Direct observations of surface fluxes of momentum, sensible, and latent heat from towers and aircraft are compared to output from the NCAR-Penn State Mesoscale Model MM5. The model flux parameterization is seen to work well if appropriate values of the roughness length z0 and moisture availability parameter M are specified. Although the surface fluxes are quite sensitive to these parameters, as found by earlier investigators, it is not obvious how to select a value for M a priori. An initial estimate of M should take into account the rainfall and cloudiness history and probably other factors. Because temperature and humidity near the surface are quite sensitive to the fluxes, it is suggested that the difference between observed and calculated air temperatures could be used iteratively to guide the choice of M.