This study investigates the relationships between airborne observations of CO2 fluxes and satellite-based vegetation indices using data from the First International Land Surface Climatology Project Field Experiment (FIFE) 1989. The CO2 flux data were obtained by flying a grid pattern at a mean altitude of 90 m over a 225 km2 grassland ecosystem, on six occasions during a 2-week period. Two satellite-based data sets were obtained from NOAA 10 advanced very high resolution radiometer (AVHRR) and one from Landsat 5 thematic mapper (TM). Linear relationships Fc=b Sr*+a were found between CO2 flux data (Fc) and foot-print-weighted simple ratios (SR*) from the various satellites. These relationships improved significantly when the footprint was adjusted with the continuously changing wind conditions and with an empirical parameter that accounts for effects of surface roughness and stability on the diffusion pattern between the surface and the airborne sampling locations. The AVHRR data collected on July 28 accounted for 53 to 81% of the variations in CO2 flux measured during the six grid flights over the site and the Landsat TM data collected on August 4 accounted for 68 to 83% of the same CO2 variations. The regression coefficient b, which is related to the photosynthetic efficiency, appears to be an excellent way to characterize the dynamic response of complex ecosystems, and the flux (Fc) at SR*=1.5 provides an estimate of soil respiration as shown through a comparison with enclosure measurements. ¿ American Geophysical Union 1995