Estimating Latent and Sensible Heat Fluxes using the Temperature Vegetation Dryness Index and MODIS Data

In this work, daily latent heat (lambdaE_d) or evapotranspiration and sensible heat (H_d) fluxes were estimated from the TVDI (Temperature Vegetation Dryness Index) modified to account for climatic gradients (TVDI_t) using MODIS data in a Spanish region (Andalusia) with strong bioclimatic gradients. The TVDI_t was correlated (R=0.90) with field measured AWC (available water content). When reinterpreting TVDI_t as the ratio between actual and potential evapotranspiration to estimate surface energy fluxes, model vs. eddy covariance data from 2 semiarid sites were reasonable for H_d (R=0.94: RMSE=25.89 Wm^-2). For lambdaE_d RMSE was low (19.46 Wm^-2) but correlations were only significant after excluding summer dates (R=0.71) when transpiration is very low. Model accuracy is currently limited by the accuracy in Rn_d estimates (<27 Wm^-2) for which results at humid sites should be better than at semiarid sites.