SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Miller, J. O.
Right arrow Articles by Daniels, W. L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Miller, J. O.
Right arrow Articles by Daniels, W. L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Miller, J. O.
Right arrow Articles by Daniels, W. L.
Related Collections
Right arrow Carbon Sequestration
Right arrow Data Management
Right arrow Soil Models
Published in Soil Sci. Soc. Am. J. 68:194-203 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

DIVISION S-5—PEDOLOGY

Soil Organic Carbon Content in Frigid Southern Appalachian Mountain Soils

Jarrod O. Millera, John M. Galbraith*,b and W. Lee Danielsc

a Research Assistant, Crop & Soil Environmental Sciences Dep., Virginia Polytechnic Institute and State Univ., B44N Smyth Hall (0404), Blacksburg, VA 24061
b Assistant Professor, Crop & Soil Environmental Sciences Dep., Virginia Polytechnic Institute and State Univ., 239 Smyth Hall (0404), Blacksburg, VA 24061
c Professor, Crop & Soil Environmental Sciences Dep., Virginia Polytechnic Institute and State Univ., 244 Smyth Hall (0404), Blacksburg, VA 24061

* Corresponding author (ttcf{at}vt.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Limited information is available on soil organic C (SOC) sequestered in frigid soils of the southern Appalachians, although the soils in frigid areas probably hold more SOC than warmer soils in the region. This study determined SOC in high-elevation (>1300 m) soils across northeast (N) and southwest (S) aspects and across three slope classes (7–15, 15–35, and 35–55%) per aspect in the Ridge and Valley of southwest Virginia. Overall, rock fragment content and bulk densities were lower, B-horizon SOC was higher, and sola were shallower than reported in similar regional studies. The A horizons were thicker on N versus S aspects, as were the sola (57 vs. 47 cm). Sola were 50% thicker on N35 to N55% sites than on S35 to S55% sites (65 vs. 44 cm). The S35 to S55% sites had the thickest litter but the thinnest A horizons, and were characterized by the highest C/N ratio and least incorporation of leaf litter. The study area average mass SOC was 112 Mg ha–1. More mass SOC was retained on S than N aspects in litter layers but less was retained in the A horizons and whole solum (99 vs. 127 Mg ha–1). There was less solum mass SOC on S7 to S15% and S35 to S55% sites than on N15 to N35% or N35 to N55% sites. The results reinforce the importance of using mass SOC rather than SOC concentration in regional SOC studies and demonstrate that steep northeast-facing slopes in frigid Appalachian landscapes have the highest mass SOC and highest potential for sequestering organic C in the soil.

Abbreviations: ET, evapotranspiration • MLRA, Major Land Resource Area • N, Northeast • S, southwest • SOC, soil organic carbon


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
SOIL C makes up the largest component of active global C stores, being twice as large as atmospheric CO2 and four times as large as biomass C reserves (Brady and Weil, 2000). Accurate quantification of soil C is necessary for detection and prediction of change in response to changing global climates. High elevation soils in the southern Appalachian Mountains have significant SOC content but have not been studied intensively and do not have a detailed soil survey because the difficult terrain and rocky soils limited their use for production agriculture. Most of the soils formed from sedimentary rock at the highest peaks in the southern Appalachians are forested and used for timber, wildlife, and watershed management. The forests on these high ridges contain northern hardwoods and conifers, resembling cool, moist, temperate forests of the northeastern USA (Shanks, 1954; Korstian, 1937; Oosting and Billings, 1951). The highest peaks in southwest Virginia have a frigid soil temperature regime, where mean annual soil temperatures are below 8°C (Lietzke and McGuire, 1987). The cool temperatures at high elevations enhance the effectiveness of the precipitation and decrease decomposition of organic material, resulting in increased SOC (Jenny et al., 1949). Witkamp (1963) found that temperature limited decomposition rates at high elevations in the humid east more than lack of precipitation. Soil organic C may also increase with elevation in the southern Appalachians because of low temperatures that slow decomposition rates (Bolstad and Vose, 2001) and increased incorporation by soil fauna (Daniels et al., 1987).

Variability of SOC in response to change in aspect and slope has been studied in the warmer mesic soil temperature regime, where SOC is higher on north than on south aspects because of the drying and heating effects of increased exposure to solar radiation on south aspects (Frank and Lee, 1966). Daniels et al. (1987) reported that A horizon organic matter content and depths were higher on north-facing slopes in the Great Smoky Mountains in soils formed from metamorphic and metasedimentary rock. Similar results were found in the other southern Appalachian studies where total SOC was greater or A horizons were darker in color on north aspects (McCracken et al., 1962; Sartz and Huttinger, 1950; Franzmeier et al., 1969). Similar findings were also reported for soils on sedimentary rock in the Appalachian Plateau of Ohio (Finney et al., 1961). Moisture effectiveness has been found to be a dominant factor in SOC sequestration, as SOC concentration is highest in concave or bowl-shaped areas (coves) along drainages that have moister conditions than convex areas (Bolstad and Vose, 2001). No studies are documented that test the effect of slope angle or aspect-slope interactions on SOC in high-elevation frigid soil temperature regime areas with sedimentary rock parent material in the southern Appalachians. The interactions of aspect and slope can mask differences due to aspect or slope alone because of their offsetting effects on effectiveness of precipitation. For example, a steep southwestern slope with high evapotranspiration (ET) and gentle slope (low runoff) may have the same effective moisture conditions as a steeper (higher runoff) northeastern slope that has lower ET.

Previous studies have estimated varying amounts of mass SOC in the southern Appalachians. Post et al. (1982) estimated a simple average of 121 Mg C ha–1 for overall cool, moist temperate forests in the region. Using NRCS National Soil Survey Laboratory Pedon Database, Kern (1994)(p. 450, Fig. 6) computed 61 to 75 Mg C ha–1 for the Appalachian and Blue Ridge Mountains and from 76 to 90 Mg C ha–1 for the remainder of Virginia based on the composition of Soil Taxonomy great groups within Major Land Resource Area (MLRA) spatial layers (Soil Survey Staff, 1990). Kern calculated SOC to 1-m depth but did not correct for rock fragment volume. Daniels et al. (1987) found up to 280 Mg C ha–1 in a mesic soil temperature regime southern Appalachian forest in North Carolina with deeper soils that had been protected from logging.

The effect of aspect and slope are less studied at cold, moist, high elevations because the overall moist climate may negate the large effects on vegetation and soils seen in drier climates, and because intensive land-use is limited. The slope percentage at which aspect becomes an important factor in SOC sequestration has not been studied in the high, cold forests in the southern Appalachian Mountains. The objectives of this study were to measure SOC on two aspects and three slope classes to determine the slope or slope–aspect interaction that describes the sites with highest potential SOC sequestration, and to determine at what percentage of slope the effect of aspect becomes important in SOC sequestration. The results should enable more accurate prediction of SOC resources in similar areas and identify the site conditions with highest potential for SOC sequestration.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Area
The study area is located in MLRA 128, the Southern Appalachian Ridge and Valley Province (Soil Conservation Service, 1981) near Burkes Garden, in Tazewell County, Virginia (Fig. 1) on a broad ridge and saddle plateau that is representative of high elevation peaks in the region. A majority of the study area lies within the designated Beartown Wilderness Area. The study area is part of the cool, temperate, moist forested life zone (Post et al., 1982) and is centered at 37° 4' 48'' N and 81° 25' 48'' W on the Hutchinson Rock 7.5' USGS quadrangle. The mean annual air temperature is 7.2°C, and average annual precipitation is 1220 mm at a weather station in an adjacent valley that is approximately 300 m lower. The study area ranges from 1200 to 1450 m and is higher than the estimated boundary of the frigid and mesic soil temperature regimes at this latitude (Mount et al., 1999). Silurian-aged, highly siliceous sandstone-conglomerate of the Tuscarora formation, along with acid sandstone and siltstone of the Juniata formation outcrop on the local ridge tops. A preliminary but uncorrelated USDA-NRCS soil survey map indicates that the dominant soils are frigid taxajuncts of mesic Hapludults and Dystrudepts. The mapped soils ranged from shallow to deep and from excessively to well drained on slopes of 2 to 75%. Shallow, excessively drained soils occurred on the steepest slopes.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 1. Frigid soil temperature regime areas of the high elevation mountains of southwest Virginia area in dark gray, with the study area outlined in black. The star in the inset shows the study area in relation to the eastern USA.

 
The native vegetation (Table 1) consists dominantly of sugar maple (Acer saccharum Marsh), red maple (Acer rubrum L.), American beech [Fagus grandifolia (Ehrh.) Little], northern red oak (Quercus rubra L.), and yellow birch (Betula alleghaniensis Britt.), with red spruce (Picea rubens Sarg.) appearing at the highest elevations (Strasbaugh and Core, 1977; Kinser, 1982; Lietzke and McGuire, 1987; Woods and Shanks, 1959). Over twenty million board feet of red spruce lumber were transported from the center of the study area between 1938 and 1942, but there has been no logging since and the area is currently unmanaged forest (Buchanan, 1982). Mixed species such as the sugar maple are commonly found on north slopes, in coves and other areas of higher moisture, while oak and pine are more common on drier sites (Mowbray and Oosting, 1968: Whittaker, 1952: Braun, 1942).


View this table:
[in this window]
[in a new window]
  		Table 1. Listing of tree species with scientific names, count of sites where they are present, and trend of occurrence between aspect–slope classes (N, northwest and S, southeast).

 
Collection of Data
Soil organic C was determined for aspect–slope classes that included N (340–90°) and S (160–270°) aspects with three slope classes 7 to 15, 15 to 35, and 35 to 55% on each aspect. Ten sites were described and sampled in areas that represented each aspect–slope class. Pixels that represented 30 x 30 m ground area for the six aspect–slope classes were identified using a level-one 30 m digital elevation model (DEM) and the DEMAT extension (downloaded from the Environmental Systems Research Institute (ESRI) website, available at http://www.esri.com) to ArcView 3.2a (ESRI, Redlands, CA) using the Horn method for rough surfaces. Clusters of three or more contiguous pixels of the same aspect–slope class were considered as potential sampling sites. Random numbers were generated to select and order 40 clusters as potential sampling sites per aspect–slope class. Each site was visited in order and the first 10 sites per aspect–slope class that met all selection criteria were sampled (Fig. 2) . The selection criteria precluded sampling in dense rhododendron stands, extremely stony or bouldery rockslides, very sandy soils, and areas that appeared highly disturbed by anthropogenic, faunal, or floral activity. These areas were not sampled to make sure that variability within SOC was due to aspect and slope differences and not due to other obvious confounding factors. The center of each sampling site was located using a handheld global positioning systems (GPS) unit, and the slope class and aspect verified before sampling.



View larger version (139K):
[in this window]
[in a new window]
  		Fig. 2. Location of the 60 sampling sites in the study area by aspect–slope class. Background is the Hutchinson Rock 7.5' USGS topographic quadrangle. The study area is demarcated by the 1200-m elevation contour.

 
Soil and site descriptions were made at each sample site between June and July 2001, and soils were most closely matched to existing soil series. Samples were taken in the (i) litter layer, (ii) A horizon, and (iii) B horizons to a depth of 1 m or bedrock contact. The upper 5 cm of the A horizon was sampled separately from the lower portion of the A horizon. The litter layer was collected within a 0.25-m2 quadrat by successively removing the Oi, Oe, and Oa horizons. The top 5 cm of the mineral soil of each pedon was collected next and then any remaining A horizon was sampled. The rest of the mineral soil profile was sampled by auger and a mixed sample of 300 g of soil was taken from the auger sample of the B horizons. Rock fragment volume was estimated in each horizon. Bulk density (Mg m–3) was measured on all mineral horizons at two selected sites per aspect–slope class using the core method with the following equation from Blake and Hartage (1986):

[1]
where ODS is the oven dry soil weight (g), V is the volume of the cylinder (cm3), RW is the rock weight (g), and RD is the rock density (Mg m–3). Bulk density (Mg m–3) was estimated for each litter layer at two selected sites per aspect-slope class using the frame method with the following equation modified from Eq. [1]:

[2]
where ODS is the oven dry (60°C) litter weight (g), A is the area of the quadrat (cm2), T is the thickness of the layer (cm), RW is the rock weight (g), and RD is the rock density (Mg m–3).

Solar irradiation and insolation of sites were estimated from tables for latitudes between 30° and 50° using latitude, slope angle and aspect (Frank and Lee, 1966).

Lab Analyses
Soil samples were ground and passed through a 2-mm sieve. Leaf litter was ground into particles <5 mm. The pH was measured with a pH meter in a 1:1 (w/w) soil to H2O mixture. Total C (g kg–1) and total N (g kg–1) were measured by combustion with an Elementar CNS analyzer (Elementar America, Inc. Mt. Laurel, NJ) according to methods described by Nelson and Sommers (1996). There were no carbonates present in any samples, so total SOC concentration was equated to SOC concentration, converted to mass Mg C ha–1 for each layer or horizon using the following formula from Bliss et al. (1995):

[3]
where OC is organic C concentration (g C kg–1 soil), Db is the bulk density (Mg m–3), T is the sampled thickness of the layer or horizon (cm), R is the rock fragment volume (%), and UCF is a unit conversion factor (0.001 kg g–1 x 0.01 m cm–1 x 10000 m2 ha–1). The mass SOC was determined for the litter layers, the A horizon, the B horizons, and the upper 1 m of soil or to bedrock if that was <1 m. Simple average mass SOC per aspect–slope class and interactions were calculated directly and not weighted by aerial extent.

Statistical Analyses
The error control design used was the completely randomized design (CRD), structured by a factorial (Lentner and Bishop, 1993). The first factor was aspect, with two levels (n = 30 for each N and S). The second factor was slope, with three levels (n = 20 for each 7–15, 15–35, and 35–55%). The six total aspect–slope classes were identified as N7, N15, N35, S7, S15, and S35. Parameter distributions were tested for normality to examine any skewing of values. Statistical Analysis Systems (SAS, Cary, NC) software was used to analyze the factorial and test for interactions, and then Fisher's Protected Least Significant Difference (LSD) test (Lentner and Bishop, 1993) was used to separate differences across aspects and slopes when F-tests were p ≤ 0.05. Pairwise comparisons were also made in SAS using the "pdiff" function to examine differences between aspect–slope classes. Bulk density values were compared between aspect–slope classes in Minitab 12.0 (Minitab, Inc., State College, PA) using the non-parametric Mann-Whitney two-sample t-test at the 90% confidence interval. Bulk density values were compared between A and B horizons (n = 12 each) in Minitab using a non-parametric two-sample sign rank test at the 90% confidence interval.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Vegetation
The native vegetation consists dominantly of sugar maple, red maple, American beech, northern red oak, and yellow birch, with red spruce appearing at the highest elevations (Strasbaugh and Core, 1977; Kinser, 1982; Lietzke and McGuire, 1987; Woods and Shanks, 1959). The sugar maple increased in frequency with slope on N aspects but decreased with slope on S aspects, while the oak and red spruce were opposite in distribution. Ohio buckeye (Aesculus glabra Willd.) was found on 7 of 10 N35 sites but not on any other sites. Yellow birch was more common on S aspects and decreased in frequency with increasing slope angle on both aspects. These results agree with those of Mowbray and Oosting (1968), Whittaker (1952), and Braun (1942) who reported that mesophytic species are commonly found on north slopes and other areas of higher moisture, while xerophytic species are more common on southwest-facing sites. The N aspects were dominantly northern red oak/Ohio buckeye, beech, and sugar maple. The S aspects were mainly red spruce, northern red oak, beech, red maple, and yellow birch.

The aboveground and root biomass and net primary productivity was not directly measured at the sampled areas, but Whittaker (1966) reported that production was higher on S than N aspect birch forest (906 vs. 668 g m–2 yr–1) in high-elevation southern Appalachian forests above 1400 m elevation in the Great Smoky Mountains. However, Whittaker reported that spruce-fir forests were more productive on N than S aspects (1024 vs. 944 g m–2 yr–1). There were no fir trees in our study, and spruce was only found in abundance on S35 sites.

Rock Fragment Volume
Rock fragment volume was lower than expected; most soils (43 of 60) had <10% rock fragments in any horizon. There were no rock fragments in the litter layers. There were no significant differences in rock fragment volume by aspect, slope, or aspect–slope classes in either mineral horizon or the whole mineral soil (Table 2). A few soils (17 of 60) had B horizons with 10 to 60% rock fragments, but only three of the soils had an average of >35% rock fragments throughout the B horizons (control section). These results are similar to studies done on undisturbed forest soils in western North Carolina (Daniels et al., 1987; Jones, 2000).


View this table:
[in this window]
[in a new window]
  		Table 2. Descriptive statistics for estimated volume percentage of rock fragments by horizon and weighted average for the mineral soil.

 
Bulk Density
Bulk density values were not different between aspect–slope classes, based on analysis of limited samples. Bulk density was uniformly low in litter layers and mineral horizons regardless of aspect or slope (Table 3). The bulk density values weighted by litter layer component thickness seem very low. The Oi horizons had very low bulk densities (0.10 Mg m–3, n = 60, SD = 0.06) because they were composed primarily of hardwood leaves, were very thin (<1 cm thick), and were sampled in June and July, about 6 mo after leaf fall. The average bulk density of the A horizons (0.72 Mg m–3, n = 12, SD = 0.21) was lower (p < 0.10) than the B horizons (1.1 Mg m–3, n = 12, and SD = 0.10).


View this table:
[in this window]
[in a new window]
  		Table 3. Descriptive statistics for bulk density by horizon.

 
The difference between A and B horizons and the relatively low values show that mass SOC prediction models that use standard bulk density values of 1.5 Mg m–3 for mineral soil horizons will report two times higher mass SOC in the A horizons and 1.4 times higher mass SOC in the B horizons than found here, and will not show the density difference between A and B horizons. Furthermore, models that use NRCS soil series NASIS database values for bulk density (Soil Survey Division, 2002) will also err because those estimates are also higher for frigid soil temperature regime soil series in southern Appalachian Mountains than the values we found. NRCS databases have estimated bulk density values for each soil series and layer. Using NRCS data for models of mass SOC estimation in frigid soil temperature regime, high-elevation southern Appalachians is therefore problematic because many of the areas are not mapped or are mapped with soil series that have a warmer mesic soil temperature regime and thus lower concentration of SOC and higher bulk density based on typical pedons from lower elevations. Bulk density values for the A horizons were similar to those found by Daniels et al. (1987), but B horizon bulk densities were lower because our soils had higher concentration of SOC in the B horizons. Litter and A horizon densities were similar to those reported by Jenkins (2002), who used the frame method to measure bulk density.

Layer and Horizon Thickness
The soils sampled in this study were shallower than those mapped by the preliminary NRCS soil survey. The average solum depth was 52 cm, and all sola were either shallow or moderately deep (Table 4). There were no differences in thickness of litter or B horizons by aspect class. However, the A horizons were thicker (p = 0.001) on N than S aspects (13.1 vs. 8.0 cm) because the N aspects, presumably due to higher soil moisture contents (Finney et al., 1961; Franzmeier et al., 1969; Mowbray and Oosting, 1968) producing more suitable conditions for microbes and detritivores that mix litter and soil materials together (Witkamp, 1963, 1966). Mowbray and Oosting (1968) reported that moister soils are also more productive sites for overall vegetation growth and litter production than drier sites, although Whittaker (1966) reported that production was higher on S than N aspect birch forest. Regardless of production differences, the moister N aspects had more SOC incorporated over time than the S aspects, resulting in thicker A horizon development. The sola were also thicker (p = 0.01) on N than S aspects (56.8 vs. 46.8 cm) because the moister N aspects would have higher leaching rates and probably less erosion than the drier S aspect sites.


View this table:
[in this window]
[in a new window]
  		Table 4. Descriptive statistics for horizon thickness and for the whole solum.

 
Slope alone had no significant effect on litter layer or horizon thickness (Table 4), but aspect x slope interactions produced significant effects. The sola on N35 sites were thickest (p = 0.05) because they had the thickest A and B horizons, and were 50% thicker than those on S35 sites. In contrast, the sola on S35 sites were thinnest (p = 0.05) primarily because the A horizons on S35 sites were one-third to one-half as thick as on other sites. The soils on S35 sites also had the thickest litter layers, because the S35 sites had a drier effective climate and the litter was not undergoing as complete decomposition by microbes or incorporation into the mineral soil by geophages. The litter on the drier S35 sites was likely higher in lignin and tannins based on tree species composition, and therefore more resistant to decomposition (Witkamp, 1963, 1966; Bal, 1970) because the vegetation was composed of more spruce and oak than the sugar maple found to dominate on the moister N aspects (Table 1). Decreased leaf litter decomposition may be due to higher C/N ratio in the Oi horizon on S35 than on any other aspect–slope class (Table 5), indicating that a higher net C input was present on S35 sites. The reduced decomposition and incorporation on the S15 and S35 sites would lead to longer residence times and accumulate to a greater thickness (Bal, 1970).


View this table:
[in this window]
[in a new window]
  		Table 5. Descriptive statistics for C/N ratio for the Oi horizon.

 
Carbon Concentrations
The average concentration of SOC in the sampled soil litter and horizon components and the whole solum (weighted average by component thickness) for all slopes and aspects is shown in Fig. 3 . There was a characteristic decline in concentration of SOC from the surface down, illustrated by comparison of concentration of SOC in each sampled component and the sola. The weighted average concentration of SOC of the three litter layers was 370 g C kg–1 soil (SD = 72, n = 60). The weighted average concentration of SOC of the two A horizon components was 70 g C kg–1 soil (SD = 20, n = 60), which was similar to studies of undisturbed forest soils in western North Carolina (Daniels et al., 1987; Jones, 2000). Concentration of SOC in B horizons was 17 g C kg–1 soil (SD = 11, n = 60), which was greater than in warmer mesic temperature regime soils reported by Daniels et al. (1987), Franzmeier et al. (1969), and Finney et al. (1961).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 3. Dark gray bars show the concentration of SOC of the solum components and whole solum (weighted average by component thickness), with standard deviation in white bars. The number of samples was 60 for all but the 13 Oa horizons and 31 lower A horizons.

 
Concentration of SOC did not vary by aspect or slope (Table 6), which is similar to other studies done in the southern Appalachians (Finney et al., 1961; Franzmeier et al., 1969; Daniels et al., 1987; Jones, 2000). Litter SOC concentration was higher on S7 than S35 sites (Table 6) because the Oe horizons on S7 were thinner than on S35 sites. The Oe horizons had lower concentration of SOC than the other litter horizons, so concentration of SOC was inversely related to the proportion of Oe horizon in the total litter.


View this table:
[in this window]
[in a new window]
  		Table 6. Descriptive statistics for concentration of C by horizon and for the whole solum. Litter and whole soil weighted by component thickness.

 
Carbon Mass
Overall, the simple average of mass SOC for the sampled well-drained soils in the study area was 112 Mg C ha–1. In comparison, Post et al. (1982) estimated 121 Mg C ha–1 for overall cool, moist temperate forests.

The A horizon thickness was a major factor in mass SOC differences (Tables 4 and 7) in this study. There was more Mg C ha–1 on S aspects in the litter layers (p = 0.05) because the litter layers were slightly thicker, especially on S35 sites where dense Oe horizons were found (Table 4 and 7). There was less mass SOC in the A horizons (p = 0.001) and the whole solum weighted average (p = 0.05) on south aspects was lower because of thinner A horizons and total depths. There was no difference in mass SOC by slope class.


View this table:
[in this window]
[in a new window]
  		Table 7. Descriptive statistics for Mg C ha–1 by horizon and for the whole solum. All soils were <1 m to bedrock.

 
Table 7 illustrates the differences in SOC between the aspect–slope classes. There was less mass SOC on the S7 and S35 sites than on N15 or N35 sites because the S7 and S35 sites had the thinnest surface horizons and total solum depths. The S15 sites were on central or lower backslope positions and had more SOC in litter (p = 0.1) and the A horizon than the other S aspect sites. Erosion potential may have been greater on the steep, dry areas such as the S35 sites than on the steep, moist N aspects because the drier S35 sites were probably more susceptible to litter removal from fires than the other sites and appeared to have less canopy cover to protect the soil surface from erosion. The litter on the S35 sites had the highest C/N ratio and probably the lowest decomposition rate, likely resulting in the longest residence time.

Soil organic C increased regularly with slope on N aspects (Table 7) as solar insolation decreased. The N15 and N35 sites would have received the least solar radiation of all sites based on estimated values for the 38° N parallel from tables by Frank and Lee (1966) and would have had the coolest, moistest effective climate. The climate on steep N aspects would be more conducive to higher vegetative inputs, greater litter incorporation, and lower erosion rates than the south aspect sites. Microbial decomposition rates slowed by lower temperatures would lead to decreased C evolution as CO2 and thus a buildup of SOC. Incorporation into the soil by earthworms and arthropods will stabilize organic matter by bonding it with mineral soil materials, also decreasing decomposition (Johnson, 1990; Wolters, 2000).

Soil Classification and Soil Organic Carbon Distribution
Thirty-one of the 60 pedons (52%) were classified as Ultisols and the rest as Inceptisols. Carbon was >0.6% in all surface horizons, and six pedons on N15 or N35 aspect-slope classes had thick and dark enough A horizons to be classified as umbric epipedons. The average B horizon SOC concentration was over 0.9 g kg–1 in 27 of the Ultisols so those were classified as Humults. The other four Ultisol pedons would likely have had >0.9 g C kg–1 in the upper 15 cm of the argillic horizon and qualified as Humults also if that part of the argillic had been sampled separately. Three pedons classified as Dystrudepts in the N15 aspect–slope class would have qualified as Umbrepts if that suborder still existed. Dominance of Umbric epipedons and Haplumbrepts has been reported on northern aspects at high elevations in the Great Smoky Mountains, part of the southern Appalachian Mountains southwest of the study area (White et al., 1993). White also reported that Pachic Haplumbrepts (now Humic Dystrudepts) occurred on northern aspects with >60% slopes, and that Umbric Dystrochrpets (now Humic Dystrudepts) occurred on southern aspects and areas that received repeated burning (Soil Survey Staff, 1990, 1999).

There were about twice as many pedons with argillic horizons (Ultisols) than without (Inceptisols) on N aspects, the opposite of what was found on S aspect classes (Table 7). These differences were consistent within aspect class regardless of slope. The N aspects were moister and this may have led to increased clay illuviation and argillic horizon formation. Even though the N aspect had higher mass SOC than the southwest, the four pedons that failed to make Humults suborder were all found on N aspects. There was no clear trend between slope class alone and argillic horizon presence. The soils on the steepest slope class had the highest amount of pedons with argillic horizons, indicating that in the moist, frigid soil temperature regime there is sufficient wetting and drying to produce argillic horizons and the runoff is low enough that erosion does not proceed faster than clay illuviation on slopes of <55%. The majority of the >35% slope pedons with argillic horizons occurred on the moister N aspects.

Haplohumults averaged 123 Mg C ha–1 and Haplumbrepts averaged 201 Mg C ha–1 (Kern 1994, Table 5). However, Hapludults and Dystrudepts were not reported. The Humults, Udults, Umbrepts, and Ochrepts averaged 126, 61, 186, and 90 Mg C ha–1 (Kern, 1994, Table 4). Based on the frequencies reported in Table 7 and the test above, the weighted average for the study area using Kern's values would have been 109 Mg C ha–1 which is slightly lower than our simple average of 121 Mg C ha–1. Kern's values should have been higher than those in this study because they were probably calculated from deeper pedons and the mass SOC was not corrected for rock fragment volume like those in this study. The average values of pedons from across the United States may have been diluted by pedons sampled in drier areas or at lower elevations.

This study's simple average of mass SOC was considerably higher than the area-weighted average shown by Kern (1994)(Fig. 6) for the Appalachian Mountains (61 to 75 Mg C ha–1) and the remainder of Virginia (76 to 90 Mg C ha–1). Area-weighted averages are often lower than simple averages because the high SOC soils often make up a small percentage of the landscape and may be omitted from small-scale soil maps if the generalization is not done carefully (Galbraith et al., 2003).


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The soils in this study were lower in rock fragments and bulk density and shallower than reported for mountain soils in previous studies, resulting in slightly lower average mass SOC storage. The lower bulk density was due to these soils having higher concentration of SOC than soils in similar studies. Carbon concentrations were very high in the B horizons compared with similar studies (Daniels et al., 1987), but the explanation is not clear without further investigation.

The effect of slope alone on mass SOC was not apparent in this study. The effect of aspect on mass SOC was probably related to the amount of solar radiation received on N aspects and possibly to erosion effects on S aspects. Differences in mass SOC by aspect were due to differences in horizon thickness rather than in concentration of SOC. Soils on N aspects (especially N35 sites) had thicker horizons and thus contained higher amounts of mass SOC because they received less solar insolation and had cooler temperatures, more effective moisture, lower erosion potential, and more geophage activity in the litter. Litter production rates were probably not as high on the moister N aspects as the drier S aspects (Whittaker, 1966), possibly because of the lower amount of sunlight received for photosynthesis. On N aspects, the litter SOC was more readily incorporated into the A horizons and protected from gaseous losses to the atmosphere during mineralization. There was also an obvious slope angle effect on N aspects where mass SOC increased with slope angle as solar insolation decreased. Soil organic C was resident in litter layers longer on S aspects because of the drier climate, lower geophage activity, and higher amount of high-lignin leaves. The effect of slope was apparent on slopes even as low as 7%. The S7 and S35 sites were located in landscape positions where erosion potential was higher than on any other sites.

This study indicates that the steepest northeast-facing sites have the highest potential to sequester SOC and should have the highest C credit value of all well-drained frigid soil temperature regime southern Appalachian soils. The steepest southwest-facing sites have very thin A horizons should be protected from disturbance that may increase erosion or litter removal. Models of global C inventory and dynamics following climate change should take both aspect and slopes over 7% into account in cold frigid areas. Inventories of SOC should give higher credit to preservation of forests on >35% slope northeast-facing sites than on more level northeast-facing sites and to northeast-facing sites over southwest-facing sites. Forest removal on the steep northeast-facing sites would result in lower SOC sequestration potential.

The difference in concentration of SOC and mass SOC demonstrates the importance of using mass SOC rather than concentration of SOC in global change models and regional SOC estimates. Carbon sequestration models and mass SOC inventory mappers should consider the soil temperature regime and the differences in mass SOC between aspect and slope combinations in the mountains. Accurate calculation of mass SOC is highly dependent on using accurate values for rock fragments, bulk density, and horizon thickness by horizon, slope, and aspect rather than using estimated soil series or average mineral horizon values that may be substantially higher than actual values and would therefore significantly over-estimate regional SOC totals.

Received for publication December 9, 2002.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
Y. Guo, P. Gong, R. Amundson, and Q. Yu
Analysis of Factors Controlling Soil Carbon in the Conterminous United States
Soil Sci. Soc. Am. J., February 27, 2006; 70(2): 601 - 612.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Miller, J. O.
Right arrow Articles by Daniels, W. L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Miller, J. O.
Right arrow Articles by Daniels, W. L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Miller, J. O.
Right arrow Articles by Daniels, W. L.
Related Collections
Right arrow Carbon Sequestration
Right arrow Data Management
Right arrow Soil Models

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome