A Multi-Sensor Methodology for Mapping Daily Evapotranspiration At High Resolution.

The use of remote sensing-based surface energy balance models to determine spatially distributed evapotranspiration (ET) is considered one of the few viable means for mapping ET from field to regional and ultimately global scales.  A key remotely sensed boundary condition used in many of these models is land-surface temperature (LST) data. However, the current generation of satellite missions does not include sensor characterized by both high spatial resolution (< 100 m) and high repeatability (daily). The new generation of sensors, such as the one proposed for the HyspIRI mission, certainly would improve the availability of high frequency data, but the fact that this mission is not planned for launch in the foreseeable future, alternative approaches need to be considered. One possible methodology to address this problem is the integration of data provided by different sensors. In this case study, the use of a multi-scale, multi-sensor approach is analyzed in which the coarse resolution GOES data (10-km) are used to map ET at the U.S. continental scale via the Atmosphere-Land EXchange Inverse (ALEXI)  model, while the Dis-ALEXI procedure disaggregates ALEXI ET to finer spatial scales using MODIS 1-km (daily) and Landsat 30-m  (16 day) sharpened LST imagery. These latter estimates are fused in order to retrieve daily 30-m ET maps over  rain fed and irrigated agricultural regions in the U.S. A comparison between the estimates obtained by temporally interpolating the Landsat data in isolation and maps derived by the fusion approach is used to quantify the value added in the introduction of coarse resolution/high frequency estimates in the modeling framework, especially when the theoretical frequency is limited by current availability of high spatial and temporal resolution LST data. This methodology, designed to use the Landsat data,  is suitable for exploiting what might be achieved with a HyspIRI type satellite mission.