Surface temperature retrieval in a temperate grassland with multiresolution sensors

Radiometric surface temperatures retrieved at various spatial resolutions from aircraft and satellite measurements at the FIFE site in eastern Kansas were compared with near-surface temperature measurements to determine the accuracy of the retrieval techniques and consistency between the various sensors. Atmospheric characterizations based on local radiosonde profiles of temperature, pressure, and water vapor were used with the LOWTRAN-7 and MODTRAN atmospheric radiance models to correct measured thermal radiances of water and grassland targets for atmospheric attenuation. Comparison of retrieved surface temperatures from a helicopter-mounted modular multispectral radiometer (MMR);(similar to 5-m ''pixel''), C-130 mounted thematic mapper simulator (TMS) (NS001, similar to 20-m pixel), and the Landsat 5 thematic mapper (TM) (120-m pixel) was done. Differences between atmospherically corrected radiative temperatures and near-surface measurements ranged from less than 1 degrees C to more than 8 degrees C. Corrected temperatures from helicopter-MMR and NS001-TMS were in general agreement with near-surface infrared radiative thermometer (IRT) measurements collected from automated meteorological stations, with mean differences of 3.2 degrees C and 1.7 degrees C for grassland targets. Much better agreement (within 1 degrees C) was found between the retrieved aircraft surface temperatures and near-surface measurements acquired with a hand-held mast equipped with a MMR and IRT. The NS001-TMS was also in good agreement with nearsurface temperatures acquired over water targets. In contrast, the Landsat 5 TM systematically overestimated surface temperature in all cases. This result has been noted previously but not consistently. On the basis of the results reported here, surface measurements were used to provide a calibration of the TM thermal channel. Further evaluation of the in-flight radiometric calibration of the TM thermal channel is recommended.