Expanding the applicability of remotely sensed soil evaporation transfer coefficient to the estimation of soil moisture over bare soil and partially vegetated areas

Soil evaporation transfer coefficient (h_a) is an effective means of estimating surface soil moisture from moderate-resolution imaging spectroradiometer (MODIS) imagery. This coefficient is a function of three variables: air temperature, land surface temperature and dry soil temperature. The first two variables can easily be obtained from different sources, whereas dry soil temperature cannot be determined as easily as the other ones, particularly over partially vegetated areas. In this paper, to enhance the capability of h_a in estimating soil moisture, we propose to use the combination of land surface temperature (LST) and normalized difference vegetation index (NDVI) to estimate dry soil temperature over partially vegetated areas as well as bare soil areas; this combination is known as LST-NDVI feature space in the literature. The underlying assumption of the proposed method is that at any given pixel over partially vegetated or bare soil areas, dry soil temperature is approximately equivalent to the maximum LST derived from LST-NDVI triangle space. The results showed that calculating h_a using dry soil temperature derived from the triangle space can result in more reliable estimation of soil moisture over bare soil and partially vegetated areas.