LITTERFALL MERCURY IN TWO FORESTED ... - Springer Link

LITTERFALL MERCURY IN TWO FORESTED WATERSHEDS AT ACADIA NATIONAL PARK, MAINE, USA KATHERINE D. SHEEHAN1 , IVAN J. FERNANDEZ2,∗ , J. STEPHEN KAHL3 and ARIA AMIRBAHMAN4 1

Sen. George J. Mitchell Center for Environmental and Watershed Research, University of Maine, Orono, ME 04469-5710; 2 Department of Plant, Soil, and Environmental Sciences, University of Maine, Orono, ME, 04469-5722, USA; 3 Center for the Environment, Plymouth State University, Plymouth, NH, 03264, USA; 4 Department of Civil and Environmental Engineering, University of Maine, Orono, ME, 04469, USA (∗ author for correspondence, e-mail: [email protected]; Tel: (207)-581-2932, Fax: (207)-581-2999)

(Received 24 May 2005; accepted 12 September 2005)

Abstract. Litterfall can be an important flux of mercury (Hg) to soils in forested landscapes, yet typically the only available data to evaluate Hg deposition is from precipitation Hg monitoring. Litterfall was collected at 39 sampling sites in two small research watersheds, in 2003 and 2004, and analyzed for total Hg. Four vegetation classes were designated in this study as hardwoods, softwoods, mixed and scrub. The mean litter Hg concentration in softwoods (58.8 ± 3.3 ng Hg g−1 ) was significantly greater than in mixed (41.7 ± 2.8 ng Hg g−1 ) and scrub (40.6 ± 2.7 ng Hg g−1 ), and significantly lower than in hardwoods (31.6 ± 2.6 ng Hg g−1 ). In contrast, the mean weighted litter Hg flux was not significantly different among vegetation classes. The lack of a significant difference in litter Hg flux between hardwoods and softwoods was attributable to the large autumnal hardwood litter Hg flux being balanced by the higher softwood litter Hg concentrations, along with the higher chronic litterfall flux throughout the winter and spring in softwoods. The estimated annual deposition of Hg via litterfall in Hadlock Brook watershed (10.1 μg m−2 ) and Cadillac Brook watershed (10.0 μg m−2 ) was greater than precipitation Hg deposition and similar to or greater than the magnitude of Hg deposition via throughfall. These results demonstrate that litterfall Hg flux to forested landscapes can be at least as important as precipitation Hg inputs. Keywords: Acadia National Park, dry deposition, flux, forested watersheds, litterfall, mercury, vegetation

1. Introduction Quantification of dry-deposition flux is a gap in our knowledge of Hg cycling in terrestrial ecosystems (Iverfeldt et al., 1996; Mason et al., 2005). Emphasis has typically been placed on the sensitivity of aquatic ecosystems to the impacts of mercury (Hg) pollution due to the potential for Hg methylation and bioaccumulation, yet a significant portion of Hg in those ecosystems is first intercepted from the atmosphere by the associated terrestrial ecosystems (Grigal, 2002). Recent research has investigated mechanisms by which mercury (Hg) is incorporated into forest foliage. Gaseous elemental Hg (Hg0 ) is by far the most prevalent Water, Air, and Soil Pollution (2006) 170: 249–265 DOI: 10.1007/s11270-006-3034-y

C

Springer 2006

250

K. D. SHEEHAN ET AL.

form of Hg in the atmosphere (Ericksen et al., 2003), but Hg0 is also relatively insoluble and unreactive (Morel et al., 1998; Rea et al., 2000; Grigal, 2002). In contrast, Hg (II) prevails in soil and foliage (Skyllberg et al., 2000). Atmospheric Hg accumulates on the surface of foliage via oxidation of Hg0 at the leaf surface, or adsorption of gaseous Hg (II) or particulate Hg at the air-water interface on the leaf surface (Bishop et al., 1998; Rea et al., 2002). Once deposited on the leaf surface, Hg (II) is either washed off during precipitation (Rea et al., 2000) or photoreduced and volatilized (Leonard et al., 1998a,b). Atmospheric Hg also accumulates in leaf tissue via stomatal uptake of Hg0 (Hanson et al., 1995; Frescholtz et al., 2003). Once internalized, gaseous Hg0 is oxidized and bound to the leaf tissue (Du and Fang, 1982). Rea et al. (2002) reported that only 25% of the modeled potential Hg0 deposition in Vermont and Michigan could account for the Hg accumulated in the foliage, supporting the possibility of a bi-directional exchange of Hg with the atmosphere suggested by other researchers (Hanson et al., 1995; Lindberg, 1996). Conversely, Greger et al. (2004) suggested that Hg assimilated in foliar tissue is not released into the atmosphere, but bound tightly. However, it is important to note that the unidirectional Hg exchange pathway Greger et al. (2004) described refers to Hg translocated from soil solution and may inadequately represent mechanisms important in gas-phase foliar uptake experiments. Foliar Hg concentration increases substantially over the growing season as a result of a combination of factors: (1) Elevated atmospheric Hg0 concentrations occur during summer months resulting from increased temperature and biologic activity, (2) high rates of physiological activity of vegetation cause Hg to be taken up through stomata, and (3) Hg is tightly bound to foliar tissues once internalized (Lindberg et al., 1992; Rasmussen, 1995; St. Louis et al., 2001; Grigal, 2002; Ericksen et al., 2003; Frescholtz et al., 2003; Grigal, 2003). Rea et al. (2002) reported Michigan and Vermont hardwood foliar Hg concentrations ranging from ∼3.5 ng g−1 in the month of May, to ∼25 ng g−1 in September. With the exception of a single month at one study site, foliar Hg concentration for each successive month was significantly greater than the previous month. A temporal trend for progressively increasing Hg concentrations in live foliage throughout the growing season was also reported by Lindberg (1996) in Tennessee. Litter Hg concentrations have been reported to be up to 60% greater than Hg concentrations in live foliage collected at the end of the growing season due to the accumulation of Hg over time, and the concentration of Hg relative to nutrients that leach and are translocated out of foliage during senescence (Lindberg, 1996; Rea et al., 1996, 2002; Tyler, 2005). Differences among plant species in the length of time for leaf exposure during the year results in important differences in their Hg uptake capability. In a review, Grigal (2003) reports that the concentration of Hg in softwood foliage is generally greater than that in hardwoods from the same location, since the needles of softwoods typically live for multiple years whereas hardwood foliage lives for a single growing season. Softwood foliage also has a greater surface area per unit mass of

LITTERFALL MERCURY IN FORESTED WATERSHEDS

251

tissue than hardwood foliage (Rassmussen et al., 1991; Kolka et al., 1999), as well as a greater surface area per unit volume (St. Louis et al., 2001), which increases its scavenging efficiency and further amplifies softwood foliar Hg concentrations. Yet for litter, researchers report mixed results. Lindberg (1996) reported a lower mean Hg concentration in softwood litter compared to hardwood litter, while St. Louis et al. (2001) reported a greater mean Hg concentration from litter collected at softwood sites than litter collected at sites with a mixture of softwoods and hardwoods. Rasmussen et al. (1991) found higher mean Hg concentrations in softwood foliage than in hardwood foliage. Median Hg concentrations were greater for softwood foliage compared to hardwood foliage in a study by Schwesig and Matzner (2000). Wide variations in reported foliar Hg concentrations underscore the need for inter-specific comparisons at the same study site, where Hg inputs and timing can be controlled. The flux of Hg to soil or water surfaces from litter depends on both plant tissue Hg concentrations and the rate of litterfall. Measuring litter Hg flux is essential to quantifying Hg mass balance in ecosystems, particularly in forested landscapes. The three predominant Hg input vectors to forested watersheds are precipitation, throughfall, and litterfall (Grigal, 2002). Precipitation Hg deposition is measured as precipitation collected for Hg analysis during events that has not been influenced by vegetation before reaching the soil or water surface. For the purpose of this investigation throughfall is defined as presented by Grigal (2002), which is precipitation collected below the forest canopy during a precipitation event, unless otherwise noted. Throughfall Hg includes Hg in precipitation, Hg dry-deposited between precipitation events and subsequently washed off the foliar surface, and the potential net of any leaching that may occur at the leaf surface. In washing experiments Rea et al. (2001) demonstrated that, unlike many nutrients, leaching or exchange of Hg at the foliar surface is negligible. Litterfall Hg is considered dry deposition since the atmosphere, rather than the soil, is the predominant source of Hg in foliage (Hanson et al., 1995; Lindberg, 1996; Bishop et al., 1998; Rea et al., 2002; Ericksen et al., 2003). Results from field research at forested sites indicate that Hg in precipitation is the smallest of the three fluxes, while litterfall is often the largest (Grigal, 2002, Miller et al., 2005). Even though annual Hg emissions due to industrial uses have declined from a peak of 10000 tons around 1970–2000 tons in the year 2000 (Hylander and Meili, 2003), emissions during the past century have resulted in Hg accumulations in the environment. As a result, 45 states in the U.S.A. have fish consumption advisories in effect for fresh and/or coastal waters (EPA, 2004). Quantifying all major vectors of atmospheric Hg deposition to forested ecosystems is vital for determining risks to biota, public health and natural resources. The first objective of this study was to quantify the concentration of Hg in litter and the magnitude of the litter Hg flux in forested watersheds at Acadia National Park (ANP), Maine, U.S.A. The second objective was to determine the influence of plant community composition on litter Hg concentration and flux. The third

252


objective was to combine data on litter Hg with other Hg research results from these watersheds to determine the importance of litter Hg flux in Hg cycling.

2. Methods 2.1. SITE

DESCRIPTION

The study watersheds, Hadlock Brook and Cadillac Brook, are located in ANP on Mount Desert Island in eastern Maine, U.S.A (Figure 1). Headwater streams drain both watersheds. Hadlock Brook watershed encompasses 47.2 ha of the southwest slope of Sargent Mountain. Cadillac Brook is located approximately five kilometers east of Hadlock Brook watershed, draining 31.6 ha of the southeastern slope of Cadillac Mountain. Hadlock Brook watershed stretches from an elevation of 137 m to 380 m at an average slope of 20% and Cadillac Brook watershed spans 122–468 m with an average slope of 28% (Nelson, 2002).

Figure 1. Location of study watersheds within Acadia National Park, Maine.


253

Due to their close proximity, the study watersheds are assumed to be subject to the same climatic conditions. The topographic contrast between the study watersheds and the adjacent Gulf of Maine enables them to intercept air masses and the pollutants they transport that traverse the Maine coast (Malm, 1999; Kahl et al., 2000; Amirbahman et al., 2004), leading to enhanced deposition of atmospheric pollutants (Jagels et al., 1989). These watersheds were selected for an assessment of Hg litterfall dynamics because of their history of biogeochemical research at these sites, and contrasting conditions of fire history and vegetation composition (Heath et al., 1992, 1993; Kahl et al., 2000). The soils of both watersheds are Lithic Udorthents and Lithic Haplorthods, and are primarily composed of glacial till with a thin O-horizon, and outcrops of Cadillac granite (Jordan, 1998; Nelson, 2002). Being geologically young and located on steep slopes, the soils are thin and discontinuous. A substantial portion of both watersheds is bare rock with patches of scrub-shrub vegetative communities (Johnson, 2002). Cadillac Brook has ∼27% of its surface area as exposed bedrock with patches of low-lying scrub-shrub vegetation, whereas Hadlock Brook has only ∼13% (Ruck, 2002). In 1947, approximately two-thirds of the Cadillac Brook watershed burned in a wildfire, while the Hadlock Brook watershed was unaffected. Erosional soil loss associated with the fire at least partly explains why the percent of soil cover and soil thickness in Cadillac Brook is less than Hadlock Brook watershed. Fire and subsequent vegetation differences also contribute to thinner O horizons in Cadillac Brook compared to Hadlock Brook watershed. The importance of early successional hardwoods in the vegetative cover in Cadillac Brook watershed is a legacy of the fire disturbance. The vegetative community in the burned area, which spans the lower third of the watershed, is composed primarily of maple (Acer rubrum, Acer pensylvanicum, Acer saccharum), aspen (Populus tremuloides, Populus grandidentata), and paper birch (Betula papyrifera). For the study period, the scrub-shrub vegetation community composition in the upper reaches of the watershed was similar to Hadlock Brook watershed. The vegetative community of the lower reaches of Hadlock Brook watershed was composed predominantly of a mix of red spruce (Picea rubens) and balsam fir (Abies balsamea), with a minor amount of grey birch (Betula populifolia).

2.2. F IELD

METHODS

2.2.1. Sample Site Selection Litterfall was sampled at a total of 19 sites in Hadlock Brook watershed and 20 sites in Cadillac Brook watershed (Figure 2). The network of 39 sites consisted of: (1) 12 core sites used for long-term sampling of litter, throughfall, and soil (Amirbahman et al., 2004; Johnson, 2002; Nelson, 2002) for the Park Research and Intensive Monitoring of Ecosystems Network (PRIMENet) study, (2) 17 throughfall study sites located along transects from ongoing throughfall studies (Nelson, 2002), and

254


Figure 2. Map of litter sample site locations within the study watersheds, Acadia National Park, Maine.

(3) 10 additional sites randomly located within under-represented vegetation types. Some of the throughfall transect study sites were omitted because they either overrepresented a vegetation type, or were not located beneath a vegetative canopy. The study sites chosen provided both a balance of replication within major vegetation types to the extent possible while linking this research to other ongoing studies in these watersheds. 2.2.2. Vegetation Classification Vegetation was identified to the species level at each site. Four broad site vegetation classes (Table I) were defined for this study as aggregates of vegetation classes developed by the U.S. Geological Survey for these watersheds (USGS, 2003). Where the map vegetation descriptors did not accurately describe the vegetation at the site due to a difference in scale, the on-site vegetation identification determined the vegetation class. Once the broad vegetation classification was established, the area of each vegetation class within the watersheds was calculated. 2.2.3. Sample Collection Acid-washed (5% nitric acid) polyethylene basins served as litterfall collectors (LFCs). Each collector measured 43.8 cm × 33.6 cm × 12.7 cm deep with 5 mm holes drilled in the bottom for drainage. LFCs were deployed in the field on

255


TABLE I The four vegetation classes used in this study and the U.S. Geological Survey vegetation classes that were aggregated for each study class Vegetation class I. Hardwood Aspen Birch Woodland/Forest Complex (forest phase) II. Softwood Conifer Swamp Woodland (white cedar phase) Mixed Conifer Woodland Spruce – Fir Forest (conifer phase) Pitch Pine Woodland III. Mixed Spruce – Fir Forest (mixed phase) Mixed Conifer – Deciduous Woodland White Pine – Hardwood Forest IV. Scrub Aspen Birch Woodland/Forest Complex (shrubland phase) Blueberry Bald – Summit Shrubland Complex

Area (ha)

Area (%)

Sites (n)

3.25

4

6

39.93

51

15

12.81

16

10

22.39

29

8

Total area (ha), percent of total area (%), and the number of sampling sites (n) for the vegetation classes are for the area of both watersheds combined.

the soil surface during the snow-free season. Litterfall collections were made by transferring litterfall from LFCs to polyethylene bags in the field. The bags were sealed, inserted into a second polyethylene bag, and transported to the laboratory on ice. Immediately following sample collection new acid-washed LFCs were deployed. When snow or ice was encountered in the LFCs, the snow or ice that contained litterfall, as determined by visual inspection, was collected with the sample. The snow/ice was then allowed to melt and drain from the plastic bags in the lab, prior to drying. Total-Hg concentration data came from the analysis of all samples collected from April 2003 to November 2003 (i.e. field season 2003), and 18 of the samples collected during the autumn of 2002. Due to practical limitations of the project, total-Hg was not measured on the remaining samples collected that included some of the autumn 2002 samples and the single collection that represented winter 2003–2004. The June and August 2003 samples were bulked by collector in order to obtain sufficient sample mass for analysis. The lowest monthly litter Hg concentration means were multiplied by the April 2004 litterfall masses, representing the winter season, to provide a conservative estimate of annual litter Hg flux to the watersheds.

256 2.3. SAMPLE


PREPARATION AND MERCURY ANALYSIS

At the laboratory, unsorted litter samples from each site were weighed, dried at 60 ◦ C in a forced draft oven, and then stored in double plastic bags. Samples were ground in a Wiley mill with a 40 mesh screen and stored double bagged at room temperature until analysis. The Wiley mill operator brushed and vacuumed the mill blades, screen, chamber and workstation to prevent cross-contamination between samples. The operator wore powder-free PVC gloves during these operations and while handling samples. Samples were digested using a water bath technique, a modification of EPA method 245.6 (Determination of Mercury in Tissue). The digestion matrix was a mixture of H2 SO4 (80% by volume) and HNO3 (20% by volume). Water bath digestion time was increased from a 1–2 h to 12 h, at 58 ◦ C, in order to improve recovery. Samples were oxidized with KMnO4 and K2 S2 O8 , followed by the addition of hydroxylamine hydrochloride to reduce excess KMnO4 . Up to 21 ml of KMnO4 was required to oxidize the sample matrix following digestion, rather than the 5 ml described in the method. Digestion batches consisted of 25 samples or less. Litterfall digests were analyzed by cold vapor atomic absorption on a Hewlett-Packard Flow Injection Mercury System model 400 (FIMS) equipped with a model AS90 auto-sampler. Quality control included sample replicates, spike recovery, reagent blanks, and laboratory standard reference materials (National Institute of Standards and Technology certified standard reference materials SRM #1575, pine needles; SRM #1547, peach leaves, SRM #1575a, pine needles).

2.4. STATISTICAL

ANALYSES AND CALCULATIONS

Statistical comparisons among vegetation types were determined by Analysis of Variance (ANOVA) at a 95% confidence level. All data were log transformed to approximate the assumptions of normal distribution of error and constant variance. R Statistical analyses were performed using SYSTAT version 11.0. All means are reported with their corresponding standard error (±SE).

3. Results and Discussion 3.1. LITTERFALL

QUANTIFICATION

Collection periods varied from 30 to 133 days in length, and therefore litterfall mass for the individual collections are presented on a per day basis to allow for comparison among collections (Figure 3). The mean rate of litterfall normalized to a per day basis ranged from 0.10 ± 0.03 to 7.38 ± 1.9 g m−2 day−1 . The greatest variation among collectors within vegetation types for individual collections of litterfall


257

Figure 3. The mean (±SE) for litterfall from each collection by vegetation class for the study period.

occurred among hardwood sites, which produced the highest litterfall masses. The smallest variation occurred among scrub sites, which generally had the lowest litterfall due to the patchiness and stature of these vegetative communities. There was an axiomatic increase in litterfall in all vegetation classes during October and November and therefore our collection program focused resources on this peak period. Litterfall in softwoods was more consistent throughout the year than in hardwoods. Over the span of the study the greatest litterfall occurred in hardwoods, due to high rates of litterfall in the autumn for this vegetation type. Softwoods do not produce a major autumn litterfall event at the scale of hardwoods, but demonstrate modest yet consistent rates of litterfall throughout the winter. This resulted in softwoods having the highest litterfall mass for the April 2004 collection (0.79 ± 0.03 g m−2 day−1 ), which represented the preceding winter. The litterfall variation for mixed vegetation sites was intermediate between hardwoods and softwoods among collectors for individual collections, and among collection periods over time. When integrated over the entire study period the overall weighted means for litterfall were 1.55 ± 0.23, 0.87 ± 0.14, 0.78 ± 0.12, and 0.36 ± 0.05 g m−2 day−1 for hardwoods, mixed, softwoods, and scrub, respectively. Although weighted mean litterfall mass for hardwoods was numerically greatest, the differences amongst hardwoods, softwoods and mixed were not statistically significant. The weighted mean litterfall mass for the scrub vegetation type was significantly lower than the other vegetation types.

3.2. Hg CONCENTRATION

OF LITTER

Descriptive statistics for the concentration of Hg in litter from all samples measured during this study are presented in Table II. Figure 4A shows mean litter Hg concentrations from this study by vegetation type. Softwood sites had significantly

258


TABLE II Descriptive statistics for Hg concentrations (ng g−1 ) in litter across all samples and by vegetation type for collections from October 2002–November 2003

All samples Hardwood Softwood Mixed Scrub

Mean

Std. Error

Median

Minimum

Maximum

n

46.9 31.6 58.8 41.7 40.6

1.9 2.6 3.3 2.8 2.7

41.3 31.0 54.9 39.0 36.2

10.7 10.7 17.2 15.4 24.9

133.4 55.6 133.4 110.8 89.2

153 26 63 39 25

Figure 4. Mean (±SE) litter Hg concentrations and Hg flux from litter for each vegetation class in the study watersheds.

higher mean litter Hg concentration (58.8±3.3 ng Hg g−1 ) than all other vegetation types. Presumably higher concentrations of Hg in softwood litter was primarily attributed to the duration of foliar exposure to atmospheric Hg prior to leaf-fall and a higher surface area for softwoods compared to hardwoods (Rassmussen et al., 1991; Rasmussen, 1995; Kolka et al., 1999; St. Louis et al. 2001; Grigal, 2003). Litter collected at hardwood sites had significantly lower Hg concentrations (31.6 ± 2.6 ng Hg g−1 ) than any of the other vegetation types, consistent with the literature. Mean litter Hg concentrations were not significantly different from each other for mixed and scrub vegetation types, but were significantly higher than hardwoods. Higher mean litter Hg concentrations in mixed and scrub vegetation types compared to hardwoods was likely due to the presence of litter from softwood species which were present at both mixed and scrub sites. Rasmussen et al. (1991) suggested that lower stature plants may absorb Hg vapor emitted from soils, but it is unlikely that


259

TABLE III Litter Hg concentrations in different vegetation types at various study locales Location

ng g−1

Vegetation type

Source

Tennessee, USA Tennessee, USA Bavaria, Germany Bavaria, Germany Ontario, Canada Ontario, Canada Vermont, USA Michigan, USA Maine, USA

105 61 33–88 46–75 33–79 25–30 47 33 40–51

Hardwood Softwood Softwood Hardwood Jack Pine Jack Pine/Birch Mixed Hardwood Mixed Hardwood Mixed

Lindberg (1996) Lindberg (1996) Schwesig and Matzner (2000) Schwesig and Matzner (2000) St. Louis et al. (2001) St. Louis et al. (2001) Rea et al. (2002) Rea et al. (2002) Johnson (2002)

this explains the higher mean litter Hg concentration at scrub sites in this study since the soil is thin and sometimes patchy at those sites. Results reported here for ANP hardwood litter Hg concentrations were similar to those in a study of Michigan hardwood litter, but somewhat lower than the mean concentration of Hg in Vermont and Tennessee hardwood litter and below the range for hardwood litter from Germany (Table III). The mean litter Hg concentration at softwood sites in ANP is similar to other softwood literature values (Table III). Previous research that measured litter Hg concentrations in these ANP study watersheds using a limited number of samples (n = 10) reported a similar but smaller range (Johnson, 2002). 3.3. Hg FLUX 3.3.1. Field Season Litter Hg Flux The mean weighted flux of Hg in litter, during field season 2003, is presented in Figure 4B. There were no significant differences in litter Hg flux among vegetation classes ( p < 0.05, Figure 4B), despite the range in mean weighted litter Hg fluxes and the significant differences evident in the concentration of Hg in litter (Figure 4A). The lack of significant differences in litter Hg flux was attributable to both the high variability in these data (particularly for hardwoods), the composite nature of the samples, and the inverse relationship between Hg concentration and litterfall mass when comparing hardwoods and softwoods. St. Louis et al. (2001) demonstrated that the relatively low mass of litterfall offsets the effectiveness of softwood Hg scavenging when comparing hardwoods and softwoods. They reported results from a softwood site with high Hg concentration but a low annual flux of Hg in litter, while litter from a hardwood site with a lower Hg concentration had a higher annual flux.

260


The Hg flux in litter was calculated for the 2003 field season (=0.57 yr) by watershed on an area-weighted basis to account for the different spatial extents of each vegetation type in each watershed. The averaged Hg flux in litter was 7.5 μg m−2 for the growing season. The 2003 field season flux of Hg in litter was 8.7 μg m−2 in Cadillac Brook watershed and 6.0 μg m−2 in Hadlock Brook watershed. The higher litter Hg flux in Cadillac reflected the greater spatial extent of hardwoods in Cadillac Brook watershed, which ultimately dominated these results due to the high hardwood litter flux in the latter part of the growing season. The high litterfall period in the autumn for hardwoods, and more consistent litterfall flux throughout the year for softwoods, resulted in important differences between these forest types for growing season, winter, and annual litter and litter Hg flux estimates. 3.3.2. Annual Litter Hg Budget Estimate Figure 5 shows an estimated annual flux of Hg in litter for this study at ANP. The flux estimate was considered conservative and based on the study period as described in the methods. Also shown in Figure 5 are other available estimates of Hg data for these watersheds that include inputs, stream export, and soil pools. The overall estimated annual mean litter Hg flux across both watersheds was 10 μg m−2 .

Figure 5. Estimates of annual Hg flux and soil pools in the study watersheds. Precipitation flux (P) for the period of this litterfall study was measured at the ANP Mercury Deposition Network site ME98 (National Atmospheric Deposition Program, 2005). The Hg soil pool and Hg stream flux (S) were measured at these sites July 1999–2000 (Johnson, 2002; Amirbahman et al., 2004). Deposition (W)∗ is defined here as an estimate of annual wet deposition that included the summation of throughfall measurements from May 2000 to November 2000 plus wet-only precipitation for the period November 1999–April 2000 as reported by Johnson (2002).


261

The annual estimates of Hg deposition in litter for each watershed were similar, in contrast to growing season comparisons discussed above. The continuous input of litter in softwoods throughout the year balanced the uniquely large event of litterfall that occurred in hardwoods at the end of the growing season. Although slightly lower, the annual Hg litter flux in ANP is similar to several relevant studies for forested ecosystems. Rea et al. (2002) reported the annual Hg litter flux was 11.4 μg m−2 for the Lake Huron Watershed, MI, composed of northern hardwoods. Similarly, St. Louis et al. (2001) calculated an annual litter Hg flux of 12.0 μg m−2 in the Experimental Lakes Area in Ontario, Canada, a northern boreal forest dominated by softwood species. Rea et al. (1996) estimated an annual litter Hg flux of 13.0 μg m−2 from 1994 data in a northern hardwood forest from the Lake Champlain Watershed, VT. For the same study site, Rea et al. (2002) determined that the annual flux of Hg in litter was 15.8 μg m−2 , from samples collected in 1995. The ANP annual Hg litter flux reported here was substantially lower than the 30.0 μg m−2 determined by Lindberg (1996) in a temperate hardwood forest in TN, but unlike ANP, their site was located near known point sources of Hg. Schwesig and Matzner (2000, 2001) reported a range of annual Hg litter fluxes from 16–32 ug m−2 in Germany, which may have been higher than at ANP since the study sites were situated in central Europe in proximity to Hg emission sources. Annual Hg litter fluxes in Scandinavia reported by Lee et al. (1998, 2000), were in the range of 18–60 μg m−2 , substantially higher than ANP. Historically, atmospheric Hg concentrations in those study areas were higher than concentrations documented in the United States, which could explain why Scandinavian litter Hg fluxes were larger than in ANP (Lindberg, 1996; Grigal, 2002). The ratio of mean annual Hg flux in precipitation (P) to Hg deposition in throughfall (W being the approximation used here) to litter Hg flux (L) is 1:1.5:1.5 overall for the study watersheds. Similar ratios were reported by Grigal (2002), 1:1.8:2.2, and Munthe et al. (1995), 1:1.5:3. These data show that only monitoring Hg in precipitation is inadequate for total deposition estimates, since total Hg deposition can be much greater than precipitation Hg deposition alone. Research conducted in warmer climates with longer growing seasons may report even higher Hg litter flux rates. The estimated annual flux of Hg in litter for each study watershed was well in excess of precipitation inputs, and comparable to throughfall deposition of Hg, calculated to include potential dry deposition of Hg as measured in throughfall during the field season and precipitation Hg inputs (as reported by MDN) for the remainder of the year (i.e., W). Dry deposition was not included in the deposition (W) figure for November1999–April 2000, and we therefore assume this estimate to be conservative. Since annual Hg export via stream water was an order of magnitude smaller than the deposition estimates reported here (L, P, and W), it is reasonable to assume that either (a) the average soil pool of 16 mg m−2 is increasing, (b) one or more important export vectors such as Hg volatilization were not measured and were significant,

262


or (c) the likelihood that both of these are true. Our data represent conservative estimates for soil pools given the limited sampling depths of Amirbahman et al. (2004), and meaningful estimates of watershed accumulation rates would require data on Hg volatilization and other factors such as disturbance history for these watersheds. Indeed, Amirbahman et al. (2004) reported details of soil differences between these watersheds and noted the soil Hg pool for Hadlock watershed was greater than the Cadillac watershed (Figure 5). They interpreted the lower Cadillac Brook soil pool as resulting from the fire of 1947 that affected only the Cadillac Brook watershed, and the influence of a more extensive hardwood vegetative cover in Cadillac compared to Hadlock since that disturbance event. Since (a) soil pools of Hg are lower in Cadillac compared to Hadlock Brook watershed, (b) labile Hg appears to reflect that soil difference in the form of greater Hg export in the Hadlock Brook stream compared to Cadillac Brook, yet (c) there was no significant difference between watersheds in litter Hg concentrations, then (d) we might assume that Hg uptake by roots was not a significant source of Hg in foliage and most of litter Hg concentrations resulted from direct transfers from the atmosphere. This is consistent with the results from the literature suggesting limited Hg uptake by plants, and perhaps only at low atmospheric Hg exposure would soil uptake be important in determining foliar Hg concentrations (Ericksen et al., 2003; Fresholtz et al., 2003). Miller et al. (2005) estimated total annual Hg deposition, defined as the sum of Hg delivered via precipitation, dry aerosol deposition, litterfall, and cloud droplet assimilation, to be within the range of 20–25 μg m−2 for this study area using a GIS-based model. Their modeled Hg input estimates agreed well with our estimate of total annual Hg input of ∼20 μg m−2 from this study.

4. Conclusions Results from this study indicated that annual Hg litter flux was not strongly influenced by the marked differences in vegetative types between watersheds, despite differences in the capacity of various vegetation types to incorporate Hg into foliage. Both watersheds appeared to have similar Hg litter fluxes on an annual basis (10.0 and 10.1 μg m−2 in Cadillac Brook and Hadlock Brook watersheds, respectively). There were clear differences in the litter Hg concentrations among plant species, which may have important ecological implications for other forms of biota. Litter Hg flux to the ANP watersheds was a significant source of Hg to these ecosystems, comparable to estimates of wet+dry deposition, and much greater than Hg deposition in precipitation. Vegetative canopies significantly increase the total Hg inputs to watersheds compared to bare rock and open area deposition. Natural resource managers concerned with potential exposure of surface waters to Hg should consider vegetated watersheds a greater concern than sparsely vegetated and barren rock watersheds, all of which are represented across the ANP landscape.


263

The flux of Hg in litter is approximately half the estimated total annual deposition of Hg to these watersheds. This suggests that watershed-scale studies of Hg mass balance need to incorporate measurements of litterfall Hg. As with the ANP wet-only Hg monitoring site (ME98), the national Mercury Deposition Network (MDN) provides valuable high quality measurements of wet-only precipitation Hg deposition but underestimates total annual Hg deposition by not including litter Hg flux. The distinction between precipitation Hg deposition and the potential magnitude of litter Hg deposition, particularly in forested landscapes, is critical in the utilization of Hg deposition monitoring data for evaluations of Hg mass balance in forested watersheds.

Acknowledgments Funding for this project was made possible by a grant from the National Park Service (Cooperative agreement H4525020030) and the Maine Agricultural and Forest Experiment Station. Maine Agricultural and Forest Experiment Station Publication Number 2827.

References Amirbahman, A., Ruck, P. L., Fernandez, I. J., Haines, T. A. and Kahl, J. S.: 2004, ‘The effect of fire on mercury cycling in the soils of forested watersheds: Acadia National Park, Maine, U.S.A.’, Water Air Soil Pollut. 152, 313–331. Bishop, K. H., Lee, Y., Munthe, J. and Dambrine, E.: 1998, ‘Xylem sap as a pathway for total mercury and methylmercury transport from soils to tree canopy in the boreal forest’, Biogeochemistry 40, 101–113. Du, S.-H. and Fang, S. C.: 1982, ‘Uptake of elemental mercury vapor by C3 and C4 species’, Environ. Exp. Bot. 22, 437–443. EPA: 2004, Factsheet, “National Listing of Fish Advisories”, U.S. Environmental Protection Agency, Office of Water and Office of Science and Technology (4305). EPA-823-F-04-016. Ericksen, J. A., Gustin, M. S., Schorran, D. E., Johnson, D. W., Lindberg, S. E. and Coleman, J. S.: 2003, ‘Accumulation of atmospheric mercury in forest foliage’, Atmos. Environ. 37, 1613–1622. Frescholtz, T. F., Gustin, M. S., Schorran, D. E. and Fernandez, G. C. J.: 2003, ‘Assessing the source of mercury in foliar tissue of Quaking Aspen’, Environ. Toxicol. Chem. 22, 2114–2119. Greger, M., Wang, Y. and Neusch¨utz, C.: 2004, ‘Absence of Hg transpiration by shoot after Hg uptake by roots of six terrestrial plant species’, Environmental Pollution 134, 201–208. Grigal, D. F.: 2002, ‘Inputs and outputs of mercury from terrestrial watersheds: A review’, Environ. Rev. 10, 1–39. Grigal, D. F.: 2003, ‘Mercury sequestration in forests and peatlands: A review’, J. Environ. Qual. 32, 393–405. Hanson, P. J., Lindberg, S. E., Tabberer, T. A., Owens, J. G. and Kim, K. H.: 1995, ‘Foliar exchange of mercury vapor: Evidence for a compensation point’, Water Air Soil Pollut. 80, 373–382. Heath, R., Kahl, J., Norton, S. and Brutsaert, W.: 1993, ‘Elemental mass balances, and episodic and ten-year changes in the chemistry of surface waters, Acadia National Park’, Final Report, National Park Service, North Atlantic Region, Boston, MA. 111 pp.

264


Heath, R. A., Kahl, J. S., Norton, S. A. and Fernandez, I. J.: 1992, ‘Episodic acidification caused by atmospheric deposition of seasalts at Acadia National Park, Maine, USA’, Water Resour. Res. 28, 1081–1088. Hylander, L. D. and Meili, M.: 2003, ‘500 years of mercury production: Global annual inventory by region until 2000 and associated emissions’, Sci. Total Environ. 304, 13–27. Iverfeldt, A., Lindberg, S. E., Karamata, S., Anoshin, G. N., Horvat, M., Laperdina, T. G., Obolenskiy, A. A., Osmonbetov, K. O., Ramel, C., Roslyakov, N. A. and Tauson, V. L.: 1996, ‘Working group on terrestrial mercury cycling’, in W. Baeyens, R. Ebinghaus and O. Vasiliev (eds), Global and Regional Mercury Cycles: Sources, Fluxes and Mass Balances, Kluwer Academic Publishers, Boston, MA, USA, pp. 543–546. Jagels, R., Carlisle, J., Cunningham, R. Serreze, S. and Tsai, P.: 1989, ‘Impact of acid fog and ozone on coastal red spruce’, Water Air Soil Pollut. 48, 193–208. Johnson, K. B.: 2002, ‘Fire and its effects on mercury and methylmercury dynamics for two watersheds in Acadia National Park, Maine’, MS Thesis, University of Maine, Orono, Maine. Jordan, G. B.: 1998, ‘Soil survey of Hancock County area, Maine’, U.S. Department of Agriculture, Natural Resources Conservation Service. Kahl, J. S., Manski, D., Flora, M. and Houtman, N.: 2000, Water Resources Management Plan Acadia National Park, 103 pp. Kolka, R. K., Nater, E. A., Grigal, D. F. and Verry, E. S.: 1999, ‘Atmospheric inputs of mercury and organic carbon into a forested upland/bog watershed’, Water Air Soil Pollut. 113, 273– 294. ˚ Verta, M., Parkman, H. and Hultberg, H.: 1998, ‘An Lee, Y. H., Bishop, K. H., Munthe, J., Iverfeldt, A., examination of current Hg deposition and export in Fenno-Scandian catchments’, Biogeochemistry 40, 125–135. Lee, Y. H., Bishop, K. H. and Munthe, J.: 2000, ‘Do concepts about cycling of methylmercury and mercury in boreal catchments stand the test of time? Six years of atmospheric inputs and runoff exports at Svartberget, northern Sweden’, Sci. Total Environ. 260, 11–20. Leonard, T. L., Taylor, G. E., Jr., Gustin, M. S. and Fernandez, G. C. J.: 1998a, ‘Mercury and plants in contaminated soils: 1. Uptake, partitioning, and emission to the atmosphere’, Environ Toxicol Chem. 17, 2063–2071. Leonard, T. L., Taylor, G. E., Jr., Gustin, M. S. and Fernandez, G. C. J.: 1998b, ‘Mercury and plants in contaminated soils: 2. Environmental and physiological factors governing mercury flux to the atmosphere’, Environ Toxicol Chem. 17, 2072–2079. Lindberg, S. E.: 1996 ‘Forests and the global biogeochemical cycle of mercury: The importance of understanding air/vegetation exchange processes’, in W. Baeyens, R. Ebinghaus and O. Vasiliev (eds), Global and Regional Mercury Cycles: Sources, Fluxes and Mass Balances, Kluwer Academic Publishers, Boston, MA, USA, pp. 359–380. Lindberg, S. E., Meyers, T. P., Taylor, G. E. Jr., Turner, R. R. and Schroeder, W. H.: 1992, ‘Atmospheresurface exchange of mercury in a forest: Results of modeling and gradient approaches’, J. Geophys. Research 97, 2519–2528. Malm, W. C.: 1999, ‘Introduction to visibility’, U.S. Department of the Interior, National Park Service, Air Resources Division, Fort Collins, CO. ISSN 0737-5352-40, 70 pp. Mason, R. P., Abbott, M. L., Bodlay, R. A., Bullock, O. R. Jr., Driscoll, C. T., Evers, D., Lindberg, S. E., Murray, M. and Swain, E. B.: 2005, ‘Monitoring the response to changing mercury deposition’, Environ. Sci. Technol. 15A (in press). Miller, E. K., Van Arsdale, A., Keeler, G. J., Chalmers, A., Poissant, L., Kamman, N. and Brulotte, R.: 2005, ‘Estimation and mapping of wet and dry mercury deposition Across Northeastern North America.’, Ecotoxicology. 14, (in press). Morel, F. M. M., Kraepiel, A. M. L. and Amyot, M.: 1998, ‘The chemical cycle and bioaccumulation of mercury’, Annu. Rev. Ecol. Syst. 29, 543–566.


265

˚ 1995, ‘Mechanisms of deposition of methylmercury and Munthe, J., Hultberg, H. and Iverfeldt, A.: mercury to coniferous forests’, Water Air Soil Pollut. 80, 363–371. National Atmospheric Deposition Program (NRSP-3): 2005, NADP Program Office, Illinois State Water Survey, 2204 Griffith Dr., Champaign, IL 61820. Nelson, S. J.: 2002, ‘Determining atmospheric deposition inputs to two small watersheds at Acadia National Park’, MS Thesis, University of Maine, Orono, Maine. Rasmussen, P. E.: 1995, ‘Temporal variation of mercury in vegetation’, Water Air Soil Pollut. 80, 1039–1042. Rasmussen, P. E., Mierle, G. and Nriagu, J. O.: 1991, ‘The analysis of vegetation for total Mercury’, Water Air Soil Pollut. 56, 379–390. Rea, A. W., Keeler, G. J. and Scherbatskoy, T.: 1996, ‘The deposition of mercury in throughfall and litterfall in the Lake Champlain watershed: A short-term study’, Atmos. Environ. 30, 3257–3263. Rea, A. W., Lindberg, S. E. and Keeler, G. J.: 2000, ‘Assessment of dry deposition and foliar leaching of mercury and selected trace elements based on washed foliar and surrogate surfaces’, Environ. Sci. Technol. 34, 2418–2425. Rea, A. W., Lindberg, S. E. and Keeler, G. J.: 2001, ‘Dry deposition and foliar leaching of mercury and selected trace elements in deciduous forest throughfall’, Atmos. Environ. 35, 3453–3462. Rea, A. W., Lindberg, S. E., Scherbatskoy, T. and Keeler, G. J.: 2002, ‘Mercury accumulation in foliage over time in two northern mixed-hardwood forests’, Water Air Soil Pollut. 133, 49–67. Ruck, P. L.: 2002, Cycling and Speciation of Mercury in Soils at Cadillac Brook and Hadlock Brook watersheds, Acadia National Park, Maine, MS Thesis, University of Maine, Orono, Maine. Schwesig, D. and Matzner, E.: 2000, ‘Pools and fluxes of mercury and methylmercury in two forested catchments in Germany’, Sci. Total Environ. 260, 213–223. Schwesig, D. and Matzner, E.: 2001, ‘Dynamics of mercury and methylmercury in forest floor and runoff of a forested watershed in Central Europe’, Biogeochemistry 53, 181–200. Skyllberg, U., Xia, K., Bloom, P. R., Nater, E. A. and Bleam, W. F.: 2000, ‘Binding of mercury(II) to reduced sulfur in soil organic matter along upland-peat soil transects’, J. Environ. Qual. 29, 855–865. St. Louis, V. L., Rudd, J. W. M., Kelly, C. A., Hall, B. D., Rolfhus, K. R., Scott, K. J., Lindberg, S. E. and Dong, W.: 2001, ‘Importance of the forest canopy to fluxes of methyl mercury and total mercury to boreal ecosystems’, Environ. Sci. Technol. 35, 3089–3098. Tyler, G.: 2005, ‘Changes in the concentrations of major, minor and rare-earth elements during leaf senenscence and decomposition in a Fagus sylvatica forest’, Forest Ecology and Management 206, 167–177. USGS: 2003, ‘Vegetation spatial database coverage for Acadia National Park vegetation mapping project’, U.S. Geological Survey, Center for Biological Informatics, Denver, Colorado.