Estimation of Spatially Distributed Evapotranspiration over Wheat Using Thermal Infrared Images and Ground-Based Radiometers

Spatially distributed evapotranspiration (ET) estimation is important for crop stress and assessment and irrigation scheduling. Using remotely sensed thermal infrared data in combination with visible-near infrared data, accurate instantaneous ET estimates are feasible. These estimates, however, need to be extended to daily time steps to have practical value at farm scales. One way to accomplish this extension is to combine hourly image observations of soil and vegetation temperatures. This approach is tested using a one-source energy balance model and data obtained from a series of morning remote sensing flights on 10 March 2005 over a wheat crop planted in Maricopa, Arizona. Aggregate remote sensing ET results from this particular day agreed well with estimates derived from soil moisture observations, but not on a plot-by-plot basis, where modeled ET variability was much less than soil moisture variability. This result is inconsistent with remotely sensed surface temperatures which do show significant spatial and temporal changes. Additional data sets will be investigated to determine how to improve ET modeling with surface temperatures.