We present a method to estimate regional carbon fluxes by modeling the changes in vegetation at flux tower sites using remotely sensed data, partitioning the flux tower signal, and making maps of regional fluxes.  We use detailed light curve analysis to partition the net ecosystem exchange (NEE) measured by a flux tower into gross primary productivity (GPP) and total ecosystem respiration (Re). These data on NEE, GPP, and Re are then summarized to the same compositing period used for daily remotely sensed data—usually 7 to 10 days. The composited flux data, remotely sensed data, and other GIS data for multiple flux tower locations and multiple years are used in data-driven models to predict carbon fluxes. Regional flux maps of grasslands show the effects of drought on this cover type, and preliminary results for croplands are also promising.