Daily High Spatial Resolution Evapotranspiration Estimation Using Multi-Satellite Data Fusion Approach in Agricultural and Forested Sites in the U.S.

Abstract:

Evapotranspiration (ET), as a major part of the water balance, is a key indicator of vegetation stress and also represents various types of water usage strategies. High spatial and temporal resolution ET mapping can provide detailed information about daily vegetation water use and soil moisture status at finer scales, which is important to water management and vegetation condition monitoring. This research employs a multi-scale ET modeling system which is based on the two source surface energy balance (TSEB) model. We discuss the utility of applying this modeling system over an irrigated agriculture area in California and a forested site in North Carolina.

The multi-scale ET modeling system integrates the Atmosphere-Land Exchange Inverse model and associated disaggregation scheme (ALEXI/DisALEXI) and fuses the ET estimations from both MODIS (1km, daily) and Landsat (30m, bi-weekly). The Spatial and Temporal Adaptive Reflective Fusion Model (STARFM) is used to retrieve high spatial and temporal resolution ET. A Data Mining Sharpener (DMS) methodology is used in the system to sharpen the native Landsat thermal infrared band (TIR) to 30m resolution. Comparing with Landsat only ET retrievals, this ET modeling system can optimize the usage of multi-satellite data, which are in different temporal and spatial resolution, to maximize the utility of high spatial and temporal ET estimation.

Daily high spatial resolution ET retrievals are compared with observations from local flux towers. Determining how model output of daily water use information can be employed in irrigation and forest management applications will be discussed.