Estimating Seasonal Carbon Fluxes over Agricultural Lands Using Satellite Remote Sensing.

Croplands are typically characterized by fine-scale heterogeneity, w hich makes it difficult to accurately estimate cropland carbon fluxes over large regions given the fairly coarse spatial resolution of high-frequency satellite observations. It is, how ever, important that we improve our ability to estimate spatially and temporally resolved carbon fluxes because croplands constitute a large land area and have a large impact on global carbon cycle. A Satellite based Dynamic Cropland Carbon (SDCC) modeling framework was developed to estimate spatially resolved crop specific daily carbon fluxes over large regions. This modeling framew ork uses the REGularized canopy reFLECtance (REGFLEC) model to estimate crop specific leaf area index (LAI) using dow nscaled MODIS reflectance data, and subsequently LAI estimates are integrated into the Environmental Policy Integrated Model (EPIC) model to determine daily net primary productivity (NPP) and net ecosystem productivity (NEP). Firstly, we evaluate the performance of this modeling framew ork over three eddy covariance flux tow er sites (Bondville, IL; Fermi Agricultural Site, IL; and Rosemount site, MN). Daily NPP and NEP of corn and soybean crops are estimated (based on REGFLEC LAI) for year 2007 and 2008 over the flux tow ersites and compared against flux tow er observations and model estimates based on in-situ LAI. Secondly, w e apply the SDCC framework for estimating regional NPP and NEP for corn, soybean and sorghum crops in Nebraska during year 2007 and 2008. The methods and results w ill be presented.