A Growing Role for Microwave Observations in Estimating Evaporation from Space

There's an increasing need for observation-based estimates of evapotranspiration (ET) that have global coverage, high spatial resolution and a high temporal resolution. Thermal infrared (TIR) approaches can meet two of these three requirements with existing satellites: i.e. high spatial resolution and global coverage at moderate resolution. The third requirement can only be partially addressed with TIR approaches because clouds prevent them from achieving a high (sustained) temporal revisit. This currently necessitates the use of temporal down-scaling methods and their associated assumptions and impact on estimation error. Cloud tolerant passive microwave (MW) retrievals have therefore been proposed to complement existing thermal infrared based ET retrieval approaches. The use of MW observations to retrieve the diurnal changes in land surface temperature is made possible by combining observations from multiple polar orbiting satellites and scaling the long-term aspects of its diurnal response to TIR-based LST products. Although the spatial resolution of the integrated MW-based LST is low (0.25 degree at present), it demonstrates a high tolerance to clouds.

A previous study tested the application of MW-based temperature product for the estimation of ET through a well-known implementation of the two-source energy balance (TSEB) approach, the Atmosphere Land Exchange Inverse (ALEXI). The analysis of that study was restricted to clear sky and showed that at the coarse spatial scale the MW-based ALEXI matches the TIR-based estimates closely in terms of 3-month inter-annual anomalies, providing a basis for further exploring a merger between the MW and TIR approaches. This present study takes the next step and will analyze the performance of ALEXI during cloudy periods, when some of its underlying assumptions have not been tested before. Using tower-based eddy-covariance ET measurements as a reference we test newly developed approaches for fusing TIR- and MW-based ET retrievals under clear to cloudy conditions. The goal of this work is to strengthen the temporal resolution and stability of global ET estimates based on the energy-balance approach, which will allow for more robust statistical analysis in the context of land-atmosphere exchange studies, even if the record length is not extended. Perhaps most importantly, it will also allow for earlier detection of agricultural drought as reflected in the ET-based drought indices.