MAPPING EVAPOTRANSPIRATION AND DROUGHT USING MULTI-SCALE THERMAL-INFRARED REMOTE SENSING DATA

Many of the natural and managed ecosystems in the western U.S. are sparsely or partially vegetated during much of the annual growing cycle. Partial vegetation cover conditions present challenges to single-source evapotranspiration (ET) algorithms based on thermal-infrared (TIR) remote sensing because they do not explicitly account for significant differences in atmospheric coupling of the soil and plant components within the thermal pixel. Errors from single-source models are exacerbated when they are applied to coarse resolution TIR data (≥1km), where small-scale vegetation and moisture features in the landscape become unresolved.

In this paper we will discuss a two-source (soil + canopy) Atmosphere-Land Exchange Inverse (ALEXI) surface energy balance model and related flux disaggregation algorithm (DisALEXI) that can be applied reliably over a range in vegetation cover conditions, from bare soil to partial cover and full canopy, and at a range in spatial scales. Techniques are being developed to integrate multi-sensor TIR imagery within this system to generate routine, regional ET maps at both high spatial and temporal resolution for water resource management applications. The system combines hourly 5-10 km resolution TIR and insolation data from the Geostationary Operational Environmental Satellites (GOES) with ~daily/1km and ~bi-weekly/100m resolution TIR images from MODIS and Landsat, respectively. It will be shown that Landsat-scale imagery is critical for capturing water-use dynamics of small hydrologic features such as irrigated fields, riparian buffers, canals and reservoirs. The potential of this modeling system for operational water management and drought monitoring will be discussed.