Hydrological Response to the 2011 Drought in Texas Using Land Surface Modeling, Remote Sensing, and GRACE
Résumé
Texas was subjected to the most extreme one-year drought on record in 2011, which had a tremendous impact on water resources statewide. This study aims to quantify evapotranspiration (ET) from land surface models (LSMs), remote sensing, and GRACE during the drought. Uncertainties in ET output from four LSMs, i.e., Noah, Mosaic, VIC, and SAC in NLDAS-2, two remote sensing-based products, i.e., MODIS and AVHRR, and GRACE-derived ET as a residual in the water budget (ET = P - R - ΔTWS) based on precipitation (P) from PRISM, monitored runoff (R), and total water storage (TWS) change from GRACE satellites were quantified using the three corner hat method that does not require a priori knowledge of the true value of ET. Water budgets were calculated using the traditional flux approach and a new storage approach in combination with the different ET products and GRACE TWS. The analyses were conducted using data from three river basins (humid - arid) primarily in Texas as case studies. Remote sensing-based ET shows markedly higher magnitudes during drought but significantly lower magnitudes at other times, particularly during wet periods than land surface model-based ET. Overestimation of ET during drought would result in overestimation of soil moisture depletion and much longer projected times for drought recovery. Uncertainties in ET are lowest in LSM ET (~5 mm/month), moderate in remote sensing MODIS- or AVHRR-based ET (10 - 15 mm/month), and highest in GRACE-based ET (20 - 30 mm/month). Uncertainties in total water storage changes from the water budget approach (ΔTWS = P-R-ET) are about half of uncertainties in GRACE-derived TWS changes for each of the basins. Future ET estimation should consider a hybrid approach that integrates LSM and satellite-based products to constrain uncertainties.