A spatially explicit scheme for tracking and validating annual landscape scale changes in soil carbon

Abstract

Short term change in the relations of land cover (LC) and soil organic carbon (SOC) are a challenge for management requiring a modeling methodology for monitoring annual transitions in the SOC. As part of this methodology, ‘annual land cover change information’, available from remote sensing is assimilated within an ecosystem modeling framework. In this research, this framework comprises of CENTURY, an agro-ecosystem model that simulates SOC and WEPP a process-based, distributed parameter model for computing soil erosion. The proposed methodology is used for studying changes in SOC from 1999 to 2006, in the Big Creek watershed, located in the Union and Pulaski Counties of Southern Illinois. Validation is performed by comparing the modeled SOC at the end of year 2006 against SOC obtained from extensive soil sampling. Two sets of hypotheses are tested, these are: (1) equality of the means of the modeled and measured SOC for the entire watershed, and (2) equality of the means of the modeled and measured SOC of the different soil series in the watershed. Extensive analysis of the measured SOC is also performed. This assessment reveals significant differences between the means of the measured SOC across LCs, whereas these differences were insignificant across soil series. Modeling results indicate increase in the SOC over eight years, with LC having a major impact in governing this change. Overall, the presented methodology, even with limited information, performs adequately. A comparison of the modeled and the measured SOC shows absence of significant differences in the means. For different soil series, both significant and insignificant differences in the means of the modeled and the measured SOC were observed. However, these mixed results are expected, as the primacy of LC in governing SOC confounds the influence of soils on SOC.