deposition of mercury, however, because of the existence of additional depositional mechanisms, e.g. dry deposition. Fitzgerald and Lamborg (2005) have noted that the generation of oxidized mercury in the gas phase may lead to significant dry depositional fluxes in addition to the fluxes associated with precipitation, while Gustin et al. (2008) have pointed out that all forms of atmospheric mercury may be dry deposited to some extent. There may even be direct gaseous adsorption on plant surfaces, such as on a peat bog (Benoit et al., 1998; Schroeder and Munthe, 1998). Thus, cores from ombrotophic peat bogs, which receive all of their inputs from the atmosphere and have time horizons extending to thousands of years, afford an indirect opportunity, at least in principle, to investigate the past atmospheric emissions and deposition fluxes or accumulation rates of mercury over much longer time periods than just the past few decades. Freshwater lake sediments have also been used for this purpose (e.g. Bindler et al., 2001) and it has recently been suggested (Biester et al., 2007) that long-term historical records of mercury deposition or accumulation contained within such sediments may be more reliable than those of peat. Firstly, as previously suggested by Biester et al. (2003), the diagenesis and degradation of organic matter in the peat column may lead eventually to the reductive release and loss, through volatility, of mercury (Hg(0)) that was previously bound strongly, as Hg(II), to thiol groups of the organic peat matrix, thereby perturbing the depositional or accumulation record. Secondly, Biester et al. (2007) have also pointed out that the archivederived accumulation rates for mercury in the 20th century are generally lower for lake sediments than for peat bogs and have suggested that the recent accumulation rate of mercury derived from peat bog analysis is overestimated because of underestimation of the age of near-surface material by 210Pb dating. In this paper, we present concentration data and calculated historical accumulation rates for mercury in 14C (half-life 5730 y)- and 210 Pb (half-life 22.35 y)-dated cores from four Scottish ombrotrophic peat bogs, for which we have previously reported data for lead, antimony and arsenic (Cloy et al., 2005, 2008, 2009). In so doing, we also consider the reasons for possible discrepancies between peat bog and lake sediment records for mercury and, in particular, test the proposition (Biester et al., 2007) that there are flaws in the 210Pb dating method as applied to peat bogs that render it unsuitable for the determination of recent accumulation rates of mercury.

5579

2. Materials and methods 2.1. Sampling sites and sample collection Fig. 1 shows the locations of the four ombrotrophic peat bogs from which the cores were collected: Turclossie Moss (TM), northeast; Flanders Moss (FM), west-central; Red Moss of Balerno (RM), eastcentral; and Carsegowan Moss (CM), southwest Scotland. The surface vegetation of the four bogs is dominated by Sphagnum species, including S. magellanicum (TM, FM, RM, CM), S. capillifolium (TM, CM), S. papillosum (FM, RM), S. palustre (TM, FM), S. molle (FM), S. fusum (FM) and S. pulchrum (CM), with the presence of other moss species such as feather moss Pleurozium schreberi (FM, RM) and ribbed bog moss Aulacomnium palustre (FM), as well as varying extents of heather (Calluna vulgaris), cross-leaved heather (Erica tetralix), cotton grass (Eriophorum vaginatum) and lichen (Cladonia portentosa) (McLeod et al., 2005; SNH, 2005). Cores were collected at various times between 2001 and 2004 (Table 1) from moss hummocks on domed parts of the bogs using either a monolith tin (M: 50 cm × 15 cm × 7 cm), which was inserted into the vertical face of a freshly dug pit, and/or a Cuttle and Malcolm (1979) corer (CM: 1 m × 5 cm × 5 cm) pushed vertically into the bog. Surface vegetation noted at the time and specific place of collection

Fig. 1. Map of Scotland showing the locations of the four peat bogs: Turclossie Moss (TM), Flanders Moss (FM), Red Moss of Balerno (RM) and Carsegowan Moss (CM). Smaller map shows location of Scotland in the Northern Hemisphere.

5580

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

Table 1 Collection details for the five cores from the four Scottish ombrotrophic peat bogs. Site

Latitude and longitude

Core identification code

Collection date (month/year)

Core length (analysed), cm

Depth of vegetation/peat interface, cm

Section thickness, cm

Cross-sectional dimensions, cm × cm

Turclossie Moss (TM) Flanders Moss (FM)

57°37′N, 2°11′W 56°09′N, 4°12′W

Red Moss of Balerno (RM) Carsegowan Moss (CM)

55°51′N, 3°20′W 55°12′N, 2°43′W

TM04M-1 FM04-1-M FM01CM-2 RM03CM-2 CM04CM-1

08/2004 10/2004 09/2001 06/2003 05/2004

47 (46) 33 (33) 100 (96) 96 (96) 106 (48)

23 13.5 3 7 9

2 3 2 2 2

15 × 12 20 × 10 5×5 5×5 5×5

of the individual hummock cores was S. capillifolium (TM04M-1), S. papillosum and A. palustre (FM04-1-M), S. palustre (FM01CM-2), P. schreberi (RM03CM-2) and S. magellanicum (CM04CM-1). The M and CM cores, ranging from 33 cm to 106 cm in length, were sliced on site into sections ranging from 2 to 3 cm in thickness and then taken to the laboratory for processing. Core identification codes and collection details are indicated in Table 1.

(i.e. ±15%), in close agreement with the ‘information only’ values of 164 ± 20 μg kg− 1 and 169 ± 7 μg kg− 1 for the ‘acid-extractable’ and total mercury concentrations, respectively (Yafa et al., 2004). The overall detection limit for the mercury analysis was typically equivalent to 20 μg kg− 1.

2.2. Sample preparation

Details of gamma spectrometry for 210Pb dating and accelerator mass spectrometry for 14C dating can be found in Cloy et al. (2005, 2008). The 210Pb dates for TM04M-1 and FM04-1-M cores were obtained using the constant rate of supply (CRS) method (Appleby and Oldfield, 1978) and for RM03CM-2 and CM04CM-1 cores by extrapolation, with the added assistance of matching 206Pb/207Pb ratios, from CRS-obtained dates for sister RM03CM-1 and CM04M cores (Cloy et al., 2008). The total inventories of unsupported 210Pb for TM04M-1, FM04-1-M, RM03CM-1, and CM04M cores were 2.62, 2.54, 2.95 and 3.38 kBq m− 2, respectively. As previously reported (Farmer et al., 2006), material representing 25 ± 7 years of accumulation was estimated to have been lost from the top of the FM01CM-2 core during sample collection in 2001. The 14C ages were calibrated to calendar age ranges using the Oxford Radiocarbon Accelerator Unit calibration program (Bronk Ramsey, 1995).

The wet peat sections were weighed, air-dried at 30 °C, reweighed and then ground using a mortar and pestle. The remaining moisture content of each air-dried peat section was determined on sub-samples by oven drying at 105 °C. Water content and bulk density profiles for the cores are in Cloy et al. (2009). Approximately 0.5–1.0 g of each air-dried peat sample selected for analysis (Table 1) was subjected to a modified acid digestion and wet oxidation method (Ure and Shand, 1974), in which the sample was heated under reflux at ∼60 °C with 10 mL concentrated nitric acid and 10 mL concentrated sulphuric acid for 2 h. After cooling and addition of 10 mL deionised water, 50–60 mL of 6% (w/v) potassium permanganate solution was added and each sample solution left overnight at room temperature to ensure complete oxidation. Subsequent addition of 1 mL of 20% (w/v) hydroxylamine hydrochloride solution reduced excess potassium permanganate, the resultant clear solutions then being filtered through Whatman 541 filter papers. After addition of 1 mL of 0.4% (w/v) Dow Corning DB 110A antifoaming agent, 1 mL of 1% (w/v) potassium dichromate solution and deionised water to make up to volume, the solutions were ready for analysis. 2.3. Sample analysis The acid-extractable mercury concentrations in the sample solutions were determined by continuous flow cold vapour atomic absorption spectrometry using a reductant solution of 0.3% (w/v) sodium borohydride solution and 5 M hydrochloric acid in a Varian VGA-76 accessory attached to a Varian SpectrAA 400 spectrometer (Varian Ltd, Oxford, UK). Standards of suitable mercury concentration (0.5–5 μg L− 1) were employed to generate the absorbance–concentration calibration curve for determination of the mercury concentration in each sample solution. Calculated mercury concentrations (μg kg− 1) in the air-dried peat samples were converted to corresponding concentrations in peat dried at 105 °C through correction for residual moisture content obtained from separately dried aliquots. Sectional samples from long cores FM01CM-02 (n = 48) and RM03CM-2 (n = 47) were analysed singly, but those from TM04M-1 (n = 23), FM04-1-M (n = 11), and CM04CM-1 (n = 24) were analysed predominantly in duplicate (n = 40) or, on occasion, in triplicate (n = 5), quadruplicate (n = 4) or singly (n = 9). For those individual peat core samples (n = 49) analysed in duplicate, triplicate or quadruplicate, the average of the relative standard deviations on the mean values of the mercury concentration was ±14%. For the 40 individual samples of the NIMT/UOE/FM/001 peat reference material analysed, the mean mercury concentration obtained was 167 ± 25 μg kg− 1

2.4. Sample dating

3. Results The mercury concentration profiles for the cores from Turclossie Moss (TM04M-1), Flanders Moss (FM04-1-M and FM01CM-2), Red Moss of Balerno (RM03CM-2) and Carsegowan Moss (CM04CM-1) are shown in Fig. 2, along with the positions of the vegetation–peat interface (i.e. the interface between the organic litter resulting from decay of living moss and the humified peat) and the calendar dates obtained from 210Pb and 14C dating. The maximum sectional mercury concentration recorded was remarkably similar for each core: 515 μg kg− 1 (1967–1973) for TM04M-1, 532 μg kg− 1 (1943–1968) for FM04-1-M, 625 μg kg− 1 (1953–1967) for FM01CM-2, 663 μg kg− 1 (1926–1945) for RM03CM-2, and 613 μg kg− 1 (1930–1944) for CM04CM-1. The onset of the major increase in mercury concentration for long core FM01CM-2 occurred at a depth of 14 cm, below which the mercury concentration averaged 148 ± 43 μg kg− 1 from 14 to 96 cm, sub-dividable into 144 ± 23 μg kg− 1 (14–42 cm), 174 ± 29 μg kg− 1 (42–82 cm) and 82 ± 32 μg kg− 1 (82–96 cm). For long core RM03CM2, the onset of the major increase occurred at a depth of 18 cm, below which the mercury concentration averaged 100 ± 23 μg kg− 1 from 18 to 96 cm, sub-dividable into 107 ± 21 μg kg− 1 from 18 to 68 cm and 85 ± 21 μg kg− 1 from 68 to 96 cm. For the shorter sequences in TM04M-1, FM04-1-M and CM04CM-1, the onset of major increase occurred at depths of 32, 27 and 24 cm, respectively, below which concentrations decreased to 156 ± 15 μg kg− 1 (32–46 cm), 56 μg kg − 1 (30–33 cm) and 203 ± 42 μg kg − 1 (24–48 cm), respectively. Above the major mercury peaks, concentrations decreased to values of 100 μg kg− 1 (2003–2004) for TM04M-1, 190 μg kg− 1 (1997–

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

5581

Fig. 2. Depth profiles of mercury concentration (±1 SD) in the five peat cores (note that the top 21 cm, corresponding to ∼ 25 years of accumulation are missing from core FM01CM-2) from the four Scottish peat bogs (cf. Table 1). Selected 210Pb dates indicate the time of the maximum mercury concentration and the approximate time of onset of major increase of mercury concentration for each core. Selected 14C dates of ca. A.D. 1 and ca. A.D. 800 are shown for FM01CM-2 and RM03CM-2, respectively. The dotted line shows the position of the vegetation–peat interface in each core (cf. text and Table 1).

2004) for FM04-1-M, 169 μg kg− 1 (1993–2003) for RM03CM-2, and 180 μg kg− 1 (2001–2004) for CM04CM-1 in the vegetation at the surface of the bog. As the top 25 years (i.e. post-1976) of material was missing from the FM01CM-2 core (Farmer et al., 2006), surface concentrations were not available for this core. A conservative element such as scandium or titanium is often used as a proxy for soil dust contributions to trace element concentrations in peat core profiles. On the basis of the measured sectional titanium concentrations for the five cores (Cloy et al., 2009) and the mercury/ titanium ratio of 1.8 × 10− 5 for the Upper Continental Crust (Wedepohl, 1995), the maximum soil dust contribution of ‘natural’ mercury did not exceed 3.2% (4.2/133), 3.6% (6.6/184), 2.8% (9.0/316), 5.7% (6.3/109) and 3.0% (4.1/136) of the total mercury concentrations for sections in the TM04M-1, FM04-1-M, FM01CM-2, RM03CM-2, and CM04CM-1 cores, respectively. 4. Discussion 4.1. Mercury accumulation rates Mercury accumulation rates (μg m− 2 y− 1), which are equivalent to atmospheric deposition fluxes if there has been no post-depositional loss or redistribution of mercury, can be calculated for the timespan of each 210Pb-dated peat core section by multiplying the sectional mercury concentration (μg kg− 1) by the sectional weight (kg) and then dividing the product by the cross-sectional area (m2) and the number of years (y) corresponding to each section as determined by 210Pb dating. The accumulation rates (fluxes) so calculated are plotted in Fig. 3. At the most northerly site, Turclossie Moss, the mercury accumulation rate in TM04M-1 increased from 21 μg m− 2 y− 1 ca. 1860 to a maximum of 51 μg m− 2 y− 1 at 1970, followed by a decline to 29 μg m− 2 y− 1 in 2004. In the central belt, at Flanders Moss, there was an increase in core FM04-1-M from 9 μg m− 2 y− 1 ca. 1849 to a maximum of 61 μg m− 2 y− 1 at 1955, followed by a decline to 30 μg m− 2 y− 1 in 2001. The other core (FM01CM-2) at Flanders Moss, however, for which the top 25 years were missing, exhibited an increase from 24 μg m− 2 y− 1 ca. 1890 to a maximum of 183 μg m− 2 y− 1 at 1948, the three-fold enhancement relative to

FM04-1-M perhaps reflecting a degree of within-bog spatial variability (cf. Bindler et al., 2004, and Section 4.2). In the eastcentral belt, at the Red Moss of Balerno, the mercury accumulation rate in RM03CM-2 increased from 47 μg m− 2 y− 1 ca. 1901 to a maximum of 77 μg m− 2 y− 1 at 1935, followed by a decline to 7 μg m− 2 y− 1 in 1999. At the most southerly site, Carsegowan Moss, there was an increase in core CM04CM-1 from 20μg m− 2 y− 1 at ca. 1874 to a maximum of 85 μg m− 2 y− 1 in 1898 (81 μg m− 2 y− 1 in 1923), followed by a decrease to a minimum of 14 μg m− 2 y− 1 in 1983 and a subsequent increase to 43–59 μg m− 2 y− 1 in the late 1990s/early 2000s. These maximum and recent accumulation rates (fluxes) can be contrasted with the average values for the ca.1800 to early 2000s, ca.1850 to early 2000s and ca.1900 to early 2000 periods (Table 2) and, indeed, in the case of the longer cores from Flanders Moss (FM01CM-2) and the Red Moss of Balerno (RM03CM-2), with the long-term average values of 4.5 μg m− 2 y− 1 (A.D. 1–1800) and 3.7 μg m− 2 y− 1 (A.D. 800–1800), respectively. In general, the average accumulation rates since A.D. 1800 are an order of magnitude greater than the long-term averages before 1800, although it should be emphasised that the latter are probably not free from anthropogenic influence as well. The accumulation rates (range 7–43 μg m− 2 y− 1) derived for the most recent years (i.e. late 1990s/early 2000s) are similar to the average values (range 13–27, mean 19 ± 5 μg m− 2 y− 1) for the 19th century, the mean recent accumulation rate of 27 ± 15 μg m− 2 y− 1 being still a factor of ∼6 ± 3 times greater than the long-term (A.D. 1– 1800) average. This derived mean value for the recent accumulation rate of mercury in Scotland is similar to the value of 35.9 μg m− 2 y− 1 directly determined in annual bulk deposition for October 1997– October 1998 collected at the Lochnagar catchment in the Cairngorms in northeast Scotland (Yang et al., 2002), but higher than the modelled wet deposition values for 1998 of up to 7–13 μg m− 2 y− 1 for the relevant areas of Scotland (Lee et al., 2001) and the measured wet deposition in 2005 of ∼3 μg m− 2 y− 1 at Banchory and Auchencorth in northeast and southeast Scotland, respectively (CEH, 2006). As noted by several workers (e.g. Benoit et al., 1998; Fitzgerald and Lamborg, 2005; Biester et al., 2007; Gustin et al., 2008), however, there may be additional inputs from dry deposition to the moss surface of a peat bog. In broad terms, the results for mercury accumulation rates in Scottish peat cores are comparable to those from other countries. The

5582

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

Fig. 3. Calculated mercury accumulation rates since the mid-19th century derived from the mercury concentration data and 210Pb-derived ages for each of the 210Pb-dated sections in the five peat cores from the four Scottish peat bogs.

mean recent accumulation rate of 27 ± 15 μg m− 2 y− 1 for the four Scottish cores is similar to the 6–11 and 19–24 μg m− 2 y− 1 for two sites in Ireland (Coggins et al., 2006), 14 μg m− 2 y− 1 for Greenland (Shotyk et al., 2003), 16 μg m− 2 y− 1 for the Faroe Islands (Shotyk et al., 2005), 17 μg m− 2 y− 1 for Dumme Mosse in Sweden (Bindler, 2003), ∼ 20 μg m− 2 y− 1 for two sites in Switzerland (RoosBarraclough and Shotyk, 2003), and, in North America, 24.5 ± 7.9 μg m− 2 y− 1 in Minnesota, USA (Benoit et al., 1998), 10–18 μg m− 2 y− 1 for three bogs in southern Ontario, Canada (Givelet et al., 2003), ∼ 2– 13 μg m− 2 y− 1 for three bogs in Maine, USA (Norton et al., 1997; RoosBarraclough et al., 2006) and 11 ± 4 μg m− 2 y− 1 for three bogs in Nova Scotia, Canada (Lamborg et al., 2002). The maximum accumulation rates for four of the Scottish cores fall within the range 51–85 μg m− 2 y− 1, again similar to those record-

ed for Ireland (40–50 and 60–70 μg m− 2 y− 1), the Faroe Islands (34 μg m− 2 y− 1), Switzerland (28.9 and 42.8 μg m− 2 y− 1), Minnesota, USA (71 ±22 μg m− 2 y− 1, Benoit et al., 1994), Maine, USA (32 μg m− 2 y− 1, Roos-Barraclough et al., 2006) and Ontario, Canada (54 and 89 μg m− 2 y− 1), while the maximum of 183 μg m− 2 y− 1 at Flanders Moss FM01CM-2 was likewise similar to maximum values in Greenland (164 μg m− 2 y− 1), Denmark (184 μg m− 2 y− 1, Shotyk et al., 2003), Ontario, Canada (141 μg m− 2 y− 1) and Maine, USA (170 and 180 μg m− 2 y− 1, Norton et al., 1997). Mean post-1890 mercury accumulation rate values of up to 33 μg m− 2 y− 1 for Swedish bogs (Bindler et al., 2004) and up to 11 μg m− 2 y− 1 for Norwegian peat (Steinnes and Sjøbakk, 2005) are, however, a little lower than the corresponding average accumulation rate values of 34–51 μg m− 2 y− 1 for the 20th century at the four Scottish bogs (Table 2). Biester et al. (2007) have

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

5583

Table 2 Mercury inventories and accumulation rates (AR) over different time periods for the five cores from the four Scottish ombrotrophic peat bogs. Location/core

Pre-1800 Inventory (mg m− 2)

Turclossie Moss TM04M-1 Flanders Moss FM04-1-M Flanders Moss FM01CM-2 Red Moss of Balerno RM03CM-2 Carsegowan Moss CM04CM-1 a

Average AR (μg m− 2 y− 1)

ca. 1800–2000

ca. 1850–2000

ca .1900–2000

Period of max. AR

Recent period

Inventory (mg m− 2)

Inventory (mg m− 2)

Inventory (mg m− 2)

Average AR (μg m− 2 y− 1)

Average AR (μg m− 2 y− 1)

Average AR (μg m− 2 y− 1)

34

51 (1967–1973)

29 (2003–2004)

47

61 (1943–1968)

30 (1997–2004)

98a

183 (1943–1953)

64 (1967–1976a)

44

77 (1926–1945)

7 (1993–2003)

51

85 (1894–1904)

43 (2001–2004)

Average AR (μg m− 2 y− 1)

5.36

4.69 26

6.26

3.53 30

5.90 31

8.08 (A.D. 1–1800) 4.5 3.72 (A.D. 800–1800) 3.7

Average AR (μg m− 2 y− 1)

9.03a

4.90 38

8.44a 51a

6.37

7.46a 67a

5.81 31

7.93

4.51 38

7.28 39

5.26 47

Top of core 1976.

placed the median value of the maximum mercury accumulation rate from peat bogs around the world over the past 150 years at ∼40 μg m− 2 y− 1. They also observed that the maximum mercury accumulation rates in peat bogs generally occurred during the period from the 1950s to the 1970s, with which the dates for the corresponding maxima at Turclossie Moss (1970) and at Flanders Moss (1948, 1955) are in accord although Red Moss of Balerno (1926–1945) occurs slightly earlier and Carsegowan Moss (1898, 1923) much earlier (Table 2, Fig. 3). Such inter-site variation could reflect different times for peak intensity of any local or regional emission sources along with the associated deposition of such locally or regionally sourced anthropogenic mercury (cf. Bookman et al., 2008, and Section 4.2). The mean inventories of mercury up to the early 2000s for the periods from ca.1800, ca.1850 and ca.1900 for the cores (n = 4, i.e. excluding FM01CM-2) are 6.48 ± 1.07, 5.92 ± 1.06 and 4.55 ± 0.75 mg m− 2, respectively. The mean values of the mercury inventories (n = 4) for the periods from ca.1800–1850, ca.1850–1900 and ca.1900–early 2000s are 0.56 ± 0.14 (0.57 ± 0.12 including FM01CM-2, i.e. n = 5), 1.37 ± 0.45 (1.29 ± 0.43, n = 5) and 4.55 ± 0.75 mg m− 2, respectively. For peat in the Lochnagar catchment, Yang et al. (2002) determined a similar total mercury inventory of 3.05 mg m− 2 for the period from 1860 to the present. Elsewhere, mercury inventories of 1.85–2.40 mg− 2 were found from 1883 to 1996 at Knockroe and 2.04–3.67 mg m− 2 at Letterfrach (1920–1996) in Ireland (Coggins et al., 2006), 0.85– 3.4 mg m− 2 in Sweden post-1890 (Bindler et al., 2004), and 3.44– 4.32 mg m− 2 at Arlberg Bog, Minnesota, USA over ca. 230 years (Benoit et al., 1998). The literature suggests that, prior to the 19th century, it can be a little difficult to discern exactly when human activities began to contribute significantly to atmospheric mercury concentrations, presumably partly because of the natural releases and deposition of mercury. For the latter, on the basis of peat core analysis, Shotyk et al. (2003, 2005) have determined 0.3–0.5 μg m− 2 y− 1 for A.D. 550–975 in Greenland and 1.27 ± 0.38 μg m− 2 y− 1 for 1520 B.C.–A.D. 1385 in the Faroe Islands. Steinnes and Sjøbakk (2005) found 0.3–0.9 μg m − 2 y − 1 for N800 y B.P. in Norway and Roos-Barraclough and Shotyk (2003) 1.0 ± 0.3 and 1.6 ± 0.4 μg m− 2 y− 1 pre-A.D. 1340 in Switzerland. Natural accumulation rates of 0.6 ± 0.2 and ∼ 1 μg m− 2 y− 1 have been reported for mercury over the period from 4000 to 500 y B.P. in Swedish peat cores (Bindler, 2003), 1.7 ± 1.3 μg m− 2 y− 1 for pre-industrial times back to 7000 y B.P. in the USA (Roos-Barraclough et al., 2006) and 1.4 ± 1.0 μg m− 2 y− 1 for 5700 B.C.–A.D. 1470 in Canada (Givelet et al., 2003). On the basis of the peat literature, Biester et al. (2007) consider the median natural accumulation rate of mercury to be ∼ 1 μg m− 2 y− 1 (range 0.6– 1.7 μg m− 2 y− 1). Thus it is likely that the pre-industrial rates of 4.5 μg m− 2 y− 1 and 3.7 μg m− 2 y− 1 found for Flanders Moss (A.D. 1–1800) and Red Moss of Balerno (A.D. 800–1800), respectively, in this study, although similar to the pre-industrial, mid-18th century

value of 4.3 μg m− 2 y− 1 for Arlberg Bog in Minnesota, USA, and, indeed, to the apparent natural rate of 3.3 ± 1.1 μg m− 2 y− 1 for 2000 B.C–A.D. 500 at Penido Vello, Spain (Martinez-Cortizas et al., 1999), have been affected by anthropogenic releases. It is noteworthy, however, that the values are close to the ∼ 5 μg m− 2 y− 1 regarded by Biester et al. (2007) as the more likely natural atmospheric flux, based on lake sediment measurements. 4.2. Intra- and inter-site variability Several workers have pointed out that within-bog variability may occur with respect to the magnitude of inventories and accumulation rates of atmospherically deposited elements determined in peat bog cores. For example, in a study of mercury in nine hummock cores from the Store Mosse bog in Sweden, Bindler et al. (2004) found that, over the past 110 years, cumulative inventories (0.85–3.4 mg m− 2) and estimated maximum accumulation rates (15–60 μg m− 2 y− 1) varied by a factor of 4 across a 2000 m2 area of the bog. At the same time, those of lead (in three cores) varied by a factor of ∼2. Similarly, in an intensive study of 25 cores from the Alport Moor ombrotrophic peat dome in the Peak District, England, Rothwell et al. (2007) reported variation by a factor of 2.2 (i.e. from 6.00–13.42 g m− 2) in the inventories of lead accumulated over the past 150 years. For the two Flanders Moss cores (FM04-1-M and FM01CM-2) reported in this study, the maximum accumulation rates varied by a factor of 3.0 (61– 183 μg m− 2 y− 1) for mercury and 2.6 (22–57 mg m− 2 y− 1) for lead (Cloy et al., 2005, 2008; Farmer et al., 2006). For the naturally occurring radionuclide 210Pb, the average depositional flux calculated from the 210Pb inventories for FM04-1-M (Cloy et al., 2005) and two other Flanders Moss cores collected in 1990 and 1996 (Farmer et al., 2006) varied by a factor of 1.8 (i.e. from 79–144 Bq m− 2 y− 1). These data for mercury, lead and 210Pb demonstrate that within-bog variability of the extent and rate of metal accumulation at Flanders Moss is comparable to that reported for peat bogs elsewhere. There may be numerous causes contributing to such within-bog or intra-site variations in metal accumulation, including variations in microtopography of the bog surface, the nature of the surface plant community and meteorological conditions, all of which can influence the interception and retention of deposition (Bindler et al., 2004; RoosBarraclough et al., 2006; Kempter and Frenzel, 2008). At the Red Moss of Balerno in 2003, separate analysis of living Sphagnum moss (S. papillosum, S. capillifolium) and feather moss (P. schreberi) samples from the bog surface showed that mercury concentrations in the former species (270 μg kg− 1) were, on average, a factor of 2.5 times greater than those of the latter (109 μg kg− 1). This is of interest given that, in contrast to the Sphagnum-dominated hummocks at the other three sites, P. schreberi was the major vegetation growing on the surface of the RM03CM-2 hummock core, which exhibited the lowest recent accumulation rate of mercury (7 μg m− 2 y− 1) of the four Scottish sites.

5584

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

Bindler et al. (2004) have pointed out that major changes in the plant community on bogs can occur over decades and it is therefore possible that both intra-site and inter-site variability in interception, retention and accumulation of atmospheric pollutants could arise over time in consequence. On the other hand, Rothwell et al. (2007), in their Peak District study of lead, which extended to multi-core investigations at two other peatland sites (Torside Clough and Upper North Grain) in addition to Alport Moor, found that within-site variability in lead pollution was dominant at the within-region, i.e. inter-site, scale. In the somewhat geographically wider-scale Scottish study, while it cannot be ruled out that intra-site variations have contributed to observed inter-site variations, it seems more likely, at least so far as variations in the timing of maximum accumulation rates of mercury are concerned, that proximity to local or regional emission sources will have played a significant part. For example, Carsegowan Moss at the southwest corner of Scotland is closer to, and has lain in the path of emissions from, the formerly heavily industrial north of England and Northern Ireland; likewise, Flanders Moss in west-central Scotland with respect to the former industrial conurbation centred on Glasgow. The Red Moss of Balerno in east-central Scotland, with its intermediate timing of 1926–1945 for the maximum mercury accumulation rate, may well have been influenced to some extent by all of these industrial centres. Turclossie Moss in the more remote northeast of Scotland will have been less susceptible than the other three sites to variations arising from proximity to specific local or regional industrial influences. 4.3. Mercury/lead (Hg/Pb) ratios and sources of mercury At depth in the Flanders Moss FM01CM-2 core, there were identifiable zones of distinct lead concentration range, e.g. 80–96 cm (‘Roman’ period, 2.9–4.2 mg kg− 1), 72–80 cm (1.1–1.7 mg kg− 1), 50– 72 cm (0.5–1.0 mg kg− 1), 36–50 cm (1.0–2.2 mg kg − 1) and 18–36 cm (3–13 mg kg− 1) (Yafa, 2004; Farmer et al., 2006), over which the mean sectional mercury concentration was relatively constant, i.e. 91 ± 39, 144 ± 28, 176 ± 24, 172 ± 46 and 148 ± 21 μg kg− 1, respectively, yielding corresponding mean sectional Hg/Pb ratios of 0.027 ±0.013, 0.105 ± 0.013, 0.233 ± 0.055 (i.e. for the zone of lowest lead concentration), 0.120 ± 0.044 and 0.027 ± 0.016, respectively. Similarly, in the Red Moss of Balerno RM03CM-2 core, there were zones of distinct lead concentration range at 64–94 cm (0.3–1.0 mg kg− 1), 46–64 cm (1.3–2.5 mg kg− 1) and 32–46 cm (3.6–25 mg kg− 1) (Cloy et al., 2008) over which the mean sectional mercury concentration also varied little, i.e. 91 ± 23, 111 ± 25 and 107 ± 18 μg kg− 1, respectively, yielding corresponding mean sectional Hg/Pb ratios of 0.163 ± 0.063 (i.e. for the zone of lowest lead concentration), 0.066 ± 0.017 and 0.013 ± 0.009, respectively. It is quite clear, therefore, that, from the bottom of the core up to a depth of 18 cm in FM01CM-2 (from which ∼ 21 cm was missing at the top of the core) and of 32 cm in RM03CM-2, it was the variation in lead, and not mercury, concentration that was the key influence upon the Hg/Pb ratios. For both TM (32–46 cm) and CM (28–48 cm) the pre-industrial mean sectional Hg/Pb ratio was 0.011 ± 0.004, similar to the corresponding values of 0.027 ± 0.016 and 0.013 ± 0.009 for FM01CM-2 and RM03CM-2, respectively. In contrast, above these depths, which corresponded to ca. 1825 in FM01CM-2 and the late 18th century in RM03CM-2, the Hg/Pb ratio was much lower (b0.01) from the early/mid 1800s until the 1970s/ 1980s at all four sites (Fig. 4), despite the considerable increase in mercury concentrations and fluxes observed from the mid-19th century to the mid-20th century (Figs. 2, 3). For example, the average sectional Hg/Pb ratios for TM04M-1 (1844–1979; 18–32 cm), FM04-1M (1832–1982; 9–27 cm), FM01CM-2 (ca. 1825–1976; 0–18 cm), RM03CM-2 (late 18th century–1982; 6–32 cm) and CM04CM-1 (1862–1987; 6–28 cm, excluding values for 16–20 cm, cf. Cloy et al. (2008)) were 0.0047 ± 0.0020, 0.0030 ± 0.0009, 0.0045 ± 0.0015,

0.0016 ± 0.0005 and 0.0026 ± 0.0006, respectively. The much lower values are attributable to the greater increase in lead concentrations in the ‘industrial’ era relative to the ‘pre-industrial’ era when compared with the corresponding increase for mercury. After the late 1970s/early 1980s, the Hg/Pb ratio in the cores increased sharply to surface values of 0.0345 for TM04M-1 (2003– 2004), 0.0127 for FM04-1-M (1997–2004), 0.0113 for RM03CM-2 (1993–2003) and 0.0207 for CM04CM-1 (2001–2004), the last core registering an even greater value of 0.0366 for 1995–2001 (Fig. 4). This increase (by an average factor of 6.65 ± 1.65) in Hg/Pb ratio reflects both the steeper decline in lead emissions to the atmosphere since 1980 (e.g. the Hg/Pb ratio in anthropogenic emissions to the atmosphere in the UK increased ten-fold from 0.0052 in 1980 to 0.0515 in 2004) and the fact that atmospheric mercury, with both significant natural and pollutant components, is more globally (rather than nationally) sourced and distributed. Therefore one would not expect the Hg/Pb ratio recorded in peat bogs (e.g. 0.0345 in 2003–2004 at Turclossie Moss) to correspond exactly to the absolute value of the ratio (0.0515 in 2004) for total anthropogenic emissions to the UK atmosphere, also bearing in mind the possible speciationrelated variations in atmospheric transport, deposition efficiency and uptake. Similarly, for mercury and arsenic (data for the latter in Cloy et al., 2009), the mean surface sectional value of 0.494 ± 0.066 for the Hg/As ratio, based on individual surface section values of 0.556 for TM04M-1, 0.452 for FM04-1-M, 0.423 for RM03CM-2 and 0.545 for CM04CM-1, was close to the average annual ratio of 0.432 ± 0.035 for UK anthropogenic emissions of mercury and arsenic, each of which declined by ca. 50% over the 1996–2004 period. As discussed above for Hg/Pb ratios, such correspondence is probably coincidental, and, in this regard, it is noticeable that the Hg/As ratios in the cores increased rapidly from ca. 1970 (when UK emissions of mercury and arsenic were respectively 9.2 and 5.3 times higher than in 2004), prior to which much lower mean sectional values of 0.114 ± 0.013 for TM04M-1, 0.100 ± 0.035 for FM04-1-M, 0.102 ± 0.014 for RM03CM-2, and 0.064 ± 0.019 for CM04CM-1 were recorded for the cores over the period from the early 1800s until the late 1960s. The implication again is that a significant proportion of mercury in recent decades is being globally distributed from both anthropogenic and natural sources by long-range transport (Fitzgerald et al., 1998; Fitzgerald and Lamborg, 2005) and to a much greater extent than other pollutant elements such as lead and arsenic. At comparatively remote Turclossie Moss, for example, the two-fold decline, at most, in the mercury accumulation rate recorded since the peak of 1970 is much less than the ∼nine-fold drop in UK mercury emissions, suggesting that a substantial proportion of the mercury being deposited at Turclossie Moss is from outwith the UK. Berg et al. (2006) have come to similar conclusions with respect to the origins of mercury deposited in Norway. In modelling studies of the total deposition of mercury in the UK in the late 1990s, Lee et al. (2001) calculated that 41% originated from the UK, 33% from Europe and 25% from the Northern Hemisphere global background, whereas Ryaboshapko et al. (2007) estimated that 50% was from the UK, 5% from Europe and 45% from global, natural and re-emission sources. The two studies differed in their value for total mercury deposition in the UK, estimated at 9.9 t y− 1 by Lee et al. (2001) and ∼ 3.5 t y− 1 by Ryaboshapko et al. (2007). It is noteworthy that the former study considered dry deposition to be responsible for 77% of the total mercury deposited in 1998. 4.4. Reliability of mercury accumulation rates derived from peat bog profiles It has recently been suggested that mercury accumulation rates in peat cores that have been dated by 210Pb could be too high because of underestimation of age by the 210Pb dating method (Biester et al,

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

5585

Fig. 4. Mercury/lead (Hg/Pb) concentration ratios versus 210Pb-derived date for each of the 210Pb-dated sections in the five peat cores since the mid-19th century.

2007). This could occur if some atmospheric particles containing 210Pb fall through surface vegetation before being trapped, thereby yielding an artificially young date for a sub-surface layer and a correspondingly elevated mercury accumulation rate for that time period. This has been advanced as one of the possible reasons for the apparent general discrepancy between recent accumulation rates of mercury as determined by analysis of dated cores from ombrotrophic peat bogs and freshwater lake sediments, the rates for the latter generally being lower. In the study reported here, the availability of 23 cm of vertical transition (in 2-cm sections) from surface living moss through decaying vegetation and organic litter to the underlying humified dark peat in the Turclossie Moss TM04M-1 core (Table 1) afforded an opportunity to investigate this theory.

The 210Pb specific activity data for TM04M-1 are shown in Fig. 5a and the near constancy of the specific activity to a depth of 22 cm, just above the vegetation–peat interface, can be interpreted as resulting from comparable rates of biological and radioactive decay (Norton et al., 1997). The 210Pb data have been used, via the CRS dating model, to generate the dates against which sectional 206Pb/ 207 Pb ratios have been plotted in Fig. 5b and also in Fig. 6 along with the mean decadal 206Pb/207Pb values since 1850 for independently dated archival Scottish Sphagnum mosses of known collection date stored in a herbarium (Farmer et al., 2002). There is generally excellent agreement between the two 206Pb/207Pb records throughout, including a close correspondence between TM04M-1 and the archival mosses over the ca. 40 years from the surface to the

5586

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

Fig. 5. Depth profiles of (a) the specific activity of unsupported 210Pb and (b) the 206Pb/207Pb atom ratio (±1 SD) of lead (Cloy et al., 2008), with selected associated dates determined by the CRS 210Pb dating method, for the Turclossie Moss core TM04M-1. The dotted line shows the position of the vegetation–peat interface in the core (cf. text and Table 1).

vegetation–peat interface. During this time there has been considerable source-related variation in the 206Pb/207Pb ratio of the lead deposited from the atmosphere and the similarity of the records, including the point of inflexion towards higher 206Pb/207Pb values resulting from the phasing out of leaded petrol and the introduction of unleaded petrol in the mid-1980s, strongly suggests that the 210Pb dating of the TM04M-1 core is reliable and that the mercury accumulation rates are ‘real’ rather than artificially produced by inaccuracies in the dating. One of the arguments advanced by Biester et al. (2007) against 210Pb dating of peat cores is that recent accumulation rate profiles of mercury and lead in some peat cores are near identical, implying, quite implausibly in the light of the recent emission histories of the two metals, that the timing of atmospheric pollution changes has been the same. This is clearly not the case for the Turclossie Moss TM04-1-M core described here (cf. Hg/Pb ratio profile since ca.1840 in Fig. 6), but the need to include living and decaying vegetation contributions to the 210Pb inventories of peat cores when using the CRS dating model (Farmer et al., 2006; Olid et al., 2008) is emphasised and strongly endorsed by the results of this study. Thus, it is possible that the typically higher values for recent accumulation rates of mercury derived from peat bogs, relative to lake sediments (Biester et al., 2007), could reflect additional mechanisms for uptake of mercury by the former, such as uptake of gaseous elemental mercury by surface moss (Fitzgerald and Lamborg, 2005), that do not exist for the latter. In addition, the surface waters of lakes may be subject to evasive losses of a significant proportion of directly deposited mercury (Biester et al., 2007). Conversely, the typically lower values for the accumulation rates of natural background mercury in peat, relative to lake sediments (Biester et al., 2007), may be attributable to some loss of mercury during long-term diagenesis of the humified organic matrix of peat (Biester et al., 2003). There is still much of interest and importance to investigate with respect to the biogeochemical processes affecting both environmental mercury (Fitzgerald and Lamborg, 2005) and peat bogs in general (Bindler, 2006; Bindler et al., 2008).

5. Conclusions Maximum mercury accumulation rates of 51–85 μg m− 2 y− 1 deduced from analysis of 210Pb-dated cores from four Scottish ombrotrophic peat bogs during the 20th century were approximately 10–20 times those calculated for the pre-industrial period from A.D. 1– 1800 and 2–3 times the average value for the four sites for the late 1990s/early 2000s. The accumulation rates and integrated post-1800 inventories were broadly similar to those recorded at comparable peat bog sites in western Europe and North America. The reliability of application of 210Pb dating to peat bog cores, a matter which has recently been the subject of debate as a potential underlying cause of differences between peat bogs and freshwater lake sediments with respect to the extent and chronology of mercury accumulation rates, was confirmed in this study by comparison of detailed stable lead isotopic profiling with independent isotopic records of known age. The need to analyse living and decaying vegetation at or near the surface, and not just the underlying peat, in historical studies of accumulation of mercury (and other metals) in peat bogs is strongly emphasised. In addition to local/ regional influences, contribution to mercury deposition and accumulation from the long-range atmospheric transport and global distribution of mercury is invoked to explain the observed temporal trends in mercury relative to those of lead and arsenic in the peat bogs and when compared with the available data for UK atmospheric emissions of these elements since 1970. It is suggested that the discrepancies between the extent of anthropogenic mercury pollution, relative to natural levels, as revealed by peat and freshwater lake sediments should be examined further with particular attention, so far as peat bogs are concerned, to deposition, uptake and retention of mercury by living plants at the surface and to the biogeochemical processes potentially responsible for loss of mercury from humified peat at depth. Acknowledgements We gratefully acknowledge permission from Scottish Natural Heritage and Scottish Wildlife Trust to collect cores from the peat

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

Fig. 6. A comparison of the 206Pb/207Pb atom ratio (±1 SD) record for the Turclossie Moss TM04M-1 core (closed circles) (cf. Fig. 5b) with the mean decadal 206Pb/207Pb atom ratio (±1 SD) record for Scottish archival herbarium moss (open circles) (Farmer et al., 2002) since the mid-19th century (bottom graph). The corresponding records of mercury accumulation rate and mercury/lead (Hg/Pb) ratio for the TM04M-1 core are shown in the middle and top graphs, along with ±1 SD values (based on ±1 SD for concentration).

bogs, the loan of a Cuttle and Malcolm corer from the Macaulay Institute, the assistance of L.J. Eades, C. Yafa and undergraduate project students J. Shimwell, D. MacDonald and A. Rasool at the University of Edinburgh in various aspects of field work, sample preparation and analysis, SUERC AMS staff for 14C determination, C. Donnelly for gamma spectrometric analysis and the UK Natural Environment Research Council for funding Joanna Cloy's PhD studentship.

References Appleby PG, Oldfield F. The calculation of 210Pb dates assuming a constant rate of supply of unsupported 210Pb to the sediment. Catena 1978;5:1–8. Benoit JM, Fitzgerald WF, Damman AWH. Historical atmospheric mercury deposition in the mid-continental U.S. as recorded in an ombrotrophic peat bog. In: Watras CJ, Huckabee JW, editors. Mercury pollution: integration and synthesis. Boca Raton: Lewis; 1994. p. 187–202. Benoit JM, Fitzgerald WF, Damman AWH. The biogeochemistry of an ombrotrophic bog: evaluation of use as an archive of atmospheric mercury deposition. Environ Res 1998;78:118–33.

5587

Berg T, Fjeld E, Steinnes E. Atmospheric mercury in Norway: contributions from different sources. Sci Total Environ 2006;368:3–9. Biester H, Martinez-Cortizas A, Birkenstock S, Kilian R. Effect of peat decomposition and mass loss on historic mercury records in peat bogs from Patagonia. Environ Sci Technol 2003;37:32–9. Biester H, Bindler R, Martinez-Cortizas A, Engstrom DR. Modeling the past atmospheric deposition of mercury using natural archives. Environ Sci Technol 2007;41:4851–60. Bindler R. Estimating the natural background atmospheric deposition rate of mercury utilizing ombrotrophic bogs in southern Sweden. Environ Sci Technol 2003;37:40–6. Bindler R. Mired in the past — looking to the future: geochemistry of peat and the analysis of past environmental changes. Global Planet Change 2006;53:209–21. Bindler R, Renberg I, Appleby PG, Anderson NJ, Rose NL. Mercury accumulation rates and spatial patterns in lake sediments from west Greenland: a coast to ice margin transect. Environ Sci Technol 2001;35:1736–41. Bindler R, Klarqvist M, Klaminder J, Förster J. Does within-bog spatial variability of mercury and lead constrain reconstructions of absolute deposition rates from single peat records? The example of Store Mosse, Sweden. Global Biogeochem Cycles 2004;18:GB3020. doi:10.1029/2004GB002270. Bindler R, Renberg I, Klaminder J. Bridging the gap between ancient metal pollution and contemporary biogeochemistry. J Paleolimnol 2008;40:755–70. Bookman R, Driscoll CT, Engstrom DR, Effler SW. Local to regional sources affecting mercury fluxes to New York lakes. Atmos Environ 2008;42:6088–97. Bronk Ramsey C. Radiocarbon calibration and analysis of stratigraphy: the OxCal program. Radiocarbon 1995;37:425–30. CEH (Centre for Ecology and Hydrology). Heavy metals report, UK heavy metal monitoring network; 2006. May. Cloy JM, Farmer JG, Graham MC, MacKenzie AB, Cook GT. A comparison of antimony and lead profiles over the past 2,500 years in Flanders Moss ombrotrophic peat bog, Scotland. J Environ Monit 2005;7:1137–47. Cloy JM, Farmer JG, Graham MC, MacKenzie AB, Cook GT. Historical records of atmospheric Pb deposition in four Scottish ombrotrophic peat bogs: an isotopic comparison with other records from western Europe and Greenland. Global Biogeochem Cycles 2008;22:GB2016. doi:10.1029/2007GB003059. Cloy JM, Farmer JG, Graham MC, MacKenzie AB. Retention of As and Sb in ombrotrophic peat bogs: records of As. Sb and Pb deposition at four Scottish sites. Environ Sci Technol 2009;43:1756–62. Coggins AM, Jennings SG, Ebinghaus R. Accumulation rates of the heavy metals lead, mercury and cadmium in ombrotrophic peatlands in the west of Ireland. Atmos Environ 2006;40:260–78. Cuttle SP, Malcolm DC. A corer for taking undisturbed soil samples. Plant Soil 1979;51: 297–300. Farmer JG, Eades LJ, Atkins H, Chamberlain DF. Historical trends in the lead isotopic composition of archival Sphagnum mosses from Scotland (1838–2000). Environ Sci Technol 2002;36:152–7. Farmer JG, Graham MC, Yafa C, Cloy JM, Freeman AJ, MacKenzie AB. Use of 206Pb/207Pb ratios to investigate the surface integrity of peat cores used to study the recent depositional history and geochemical behaviour of inorganic elements in ombrotrophic bogs. Global Planet Change 2006;53:240–8. Fitzgerald WF, Lamborg CH. Geochemistry of mercury in the environment. In Environmental geochemistry (ed. Lollar BS), Vol 9, Treatise in geochemistry (eds. Holland HD, Turekian KK), Elsevier-Pergamon, Oxford, 2005, pp. 107–48. Fitzgerald WF, Engstrom DR, Mason RP, Nater EA. The case for atmospheric mercury contamination in remote areas. Environ Sci Technol 1998;32:1–7. Givelet N, Roos-Barraclough F, Shotyk W. Predominant anthropogenic sources and rates of atmospheric mercury accumulation in southern Ontario recorded by peat cores from three bogs: comparison with natural “background” values (past 8000 years). J Environ Monit 2003;5:935–49. Gustin MS, Lindberg SE, Weisberg PJ. An update on the natural sources and sinks of atmospheric mercury. Appl Geochem 2008;23:482–93. Kempter H, Frenzel B. Titanium in ombrotrophic Sphagnum mosses from various peat bogs of Germany and Belgium. Sci Total Environ 2008;392:324–34. Klaminder J, Renberg I, Bindler R. Isotopic trends and background fluxes of atmospheric lead in northern Europe: analyses of three ombrotrophic bogs from south Sweden. Global Biogeochem Cycles 2003;17:1019. doi:10.1029/2002GB001921. Lamborg CH, Fitzgerald WF, Damman AWH, Benoit JM, Balcom PH, Engstrom DR. Modern and historic atmospheric mercury fluxes in both hemispheres: global and regional mercury cycling implications. Global Biogeochem Cycles 2002;16:1104. doi:10.1029/2001GB001847. Lee DS, Nemitz E, Fowler D, Kingdon RD. Modelling atmospheric mercury transport and deposition across Europe and the UK. Atmos Environ 2001;35:5455–66. Lohman K, Seigneur C, Gustin M, Lindberg S. Sensitivity of the global atmospheric cycle of mercury to emissions. Appl Geochem 2008;23:454–66. Martinez-Cortizas A, Pontevedra-Pombal X, Garcia-Rodeja E, Nóvoa-Muñoz JC, Shotyk W. Mercury in a Spanish peat bog; archive of climate change and atmospheric metal deposition. Science 1999;284:939–42. McLeod CR, Yeo M, Brown AE, Burn AJ, Hopkins JJ, Way SF, editors. The habitats directive: selection of special areas of conservation in the UK. 2nd edn. Joint Nature Conservation Committee, Peterborough; 2005. www.jncc.gov.uk/SACselection. NAEI. UK Emissions of Air Pollutants 1970 to 2004. 18th annual report, UK National Atmospheric Emissions Inventory, National Environmental Technology Centre; 2006. Norton SA, Evans GC, Kahl JS. Comparison of Hg and Pb fluxes to hummocks and hollows of ombrotrophic Big Heath Bog and to nearby Sargent Mt. Pond, Maine, USA. Water Air Soil Pollut 1997;100:271–86. Nriagu JO. A global assessment of natural sources of atmospheric trace metals. Nature 1989;338:47–9.

5588

J.G. Farmer et al. / Science of the Total Environment 407 (2009) 5578–5588

Nriagu JO, Pacyna JM. Quantitative assessment of worldwide contamination of air, water and soils by trace metals. Nature 1988;333:134–9. Olid C, Garcia-Orellana J, Martínez-Cortizas A, Masqué P, Peiteado E, Sanchez-Cabeza J-A. Role of surface vegetation in 210Pb dating of peat cores. Environ Sci Technol 2008;42: 8858–64. Pacyna EG, Pacyna JM, Fudala J, Strzelecka-Jastrzab E, Hlawiczka S, Panasiuk D. Mercury emissions to the atmosphere from anthropogenic sources in Europe in 2000 and their scenarios until 2020. Sci Total Environ 2006a;370:147–56. Pacyna EG, Pacyna JM, Steenhuisen F, Wilson S. Global anthropogenic mercury emission inventory for 2000. Atmos Environ 2006b;40:4048–63. Roos-Barraclough F, Shotyk W. Millennial-scale records of atmospheric mercury deposition obtained from ombrotrophic and minerotrophic peatlands in the Swiss Jura Mountains. Environ Sci Technol 2003;37:235–44. Roos-Barraclough F, Givelet N, Cheburkin AK, Shotyk W, Norton SA. Use of Br and Se in peat to reconstruct the natural and anthropogenic fluxes of atmospheric Hg: a 10000-year record from Caribou Bog, Maine. Environ Sci Technol 2006;40:3188–94. Rothwell JJ, Evans MG, Lindsay JB, Allott THE. Scale-dependent spatial variability in peatland lead pollution in the southern Pennines. Environ Pollut 2007;145:111–20. Ryaboshapko A, Bullock Jr OR, Christensen J, Cohen M, Dastoor A, Ilyin I, et al. Intercomparison study of atmospheric mercury models: 2. Modelling results vs. long-term observations and comparison of country deposition budgets. Sci Total Environ 2007;377:319–33. Schroeder WH, Munthe J. Atmospheric mercury — an overview. Atmos Environ 1998;32: 809–22. Shotyk W, Weiss D, Appleby PG, Cheburkin AK, Frei R, Gloor M, et al. History of atmospheric Pb deposition since 12, 370 14C years BP from a peat bog, Jura Mountains, Switzerland. Science 1998;281:1635–40.

Shotyk W, Goodsite ME, Roos-Barraclough F, Frei R, Heinemeier J, Asmund G, et al. Anthropogenic contributions to atmospheric Hg, Pb and As accumulation recorded by peat cores from southern Greenland and Denmark dated using the 14C “bomb pulse curve”. Geochim Cosmochim Acta 2003;67:3991–4011. Shotyk W, Goodsite ME, Roos-Barraclough F, Givelet N, Le Roux G, Weiss D, et al. Accumulation rates and predominant atmospheric sources of natural and anthropogenic Hg and Pb on the Faroe Islands. Geochim Cosmochim Acta 2005;69:1-17. SNH (Scottish Natural Heritage). http://www.snh.org.uk/Peatlands/wc-RaisedBog.asp, 2005. Steinnes E, Sjøbakk TE. Order-of-magnitude of increase of Hg in Norwegian peat profiles since the onset of industrial activity in Europe. Environ Pollut 2005;137:365–70. Ure AM, Shand CA. The determination of mercury in soils and related materials by coldvapour atomic absorption spectrometry. Anal Chim Acta 1974;72:63–77. Wedepohl KH. The composition of the continental crust. Geochim Cosmochim Acta 1995;59:1217–32. Yafa, C. The accurate analysis and environmental geochemistry of inorganic elements in peat bogs. PhD Thesis, Univ Edinburgh, 2004, 361 pp. Yafa C, Farmer JG, Graham MC, Bacon JR, Barbante C, Cairns WRL, et al. Development of an ombrotrophic peat bog (low ash) reference material for the determination of elemental concentrations. J Environ Monit 2004;6:493–501. Yang H, Rose NL, Battarbee RW, Boyle JF. Mercury and lead budgets for Lochnagar, a Scottish mountain lake and its catchment. Environ Sci Technol 2002;36:1383–8.