Surface fluxes of sensible and latent heat over a tall grass prairie in central Kansas, as measured by 22 surface stations during FIFE 1987, are compared with values gained indirectly by simple linear extrapolation of aircraft-measured flux profiles to the surface. The results of 32 such comparisons covering the period June 26 to October 13, 1987, indicate that (1) the flux profiles were generally linear, as expected in a convective boundary layer; 2) in general, aircraft-derived surface fluxes were 30% less than the surface averages for sensible heat and 10% less for latent heat; (3) high-pass filters, with cutoffs at 5 km, acted to reduce the uncertainty of the aircraft flux estimates, without changing the magnitude or sign of the disagreement between the aircraft and the surface fluxes; (4) undersampling of high frequencies by the aircraft could have caused as much as a 15% underestimate of the surface sensible heat fluxes; and (5) the aircraft and surface Bowen ratios also disagreed, with the greatest disagreement occurring during the dry periods with decreased plant activity (IFC 4). ¿ American Geophysical Union 1992