Fluxes of carbon dioxide, water vapour, and sensible heat obtained over a grassland ecosystem, during the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE), using an aircraft- and two tower-based systems are compared for several days in 1987 and in 1989. The tower-based cospectral estimates of CO2 (WC), sensible heat (WT), water vapour (Wq), and momentum (WU), expressed as a function of wavenumber K times sampling height z, are relatively similar to the aircraft-based estimates for K¿z>0.1. A measurable contribution to the fluxes is observed by tower-based systems at K¿z<0.01 but not by the aircraft-based system operating at an altitude of approximately 100 m over a 15¿15 km area. Using all available simultaneous aircraft and tower data, flux estimates by both systems were shown to be highly correlated. As expected from the spatial variations of the greenness index, surface extrapolation of airborne flux estimates tended to lie between those of the two tower sites. The average fluxes obtained, on July 11, 1987 ,and August 4, 1989, by flying a grid pattern over the FIFE site agreed with the two tower data sets for CO2, but sensible and latent heat were smaller than those obtained by the tower-based systems. However, in general, except for a small underestimation due to the long wavelength contributions and due to flux divergence with height, the differences between the aircraft- and tower-based surface estimates of fluxes appear to be mainly attributable to differences in footprint, that is, differences in the area contributing to the surface flux estimates. ¿ American Geophysical Union 1992