Enhanced Resolution of Evapotranspiration from Riparian Systems and Field Edges by Sharpening the Landsat Thermal Band

Evapotranspiration (ET) is the major consumptive use of irrigation water, and thus, spatial and temporal quantification of ET is important to agricultural water management. In water rights management and precision irrigation, information on ET is desirable on a field or subfield scale. Therefore, fine resolution satellite imagery and ET products derived from the satellite imagery are highly desirable. However, for a majority of satellites, spatial resolution of the longwave (thermal) band(s) is coarser than the coincident shortwave bands, which creates a compatibility and correspondence issue for the data used for energy balance (EB) and increases the net pixel resolution and reduces fidelity of ET calculations. Typically, surface temperature (T_s) and vegetation indices (VI) are closely correlated, especially under conditions of high availability of soil water, due to the effects of evaporative cooling by vegetation, especially some days after wetting events, when the exposed soil surface is dry and warm relative to vegetation. This physical relationship can be exploited to sharpen T_s using VI if the assumption of cooled surface by vegetation holds. This paper describes a technique developed for application with the METRIC and SEBAL EB procedures that uses the concept of hot and cold thermal conditions associated with dry and wet surface conditions during calibration as a means to distribute T_s at a subpixel scale using subpixel VI. The result is an image of T_s that has the same spatial resolution of the short wave images. Therefore, the resolution of ET images created during the METRIC or SEBAL processes can have equally high resolution. Applications made in Idaho and New Mexico using sharpened thermal imagery are described.