Regional evapotranspiration estimates using the relative soil moisture ratio derived from SMAP products

摘要

Evapotranspiration (ET) is the main water loss in water balance, particularly in arid and semiarid regions. In the last decades, remote sensing technology has been providing valuable information to monitor ET at large scale. In this study, a modification of the relative evapotranspiration (F) parameter introduced by Venturini et al. (2008) is presented. The surface actual water vapor pressure (e(s)) is estimated incorporating surface texture information and the soil moisture (SM) content, as proposed by Deardorff (1978). The introduction of SM content in e(s) estimation led to a physically meaningful estimation of the traditional F formulation. As a result, the original Venturini et al.'s ET application is simpler, more flexible and robust. The results presented here were calculated from the Soil Moisture Active Passive (SMAP) products with two different spatial resolutions 36 and 3 km, to compute regional ET. The proposed ET model was validated using ground observations from Bowen Ratio stations and Eddy Covariance towers. The ET results with a spatial resolution of 36 km showed good agreement with ground data, with a root mean square error varying from 0.91 to 1.22 mm/d. The recently released SMAP SM product at 3 km produces ET estimates with a root mean square error of 0.63 mm/d for autumn-winter season. The new ET model results were also compared with the well-known ET MYD16A2 product, making evident that the proposed approach outreaches the Moderate Resolution Imaging Spectroradiometer (MODIS) methodology. The preliminary results presented with our simple, flexible, universal and robust ET equation using SMAP SM, are promising.