Estimation of soil moisture using trapezoidal relationship between remotely sensed land surface temperature and vegetation index

摘要

The trapezoidal relationship between land surface temperature (T_s) andVegetation Index (VI) was used to estimate soil moisture in the presentstudy. An iterative algorithm is proposed to estimate the vertices of theT_s ~ VI trapezoid theoretically for each pixel, and then WaterDeficit Index (WDI) is calculated based on the T_s ~ VI trapezoidusing MODIS remotely sensed measurements of surface temperature and enhancedvegetation index (EVI). The capability of using WDI based on T_s ~ VItrapezoid to estimate soil moisture is evaluated using soil moistureobservations and antecedent precipitation in the Walnut Gulch ExperimentalWatershed (WGEW) in Arizona, USA. The result shows that, the T_s ~ VItrapezoid based WDI can capture temporal variation in surface soil moisturewell, but the capability of detecting spatial variation is poor for such asemi-arid region as WGEW.