Skip to main content
James Thompson, PhD
Professor of Soil Science
Davis College
WVU Home

Projects

Soil Carbon Storage and Dynamics

  • Modeling soil organic carbon at the landscape-scale using legacy data and pedometric techniques.

Spatial Patterns of Soil Organic Carbon in Forest Soils: Effects of Landscape Attributes and Forest Management Practices

E. S. Anderson, Dep. of Soil Science, North Carolina State Univ.
J. A. Thompson, Div. of Plant & Soil Sciences, West Virginia Univ.

While the sources of carbon dioxide to the atmosphere are well documented, the role of the various carbon dioxide sinks is not as certain. Although a number of studies have investigated soil organic carbon (SOC) sequestration in agricultural and prairie soils, relatively few studies have investigated SOC storage in forest soils, even though forest soils currently store 40% of global soil carbon. Soils have differing potential to sequester SOC depending on a number of inherent factors including soil nutrient status, pH, texture, temperature, and water content. Environmental and human factors such as topography and management history have also been shown to be effective predictors of SOC storage. To predict where in the landscape the greatest potential lies in sequestering SOC we need to first understand how landscape and management factors influence SOC storage at the site, landscape, and the watershed scale. The specific objectives of this research are to:

  • Elucidate relationships among landscape factors, forestry management practices and the spatial variability of SOC in aggrading Southern pine stands of the Lower Coastal Plain
  • Parameterize and validate quantitative soil-landscape models considering important landscape and management factors that allow prediction of SOC at watershed and regional scales
  • Quantify total SOC on an areal basis for each watershed

With an understanding of the important factors that lead to SOC sequestration in forest soils we will be able to predict where in the landscape to focus management strategies to achieve our future carbon sequestration goals. Conservation and enhancement of SOC pools is of importance because of it positive influence on forest growth of long-term sustainability of soil. Knowledge of spatial patterns of SOC will aid in forest management decisions regarding sustainable forest productivity and growth. Additionally, watershed scale spatial models of SOC will provide a more precise measurement of local carbon inventories.

Landscape and Seasonal Influences on Soil Respiration Rates in Forest Soils of Southeastern Kentucky

A. C. Abnee, Dep. of Agronomy, Univ. of Kentucky
J. A. Thompson, Div. of Plant & Soil Sciences, West Virginia Univ.
R. K. Kolka, North Central Res. Sta., USDA Forest Service

Data on carbon fluxes such as soil respiration are needed to develop strategies for increased carbon sequestration and reduced levels of atmospheric trace gases. Soil respiration is driven by proximal factors (e.g., soil temperature and soil moisture), which affect soil respiration by regulating organic matter production and decomposition, and distal factors (e.g., topography), which affect soil respiration by influencing proximal factors. The objectives of this study were: (i) to quantify carbon dioxide flux from forest soils; (ii) to relate carbon dioxide flux to proximal and distal factors; and (iii) to develop predictive soil-landscape models of soil respiration across a forested watershed. Sampling points were selected using a random stratified approach, with strata established based on slope aspect (NE and SW), slope shape (convergent and divergent), and slope position (upper, middle, and lower backslope), with four replicate sample points for each combination of landscape factors. Carbon dioxide flux from the soil surface was measured using a portable soil respiration system. Data were collected monthly for 12 months to examine both seasonal and landscape trends. Soil samples collected at each point were analyzed in the laboratory to determine potential soil respiration rates using static-incubation gas chromatography. Correlation analysis was used to relate field respiration rates to landscape attributes. We also generated empirical models using stepwise regression to examine relationships between respiration rates and both proximal (soil physical and chemical properties) and distal factors (terrain attributes calculated from a 30-m digital elevation model). Both potential and in situ soil respiration rates were greater on the NE-facing slopes than the SW-facing slopes. The models that we developed and validated explain up to 52% of variability in measured soil respiration, although model relationships changed over time. In general, soil temperature and slope aspect were the best predictors of soil respiration rates on the NE-facing slopes, while slope curvature, slope gradient, and upslope contributing area were the best predictors soil respiration rates on the SW-facing slopes.

Soil Carbon Storage Estimation in Central Hardwood Forest Watersheds using Quantitative Soil-Landscape Modeling

J. A. Thompson, Div. of Plant & Soil Sciences, West Virginia Univ.
R. K. Kolka, North Central Res. Sta., USDA Forest Service

Carbon sequestration and storage in soils is important to forest ecosystems. Decomposed woody debris and leaf litter add organic matter to the soil and increase soil cation exchange capacity, enhance soil aggregation, and increase aeration and water holding capacity. Moreover, forest soils of the central hardwoods region may serve as important carbon sinks for ameliorating excess atmospheric carbon dioxide. When assessing watershed and regional scale soil carbon storage, spatial extrapolation is an important issue. A stratified “measure-and-multiply” approach does not account for inherent variability within map units, thus limiting the transferability of such predictions, especially to areas of different scale, or areas where the map units change significantly. Our objective is to develop quantitative soil-landscape models that ascertain relationships between soil carbon and topographic variables derived from digital elevation models (DEM). We measured the amount of carbon stored in the soil to a depth of 30 cm in triplicate 6.25-cm cores at 104 forest inventory plots established in a 380-m grid pattern across a 1500 ha watershed. We also acquired U. S. Geologic Survey 30 m DEM data for entire study area, and calculated various terrain attributes, including slope gradient, slope aspect, slope curvature, and upslope contributing area. We generated multivariate statistical models to describe the variability of soil carbon using forward stepwise selection and robust linear regression. Because of the strong effect of slope aspect on soil formation in these landscapes, we stratified the data into four aspect classes (NE, SE, SW, and NW). Coefficients of correlation® between measured soil carbon and individual terrain attributes were statistically significant, but low (less than ±0.50). However, we developed and validated empirical models that explain up to 69% of variability in soil carbon using only three or four terrain attributes as predictors. This approach can resolve variability of soil carbon within map units, represent continuous variability of soil carbon across landscapes, and may be transportable to similar landscapes.

Land Use & Water Quality

  • Fragipan influence on hydropedological properties of benchmark soilscapes in West Virginia
  • Organic phosphorus fractionation in West Virginia benchmark soils
  • Seasonal Infiltration and Subsurface Water Dynamics across Benchmark Soil Catenas of Eastern West Virginia
  • Phosphorus Sorption Capacity of West Virginia Soils: Spatial Assessment and in situ Leachability
  • Evaluation and monitoring of mitigation wetlands associated with Corridor H

Water Quality Implications of Urban Development in Mixed Use Watersheds

C. B. Coulter, Dep. of Agronomy, Univ. of Kentucky
R. K. Kolka, North Central Res. Sta., USDA Forest Service
J. A. Thompson, Div. of Plant & Soil Sciences, West Virginia Univ.

Water quality and nonpoint source (NPS) pollution are important issues in many areas of the world, including the Inner Bluegrass Region of Kentucky where urban development is changing formerly rural watersheds into urban and mixed use watersheds. In watersheds where land use is mixed, the relative contributions of NPS pollution from rural and urban land uses can be difficult to separate. To better understand NPS pollution sources in mixed use watersheds, surface water samples were taken at three sites that varied in land use to examine the effect of land use on water quality. Within the group of three watersheds, one was predominately agriculture (Agricultural), one was predominately urban (Urban), and a third had relatively equal representation of both types of land uses (Mixed). Nitrogen (N), phosphorus (P), total suspended solids (TSS), turbidity, pH, temperature, and streamflow were measured for one year. Comparisons are made among watersheds for concentration and fluxes of water quality parameters. Nitrate and orthophosphate concentrations were found to be significantly higher in the Agricultural watershed. Total suspended solids, turbidity, temperature, and pH, were found to be generally higher in the Urban and Mixed watersheds. No differences were found for streamflow (per unit area), total phosphorus, and ammonium concentrations among watersheds. Fluxes of orthophosphate were greater in the Agricultural watershed that in the Urban watershed while fluxes of TSS were greater in the Mixed watershed when compared to the Agricultural watershed. Fluxes of nitrate, ammonium, and total phosphorus did not vary among watersheds. It is apparent from the data that Agricultural land uses are generally a greater source of nutrients than the Urban land uses while Urban land uses are generally a greater source of suspended sediment.

Terrain Analysis & Soil-Landscape Modeling
  • Soil Spatial Prediction using Multi-scale Terrain Analysis in the Appalachian Mountains of West Virginia
  • Multi-Scale Terrain Analysis of Gridded DEM

Soil-Landscape Modeling for Defining Landform Management Segments Transferable Across a Physiographic Region

J. A. Thompson, Div. of Plant & Soil Sciences, West Virginia Univ.
E. M. Pena-Yewtukhiw, Dep. of Agronomy, Univ. of Kentucky
J. H. Grove, Dep. of Agronomy, Univ. of Kentucky
C. E. Kiger, Dep. of Agronomy, Univ. of Kentucky

Common methods for characterizing within-field soil variability for precision agriculture applications, such as grid sampling, are data dependent and do not account for the primary causes of soil variability—processes of soil formation. Instead of interpolating measured values, a more valuable outcome is to link observed patterns to processes of soil development. We have collected high-resolution digital elevation models (DEM) for multiple fields across the Pennyroyal physiographic region of central and western Kentucky, and used these DEM to direct soil sampling using a stratified random sampling design. In each field, 30 discrete soil samples were extracted for morphological, physical, and chemical characterization. We are examining the inherent differences in terrain attributes among fields, and are developing soil-landscape models that can predict the spatial patterns of A-horizon thickness, soil organic carbon content, and clay content from spatial patterns of terrain attributes derived from the DEM. We also examine whether these soil-landscape relationships are similar among landscapes of similar soils within a physiographic region. This work may provide producers with a means to quickly and easily assess the nature of soil variability in their fields and divide the field into meaningful management zones.

LIDAR Density and Linear Interpolator Effects on Elevation Estimates

E. S. Anderson, Dep. of Soil Science, North Carolina State Univ.
J. A. Thompson, Div. of Plant & Soil Sciences, West Virginia Univ.
R. E. Austin, Dep. of Soil Science, North Carolina State Univ.

Linear interpolation of irregularly spaced LIDAR elevation data sets is needed to develop realistic spatial models. We evaluated inverse distance weighting (IDW) and ordinary kriging (OK) interpolation techniques and the effects of LIDAR data density on the statistical validity of the linear interpolators. A series of 10 forested 100-ha LIDAR tiles on the Lower Coastal Plain of eastern North Carolina was used. An exploratory analysis of the spatial correlation structure of the LIDAR data set was performed. Weighted nonlinear least squares (WNLS) analysis was used to parameterize best-fit theoretical semivariograms on the empirical data. Tile data were sequentially reduced through random selection of a predetermined percentage of the original LIDAR data set, resulting in data sets with 50, 25, 10, 5, and 1% of their original densities. Cross-validation and independent validation procedures were used to evaluate root mean square error (RMSE) and kriging standard error (SE) differences between interpolators and across density sequences. Review of errors indicated that LIDAR data sets could withstand substantial data reductions yet maintain adequate accuracy (30 cm RMSE; 50 cm SE) for elevation predictions. The results also indicated that simple interpolation approaches such as IDW could be sufficient for interpolating irregularly spaced LIDAR data sets.

Horizontal Resolution and Data Density Effects on Remotely Sensed LIDAR-Based DEM

E. S. Anderson, Dep. of Soil Science, North Carolina State Univ.
J. A. Thompson, Div. of Plant & Soil Sciences, West Virginia Univ.
D. A. Crouse, Dep. of Soil Science, North Carolina State Univ.
R. E. Austin, Dep. of Soil Science, North Carolina State Univ.

Production of different horizontal resolution DEM with the same vertical precision from the same data source without generalizations associated with resampling regimes is important for predicting scale dependent environmental variables. The use of light detecting and ranging (LIDAR) elevation data sets offers the flexibility needed to produce multiple horizontal resolutions of DEM from the same data source. A series of 61 LIDAR tiles (100 ha) were collected from the North Carolina Flood Mapping Program covering the spatial extent of the Hofmann Forest in the Lower Coastal Plain of Eastern North Carolina. The LIDAR data set was reduced to 50%, 25%, 10%, 5%, and 1% of the original density. We created 5-m, 10-m, and 30-m DEM with 0.1 m vertical precision were created for each density level and used paired t-test to determine if the true mean of their differences were equal to zero. Differences indicated that for the 30-m DEM, LIDAR data sets could be reduced to 10% of their original data density without statistically altering the produced DEM. However, the 10-m DEM could only be reduced to 25% of the original data set before statistically altering the DEM. Data reduction was more limited for the 5-m DEM with possible reduction only to 50% of their original density without producing statistically different DEM. Our evaluation provides some indication as to the minimum required LIDAR data density to produce a DEM of a given horizontal resolution. However, evaluation of additional horizontal resolutions and additional density reduction is required to provide a clearer understanding of the effect of LIDAR data density.