Assessing Carbon Dynamics in Agriculture Using Remote Sensing

1. Introduction Increasing atmospheric concentrations of CO 2 and other greenhouse gases is a global concern. Agricultural activities contribute to CO 2 and N 2 O emissions through combustion of fossil fuels, soil organic carbon (SOC) decomposition, and biomass burning. Although green plants convert CO 2 into carbohydrates and biomass, most of the CO 2 that green plants absorb reenters the atmosphere through respiration of plants and animals and through microbial decomposition. Depending on land use and management, soil can function as either a source or sink for atmospheric CO 2. Based on the large decreases in soil organic carbon when native forests and grasslands were converted to agriculture, the potential for C sequestration in soils is very large (Lal et al., 1999). Carbon can be stored in the soil in either living organisms or in their residues in a form which resists further biological degradation. Models can predict net carbon sequestration for different soil types and land management. Numerous models of C dynamics have been published and range from complex research-oriented models to simple empirical applications-oriented models (Ma and Shaffer, 2001). The complex research models emphasize the underlying biological, chemical, and physical processes that control C flows, but tend to be point-based because of their detailed input data requirements. The simple empirical models correlate ecosystem-scale processes with parameter that are readily measured in the field and, as a result, may gloss over some important functional relationships. The linkage of process models to geographic information systems (GIS) for spatially distributed fields or watersheds has blurred the spatial scale distinction between research and application models. Lack of data to support these process models across a wide range of soil and land management scenarios continues to be a major issue limiting their usability. Robust approaches for extending C models from local to regional and global scales have not been identified and evaluated. Recent advances in remote sensing of vegetation and soils can potentially provide some of the biophysical parameters needed by various C models to predict C dynamics across landscapes. In this paper, we briefly 1) review current status for remote sensing of crops and soils, and 2) examine the potential role of remote sensing for assessing C dynamics in agriculture.