EVAPOTRANSPIRATION ESTIMATION USING A NORMALIZED DIFFERENCE VEGETATION INDEX TRANSFORMATION OF TWO SATELLITES DATA IN ARID MOUNTAIN AREAS Ayoub Almhab Department of Remote Sensing Faculty of Geoinformation science and engineering Universiti Teknologi Malaysia e-mail: aalmhap@maktoob.com Ibrahim Busu Department of Remote Sensing Faculty of Geoinformation science and engineering Universiti Teknologi Malaysia ibusu@fksg.utm.my Abstract Evapotranspiration (ET) was estimated using a normalized difference vegetation index (NDVI) of satellite data on central Yemen Mountains. A procedure was developed which equated the index to crop coefficients. Evapotranspiration estimates for fields for three dates of Landsat Thematic Mapper data were highly correlated with ground estimates. Service area estimates using landsat Thematic Mapper (TM) and NOAA Advanced Very High Resolution Radiometer (AVHRR) data agreed well with estimates based on National Water Resources Authority (NWRA) gauging station data. Comparisons of ET results with traditional ET models show good agreement. Sensitivity analyses show that the model is accurate even without atmospheric correction. Keywords: Evapotranspiration; NWRA; Yemen; NDVI; TM; AVHRR 1. 0 INTRODUCTION Evapotranspiration (ET) is integral to the hydrological and climatic processes of the Earth and its atmosphere. The current emphasis on global change research makes it important to understand and measure ET and its effects on Earth processes. Much prior research has focused on the microclimatic aspects of ET to develop the theoretical basis for understanding this process. Given the necessity for assessment of large areas in global change research, methods need to be developed that will address regional, continental and global ET measurements in some quantitative manner. Remotely sensed data are suitable to assess large areas, and considerable effort has been made to characterize vegetation using satellite data. Basic relationships exist between spectral reflectance and vegetative characteristics. These relationships allow the use of spectral transforms to define biophysical parameters for plants. The normalized difference vegetation index (NDVI) has been shown to be related to plant canopy variables, which also relate to ET. Ritchie & Burnett (1971), Wiegand et al. (1979), Holben et al. (1980), Tucker (1979), and Weiser et al. (1986) have shown a relationship between NDVI and green leaf area index (LAI), photosyntheticaliy active biomass, and percentage green cover for grass, soybean and corn canopies. The LAI has been related to ET by Stern (1965) for safflower and by Hinkle et al. (1984) for corn and soybeans. Hinkle et al. (1984), using three different plant densities of corn, found that ET was proportional to density only until the LAI reached 2.7. They further showed that at an LAI of 2.7 the crop coefficient became nearly constant until the crop began to senesce. A crop coefficient is an empirical ratio of measured crop ET to