Intercomparison of Remote Sensing Models for Estimating Actual Daily and Seasonal Evapotranspiration
Abstract:Quantifying consumptive water-use of irrigated agriculture for the Western US is highly important as it accounts for about 85% of the total water withdrawals and 74% of the irrigated acres in the US. Consumptive water-use is mainly based on actual evapotranspiration (ET) estimates for the different agricultural areas. Recent advances in remote sensing have showed its ability to provide spatially distributed estimates of ET at different temporal and spatial scales. Considering that there are presently many different remote sensing of ET models, the U.S. Geological Survey has commissioned a study to investigate some of these models in an effort to develop guidelines and specification for their use by water managers in the Western US. For such this analysis provides results of an inter-comparison of multiple remote sensing of ET models. The models being compared include the Remote Sensing of Evapotranspiration (ReSET), Mapping Evapotranspiration with Internalized Calibration (METRIC), Surface Energy Balance System (SEBS), Operational Simplified Surface Energy Balance (SSEBop), and the Disaggregated Atmosphere-Land Exchange Invers (DisALEXI) models.
These models were tested over irrigated agricultural fields at the Palo Verde Irrigation District (PVID), California, covering a total area of about 500 km2 . Multiple Landsat 5 scenes were used during the summer of 2008 to estimate daily and seasonal actual ET. Estimates of daily actual ET were compared with ground-based Bowen ratio measurements during the Landsat overpass dates. Analysis of water balance is also carried out to highlight the performance of the different models in providing seasonal consumptive water-use.