TEMPORAL AND SPATIAL VARIATION OF Hg AND Pb CONCENTRATIONS IN NORTHEASTERN U.S.-CANADIAN SUBALPINE FOREST FLOOR SOILS

Gordon C. Evans¹, S. A. Norton², I. J. Fernandez³, J. S. Kahl^2,4, and D. Hanson⁵, ¹ East Lansing, MI 48823; ²corresponding author, Department of Geological Sciences, University of ME, Orono, ME 04469-5790 Norton@Maine.Edu; ³Department of Plant, Soil, and Environmental Sciences, University of Maine, ⁴ Water Research Institute, University of Maine, ⁵Cambridge, MA 02140.

ABSTRACT

Forest floor soil from high elevation spruce-fir sites from southern Vermont, U.S. to the Gaspé Peninsula, Québec, Canada was sampled in 1979 and re-sampled in 1996. Soil pH, LOI, Hg, and Pb were determined for litter and the underlying (F +H) horizon from both years. For 1979, the concentration of Hg in litter ranged between 150 and 300 (g/kg, with no regional trend. By 1996, Hg concentrations were 25 to 50% lower in litter. The concentration of Pb in 1979 litter samples decreased significantly from 160 to 200 mg/kg in southern Vermont to 60 mg/kg in Québec. By 1996, the Pb concentration ranged between 32 and 66 mg/kg with no regional trend. Decreases at sites were proportional to the absolute value in 1979 for Pb and Hg. Litter chemistry indicates a decline in atmospheric deposition from 1979 to 1996 for Hg and Pb and probably higher Hg and Pb accumulation rates in the past to the southwest. The concentrations for Pb in the underlying (F + H) horizon irregularly decrease from southwest to northeast, suggesting that earlier deposition of Pb was higher in the southwest.

INTRODUCTION

The distribution and concentrations of trace metals in New England soils (Hanson et al., 1982), and lake sediment and peat (Norton et al., 1997) provide evidence for changing anthropogenic emissions of metals to the atmosphere and their subsequent deposition. The sources for much of the pollution reaching the northeastern U.S. and Canada lie principally to the west and southwest. The U. S. Clean Air Act (CAA) of 1970, the CAA Amendments of 1990, and the removal of tetra-ethyl Pb from gasoline have caused significant reductions of emissions of Pb. Estimated emissions of Pb declined from 75,000 to 5,000 short tons from 1980 to 1995 (e.g., Eisenreich et al., 1986). Declines in emissions and deposition of Hg in New England over a similar period have been estimated from paleolimnological records and time series studies of soils (Norton et al., 1997). Analyses of forest soils have been commonly utilized to assess spatial trends of pollutant deposition (Hanson et al., 1982; Johnson et al., 1982; Friedland and Johnson, 1985; Herrick and Friedland, 1990). High elevation soils are particularly sensitive to changes in atmospheric chemistry due to higher wet deposition and deposition of pollutants by wet and dry deposition, fog, and rime ice (Schlesinger and Reiners, 1974). For example, Weathers et al. (in press) conclude that deposition of Pb (at least) in forested terrane is strongly dependent on vegetation type, aspect, topography, and elevation. Fluxes of Pb, and consequently other chemicals, to the forest floor may vary locally by a factor of more than three. Most watershed-scale evidence indicates that a large percentage of atmospherically deposited Pb is retained in organic soil.

Mercury is of particular concern in the northeastern U.S. and the Canadian Maritime Provinces due to elevated concentrations of methyl-Hg in fish. Most studies suggest that 90 to 95% of Hg deposited on forested catchments is retained by the soils. Haines (pers. comm.) estimated that 90 to 95% of Hg deposited from the atmosphere in throughfall at the Bear Brook Watershed in Maine is currently retained.

This paper evaluates the concentrations of Hg and Pb in high elevation forest floor litter collected in 1979 and from the same locations in 1996 on a transect from southern Vermont to the Gaspé Peninsula, Québec.

METHODS

Site selection and sampling: The 1979 study sampled 14 mountain sites. Sites 1-3 are in Vermont, 4-6 are in New Hampshire, 7-12 are in Maine, 13 is in New Brunswick, and 14 is in Québec (sites are arranged along the transect on Figure 1). Seven mountains were resampled in 1996, six in Maine, and one in Vermont. An alternate site (Mt. Richardson, 1996) was selected in Québec due to restricted access at Mt. Cartier (1979). Sampling sites for 1979 and 1996 were in the sub-alpine spruce-fir zone, just below the tree line or just below the summit in areas of dense conifers. Trees were typically even-aged (>50 years) and 4-6 m tall. Sample points were relatively horizontal and free of understory vegetation. All sites were on southwest- or west-facing slopes. The sampling elevations for 1996 approximated those of 1979 (± 10 m). Litter samples consisted of undecomposed material which easily separated from the F and H horizons in the forest floor. Samples for both surveys were collected at three elevations, separated by 10 m elevation. At each elevation, five sampling points were selected, typically within a radius of 3 m. Thus each mountain yielded 15 litter samples. For 1996, a 15x15 cm template was used to ensure each block of soil came from a uniformly-sized area.

Laboratory processing: Samples from 1979 were dried in field collection paper bags at 50^oC. Samples from 1996 were dried in glass trays at 35^oC. All samples were ground to <40 mesh in a stainless steel Wiley Mill. Homogenized sub-samples (both surveys) were re-dried at 35^oC prior to weighing out aliquots for chemical analyses, and determination of loss-on-ignition (LOI). The pH of soils was done in 0.01 M CaCl₂ and determined with an Orion 8104 Combination pH probe. For Hg, all litter samples from 1996 were analyzed individually. For the 1979 Hg samples, we combined equal masses of each litter sample from a site (typically 15 samples) to make one composite sample and then sampled the composite. Thus for each mountain, sample n = 15 for 1996 and n = 1 for 1979 for Hg. All samples for Hg analysis were digested following modified USEPA methods and analyzed by cold-vapor atomic absorption (1979) or fluorescence (1996). The two methods yielded virtually identical results for certified reference materials.

For determination of Pb, Hanson et al. (1982) used HNO₃ and perchloric acid (HClO₄) to digest the 1979 organic soil. The 1996 samples were digested following a modification of USEPA Method 3050. The concentrations of Pb were determined by AAS on a Perkin-Elmer Flame AA (1979) and 4100ZL Furnace AA (1979 and 1996). Re-analysis of 1979 samples demonstrated that 1996 methods yielded a +5% linear bias (i.e., Hansen et al. data for the same samples were 5% lower (R2 = 0.97)). We corrected 1979 values, that were not remeasured, for this bias.

RESULTS and DISCUSSION

Mercury concentrations in 1979 litter ranged from 155 to 305 (g/kg (all results are based on dry weight) and varied erratically along the 1979 transect with a slight tendency to lower values in Maine (sites 7-12) and New Brunswick (site 13) (Figure 1). The 1996 values for spatially coincident sites are all lower, the difference ranging from 15 to 50%.

Concentrations of Pb in litter in 1979 ranged from nearly 200 mg/kg in Vermont to 60 in Québec with a statistically significant decline from southwest to northeast. Sites with higher concentrations also had more variable concentrations (Figure 1). By 1996, concentrations ranged from 23 to 66 mg/kg and there was no regional gradient for the concentration of Pb in litter. Reductions in concentration ranged from 85% in Vermont to less than 50% in Québec where absolute values were low in 1979. The relatively low ash % (100%-LOI) for all samples of litter in both surveys indicates that virtually all of the Hg and Pb may be attributed to accumulation in the organic soil because the concentration of Pb in the bedrock is typically less than about 20 mg/kg inorganic matter. Hg is also very low in the bedrock (but typically not determined). Hg in lake sediments is strongly associated with the organic fraction of the sediment, not the inorganic matter. Thus, inorganic soil matter dilutes the concentration of Hg and Pb in the forest floor soil.

Litter pH for 1996 samples increased from 3.5 in Vermont to 3.9 in Québec, paralleling the precipitation pH gradient of 1996. Values for pH have increased significantly at most sites since 1979, particularly in Vermont, where mean annual H⁺ deposition decreased from 832 to 309 µeq/m2. The pH of all (F + H) samples was lower than in litter, near pH 3, and was unchanged from 1979 to 1996.

Concentrations of Hg in the (F + H) composite 1996 samples decline irregularly from 461 in Vermont (site 2) to 291 g/kg in Québec from Site 2 to 14b. Lower values occur in Maine. Pb declined from 318 to 95 mg Pb/kg. In contrast, there is no regional trend for Hg and Pb in litter in 1996.

Ash% ranges from 4.76 to 10.3% in (F + H) samples. The general downward increase in ash% from L to (F+H) probably results from two processes: a relative increase in amount of mineral soil mixed with the organic matter, and preferential loss of organic matter as a result of decay. Both Hg and Pb are strongly retained by organic matter, so if the H layer material, totally or partly, pre-dates increased anthropogenic atmospheric inputs, concentrations would be lower in the H layers. Additionally, the bedrock at all sites has concentrations of Hg and Pb lower than in the organic soil and thus mineral soil dilutes the concentration of Hg and Pb deposited from the atmosphere. This is particularly clear for Pb. Emissions of Pb have decreased dramatically since 1970, and deposition of Pb has decreased (Norton et al., 1997) so the newer material in the L layer should have lower Pb concentrations. Concentrations of Pb may increase in the (F + H) layer as organic matter decays and the Pb becomes relatively enriched, and older Pb-rich litter is converted into (F + H). The Pb and Hg concentrations for the (F + H) layers are higher than for litter for all sites.

CONCLUSIONS

In 1996, forest floor soils on six Maine peaks, plus one each in Vermont and Québec were re-sampled to evaluate whether reduced emissions of SO₄ , Pb, and other atmospheric pollutants have resulted in changes in sub-alpine forest soil chemistry since 1979. From 1979 to 1996, precipitation pH in Maine had increased from about 4.3 to 4.6, and Pb in air had decreased to 0.03 µg/m³, two percent of the National Ambient Air Quality Standard, by 1993.

Concentration of Hg in litter varied erratically from about 150 to 300 ug/kg along the transect in 1979. Hg values did not vary significantly along the transect in 1996 and were consistently significantly lower by 25 to 50%. This finding is consistent with that of Engstrom et al. (1997) and Norton et al. (1997) based on paleolimnological assessments. The 1979 concentration of Pb in litter decreased significantly from southwestern New England sites (160 to 200 mg/kg) to about 60 mg/kg in Quebec. Reduced atmospheric deposition of Pb since 1979 is reflected in the litter chemistry, where Pb concentrations for 1996 ranged from 32 to 66 mg/kg across all sites, with no regional trend. Lead concentrations decreased from 195 to 37 mg/kg at Site 2 in Vermont from 1979 to 1996. Smaller percentage decreases for Pb occurred at all other sites, approximately in proportion to the absolute concentration for 1979.

REFERENCES

Eisenreich SJ, Metzer NA, Urban NR (1985), Environ. Sci. Tech. 20:171-174.

Engstrom DR, Swain EB (1997), Environmental Science and Technology. 31:960-967.

Friedland AJ, Johnson AH (1985), J. Environ. Quality. 14: no. 3, 332-336.

Hanson DW, Norton SA, Williams JS (1982), Water, Air, Soil Pollution. 18:227-239.

Herrick GT, Friedland AJ (1990), Water, Air, Soil Pollution. 53:151-157.

Johnson AH, Siccama TG, Friedland AJ (1982), J. Environ, Qual. 11: no. 4, 577-580.

Norton SA, Evans GC, Kahl JS (1997), Water, Air, Soil Pollution. 100:271-286.

Schlesinger WH, Reiners WA (1974), Ecology. 55:378-386.

Weathers KC, Lovett GM, Likens GE, Lathrop R, in press, Ecol. Applic.

Figure Caption

Figure 1: Changes in the concentrations of Hg (triangles) and Pb (squares) in forest floor litter from southwestern Vermont, U.S. (Site 1) to the Gaspé Peninsula, Québec, Canada based on samples from 1996 (open symbols) and 1979 (filled symbols). The standard deviation are given for 1996/79 data where (n = 15).