HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (35) Citing Articles via Google Scholar Google Scholar Articles by Huntington, T.G. Articles by Blum, A.E. Search for Related Content PubMed Articles by Huntington, T.G. Articles by Blum, A.E. GeoRef GeoRef Citation Agricola Articles by Huntington, T.G. Articles by Blum, A.E.

DIVISION S-7-FOREST & RANGE SOILS

Calcium Depletion in a Southeastern United States Forest Ecosystem

^a U.S. Geological Survey, 3039 Amwiler Rd. Suite 130, Atlanta, GA 30360 USA
^b U.S. Geological Survey, 10 Bearfoot Rd., Northborough, MA 01532 USA
^c Dep. of Civil Engineering, 220 Hinds Hall, Syracuse Univ., Syracuse, NY, 220 Hinds Hall, Syracuse, NY 13244 USA
^d Dep. of Biology, Emory Univ., Atlanta, GA 30322 USA
^e U.S. Geological Survey, 3215 Marine St., Boulder, CO 80303 USA

thunting{at}usgs.gov

	ABSTRACT

TOP ABSTRACT INTRODUCTION Materials and methods NOTES Results and discussion Conclusions REFERENCES

 
Forest soil Ca depletion through leaching and vegetation uptake may threaten long-term sustainability of forest productivity in the southeastern USA. This study was conducted to assess Ca pools and fluxes in a representative southern Piedmont forest to determine the soil Ca depletion rate. Soil Ca storage, Ca inputs in atmospheric deposition, and outputs in soil leaching and vegetation uptake were investigated at the Panola Mountain Research Watershed (PMRW) near Atlanta, GA. Average annual outputs of 12.3 kg ha^-1 yr^-1 in uptake into merchantable wood and 2.71 kg ha^-1 yr^-1 soil leaching exceeded inputs in atmospheric deposition of 2.24 kg ha^-1 yr^-1. The annual rate of Ca uptake into merchantable wood exceeds soil leaching losses by a factor of more than five. The potential for primary mineral weathering to provide a substantial amount of Ca inputs is low. Estimates of Ca replenishment through mineral weathering in the surface 1 m of soil and saprolite was estimated to be 0.12 kg ha^-1 yr^-1. The weathering rate in saprolite and partially weathered bedrock below the surface 1 m is similarly quite low because mineral Ca is largely depleted. The soil Ca depletion rate at PMRW is estimated to be 12.7 kg ha^-1 yr^-1. At PMRW and similar hardwood-dominated forests in the Piedmont physiographic province, Ca depletion will probably reduce soil reserves to less than the requirement for a merchantable forest stand in 80 yr. This assessment and comparable analyses at other southeastern USA forest sites suggests that there is a strong potential for a regional problem in forest nutrition in the long term.

Abbreviations: CEC, cation-exchange capacity • DBH, diameter at breast height • NADP/NTN, National Acid Deposition Program/National Trends Network • PMRW, Panola Mountain Research Watershed • XRF, x-ray fluorescence spectroscopy

	INTRODUCTION

TOP ABSTRACT INTRODUCTION Materials and methods NOTES Results and discussion Conclusions REFERENCES

 
THERE IS INCREASING EVIDENCE that soil Ca stores in aggrading forests on formerly cultivated lands in the southeastern USA are being steadily depleted through biomass uptake and soil leaching (Wells and Jorgensen, 1975; Johnson et al., 1988b; Binkley et al., 1989a; Johnson and Todd, 1990; Knoepp and Swank, 1994; Richter et al., 1994). The potential for Ca depletion in eastern U.S. forests has been suggested in earlier reviews (Mann et al., 1988; Federer et al., 1989; Shortle and Bondietti, 1992) and modeling studies (Binkley et al., 1989b), but few studies have quantified inputs and outputs in relation to soil reserves.

Southeastern forest soils are especially prone to Ca depletion for several reasons. First, these soils generally developed in base-poor igneous and metamorphic rocks. Second, these soils typically are old, having weathered in place for several hundred thousand to 2 million years, during which time most Ca-bearing minerals within the rooting zone were depleted (Calvert et al., 1980; Pavich, 1989, Pavich et al., 1989; Daniels and Hammer, 1992; Markewich et al., 1994). Because of this depletion of Ca-bearing minerals in soils and saprolite, the potential for replenishment of soil exchangeable Ca through mineral weathering is low. Biological cycling maintains adequate nutrients for tree growth, but once biological cycling is disrupted, through forest harvesting, bases can be rapidly lost due to leaching (Buol et al., 1980). Third, historical land use during the 1800s and early 1900s generally included an extensive period of exploitative agricultural practices that resulted in extensive loss of topsoil (Hoover, 1949; Nelson, 1957; Brender, 1974; Trimble, 1974) and the removal of a substantial amount of the exchangeable soil Ca stores present at the time of European settlement.

Acidic deposition accelerates Ca losses because increased acidity solubilizes Al, which displaces Ca from soil exchange sites, and Ca is then readily leached in association with SO₄ (Galloway et al., 1983; Reuss and Johnson, 1986). Sulfate retention in soils can delay the onset of acidification by reducing the flux of SO₄ through the soil. Many southeastern USA soils retain a large proportion of atmospherically deposited SO₄ (Shriner and Henderson, 1978; Rochelle et al., 1987); however, the soil's capacity to retain SO₄ is finite so that a substantial amount of SO₄ is leached (Swank and Crossley, 1988; Johnson and Van Hook, 1989; Huntington et al., 1994). Long-term trends in stream water chemistry and SO₄ deposition at several forested watersheds at the Coweeta Hydrologic Laboratory in North Carolina and in the Shenandoah National Park in Virginia support the hypothesis that soil retention of atmospherically derived SO₄ has decreased in recent years (Ryan et al., 1989; Johnson et al., 1993).

Precipitation SO₄ concentrations and SO₄ deposition have decreased throughout much of the southeastern USA in recent decades (Lynch et al., 1995, 1996). However, even at current rates of SO₄ deposition, watershed model simulations indicate that SO₄ leaching may contribute to chronic soil Ca leaching losses (Hooper and Christophersen, 1992; Johnson et al., 1995). Decreases in S deposition can reduce the rate of base cation leaching over the long term, but ongoing chronic SO₄ deposition at rates well above preindustrial levels will continue to adversely affect leaching of soil Ca.

Soil acidification is also influenced by atmospheric deposition of Ca because this form of Ca input can replenish Ca leached or lost due to tree uptake. Atmospheric Ca deposition has been declining in recent years throughout the eastern USA (Swank and Waide, 1988; Hooper and Peters, 1989; Hedin et al., 1994; Lynch et al., 1995, 1996), which may also decrease Ca availability in soils. The reasons for these declines in Ca deposition are not fully understood but may be related to implementation of emission controls, including electrostatic precipitators that remove particulates containing Ca, fewer unpaved roads, and reduced wind erosion (Hedin et al., 1994).

The net effect of chronic SO₄ loading, reduction in SO₄ retention, decline in Ca deposition, and the uptake of Ca in vegetation is hypothesized to be the depletion of soil Ca. There is a growing body of work that supports this hypothesis, including direct measurements of soils in long-term studies. Acidic deposition-induced declines in soil Ca have been proposed in recent reports in Europe (Bergkvist, 1986; Skeffington and Brown, 1986; Graveland et al., 1994; Kirchner and Lydersen, 1995; Wesselink et al., 1995) and the USA (Johnson et al., 1994; Richter et al., 1994; Lawrence et al., 1995; Likens et al., 1996). In other studies, tree uptake is thought to be largely responsible for observed declines in soil Ca (Johnson et al., 1988a; Johnson and Todd, 1990; Knoepp and Swank, 1994).

Little is known about the impacts of timber harvesting and leaching of Ca on the maintenance of soil fertility and sustainability of forest productivity because there have been only a few harvest rotations in the eastern USA. The conventional view is that nutrient limitation of forest growth in the southeastern USA is typically related to deficiency of N or P, and in some cases, K (Binkley et al., 1989b; North Carolina State Forest Nutrition Cooperative, 1995). However, there are compelling reasons why Ca is likely to become limiting in some forests. In most intensively studied forests it can be shown that the atmospheric deposition of N supplies a substantially higher proportion of the wood requirement for N than is true for Ca (e.g., Johnson and Lindberg, 1992).

Watershed studies provide a useful framework for understanding biogeochemical mass balance because a variety of biotic and abiotic processes that affect element budgets can be quantified (Velbel, 1985, 1988; Bricker et al., 1994). In this study, we examined the biogeochemical mass balance for Ca at one intensively studied watershed in the Piedmont physiographic province near Atlanta, GA, to assess Ca status of the ecosystem. The objective was to quantify soil Ca storage and ecosystem inputs and outputs to determine whether soil reserves were being depleted.

	Materials and methods

TOP ABSTRACT INTRODUCTION Materials and methods NOTES Results and discussion Conclusions REFERENCES

 
Site Description
The PMRW is a first-order 41-ha forested catchment located in the Piedmont physiographic province 25 km southeast of Atlanta, GA (84°10'W, 33°37'N) (Huntington et al., 1993). The forest consists of predominantly even-aged deciduous or mixed deciduous and coniferous stands, and a smaller proportion of predominantly coniferous stands. The communities on mesic sites are dominated by mockernut hickory [Carya tomentosa (Poiret) Nutt.], pignut hickory [C. glabra (Miller) Sweet], northern red oak (Quercus rubra L.), white oak (Quercus alba L.), and tulip poplar (Liriodendron tulipifera L.) (Carter, 1978; Skeen et al., 1980). The dominant coniferous species is loblolly pine (Pinus taeda L.), which occurs mainly in more xeric environments and in a mid successional stand. Hardwood or hardwood dominated stands predominated on 70% of the forested part of the watershed and the remainder was coniferous or coniferous dominated.

The forest composition and age structure reflect past land use and periods of agricultural abandonment typical for the Piedmont region in Georgia (Nelson, 1957; Brender, 1974). There is an 3-ha granite outcrop within the watershed. The outcrop contains several small islands of soil and vegetation. The climate is warm temperate subtropical with a mean annual average temperature of 16°C (Hodler and Schretter, 1986) and long-term average annual precipitation of 1.24 m (Carter and Stiles, 1983). During the period October 1986 through September 1997, rainfall averaged 122.3 cm yr^-1 and runoff averaged 36.5 cm yr^-1. The stream is perennial and during the period 1986 through 1998 the minimum, maximum, and mean flows were 0.57, 973, and 4.74 L s^-1 respectively.

Historical records of regional land use and tree ring analysis at PMRW suggest that most of the timber was originally cut ca. 1820 and that the land was farmed until the early 1900s. A tree diameter survey conducted in 1992 indicated a broad range in tree diameters for all species. There was a relatively uniform distribution of larger (38–53 cm diam. at breast height [DBH]) canopy dominant trees indicating that, on average, there was one 38- to 53-cm DBH tree for every 259 m². Tree ring analyses from cores of 21 trees show the average age for these canopy dominant trees to be 60 to 80 yr in 1992. Tree ring analyses also showed that a small percentage of the trees were 90 to 125 yr old. The occurrence of these older trees indicates that a small area (<6%) within PMRW has not been cultivated since the late 1800s. There is also an area (6 ha) in the northeastern part of the watershed that is predominantly in P. taeda where tree age is substantially less than in the rest of the watershed. The variations in tree diameter, tree age, and vegetation composition together with remnants of barbed wire fencing suggest a complex pattern of agricultural use and abandonment.

Soils are predominantly Typic Kanhapludults—Madison (fine, kaolinitic, thermic Typic Kanhapludult), Pacolet (fine, kaolinitic, thermic Typic Kanhapludult), Cecil (fine, kaolinitic, thermic Typic Kanhapludult)—developed in residuum and colluvium, intergrading to Typic Hapludults (Rion [fine-loamy, mixed, semiactive, thermic Typic Hapludult]), Typic Dystrudrepts (Ashlar [ coarse-loamy, mixed, semiactive, thermic Typic Dystrudept]), and Lithic Udipsamments (Wake [mixed, thermic Lithic Udipsamment]) developed in colluvium or in highly eroded landscape positions (Fig. 1) . Alluvial soils include Aquic Hapludults (Altavista [fine-loamy, mixed, semiactive, thermic Aquic Hapludult]), Typic Fluvaquents (Bibb [coarse-loamy, siliceous, active, acid, thermic Typic Fluvaquent]), and Typic Udifluvents (Congaree [fine-loamy, mixed, active, nonacid, thermic Typic Udifluvent]). Differences between the principal soil series in the uplands are of minor importance to Ca cycling because all are highly weathered, low in cation-exchange capacity (CEC), acidic, and well drained. Ashlar soils differ from the predominant Madison, Pacolet, and Cecil soils in that they do not contain a diagnostic argillic (illuvial, clay-rich) horizon and they are generally more coarsely textured and excessively well drained. Wake soils are shallow and sandy textured and occur mostly as a complex with Ashlar soils. Alluvial soils occur along streambanks and in a small, poorly drained floodplain, near the confluence of three tributaries. The soils at PMRW are similar in chemical properties to predominant soils of the Piedmont region in the southeastern USA (Buol, 1973).

View larger version (54K): [in this window] [in a new window]   Fig. 1 Panola Mountain Research Watershed site map showing soil series, water, soil, and precipitation sampling sites and streams

Chemical weathering of granodiorite at Panola probably supplies only a small amount of Ca for tree uptake based on chemical composition of soil, saprolite, and partially weathered bedrock and on slow weathering rates estimated for plagioclase feldspar in this environment. A Panola Granite weathering profile from a ridge-top position obtained from a drill core that successively penetrated soil, saprolite, and bedrock to a depth of 12 m was analyzed for Ca composition and Ca release (White et al., 1999). In this profile soil extended to 1 m, saprolite to 4.5 m, and weathered granodiorite to 11 m. Major element composition determined by x-ray fluorescence spectroscopy (XRF) indicated that calcite (approximated by coulometric titration and acid digestion and gas chromatography) and plagioclase (approximated by total CaO) are largely depleted in soil, underlying saprolite, and 2 m down into the underlying partially weathered granodiorite (White et al., 1999). Calcite and plagioclase approach unweathered granodiorite composition across a sharp weathering front at a depth of 8 to 10 m. The calcite weathering front is offset downwards from the plagioclase weathering front by <1 m, implying concurrent weathering with plagioclase weathering dominating, because of its much larger mass (White et al., 1999). These parallel weathering fronts imply that solid–solution contact and flushing of accumulated weathering products limit weathering rates.

Nonexchangeable soil and saprolite (residue following 1 M NH₄OAc extraction) at a site adjacent to that described by White et al. (1999) and two hillslope sites at PMRW were analyzed by pelletizing using lithium metaborate fusion, dissolving in HCl, and XRF. The results of this analysis (Alex Blum, USGS, 1994, unpublished data) help to quantify the proportion of mineral Ca remaining in these profiles. The bedrock has a CaO percentage of 1.93% (White et al., 1999) and a bulk density of 2.65 Mg m^-3, whereas in the soil and saprolite to a depth of 4.6 m the CaO percentage varies between 0.14% at the surface and 0.047% near the base and the bulk density varies from 1.2 to 1.6 Mg m^-3. Based on the measured CaO percentages and measured bulk density we calculated that on average for the watershed 2.5% of the original mineral Ca remains in the surface 1 m. Because the deeper saprolite is even more depleted than the surface soil, each successive meter below the surface 1 m has less Ca than the surface 1 m. Depletion of nonexchangeable, mineral Ca at PMRW is consistent with other studies in the southeastern USA (Calvert et al., 1980; Pavich et al., 1989; Daniels and Hammer, 1992; Markewich et al., 1994).

Calcium Mass Balance
The net change in soil exchangeable Ca inventory was operationally defined as the sum of inputs and outputs to the surface 1 m of soil for all upland portions of the watershed. The upland portions include all areas except the alluvial soils near the watershed outlet where the water table remains within the tree rooting zone throughout the year. This area represents 6% of the watershed surface area, and the soils are mapped as Bibb, Congaree, Altavista, and Toccoa Series (Fig. 1). In the upland areas, tree roots do not have access to groundwater underlying these alluvial soils. The Ca concentration in groundwater underlying the alluvial soils is somewhat higher than the stream water, soil water, and ephemeral groundwater in the uplands. The reservoir of groundwater underlying the alluvial soils probably contains weathering products derived from the weathering front within the bedrock. Since trees growing in this area can obtain Ca from this source, the effective rate of Ca weathering and soil resupply is undoubtedly much higher than in the uplands where trees probably do not have access to the products of weathering in a zone beginning 2 m into bedrock and well above the water table.

The change in Ca inventory for the upland soils was expressed as a linear combination of fluxes on an annual basis in units of kilograms per hectare per year

(1)

where is the change in soil exchangeable Ca (kg ha^-1 yr^-1); Q_D is the flux due to atmospheric deposition; Q_W is the flux due to weathering of primary minerals remaining in the soil, saprolite, and partially weathered bedrock to a depth of 1 m; Q_L is the flux due to leaching of soil Ca; and Q_U is the flux due to tree uptake into merchantable timber.

Deposition and Precipitation Monitoring
The U.S. Geological Survey has studied hydrologic, geochemical, and biological processes at PMRW since 1985. Continuous measurements of elemental inputs, outputs and internal cycling provide the mass balance constraints necessary to estimate element budgets. Calcium in wet deposition at PMRW was measured in rainfall using two Aerochem Metrics 301 automatic wetfall collectors (Aerochem Metrics, Bushnell, FL) (Huntington et al., 1994).¹ The location of one collector is shown on Fig. 1; the other collector was located in a clearing 300 m east-northeast of the watershed outlet. Precipitation samples were collected weekly and analyzed by ion chromatography or direct coupled plasma spectroscopy. Solute concentrations were multiplied by weekly rainfall amounts to calculate weekly deposition amounts. Rainfall quantity was recorded using tipping bucket rain gauges and data loggers. Procedures for estimation of wet deposition closely follow those of the National Acid Deposition Program/National Trends Network (NADP/NTN) (Peden, 1986; James, 1993).

The estimation of dry deposition of Ca is important because it is a large proportion of the total atmospheric deposition. However, there is a substantial amount of uncertainty in the methodology for measurement of dry deposited Ca (Lindberg et al., 1988; Church et al., 1989; Lindberg et al., 1990). The measurement and modeling techniques available for estimation of dry deposited Ca rely on a number of assumptions and computational steps. Coarse particulates are thought to be enriched in Ca relative to finer particulates, but techniques for estimating dry deposition of coarse particulates are perhaps the most uncertain because of the difficulty of scaling between surrogate surfaces and complex forest canopies. In this study we used 1.7 for the ratio of dry-to-wet deposition of Ca because it represents a reasonable approximation based on earlier estimates published for the southeastern USA (Church et al., 1989; Johnson and Lindberg, 1992). Using this ratio, wet deposition is multiplied by 2.7 to estimate total atmospheric deposition. The equation used to calculate Ca flux in atmospheric deposition was:

(2)

or the sum for all weeks where P_v is the weekly precipitation volume and P_c is the Ca concentration in precipitation for that week and 2.7 is the factor to estimate total deposition from wet-only deposition.

Weathering Flux
The weathering rate for the watershed as a whole has been estimated to be 4.8 kg Ca ha^-1 yr^-1 (White and Blum, 1995). This estimate is based on the geochemistry of stream discharged at the watershed outlet that has been corrected for precipitation inputs. Because it integrates the entire watershed it includes weathering of amphibolite that occurs only in a small riparian area near the watershed outlet. This estimate ignores the contribution of soil exchangeable Ca to stream water and therefore is an overestimate of the true weathering rate. In the upland portions of the watershed, the weathering front, and presumably the primary source of weathering Ca, is >3 m beneath the saprolite–bedrock interface.

For this study we chose to estimate the mineral weathering flux in the soil, saprolite, and partially weathered bedrock that constitutes the effective tree rooting volume because we think it is unlikely that tree roots penetrate through to the active weathering front. Based on the geochemical analyses described above we assumed that 2.5% of the original primary mineral Ca remained within the surface 1 m and that this material was weathering at the same rate as that estimated for the watershed as a whole. This is a conservative estimate, from the Ca depletion perspective, because the more readily weatherable mineral forms have been depleted. Laboratory experimental simulated weathering studies of granitoid rock materials from Panola indicate a very rapid decline in the release of Ca over time (>1 yr), suggesting that after the removal of the more readily available Ca in fresh and weathered bedrock further release is about two orders of magnitude slower (White et al., 1999). The flux of Ca due to weathering in the surface 1 m was calculated assuming that only 2.5% of the total watershed weathering were derived from this layer since it contained only 2.5% of the original mineral Ca. Based on these assumptions, we estimated that 0.12 kg Ca ha^-1 yr^-1 were released due to weathering in the surface 1 m. Weathering within deeper soil, saprolite, and partially weathered bedrock above the weathering front would be <0.12 kg Ca ha^-1 yr^-1 m^-1 of regolith since these materials have similar or lower CaO percentages than the surface 1 m (White et al., 1999).

Vegetation Uptake
Net annual Ca uptake in merchantable wood was estimated from measurements of merchantable biomass divided by tree age and multiplied by Ca concentration of bole wood for the predominant tree species at PMRW. Merchantable wood is the volume of stemwood plus bark that would be removed from the site in a stem-only tree harvest. Merchantable biomass was determined from a 1992 survey of tree diameter (DBH) and allometric regression equations to estimate stemwood biomass and bark from DBH. For the major deciduous tree species we used equations developed by Clark et al. (1985, 1986). The equations for southern pine tree species were provided in written communication (Alexander Clark III, USDA Forest Service, Athens, GA, 1992). The equations for pine biomass were:

(3)

(4)

The pine equations were for green weight, and we assumed that dry weight was equal to 53% of green weight. The tree survey included 67 plots with 0.04-ha area each. Tree diameter was measured using calipers for trees under 20-cm DBH and standard DBH measuring tape for trees 20 cm. For these surveys, trees were defined as live, merchantable stems 10 cm DBH. Biomass associated with shrubs and saplings averaged 7% of total woody stem biomass for the deciduous forest at Panola Mountain (Carter, 1978). Shrub and sapling biomass was not included in the calculation of merchantable wood.

Calcium concentrations in bole wood were estimated using values reported for appropriate tree species (Switzer and Nelson, 1972; Johnson et al. 1988a, 1988b; Johnson and Henderson, 1989; Johnson and Todd, 1990). A watershed-average value of 6.2 g kg^-1 Ca concentration in merchantable wood was assumed for hardwood species. A watershed-average value of 1.0 g kg^-1 Ca concentration in merchantable wood was assumed for southern pine species. Calcium removal during harvest would be underestimated if the forest were harvested by whole-tree harvesting methods since this would remove more Ca in branches. The equation used to calculate the flux of Ca into merchantable biomass for each plot was:

(5)

where B_P and B_H are the rates of merchantable biomass accumulation in pine and hardwood, respectively (Mg ha^-1 yr^-1); and C_BP-Ca and C_BH-Ca are the Ca concentrations in pine and hardwood stemwood biomass respectively. Both C_B-Ca and B are calculated on a dry weight basis.

Leaching Fluxes
Calcium leaching losses were estimated using the following equation.

(6)

where Q_R is the mean annual measured runoff (corrected by subtracting runoff from the 3-ha granite outcrop) and C_S-Ca is the average Ca concentration for water draining out of the rooting zone. Water draining out of the rooting zone is assumed to have leached through the soil and acquired a Ca concentration equivalent to the average measured concentration for waters draining upland soils. Soil water flux was considered to be equivalent to runoff (corrected for the outcrop) since overland flow occurs only on the steepest slopes during the most intense storms and since the watershed is assumed to be tight so that no groundwater is exported under the weir. Average Ca concentration for water draining the rooting zone was multiplied by the water flux to calculate the average annual amount of Ca leached. The average concentration for water draining the rooting zone was based on analysis of hundreds of water samples collected during storms between 1986 and 1996. These samples were collected from soil, ground, and surface waters draining the uplands. On the basis of cores drilled into underlying bedrock and bedrock exposures, we think that all of these locations are underlain by granitic parent materials. For most of these locations, flow is ephemeral, stopping after storm waters have drained out of contributing soil and saprolite. The exceptions are surface water sites SW210 and SW420 draining upland portions of the catchment that maintain flow during most of the year. The locations of the collectors are shown in Fig. 1. Soil water was collected with porous cup lysimeters where the collection area was centered at 91-cm soil depth. Vacuum was applied prior to storms and samples were pumped out of the lysimeters following storms. Groundwater samples were collected from shallow (1.5–2 m) wells resting on bedrock. The wells were screened for 30 cm above the saprolite–bedrock interface so that they collect only water that has drained through >1 m depth of soil and represents the local groundwater at the time of sampling. Groundwater samples were collected from the wells after three well volumes had first been pumped and discarded. Surface water samples were collected both manually and using automatic samplers during storms. More detailed descriptions for sample collection and analysis are reported elsewhere (Huntington et al., 1994).

Soil Exchangeable Calcium Inventory
Samples from the mineral soil were collected at 35 sites located throughout the watershed representing different landscape positions, drainage classes, and vegetation types (Fig. 1). Soil samples were collected prior to the current soil mapping (Fig. 1) so they were not stratified based on the mapped soil series. Sampling positions were randomly located along transects, or selected randomly from a pool of all possible points within landscape or vegetation classes. Locations for transects were selected randomly from all possible locations representing those landscape gradients. Potential sampling sites were rejected if they fell within 1.5 m of a tree trunk. For 32 sites sampled in 1991 and 1992, samples were collected by soil horizon. Soil pits 50 cm in diameter were excavated to identify soil horizons and to collect samples. Bulk density was measured with a standard technique using a cylindrical coring device driven vertically downward into the midrange of each horizon (Blake and Hartge, 1986). In 1991, three additional sites were sampled by depth strata rather than by soil horizon. At each of these three sites, 8 to 10 samples were collected in the surface 1 m using a 5.7-cm-diam. bucket auger. Bulk density was measured with the same technique as described above.

Coarse (>2 mm) root volume was not quantified in this study but was estimated to be <0.2% of the total volume in the surface 1 m based on belowground biomass of 53 Mg ha^-1 (Huntington, 1995) and an average specific gravity of 0.4 g cm^-3. Root volumes estimated for a northern hardwood forest in New Hampshire (Fahey et al., 1988) and for southern hardwood forests in North Carolina (Monk and Day, 1988) were <1% of total soil volume in the surface 1 m.

Samples were collected by horizon from all four faces of the excavated pits. Samples from each pit face were aggregated to form a bulk composite horizon sample that was representative of the entire range of horizon depth. Soils were air dried and sieved to pass a 2-mm sieve. We estimated that rock coarse fragment volume (>2 mm) totaled <1% of the volume of material excavated from all pits with the exception of two pits closest to the margin of a large rock outcrop where exfoliation was evident. In these two pits rock fragment volume probably exceeded 30% but was not quantified; bulk density measurements were made where no rock was encountered. Based on the soil map (Fig. 1) that indicates large areas of Ashlar–Wake complex, which is classified as very bouldery, rock fragment volume may be substantially greater in some areas within the watershed than where we sampled. However, we collected many samples in the area that was mapped Ashlar–Wake complex and we did not find significant rock fragment volume except for the two pits mentioned. In this study, we did not quantify rock fragment volume so pool sizes for exchangeable cations are overestimates of base cation content in areas with appreciable rock fragment volume.

Soil exchangeable cations (Ca, Mg, K, Na, and Al) were measured using unbuffered 1 M NH₄Cl extraction following a mechanical vacuum extraction method described by Huntington (1996). Cation concentrations in extracts were measured by direct coupled plasma optical emission spectroscopy. Soil mass for each horizon (or depth interval) was calculated from horizon thickness and bulk density. Concentrations of exchangeable cations were multiplied by soil mass to compute elemental contents by horizon or depth strata. Exchangeable element contents by horizon were converted to depth strata by simple linear interpolation so that all of them could be compared on a depth basis. In some shallow soils bedrock was encountered between 44 cm and 1 m. If there was no soil in a specified depth strata, element content for that strata was set to zero so that the calculation of total element content within the surface 1 m was for all 35 pedons. Cation-exchange capacity was estimated as the sum of exchangeable base cations plus Al. Base cations and Al were determined from the unbuffered 1 M NH₄Cl extraction and reported in units of centimoles of charge per kilogram. Base saturation was determined as the sum of base cations (Ca, Mg, K, and Na) divided by the CEC and expressed as a percentage. Calcium saturation was determined as exchangeable Ca divided by CEC times 100. Soil pH was determined in a mixture of 5 g soil and 5 cm^-3 of 0.01 M CaCl₂ after stirring intermittently for 30 min (McLean, 1982). Particle-size distribution for each horizon was determined by pipette and sieving (Kilmer and Alexander, 1949).

Forest floor samples were collected at 12 locations under deciduous and 12 locations under coniferous forest plots during 1993 and 1994. The forest plots were selected from those that had been sampled for litterfall and throughfall in an earlier study (Cappellato et al., 1993). Forest floor mass varies seasonally, particularly in the deciduous plot because of low mass and rapid rate of decomposition and incorporation of litter, which is characteristic of these mull soils (Buol et al., 1980). Forest floor samples (combined Oi, Oe, and Oa) (0.0625 m²) were collected during all seasons of the year. Mass was calculated for each collection and these values were averaged to determine the annual average forest floor mass under each forest type. Calcium concentration in the forest floor was not determined but, based on published values, was assumed to be10 g kg^-1 under hardwoods and 7.1 g kg^-1 under conifers (Johnson and VanHook, 1989; Johnson and Todd, 1990; Johnson and Lindberg, 1992; Richter et al., 1994). Litterfall was measured on hardwood (1994–1997) and coniferous plots (1992) using a minimum of eight baskets (55.5 by 41.5 cm) per plot. Litter baskets were placed randomly in 10 by 20 m plots.

	Results and discussion

TOP ABSTRACT INTRODUCTION Materials and methods NOTES Results and discussion Conclusions REFERENCES

 
Calcium Deposition Fluxes and Trends
Atmospheric wet deposition of Ca at PMRW averaged 0.83 kg ha^-1 yr^-1 for the period 1986 to 1996 (standard deviation was 0.18 kg ha^-1 yr^-1). The range in deposition during this period was 0.63 to 1.17 kg ha^-1 yr^-1. Dry deposition was estimated as 1.7 times wet deposition, yielding a total atmospheric deposition of 2.24 kg ha^-1 yr^-1. Throughout the southeastern USA, there is substantial variation in wet deposition of Ca as measured in the NADP/NTN monitoring network (Lynch et al., 1995). Wet deposition varied from 0.58 to 2.12 kg ha^-1 yr^-1, with a mean of 1.23 kg ha^-1 yr^-1 for 32 southeastern U.S. sites having varying periods of record through 1994. There was a small but significant downward trend in precipitation Ca concentration at PMRW from 1986 through 1994. During that period precipitation Ca concentration decreased at an average rate of 0.149 µmol kg^-1 yr^-1. In 1995 and 1996 Ca concentration in precipitation increased so that the longer-term trend for the period 1986 to 1996 was not significant.

Although N has historically been considered the element most limiting to forest growth in the southeastern USA, recent element cycling studies indicate that Ca is more likely to be significantly depleted by harvesting and leaching (Johnson and Lindberg, 1992). Comparing the biogeochemistry of N and Ca at PMRW suggests that in the long term Ca may become more limiting than N. At PMRW in the Georgia Piedmont, atmospheric deposition of 4.3 kg N ha^-1 yr^-1 (1985–1997) provides 75% of the estimated wood requirement for N. In contrast, the estimated atmospheric deposition of Ca of 2.24 kg ha^-1 yr^-1 provides only 25% of the wood requirement. Furthermore, with the exception of high elevation forests, very little N is exported from southeastern forests compared with Ca, so that nearly all of the atmospherically deposited N becomes available for tree growth. For example, at PMRW only 0.1 kg N ha^-1 yr^-1 is lost due to leaching to the stream and exported from the upland portions of the watershed, whereas 2.7 kg Ca ha^-1 yr^-1 is exported.

Calcium Uptake into Merchantable Biomass
Dry weight of merchantable biomass in 1992 was estimated to be and Mg ha^-1 for hardwood and pine species, respectively, for a total biomass of Mg ha^-1. Using 6.2 g kg^-1 Ca for hardwoods and 1.0 g kg^-1 Ca for pines, we estimate and kg Ca ha^-1 storage in merchantable biomass. On the basis of an assumed average tree age of 70 yr in 1992, we estimate the annual net wood increment of Ca uptake into merchantable biomass to be and kg ha^-1 yr^-1 for hardwood and pine species, respectively, for an average rate of 12.3 (SE = 0.76) kg ha^-1 yr^-1. This biomass uptake rate is somewhat lower than that reported for predominantly oak–hickory forests at Walker Branch in Tennessee and similar to that reported for southern hardwoods at Coweeta, NC (Johnson and Lindberg, 1992).

Calcium Leaching Fluxes
Calcium concentrations in soil tension lysimeters, groundwater wells in footslope positions, and stream water were consistently in the range 15 to 25 µmol kg^-1 with mean of 20 µmol kg^-1 (Fig. 2) . During the period October 1986 through September 1997, rainfall averaged 122.3 cm yr^-1 and runoff averaged 36.5 cm yr^-1. Assuming that all of this water leached through the soil and acquired an average Ca concentration of 20 µmol kg^-1, the Ca leaching loss over this period averaged 2.71 kg ha^-1 yr^-1.

View larger version (19K): [in this window] [in a new window]   Fig. 2 Box plot showing Ca concentration mean, quartiles, deciles, and outliers in groundwater (GW), soil water (VW), and streamwater (SW) at various stations in the Panola Mountain Research Watershed, Georgia

Soil Exchangeable Calcium
The pool of soil exchangeable Ca in the surface 1 m of mineral soil at PMRW in 1992 was estimated to be (Table 1) . The 0- to 10-cm depth strata contained about 2.5 times as much Ca as the 10- to 20-cm strata (Table 1), reflecting the higher exchangeable Ca concentrations in the A and A/B horizons compared with those in underlying horizons (Table 2) . In spite of the fact that exchangeable cation concentrations are substantially lower in deeper, B horizons, than in the surface horizons, there is almost twice as much Ca stored between 20 cm and 1 m than in the surface 0- to 20-cm depth strata because of the greater soil mass in this strata. The high spatial heterogeneity in pool size was mostly due to heterogeneity in soil depth. Soil thickness varied substantially even within a given landscape position. Shallow soils are common on ridges, but deeper soils also occur on ridgetops. Typically, the shallowest soils occur in midslope positions on steeper slopes and the deepest soils occur in footslope positions or in the floodplain alluvium near the watershed outlet.

View this table: [in this window] [in a new window]   Table 1 Exchangeable cation pools by soil depth strata at the Panola Mountain Research Watershed.

View this table: [in this window] [in a new window]   Table 2 Average horizon thickness, horizon depth, concentration of exchangeable cations, cation-exchange capacity (CEC), base saturation, and Ca saturation.

There is no evidence for differences in clay mineralogy between horizons (Shanley, 1989; Nixon, 1981), so it is likely that higher soil organic matter in the A horizon (Table 3) is responsible for higher CEC in the A horizon. These relationships with soil depth (Tables 2 and 3) reflect the fact that biological cycling maintains higher Ca concentration in the A horizon as is true for base cations in Paleudult soils predominating the Atlantic Coastal Plain (Buol, 1973). The predominant soils at Panola Mountain are Hapludults rather than Paleudults, but because they are depleted of Ca-bearing weatherable minerals Ca has been conserved through biological cycling and possibly the residual effects of past liming practices.

View this table: [in this window] [in a new window]   Table 3 Bulk density, texture, organic C and N concentration, and pH by soil horizon at the Panola Mountain Research Watershed

Annual litterfall (excluding bolewood) is 6.0 Mg ha^-1 under deciduous forest and 3.6 Mg ha^-1 under coniferous forest at Panola. The mean residence time for litter (excluding bolewood) in the forest floor under the hardwood forest was 1 yr compared with 12 yr in the coniferous forest. Based on the relatively small pool of Ca in the forest floor and its rapid turnover and replenishment, this pool is probably not subject to chronic depletion as is the exchangeable pool in the mineral soil.

Cores drilled into the rock and exposures of rock in local roadcuts indicate that the hard rock is fractured along both flat-lying and steeply inclined planes. The less common near-vertical fractures connect the flat-lying fractures with the overlying saprolite and soil horizons. Near-horizontal, water-bearing fractures have been recognized in granite bodies as deep as 1 km. Within 100 m of the land surface, the rock bordering these fractures may be locally saprolitized, with thicknesses of the weathered zone ranging from a few centimeters to at least 1 m (David Prowell, USGS, Atlanta, GA, 1998, personal communication). Evidence from drilling and crest stage gauges installed to a depth of 3.6 m at a ridge location indicates that the water table does not rise into these fractures and therefore cannot be a mechanism that transports Ca from deeper strata to the main rooting zone. Tree roots that penetrate these fractures to any significant extent could acquire Ca from fresh weathering surfaces and transport it to the land surface. It is not known whether such deep rooting provides appreciable Ca for forest growth.

Calcium Input–Output Budget
Biogeochemical mass–balance calculations for Ca at PMRW averaged for the 70-yr life of the current forest stand indicate that the rate of Ca depletion by vegetation uptake and soil leaching is several times greater than the rate of input through atmospheric deposition (Fig. 3) . Based on the average annual wood increment and the Ca content of merchantable wood of the species occurring at PMRW, the estimated rate of Ca uptake into merchantable wood is 12.3 kg ha^-1 yr^-1. The estimated rate of soil leaching, 2.71 kg ha^-1 yr^-1, is slightly greater than the rate of atmospheric deposition of Ca in precipitation and dry deposition, 2.24 kg ha^-1 yr^-1. The rate of Ca inputs through weathering of primary minerals was not measured directly in this study but was estimated to be 0.12 kg ha^-1 yr^-1 assuming that mineral Ca remaining within the surface 1 m was weathering at the same rate as bedrock at the weathering front per unit mass of Ca remaining. The net rate of Ca depletion from the soil exchange complex is nearly equivalent to tree uptake into merchantable timber. At PMRW the long-term rate of Ca depletion from the soil exchangeable pool is estimated to be 12.7 kg ha^-1 yr^-1. Continuous depletion at this rate would reduce soil reserves to 762 kg ha^-1, the approximate requirement for a merchantable stand (60 yr old) of mixed southern hardwood and southern pines at this site, in 82 yr.

View larger version (31K): [in this window] [in a new window]   Fig. 3 Schematic showing pools and fluxes of Ca at the Panola Mountain Research Watershed

View larger version (19K): [in this window] [in a new window]   Fig. 4 Distribution of soil exchangeable Ca pool vs. soil Ca depletion rate for Panola Mountain Research Watershed and several other intensively studied sites in the southeastern USA. American beech, Fagus grandifolia Ehrh. (A) (Johnson and Lindberg, 1992); chestnut oak, Quercus prinus L. (B) (Johnson and Todd, 1990); mixed conifer and southern hardwood (C) (Johnson and Todd, 1990); Quercus–Carya spp. (D) (Johnson and Todd, 1990); Quercus spp. (E) (Rick Web, Univ. of Virginia, 1995, personal communication); loblolly pine, Pinus taeda L. (F) (Richter et al., 1994); mixed southern hardwood (G) (Johnson and Lindberg, 1992); P. taeda (H) (Huntington, 1996); P. taeda (I) (Johnson and Lindberg, 1992); mixed conifer and southern hardwood (J) (this study); P. taeda (K) (Johnson et al., 1988b); P. taeda (L) (Johnson and Lindberg, 1992); slash pine, P. elliotti Engelm. (M) (Johnson and Lindberg, 1992); eastern white pine, P. strobus L. (N) (Johnson and Lindberg, 1992); red spruce, Picea rubens Sarg. (O) (Johnson and Lindberg, 1992)

The recovery of water quality and soil base cation status that are expected to occur following declines in SO₄ deposition may slow because of declines in Ca deposition (Lawrence et al., 1999). This will be most noticeable in areas of base-poor soils and low rates of mineral weathering. Recovery of base cation status requires replenishment of soil exchangeable cations previously lost via leaching and vegetation uptake, which is largely dependent on atmospheric deposition inputs.

Uncertainties in Calcium Flux Estimates
There are uncertainties in each component of the Ca mass balance that are important to an overall assessment of the reliability of the conclusions regarding net soil Ca depletion and the implied threat to sustainability of forest productivity. In this and other intensive case studies, a number of assumptions are required to construct input and output budgets. An examination of several of the more critical uncertainties is useful for understanding how robust the conclusions of this study are, the regional scope of their applicability, and needs for further study.

Uncertainty in the rate of Ca inputs through mineral weathering to the rooting environment of trees is one of the most critical aspects of this elemental mass balance. Although 98.5% of weatherable mineral Ca has been depleted from the soil and saprolite at PMRW, the amount remaining (9500 kg ha^-1) is still substantial and comparable with that found at other intensively studied sites in the southeastern USA (Johnson and Lindberg, 1992). Uncertainty in the rate of weathering resupply hinges on the significance of the weathering of this residual plagioclase feldspar and on whether there are any mechanisms by which trees can recover Ca from the active weathering front that is well into the bedrock. Several lines of evidence suggest that the rate of weathering of the residual plagioclase feldspar is probably insufficient to resupply a significant amount of Ca in comparison with amounts of Ca lost via uptake and leaching.

Based on the weathering rate calculated for the watershed as a whole, which contains amphibolite near the outlet only, the Ca released from the depleted portion of the profile is very small. At PMRW it has been estimated that the weathering flux is 4.8 kg Ca ha^-1 yr^-1 (White and Blum, 1995). This estimate may be an overestimate because it is based on discharge at the watershed outlet and includes weathering of amphibolite that is not present in most of the watershed and because it ignores net losses due to soil ion exchange reactions that leach Ca from soils. If 2.5% of the weathering were derived from the soil, saprolite, and partially weathered bedrock, which is consistent with the distribution of Ca in this profile, then only 0.12 kg Ca ha^-1 yr^-1 would be released. The remainder of the Ca released in the watershed would be derived from the weathering front, which at Panola begins at 2 m below the saprolite–bedrock interface (White et al., 1999). At Panola, and at most Piedmont landscapes, it is doubtful that the trees on hillslopes and ridges have access to the weathering products from this weathering front. Piezometers and crest stage gauges monitored between 1992 and 1994 at PMRW at the ridge location and on the hillslopes below the ridge never recorded any free water at the saprolite–bedrock interface on ridge and hillslope positions, suggesting that flowpaths draining this weathering front only enter the stream system in the riparian zone.

Several long-term studies at southeastern U.S. sites have demonstrated declines in exchangeable soil Ca pools, indicating that weathering and deposition inputs together are insufficient to replenish losses through vegetation uptake and leaching (Johnson et al., 1988a, 1991; Richter et al., 1994; Knoepp and Swank, 1994). In contrast to these studies a recent study in Tennessee found no decrease in soil exchangeable Ca during 15 yr following a whole-tree harvest (Johnson and Todd, 1998). In that study Ca uptake in vegetation and leaching was large (50% of initial exchangeable soil pool), suggesting the possibility that weathering of residual minerals or dolomitic parent material was able to resupply the exchangeable pool through deep rooting, although no data were collected to evaluate these hypotheses.

In the northeastern USA, parent materials are typically much younger and contain more weatherable Ca than in the southeastern USA (April and Newton, 1992); therefore, weathering probably plays a more significant role in cation resupply. However, studies in the Adirondack Mountains of New York (Johnson et al., 1994) and at Hubbard Brook in the White Mountains of New Hampshire (Likens et al., 1996) have indicated declines in soil Ca reserves in these forest soils, as well as suggesting that weathering rates are insufficient to resupply Ca. Investigations in Europe also found decreases in soil exchangeable Ca attributed to both acidic deposition-induced leaching losses and tree uptake (Bergkvist, 1986; Skeffington and Brown, 1986; Ulrich, 1991; Falkengren-Grerup and Tyler, 1992; Graveland et al., 1994; Wesselink et al., 1995; Lapenis et al., 2000). The presence of variable amounts of exchangeable and weatherable mineral sources of Ca in these studies suggests that weathering rates are too slow to compensate for aggregate losses.

Evidence of soil Ca depletion in northeastern U.S. forests comes from the use of Sr isotope ratios as tracers of atmospheric vs. mineral sources of Ca. In one study 50 to 60% of the soil exchangeable Ca and stemwood Ca in red spruce, balsam fir [Abies balsammea (L.) Mill.], and heart leaf paper birch (Betula cordifolia Reg.) stands in the Adirondack Mountains, New York, was of atmospheric origin, suggesting depletion of Ca from mineral sources (Miller et al., 1993). Recent studies of the temporal pattern of Sr isotope ratios in red spruce at Cone Pond in New Hampshire have indicated that in recent decades trees appear to have shifted sources of cations towards proportionately more being derived from atmospheric sources vs. mineral sources (T. Bullen, USGS, 1999, personal communication).

Trends in stream water Ca concentration in recent decades are also consistent with declining soil Ca inventories. If changes in stream water Ca concentration reflect changes in soil exchangeable inventories rather than changes in weathering rates, then declines in concentration suggest decreases in exchangeable Ca concentration. Evidence from long-term monitoring studies at PMRW and several other locations in the eastern USA indicates that Ca concentrations in stream water have decreased in recent years (Swank and Waide, 1988; Aulenbach et al., 1996; Likens et al., 1996; Mast and Turk, 1999; Clow and Mast, 1999). These declines have been associated with Ca depletion through vegetation uptake, soil leaching, and declines in Ca deposition (Likens et al., 1998). There is also strong evidence for declining stream water Ca concentration in the Adirondack and Catskill Mountains in New York (Stoddard et al., 1999) and the Shenendoah Mountains in Virginia (Rick Webb, 1998, personal communication).

It is unlikely that tree roots can penetrate to the weathering front to recover Ca since it is 2 m into partially weathered, competent rock. It is not certain what the flowpaths are from the weathering front to the stream, but they are likely through the framework of scarce vertical fractures and more common flat-lying fractures in the bedrock. Only those tree roots that can penetrate the occasional vertical fractures would have access to the flowpaths down gradient from the weathering front. Root distribution was not quantified in this study, but only a very small fraction (<0.1%) of tree roots in temperate deciduous forests normally occur below 2 m (Jackson et al., 1996).

Net fluxes into trees are quite variable depending on tree species, site environmental conditions, and stand age. In rare cases where forest harvesting is prohibited, net fluxes may be zero if the ecosystem is at steady state or even negative (released from the biomass) if the ecosystem is degrading. An important uncertainty in the long-term rate of tree uptake is the possible adaptation by forests to declining soil Ca status. Trees may adapt to accumulate less Ca as soil exchangeable and soil solution Ca concentrations decline (Johnson et al., 1995). Aggrading forests may adapt by changing species composition in favor of trees that tolerate lower Ca supplies. Under conditions of Ca limitation, tree growth rates would decline, resulting in reduced soil depletion rates. Following timber harvesting or other disturbances naturally regenerating forest stands are likely to adapt by changing species composition in favor of species with smaller Ca requirements. Declines in soil Ca may also lead to reduced tree resistance to pests and pathogens, resulting in greater infestation and mortality (Mclaughlin and Wimmer, 1999) and lower net Ca uptake.

Calcium flux into merchantable tree biomass represents a large component in the Ca budget for most forests that have been extensively studied (Johnson and Todd, 1990; Johnson and Lindberg, 1992; Richter et al., 1994). As the data from this study and other studies indicates, Ca losses through tree uptake frequently exceed fluxes in soil leaching (Johnson and Todd, 1990; Johnson and Lindberg, 1992; Richter et al., 1994). Model simulations using the nutrient cycling model NUCM (Johnson et al., 1995) and geochemical mass balance analyses (Taylor and Velbel, 1991) have illustrated the importance of tree uptake in decreasing exchangeable soil Ca pools in the southeastern USA.

There is no systematic soils inventory at a regional scale with which to assess how widespread the Ca depletion problem might be, but there are small data sets that give some indications. The soils database developed by the Direct Delayed Response Project (DDRP) (Turner et al., 1993) indicates that 70% of the soils in the Southern Blue Ridge Province had exchangeable Ca inventories of 1000 kg ha^-1 or less, which is about one-half that at Panola and midrange for the sites shown in Fig. 4. Similarly, in a forest soil database in southeastern USA (Lewis and Conkling, 1994) for southern pine ecosystems 50% or more of the soils in the Coastal Plain subregion had exchangeable Ca inventories of 1000 kg ha^-1 or less and 50% or more of the soils in the Piedmont subregion had exchangeable Ca inventories of 2500 kg ha^-1 or less.

Calcium depletion on a regional scale is of major concern for the health and productivity of forest ecosystems. It has been proposed that Ca depletion is having an adverse impact on forest productivity for northern hardwood forests (Likens et al., 1996). Calcium depletion is also implicated in the decline in red spruce (Shortle et al., 1995, 1997) and sugar maple (Acer saccharum Marshall) (Long et al., 1997). McLaughlin and Wimmer (1999) have hypothesized that Ca supply exerts a significant control on forest structure and function and that forest ecosystem adaptation to reduced Ca will become increasingly evident as Ca is further depleted. Calcium regulates many physiological processes that are directly related to growth, and Ca is involved in response to stress from pests, pathogens, winter injury, and oxidant damage. Numerous indications of physiological adaptations to limited Ca supply, forest ecosystem dysfunction at many levels under reduced Ca, and positive responses of a variety of indicators of forest health to Ca additions (McLaughlin and Wimmer, 1999) support this hypothesis.

	Conclusions

TOP ABSTRACT INTRODUCTION Materials and methods NOTES Results and discussion Conclusions REFERENCES

 
The assessments of Ca inputs and outputs at PMRW and other intensively studied sites in southeastern and mid-Atlantic regions of the USA indicate a broad regional pattern of Ca depletion in forest ecosystems. Soils at PMRW have relatively low Ca reserves in exchangeable forms in comparison with tree requirements to produce merchantable timber. The preponderance of evidence suggests that the weathering rate of nonexchangeable Ca in primary minerals within the rooting zone is insufficient to compensate for the losses due to tree uptake and soil leaching. This assessment suggests that Ca depletion at PMRW and other similar forest ecosystems of the southeastern USA has the potential to threaten the long-term sustainability of forest productivity. Investigating and monitoring this potential threat will require support for long-term multidisciplinary studies implemented at various spatial scales. Existing small watershed studies supported by the U.S. Geological Survey, U.S. Forest Service, National Park Service, Department of Energy, and National Science Foundation, in conjunction with monitoring programs such as the NADP/NTN, Forest Inventory and Analysis, and the Forest Health Monitoring programs, could provide the framework for a comprehensive assessment.North Carolina State Forest Nutrition Cooperative 1995

	ACKNOWLEDGMENTS

 
The U.S. Geological Survey's Water Energy and Biogeochemical Budgets Program for small watershed research and the U.S. Geological Survey's Acid Rain Program supported this analysis. Larry West (University of Georgia, Athens, GA), David Prowell (USGS, Atlanta, GA), and Helaine Markewich (USGS, Atlanta, GA) provided helpful insights regarding the mineralogy, weathering, geomorphology, and underlying bedrock fracturing at Panola Mountain. Rick Webb (University of Virginia) provided unpublished data from the Shenandoah Mountains and also a helpful manuscript review. Alisa Mast (USGS, Denver, CO) provided a helpful manuscript review. Larry West (University of Georgia, Athens, GA) provided soil particle-size analyses.

	NOTES

TOP ABSTRACT INTRODUCTION Materials and methods NOTES Results and discussion Conclusions REFERENCES

 
¹ Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Geological Survey.

Received for publication May 4, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION Materials and methods NOTES Results and discussion Conclusions REFERENCES

 

• April R., Newton R. Mineralogy and mineral weathering. In: Johnson D.W., Lindberg S.E., eds. Atmospheric deposition and forest nutrient cycling. New York: Springer-Verlag, 1992:378-425.
• Atkins R.L., Higgins M.W. Superimposed folding and its bearing on geologic history of the Atlanta, Georgia, area. In: Frey R.W., ed. Excursions in southeastern geology, Vol. 1. Atlanta: American Geological Inst, 1980:19-40.
• Aulenbach B.T., Hooper R.P., Bricker O.P. Trends in the chemistry of precipitation and surface water in a national network of small watersheds. Hydrol. Process. 1996;10:151-181.
• Bergkvist B. Leaching of metals from a spruce forest soil as influenced by experimental acidification. Water Air Soil Pollut. 1986;31:901-916.
• Binkley D., Driscoll C.T., Allen H.L., Shoenberger P., McAvoy D. Acidic deposition and forest soils: Context and case studies in the southeastern U.S. New York: Springer-Verlag, 1989 a Ecological Studies no. 72..
• Binkley D., Valentine D., Wells C., Valentine U. An empirical model of the factors contributing to 20-yr decrease in soil pH in an old-field plantation of loblolly pine. Biogeochemistry 1989;8:39-54 b.
• Blake G.R., Hartge K.H. Bulk density. In: Klute A., ed. Methods of soil analysis. Part 1, 2nd ed Madison, WI: ASA and SSSA, 1986:363-375 Agron. Monogr. 9..
• Brender E.V. Impact of past land use on the lower Piedmont forest. J. For. 1974;72:34-36.
• Bricker O.P., Paces T., Johnson C.E., Sverdrup H. Weathering and erosion aspects of small catchment research. In: Moldan B., Cerny J., eds. Biogeochemistry of small catchments: A tool for environmental research. New York: John Wiley and Sons, 1994:85-105 SCOPE.
• Buol, S.W. 1973. Soils of the southern states and Puerto Rico. North Carolina State Univ. and SCS USDA. Southern Coop. Ser. Bull. 174.
• Buol S.W., Hole F.D., McCracken R.J. Soil genesis and classification, 2nd ed Ames: Iowa State Univ. Press, 1980.
• Calvert C.S., Buol S.W., Weed S.B. Mineralogic characteristics and transformations of a vertical rock-saprolite-soil sequence in the North Carolina Piedmont I. Profile morphology, chemical composition, and mineralogy. Soil Sci. Soc. Am. J. 1980;44:1096-1103.[Abstract/Free Full Text]
• Cappellato R., Peters N.E., Ragsdale H.L. Acidic atmospheric deposition and canopy interactions of adjacent deciduous and coniferous forests in the Georgia Piedmont. Can. J. For. Res. 1993;23:1114-1124.
• Carter, M.E.B. 1978. A community analysis of the Piedmont deciduous forest of the Panola Mountain State Conservation Park: Atlanta, Georgia. M.S. thesis, Emory Univ., Atlanta.
• Carter R.F., Stiles H.R. Average annual rainfall and runoff in Georgia, 1941–1970: Georgia Geological Survey, Hydrologic Atlas 9, 1 sheet. Reston, VA: U.S. Geological Survey, 1983.
• Church M.R., Thornton K.W., Shaffer P.W., Stevens D.L., Rochelle B.P., Holdren G.R., Johnson M.G., Lee J.J., Turner R.S., Cassell D.L., Lammers D.A., Campbell W.G., Liff C.I., Brandt C.C., Liegel L.H., Bishop G.D., Mortenson D.C., Pierson S.M., Schmoyer D.D. Direct/delayed response project: Future effects of long-term sulfur deposition on surface water chemistry in the Northeast and Southern Blue Ridge province Volume 1: Executive Summary, project approach and data sources. Corvallis, OR: USEPA, 1989 EPA/600/3-89/061a, + app..
• Clark A., III, Burgan T.M., Field R.C., Dress P.E. Users manual for total-tree multi-product cruise program. USDA Forest Service, Southeast Forest Experiment Station. Asheville, NC: USDA Forest Serv, 1985 General Technical Rep. SE-31. Southeast Forest Exp. Stn.,.
• Clark A., III, Phillips D.R., Frederick D.J. Weight, volume, and physical properties of major hardwood species in the Piedmont. USDA Forest Service, Southeast Forest Experiment Station. Asheville, NC: Forest Serv, 1986 General Technical Rep. SE-255. Southeast Forest Exp. Stn., USDA.
• Clow D.W., Mast M.A. Long term trends in streamwater and precipitation chemistry at five headwater basins in the Northeastern United States. Water Res. Res. 1999;35:541-554.
• Daniels R.B., Hammer R.D. Soil geomorphology. New York: Wiley, 1992.
• Fahey T.J., Hughes J.W., Pu M., Arthur M.A. Root decomposition and nutrient flux following whole-tree harvest of northern hardwood forest. For. Sci. 1988;34:744-768.
</