Estimating surface fluxes over the SGP site with remotely sensed data

The estimation of surface energy fluxes requires the combination of data from several sources. Typically land use and vegetation cover are obtained from visible and near infrared (VNIR) data, while a surface state variable, the surface temperature, is obtained from thermal infrared (TIR) data. Approaches to combine these data for the estimation of surface energy fluxes were studied as part of the SGP97. Toward this end TIR and VNIR images for four days were collected over the El Reno and ARM Cart sites of SGP using the TIMS (Thermal Infrared Multispectral Scanner) TIMS (Thematic Mapper Simulator) instruments. At both sites intensive ground measurements are available. Most of the imagery were acquired at an altitude of 5 km yielding a pixel resolution of approximately 12 meters. One flight line was flown at a 1.5 km altitude yielding 4 meter resolution. The observed brightness temperatures were corrected for atmospheric effects using MODTRAN and nearby, in both space and time, radiosoundings. Comparisons with bulk water temperatures of small ponds are within 0.5 to 1C. The Temperature Emissivity Separation (TES) algorithm developed for use with ASTER data was applied and the results compared with laboratory measurements of the soil emissivity. The resulting surface temperatures are used in a two source (soil and vegetation) model to estimate surface fluxes on an areal basis. The concurrent VNIR imagery were used to estimate the fractional vegetation cover. The resulting fluxes were in good agreement with ground measurements. The spatial scales of surface temperatures and the resulting fluxes were also analyzed