THE SURFACE HEAT-FLUX AS A FUNCTION OF GROUND COVER FOR CLIMATE MODELS

Surface heat fluxes were examined as a function of surface properties and meteorological conditions in a 100 km x 100 km grid square at 1-km spatial resolution centered at the location of the First ISLSCP (Inte rnational Satellite Land Surface Climatology Project) Field Experiment (FIFE), the Forest Ecosystem Dynamics site in central Maine, and a se miarid rangeland site around Walnut Gulch, Arizona. This investigation treats the surface heat Aux variability within a GCM grid box to prov ide insight into methods for treating that variability in climate mode ls. The heat fluxes were calculated using NOAA AVHRR and available met eorological data. The average heat fluxes that were estimated using th e various area ground-cover representations were compared with the ens emble average heat fluxes for the entire area, which were assumed to b e the best representation of the heal fluxes for the areas. Average he at fluxes were estimated for the entire 100 km x 100 km area based on a single ground-cover representation, and the mean error for the area sensible heat flux was about 10% and for the area latent heat flux, 21 %. The estimation error was reduced, and in some cases significantly r educed, when the area heat fluxes were estimated by partitioning the a rea according to significant ground cover. The most significant effect of the partitioning was on the latent heat flux estimates.