Increased levels of atmospheric GreenHouse Gas (GHG) emissions, especially CO2, have necessitated mitigatory action for GHGs both at international and national scale. Long term carbon storage or carbon sequestration in agricultural soils has been identified to help reduce CO2 emissions to the atmosphere. Carbon inputs through crop residues and the decomposition process in soils are the key components in determining the ability of a soil to sequester C. We estimated the county level crop residue carbon inputs in the US, and modeled and studied the inter-annual variability in carbon dynamics in US agricultural soils over a sixteen year period, from 1982-1997. The first part of the study evaluated the potential use of existing national crop production databases in estimating soil carbon inputs. Past studies have found reliable relationships between the crop yields and above ground biomass for different crops. Therefore carbon inputs to soils from crop residues could be estimated using available crop yield data. Crop yields are recorded in two databases maintained by the US Department of Agriculture: the National Agricultural Statistics Survey (NASS), and the Census of Agriculture (Ag Census). The NASS crop yield estimates are produced annually and Ag Census data are produced every 5 years. Since both databases contain missing values at county-level, we developed interpolation methods to create complete crop yield and acreage databases for the period 1982-1997, to be used in estimating carbon inputs from the major crops of the US. Initially, county-level data gaps in NASS yields were filled using regression analyses with the Ag Census data, and the remaining gaps were filled using mixed linear effect models that had environmental and irrigation- variables as covariates. The gaps in the county- level acreage data reported by NASS were filled using the acreage information from Ag Census, and mixed linear effect models incorporating several weather and economic variables as covariates. We then estimated crop residue carbon inputs to soils using the completed yield and acreage information, and modeled interannual variability of carbon dynamics in the US agricultural soils. In this effort, we studied the trends in crop yields and overall soil carbon storage, and related this variation to variation in climate, weather and other environmental factors.