| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA Forest Service, 2730 Savannah Hwy., Charleston, SC 29414 USA
b Desert Research Institute, P.O. Box 60220, Reno, NV 89506 USA
c Jr., Environmental and Resource Sciences Division, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN 27831 USA
trettinc{at}cofc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResultsDiscussionConclusions and perspectivesREFERENCES |
|---|
Abbreviations: ICP, inductively coupled plasma spectometry • WBW, Walker Branch Watershed
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusions and perspectivesREFERENCES |
|---|
In contrast to suggestions that soil nutrient pools are buffered from short-term changes, several recent, long-term studies have shown that soil change can occur on a scale of decades or even on a seasonal basis under certain conditions (reviewed by Johnson et al., 1991). Significant reductions in the pools of exchangeable base cations for instance, have been noted during periods of one to three decades in European and North American forests (Ahokas, 1997; Binkley et al., 1989; Drohan and Sharpe, 1997; Johnson et al., 1994, 1988; Falkengen-Grerup and Eriksson, 1990; Falkengen-Grerup and Tyler, 1992; Hallbäcken and Tamm, 1986; Knoepp and Swank, 1994; Mandersheid et al., 1995; Richter et al., 1994). Surface soil concentrations may vary on a seasonal basis (Haines and Cleveland, 1981; Johnson et al., 1988). These changes were attributed to sequestration of base cations (especially Ca) in biomass and to leaching, the latter of which is often accelerated by acid deposition.
A long-term study of soils on the Walker Branch Watershed (WBW), in Tennessee, was among the first to suggest that soil Ca depletion from uptake could be a major factor constraining sustainable production if stands were harvested without regard to conserving nutrients (Johnson et al., 1988). Other work on those soils showed that cation leaching was also a major factor affecting the soil cation balance (Johnson and Todd, 1990). Long-term studies on the Coweeta watersheds and a South Carolina piedmont soil have also shown uptake and leaching to result in reductions in soil nutrients (Knoepp and Swank, 1994; Richter et al., 1994). Long-term, chronosequence studies have also been important for understanding changes in soil processes following disturbance such as agriculture. Those studies have typically found a reduction in nutrients and C on managed sites, and then a "recovery" following reversion of the sites to trees. The control or reference sites are typically "undisturbed" forests, with the assumption that the soil conditions are relatively constant. Studies at the Walker Branch and Coweeta watersheds as well as elsewhere (e.g., Haines and Cleveland, 1981) have demonstrated that soil chemical properties are not constant in undisturbed stands; accordingly, understanding the inherent temporal dynamics of soil conditions is fundamental to assessing long-term changes as a result of external stressors (e.g., management or climate change).
We present an assessment of nutrients and C in soils and vegetation within eight sites for a 21-yr period on the WBW. Our objective is to characterize the temporal changes in the distribution of nutrients and C within the sites and to consider the changes with respect to forest type, soils, and geomorphic position. An earlier study of the WBW established the potential for nutrient depletion in these undisturbed forests, and considered the implications of assessing long-term soil changes at the landscape level following 10 yr of observation (Johnson et al., 1988). This study reinforces the findings from the previous work on the watershed and offers evidence to suggest that interactions of soil moisture regime and deep rooting may affect interpretations of nutrient cycling based on characterization of the surface soils. We also confirm that parent material and geomorphic position are important in affecting soil properties with time (e.g., <20 yr).
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusions and perspectivesREFERENCES |
|---|
300 to 350 m elevation, has 14.5°C mean annual temperature, and receives 151 cm average annual precipitation of which 43–48% is estimated to undergo evapotranspiration (Henderson et al., 1978; Johnson et al., 1988). In 1967, 298 vegetation inventory plots were established on the WBW per the protocol of Harris et al. (1973). Twenty-four of those plots were selected in 1972 to represent four major forest types (pine, yellow poplar, oak–hickory, and chestnut oak) for an intensive study of nutrient cycling. As part of that study, standing biomass, forest floor, and soils were sampled from four to six subplots at each location (Henderson and Harris, 1975; Henderson et al., 1978). Eight of those plots have served as the basis for assessing long-term changes in soil chemical properties and nutrient cycling (Johnson et al., 1988), and were resampled for this study.
The eight long-term soil study plots, although originally selected on the basis of cover type, represent different soil and geomorphic settings that are characteristic of the watershed locally, and more generally of the ridge and valley province of the Appalachians (Table 1) . Three plots (Plots 179, 107, and 237) occur on soils formed in saprolite weathered from dolomite. They are characterized by a silt loam surface overlying a clay loam argillic horizon. The chert content is variable, ranging from 5 to 35%. Water infiltration is rapid, but perching may occur above the argillic. Two plots (Plots 98 and 281) represent the cove position that occurs near the base of the slope. The parent materials of the soils on these plots are derived from creep materials and underlying residuum. The soils are characterized by a thick, dark A horizon, tend to be moist, and are enriched by hydrologic inputs from up slope. Plot 26 is typified by saprolite weathered from dolomite, on the side slope of a doline. A doline is a collapse structure analogous to a sinkhole but without an open swallow hole. Plot 26 was originally selected to represent the pine-cover type, but because of mortality, the current vegetation is dominated by yellow poplar. Colluvial soils are typified by Plot 42. These soils are formed in deep colluvium derived from residual soils. The soils occur primarily in upper slope positions and may contain a restrictive fragic layer that can perch water. Chert content ranges from 15 to 50% by volume. Finally, soils that are formed in chert beds, 30 to 60% chert content by volume, are characterized by Plot 91. These soils occur in upper slope positions and tend to be dry due to the high coarse-fragment content.
|
Six 1-m2 subplots within the 12-m2 soil plot on each site were used for sampling. Subplots were randomly selected from the 12 by 12 m grid, excepting previously sampled subplots from the selection pool. Within each subplot, all wood of >2.5-cm diameter was collected. Wood of <2.5-cm diameter and the Oi and Oa horizons were sampled from a 0.25-m2 ring located in the center of each subplot. Oi litter was defined as recognizable by species, while Oa was more highly decomposed forest floor (Oe material was included in the Oa sample). Using a 7.5-cm diameter bucket auger, soil cores were collected from the center of the subplot in 15-cm increments to a depth of 60 cm. The 60-cm depth was selected originally on the basis that most roots were present above that depth.
Twigs, bark, and leaves were separated from Oi and Oa samples before drying. Wood, woody litter, and O horizon samples were oven dried to constant weight at 100°C, while soil samples were dried at 60°C. Soil was separated from the coarse fragments with a 2-mm sieve. Forest floor samples were ground in a Tecator Cyclotec sample mill (Tecator, Herndon, VA). All soil and forest floor samples were stored in archival containers.
Samples collected in 1993 and archived samples from 1972 and 1982 were analyzed as follows. Soils were analyzed for pH using 1:1 soil solution in both water and 0.01 M CaCl2. Carbon and N were determined using a Perkin Elmer 2400 CHN Analyzer (Perkin Elmer, Norwalk, CT). Exchangeable Ca2+, Mg2+, K+, Na+, and Al3+ were extracted with 1 M NH4Cl and analyzed by inductively coupled plasma spectroscopy (ICP). Extractable P was obtained using dilute acid (0.5 M HCl) F (Olsen and Dean, 1965). Forest floor samples were analyzed for total N and C using a Perkin-Elmer 2400 CHN Analyzer (Perkin Elmer). Cation concentrations in the forest floor samples were obtained from an ICP scan following ashing and solubilization. All laboratory analyses included blanks, standards, and 10% duplication. Concentration data from mineral soils were converted to content using bulk density that had be adjusted for coarse fragment content from each soil horizon (Peters et al., 1970). Forest floor concentration data was converted to content using the actual mass of the sampled volume.
Design and Analytical Considerations
The principal factors affecting temporal assessment of soil data are season, sampling protocols, and laboratory methods. A previous study on the WBW (Johnson et al., 1988) documented significant seasonal fluctuations in soil chemical properties, hence the need to normalize season when assessing long-term trends. Since the sampling of the Walker Branch long-term soil assessment plots has taken place during the same month in each of the sampling years, seasonal variation should be minimized. Certainly there is some intraseasonal variation, but it is likely to be much smaller than the interseasonal variation that reflects larger shifts in temperature and moisture, especially in deeper soil horizons. Variation due to sampling was also minimized by using protocols that were identical to those used in the previous study (Johnson et al., 1988). Finally, variation in laboratory methods and techniques may cause significant variation in the analysis of a soil sample. To minimize this factor, samples from 1972 and 1982 were reanalyzed according to current methods. Accordingly, the temporal changes found in chemical soil properties should reflect responses to inherent ecosystem processes.
This study is based on plots that were originally selected as a subset of inventory plots to characterize selected forest types. Accordingly, replication was not a consideration in the original work. For purposes of this study, and the previously reported work from 1982, individual plots were considered as the experimental units. Analyses were conducted using a two-way analysis of variance (ANOVA) with year and plot as the treatment factors. Comparisons among years for individual plots was done using unpaired t-tests. Tests of significance were conducted at P = 0.05.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusions and perspectivesREFERENCES |
|---|
Changes in Element Concentrations within the A and B Horizons
Surface and Subsurface Mineral Horizons
The concentration of Ca and Mg within the A horizon (0–15 cm) has been relatively stable during the 21-yr measurement period (Fig. 1)
. Some interdecade variation in base cation concentration is evident, but none of the surface mineral soil horizons exhibited either increasing or decreasing trends. Aluminum also did not exhibit significant changes. The content of cations within the A horizons does not appear to be strongly affected by soil type or physiographic position, although the cove soils (Plots 98 and 281) tended to have higher base cation content and lower Al than the other soils. In contrast, P concentrations were similar across soils, and most exhibited a reduction in P during the last 11-yr measurement period. The reduction in P was particularly pronounced on the saprolitic (Plots 107, 179, and 237) and the colluvial (Plot 42) soils, which experienced a 25 to 50% decline. Carbon and N tended to increase in the A horizon of all soils except for the saprolitic soils (Plots 107, 179, and 237), although the change was usually not statistically significant. The change in N relative to C was greater for the colluvium (Plot 42), chert (Plot 91), and doline (Plot 26) soils, resulting in a decrease in the C/N ratio. Conversely, the small change in N relative to C in the saprolitic soils resulted in an increase in C/N ratio. The cove soils (Plots 98 and 281) did not exhibit a change in C/N ratio.
|
|
|
|
|
Ecosystem Response
Carbon and nutrient pools for the forest floor and biomass of the eight plots in 1972, 1982, and 1993 are presented in Tables 2 and 3 , respectively. The net increment by ecosystem component (soil to a depth of 60 cm, forest floor, and biomass) during the last 11 yr is presented in Fig. 5
. Neither vegetation uptake nor biomass increment accounted for the reduction of Ca and Mg on cove soils (Plots 98 and 281) or Plot 237 of the saprolitic soils. In each of those soils, base cation loss was greater than accumulation in the vegetation or forest floor. The greatest positive increments in above-ground accumulation of Ca were on the saprolitic plots (Plots 107, 179, and 237). Plots 107 and 179, the ridge members in this group, had very low levels of extractable soil Ca, with the total pool significantly less than the net increment. The positive net increase in biomass Ca resulted in a net ecosystem gain for Ca on these plots. The colluvial site (Plot 42) was the only one to exhibit a net gain in Ca as a result of increases in extractable soil Ca. Sequestration in aboveground biomass did not appear to be a primary factor accounting for the decline in soil extractable P, except for Plot 107. As a result, five of the eight plots exhibited reductions in P during the last 11 yr.
|
|
37% of the C storage in these forested ecosystems, with the aboveground biomass comprising the majority. We recognize that these estimates do not include root biomass and that the soil would contain a greater proportion of C if root biomass were included. Similarly, if a greater soil volume were included (e.g., >60 cm), the soil C pool would increase. |
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusions and perspectivesREFERENCES |
|---|
Leaching is the primary mechanism attributed to reductions in soil Ca pools across a diverse set of Ultisols in the southeastern USA (Knoepp and Swank, 1994; Richter et al., 1994). Leaching was also attributed as the cause of Mg loss on the WBW following the first 10-yr assessment. Results from this study also suggest that leaching is an important cause of Ca loss. Sulfate is the primary anion associated with cation leaching in the southeastern USA (Johnson and Todd, 1990; Richter et al., 1994), although other anions and DOC may also be involved. While no leaching data were collected in this study, during the last 11-yr period the loss of soil Ca relative to fluxes into the forest floor or biomass was large for plots located in depressional settings (Plots 98, 237, and 281). Such low-lying topographic features function as a collection point for water, thereby increasing the potential for leaching. Leaching rates for Ca may range between 1 and 15 kg ha-1 yr-1 in southeastern Ultisols (Johnson and Todd, 1987; Richter et al., 1994). Accordingly, the losses measured in soils occurring in depressional positions greatly exceeded what would be expected. The only other site showing a net reduction in Ca was on the chert soil (Plot 91), and the cause was a reduction in biomass storage. Although the 45- to 60-cm depth layer showed a reduction in Ca, suggesting either leaching or uptake, the net soil increment (0–60 cm) was positive for plots representing the colluvium (Plot 42), chert (Plot 91), and one of the saprolitic soils (Plot 107). These data emphasize the need to consider adjacent soil layers within the primary rooting zone. The role of roots in Ca flux in these soils has not been considered, but it is likely to be an important factor. As previously noted, deep rooting is suspected as the mechanism for tapping deep subsoil sources of Ca. Correspondingly, retranslocation within the rooting zone may also be a means of affecting soil nutrient pools. Weathering is another mechanism for Ca release into the soil; however, the Ultisols on Walker Branch are highly weathered and weathering inputs are probably quite small (Johnson and Henderson, 1989).
There is little correspondence between the soil changes and the original forest types. Instead the soils on the ridge tops (Plots 91, 107, and 179) have lower levels than those in the mid-slope and depressional settings. The ridge soils are highly weathered, with no cation inputs other than atmospheric deposition. As a result, these soils showed little change during the last 11 yr. That response is expected to continue on those sites until the vegetation, with large Ca and Mg reserves, is recycled through the surface soil.
Acidification
Soil acidification usually occurs in agrading forests as a result of cation uptake and leaching losses (Binkley and Richter, 1987). Studying soil properties across 34 yr, covering conversion of an old field site to a loblolly pine (P. taeda L.) plantation, Markewitz et al. (1998) reported an increase in acidity that was primarily attributable to vegetation uptake, but that acid deposition was also an important factor, contributing an estimated 32% of the increased acidity. The Walker Branch soils did not show any consistent signs of acidification. The cove soils (Plots 98 and 281) did show a decline in pH and an increase in Al from the initial measurements in the B. The decline in pH is considered a result of cation leaching and Al release. Increased Al in the 45- to 60-cm B layer compared with the values from the initial measurements were also evident in Plots 26, 42, and 91, and has been attributed to the presence of polyhydroxy Al in the vermiculite clay of those soils (Johnson et al., 1988).
Phosphorus
Yanai (1992) developed a P budget for a northern hardwood forest where net P uptake by trees resulted in a net soil P decrease of 1.4 kg P ha-1 yr-1. Reduction of P was primarily in the mineral soil, although the forest floor had a greater quantity of available P. The cumulative reductions measured in the upper 60 cm of soil on WBW during the last 11-yr period were within the range expected, assuming a net uptake of 1.4 kg P ha-1 yr-1. However, P reductions in the soil and forest floor were not balanced by uptake in most of the stands on the WBW. Accordingly, there has been a net reduction in P representing a significant component of the extractable soil P. The reduction of soil P corresponded with those plots exhibiting changes in soil C. Yanai (1992) calculated that approximately one-third of the mineralizable soil P was in the organic fraction; accordingly, loss of that organic matter pool would further reduce potentially mineralizable P. Given that P leaching is minimal in these forest systems, the most probable explanation for the decline in soil extractable P is adsorption onto Fe and Al hydrous oxides. While we did not determine the fate of the P loss from the WBW plots, the net reduction in extractable soil P pools suggests the potential P limitations to forest growth in the future unless sources are available from deeper soil depths or weathering. Correspondingly, the high proportion of the total ecosystem P contained in the biomass reflects the dependence of these forests on internal recycling of P.
Nitrogen
The first 10-yr assessment of soil changes on the WBW reported a small decline in soil N in the B horizons (Johnson et al., 1988). We also found a reduction in N across most soils, and the loss for some was great (900–1300 kg ha-1 yr-1); however, despite the apparent large reductions, the reductions were not statistically significant except for Plots 26 and 237. For the other plots, within-plot variability was sufficiently large to preclude the detection of statistically significant differences. Few other studies have reported on the long-term changes of N in forest soils. Recently, a study on the Coweeta north and south reference watersheds also measured reductions in soil N during a 20-yr period (Knoepp and Swank, 1997). That study attributed the soil N loss to either leaching or transfers to soil pools below their 30-cm sampling depth. The reductions in soil N both in this study and the Coweeta study are contrary to many models that suggest a much more conservative soil N pool. For example, a model developed for Coweeta projected soil N losses of 4 kg ha-1 yr-1, which is much less than the actual measured soil changes (Knoepp and Swank, 1997). The reductions in soil N at Coweeta were more pronounced on the north reference than the south watershed, suggesting to the authors that landscape position (e.g., precipitation, soil moisture, and particulate matter) had a significant effect on soil N pools. In the WBW, position did have a large affect on N pool size, with the highest pools in the cove (Plots 98 and 281) soils. However, changes in soil N were evident in soils with northeast (Plot 237) and southwest (Plot 26) exposures. Accordingly, the landscape position effect is not simple or direct.
Carbon
Soil C levels were expected to be constant or increasing given the undisturbed condition of the forest. The reduction in soil C, albeit quite variable for some soils, was unexpected. Lower C levels could suggest: (i) potentially important changes in the C balance of the site or active C pools (e.g., forest floor or soil organic matter), (ii) the importance of temporal variability in soil C pools, or (iii) a change toward a new equilibrium level. However, the overall differences noted among years for the 0- to 60-cm soil depth most likely reflects the variability and difficulty in detecting differences with relatively few samples for most sites. For those plots exhibiting relatively large reductions (Plots 26, 98, 281, 179, and 237), either a change in the organic matter inputs (e.g., production) or outputs (e.g., decomposition) are the primary mechanisms controlling soil C. Production levels did not correspond with changes in soil C pools, that is, sites exhibiting both high and low aboveground C sequestration rates exhibited loss of soil C. Carbon leaching is a small flux compared with CO2 evolution (Johnson and Henderson, 1989); thus it is likely that decomposition is the primary factor affecting any reduction in soil C. Measurements from control stands on the southern reference watershed at Coweeta indicated declining soil C levels (Knoepp and Swank, 1997), but significant (>50%) interannual variation in soil C in A (0–10 cm) and B (10–30 cm) horizons effectively obscured definitive long-term trends or detection in net change in their 17-yr measurement period. The annual values exhibited a sinusoidal pattern, with peaks and troughs occurring within 3 to 5 yr in both horizons. Knoepp and Swank did measure a reduction in C in the A horizon in the northern reference forest during a 20-yr period, but due to the intraannual variability observed on the southern reference, they were uncertain about establishing a trend. Earlier work by Haines and Cleveland (1981) also reported significant seasonal variation in soil organic matter for several forest types. Testing a model on the development of soil C pools, Knoepp and Swank (1997) constructed a 70-yr chronosequence, originating with a clear-cut forest condition. They found that soil C pools would increase and then decline with time. A similar response was also reported by Snyder and Harter (1984) in northern hardwoods. However, other patterns have also been observed. Mattson and Smith (1993) reported that soil C pools did not change following cutting. In another study on the Oak Ridge Reservation adjacent to the WBW, Johnson and Todd (1998) reported net increases in soil C after 15 yr on sites that had been whole-tree and sawlog-only harvested, and the reference site for that study also exhibited a net soil C gain (11 Mg C ha-1) in the upper 45 cm after 15 yr (1980–1995). That gain on a mixed oak site occurred on soils similar to those in this study. The differing response between these two sites on the Oak Ridge Reservation may be due to differences in slope, aspect, stand age, and development, as well as the sampling time and interval.
Other studies on long-term soil C changes have reported significant temporal changes that may be attributable to land management practices, vegetation type, temperature, precipitation, and soil physical and chemical properties. However, identifying the causative factor is often difficult because not all of the factors have been considered during the course of the study. While the role of temperature and soil moisture in regulating organic matter turnover in the short-term is well established (Olson, 1963), studies using long-term data have not attempted to consider the cumulative effect of the temperature and moisture factors on long-term soil C changes. Nor have studies of changes in soil C pools considered which organic matter fractions are contributing to the loss or change. Most organic matter fractionation schemes recognize labile and recalcitrant pools. The recalcitrant pools are considered to be "stable", with a turnover time typically measured in 100 to 1000 yr. Accordingly, the change in soil C measured in this study and others should reflect the change in the labile organic matter pool(s). If verified, seasonal and interannual variation in soil C pools would be expected as a result of changes in both biotic and abiotic factors that control organic matter turnover. Another limitation in this study, and others assessing the effects of vegetation on soil C pools, is that belowground biomass production is not measured, thereby complicating the assessment of the sources of organic matter for the soil C pool. Other factors known to affect soil C pools include acidic precipitation, where competition for anion exchange sites between SO-4 and DOC result in a reduction in soil C (Moore, 1997). However, mineralization, rather than leaching or changes in inputs, is probably the primary cause of reductions in soil C pools. Himes (1997) suggested that changes in soil C pools should induce changes in N and P pools corresponding with sequestration or mineralization. For most plots in this study, there was relative correspondence to the C/N (14:1) and C/P (70:1) ratios expected in soils when there was a reduction in soil C. On Plot 91, which exhibited a gain in soil C, there was great deviation from the reference norm, with a C/N gain of 7 and a C/P gain of 15.
This study and the relatively few other long-term studies on forest soil C pools demonstrate the variable and dynamic nature of the pools. As a result, there is danger in presuming whether mineralization or sequestration is occurring at any given point in time. It is well established that conversion of wildland soils to cultivation results in significant reductions in soil C pools (Mann et al., 1988). Similarly, reforestation of agricultural or disturbed sites is presumed to result in soil C gains. However, C pools in a mature forest soil must have some dynamic equilibrium range that is controlled by physical and chemical soil properties, biomass production and allocation, nutrients, and ambient conditions. The data from Walker Branch and Coweeta watersheds suggest that range may be quite large, and subject to both short- and long-term variation. Recognizing the dynamic nature of soil C pools in what appears to be a stable forest environment highlights the need for caution when assessing soil C sequestration potentials, modeling soil C fluxes, or assessing the role of soils in long-term C sequestration.
|
Conclusions and perspectives |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusions and perspectivesREFERENCES |
|---|
A major finding after the first 10 yr of this study was the potential for cation depletion in these soils, particularly if vegetation were removed from the site (Johnson et al., 1988). That work prompted considerable interest and concern regarding the potential effects of whole-tree harvesting on sustainable nutrient supplies. As a result, numerous studies have been conducted confirming the potential for productivity decline as a result of nutrient removal in biomass. However, a major limitation of most of the forest soil assessments is that usually only the upper 50 to 100 cm of soil are considered. This practice is rationalized by the fact that the largest proportion of the roots is found near the surface. Interestingly, one of the important distinctions that forest soil scientists make about forest soils is that the trees are long lived and deep rooted, thereby affecting soil processes. While inferred, the effects of deep rooting on nutrient cycling have not been extensively studied. Our results suggest that deep rooting is a major factor controlling cation uptake in mature forests: although the upper 60 cm showed cation loss, productivity was sustained, apparently by deep soil pools. Similarly, cation cycling through deep rooting is presumably responsible for maintaining the cation balance on those sites not exhibiting cation loss. The capacity of deep soil supplies to satisfy present nutrient cycling demands has not been widely recognized. Both the role of rooting and deep soil pools need to be considered in future nutrient cycling studies.
The original nutrient cycling study was designed to consider differences among cover types within the WBW. As a result of natural stand development and succession, the cover types have changed during the past 21 yr, reflecting the ephemeral nature of forest types as the basis for identifying sites. Site classification systems that integrate soils, understory, and overstory have been used effectively to identify sites or ecosystem units that have similar functions. Our work reemphasizes the inadequacy of forest cover type as a common denominator for assessing nutrient cycling or other soil processes. Although, cover type and species distribution do tend to correspond with soils (Lietzke, 1994), reflecting both nutrient and moisture regimes.
Agrading forests are considered to be net C sinks, and mature forests are usually in equilibrium. Our results indicate that forests may be a net source of C when net aboveground productivity is low. Loss of soil C, the factor causing the net release of C from the ecosystem, is presumed to be primarily a result of decomposition. Temporal variability in soil C pools is known to be significant. Whether the results from three measurement points in this study during a 21-yr period represent a consistent trend is uncertain. Accordingly, considerably more work is needed to assess long-term soil C pools, organic matter fractions, and factors controlling organic matter dynamics in mineral soil horizons.
Originally, three soils and four cover types were recognized for this long-term study of nutrient cycling, suggesting a relatively simple system. The cumulative findings for the 21-yr assessment period highlight the complexity of this upland landscape and the nuances in the interactions of vegetation, soils, and geomorphic position. These results further highlight the danger in extrapolating findings from one site or study to other sites or regions. Within the WBW, sites varied from a net source to a sink for cations, C, P, and N. Accordingly, any generalization derived form one particular site could be wrong for another site on the watershed. Unfortunately, we have not been able to directly determine the causes of changes in nutrient and C pools; that understanding would be useful in developing a more generalizable approach for thoroughly assessing long-term nutrient cycling. However, our consideration of the causative factors is well grounded in the literature and research on the WBW. Hence, this work should provide a basis for testing ideas and models regarding sustainable productivity and nutrient cycling.Falkenggren-Grerup Eriksson 1990; Falkenggren-Grerup Tyler 1992; Hallb1.gif" BORDER="0">cken Tamm 1986
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 8, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusions and perspectivesREFERENCES |
|---|
cken L., Tamm C.O. Changes in soil acidity from 1927 to 1982–84 in a forest area of south-western Sweden. Scand. J. For. Res. 1986;1:219-232.This article has been cited by other articles:
|
|
 |
|
  D. W. Johnson, D. E. Todd Jr., C. F. Trettin, and J. S. Sedinger Soil Carbon and Nitrogen Changes in Forests of Walker Branch Watershed, 1972 to 2004 Soil Sci. Soc. Am. J., August 27, 2007; 71(5): 1639 - 1646. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. H. Leggett and D. L. Kelting Fertilization Effects on Carbon Pools in Loblolly Pine Plantations on Two Upland Sites Soil Sci. Soc. Am. J., January 6, 2006; 70(1): 279 - 286. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Courchesne, B. Cote, J. W. Fyles, W. H. Hendershot, P. M. Biron, A. G. Roy, and M.-C. Turmel Recent Changes in Soil Chemistry in a Forested Ecosystem of Southern Quebec, Canada Soil Sci. Soc. Am. J., June 28, 2005; 69(4): 1298 - 1313. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Kelly and P. A. Mays SOIL CARBON CHANGES AFTER 26 YEARS IN A CUMBERLAND PLATEAU HARDWOOD FOREST Soil Sci. Soc. Am. J., April 11, 2005; 69(3): 691 - 694. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. S. Onsoy, T. Harter, T. R. Ginn, and W. R. Horwath Spatial Variability and Transport of Nitrate in a Deep Alluvial Vadose Zone Vadose Zone J., February 1, 2005; 4(1): 41 - 54. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. W. Johnson, D. E. Todd Jr., and V. R. Tolbert Changes in Ecosystem Carbon and Nitrogen in a Loblolly Pine Plantation over the First 18 Years Soil Sci. Soc. Am. J., September 1, 2003; 67(5): 1594 - 1601. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P.S. Homann, B.T. Bormann, and J.R. Boyle Detecting Treatment Differences in Soil Carbon and Nitrogen Resulting from Forest Manipulations Soil Sci. Soc. Am. J., March 1, 2001; 65(2): 463 - 469. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||
