Regional scale evapotranspiration estimation using vegetation index and surface temperature from NOAA satellite AVHRR data

Measurement of the water vapor cycle of the land surface is crucial in improving the management of the limited fresh water resources. Despite the importance of evapotranspiration in the hydrologic cycle, its magnitude, spatial and seasonal distributions, and its relation to environmental variables remains relatively unknown. In the state of Florida, evapotranspiration (ET) is second only to precipitation in magnitude within the hydrologic budget.; This research began with the examination of the spectral responses of surface objects in Florida by the use of a hand-held radiometer. The relationship between the field radiometry results and satellite remote sensing data was then examined. The optimum spatial resolution between the fine scale radiometer and the coarse scale satellite data was also explored. Surface temperature and vegetation index were extracted from the NOAA satellite AVHRR data. The concept of a Normalized Difference Vegetation Index (NDVI) and Diurnal Surface Temperature Variation (DSTV) triangle was introduced, based on the pattern of ground objects located in the NDVI-DSTV scattergram. The NDVI-DSTV triangle relationship was used to calculate the vegetation and moisture coefficient (VMC) for fractional contribution from various vegetation cover and soil moisture backgrounds. A geographic information system (GIS) was utilized to manage the satellite and ground-based data. Regional ET was then estimated from the data layers in the GIS.; Results from the hand-held radiometer study indicate that spectral response can be used to determine the most appropriate spectral bands and season for vegetation discrimination. Correlation study of radiometer and satellite observations indicates that they relate to each other most strongly in the format of vegetation index. As to satellite imagery, to identify the land use activities the optimum spatial resolution is 300 m and 600 m for infrared and thermal bands, respectively. The concept of NDVI-DSTV triangle was successfully verified with both ground and satellite observations and the relationships of NDVI-DSTV determined the fractional combination from agriculture, wet and dry soil surfaces. Results of the calculated VMC clearly displayed the fractional contributions of crop coefficients from vegetation, wet and dry soils. Results of ET estimation show that this algorithm did provide an efficient method for studying the spatial variation in ET and making useful estimates of ET in south Florida.