Soil carbon relationships with terrain attributes, electrical conductivity, and a soil survey in a coastal plain landscape

Soil organic carbon (SOC) estimation at the landscape level is critical for assessing impacts of management practices on C sequestration and soil quality. We determined relationships between SOC.. terrain attributes. field scale soil electrical conductivity (EC), soil texture and soil survey map units in a 9 ha coastal plain field (Aquic and Typic Paleudults) historically man, P aged by conventional means. The site was composite sampled for SOC (0-30 cm) within 18.3 X 8.5-m grids (n = 496). and two data sets were created from the original data. Ordinary kriging. co-kriging, regression kriging and multiple regression were used to develop SOC surfaces that were validated with an independent data set (n = 24) using the mean square error (MSE). The SOC was relatively low (26.13 Mg ha(-1)) and only moderately variable (CV = 21%), and showed high spatial dependence. Interpolation techniques produced similar SOC maps but the best predictor was ordinary kriging (MSE = 9.11 Mg-2 ha(-2)) while regression was the worst (MSE = 20.65 Mg-2 ha(-2)). Factor analysis indicated that the first three factors explained 57% of field variability; compound topographic index (CTI), slope, EC and soil textural fractions dominated these components. Elevation, slope, CTI, silt content and EC explained up to 50% of the SOC variability (P less than or equal to 0.01) suggesting that topography and historical erosion played a significant role in SOC distribution. Field subdivision into soil map units or k-mean clusters similarly decreased SOC variance (about 30%). The study suggests that terrain attributes and EC surveys can be used to differentiate zones of variable SOC content. which may be used as bench marks to evaluate field-level impact of management practices on C sequestration.