The pedologic C pool comprises of soil organic C (SOC) and soil inorganic C (SIC) components. Of the two components, the SOC pool is highly reactive and is a strong determinant of numerous ecosystems services. Estimates of SOC pool and their spatial variability in terrestrial ecosystems are essential to better understand the global C cycle, to estimate the soil C sink capacity, and to quantify the amount of SOC sequestered in a defined time period. But the amount of C stored in the soil per unit area is highly variable as the magnitude of SOC pool at a location depends on a range of factors such as soil type, land use, annual input of biomass C, topographic features, and climatic conditions. These factors differ among locations and ecoregions. Consequently, several approaches are needed to develop a reliable estimate of SOC pool at different spatial scales. Therefore, the overall goal of this study was to understand the storage and dynamics of SOC pool at a regional scale. Specific objectives were to; develop methodology to quantify the SOC pool within different depth intervals at a regional scale, use environmental variables for regional scale SOC predictions, and assess the effect of tillage practices on the storage and dynamics of SOC in contrasting agricultural soils.
Three studies were conducted to meet the above mentioned objectives in Midwestern United States (Ohio, Michigan, Indiana, Kentucky, Pennsylvania, West Virginia and Maryland). Soil legacy databases maintained by National Soil Survey laboratory, Pennsylvania State University, The Ohio State University, and field collected soil samples were used in this study. Environmental variables covering the study area were collected from secondary databases. Soil and environmental databases were assembled in geographic information system to develop spatially explicit models. Various univariate and multivariate mathematical, statistical and geostatistical methods including SOC profile depth distribution functions, ordinary kriging, regression kriging, analysis of variance, multiple linear regression, and geographic weighted regression techniques were used to synthesize meaningful conclusions about the SOC sequestration and dynamics at a regional scale.
Results indicated that SOC pool estimates for regional scales within desired depth intervals can be made by using the exponential soil depth functions at SOC profiles and interpolating the coefficients of exponential functions. This method of predictive mapping is especially useful in scenarios where there are missing observations for some horizons as they can be interpolated using the exponential equations. Similarly, by converting conventional till to no till agriculture, some of the depleted historic SOC pool can be resequestered. In addition to environmental concerns, such a strategy can also create economic opportunities for farmers through C trading. Likewise, by using the range of spatial autocorrelation in SOC data in a geographic weighted regression (GWR) framework, better estimates of SOC pools can be made at large spatial scales. Though it is unlikely that a single model can be developed to be applicable to all soil landscapes in regional scale studies, GWR approach can play a vital role in improving the prediction ability of SOC pools across the regional scales and this methodology can be used readily by the land managers.