Search for a command to run...
Evapotranspiration (ET) mapping at the Landsat spatial resolution (100 m) is essential to fully understand water use and water availability at the field scale. Water use estimates in the Colorado River Basin (CRB), which has diverse ecosystems and complex hydro-climatic regions, will be helpful to water planners and managers. Availability of Landsat 8 images, starting in 2013, provides the opportunity to map ET in the CRB to assess spatial distribution and patterns of water use. The Operational Simplified Surface Energy Balance (SSEBop) model was used with 528 Landsat 8 images to create seamless monthly and annual ET estimates at the inherent 100 m thermal band resolution. Annual ET values were summarized by land use/land cover classes. Croplands were the largest consumer of “blue” water while shrublands consumed the most “green” water. Validation using eddy covariance (EC) flux towers and water balance approaches showed good accuracy levels with R2 ranging from 0.74 to 0.95 and the Nash–Sutcliffe model efficiency coefficient ranging from 0.66 to 0.91. The root mean square error (and percent bias) ranged from 0.48 mm (13%) to 0.60 mm (22%) for daily (days of satellite overpass) ET and from 7.75 mm (2%) to 13.04 mm (35%) for monthly ET. The spatial and temporal distribution of ET indicates the utility of Landsat 8 for providing important information about ET dynamics across the landscape. Annual crop water use was estimated for five selected irrigation districts in the Lower CRB where annual ET per district ranged between 681 mm to 772 mm. Annual ET by crop type over the Maricopa Stanfield irrigation district ranged from a low of 384 mm for durum wheat to a high of 990 mm for alfalfa fields. A rainfall analysis over the five districts suggested that, on average, 69% of the annual ET was met by irrigation. Although the enhanced cloud-masking capability of Landsat 8 based on the cirrus band and utilization of the Fmask algorithm improved the removal of contaminated pixels, the ability to reliably estimate ET over clouded areas remains an important challenge. Overall, the performance of Landsat 8 based ET compared to available EC datasets and water balance estimates for a complex basin such as the CRB demonstrates the potential of using Landsat 8 for annual water use estimation at a national scale. Future efforts will focus on (a) use of consistent methodology across years, (b) integration of multiple sensors to maximize images used, and (c) employing cloud-computing platforms for large scale processing capabilities.