Improving Irrigation Water Management along the Lower Colorado River, USA, Using Sentinel 1 & 2

Irrigated farms are major users of Colorado River water and are important contributors to the U.S. Southwestern economy. However, extreme drought and over-allocation of river supplies now threaten their livelihood and sustainability. One way to improve effective and efficient water use by these farms is by using satellite-based remote sensing. While these observations have been known and reported for decades, their utility has been transformed. Greatly improved spatial resolution, spectral sampling, and re-visit periodicities, are now enabling close to real time monitoring of crop growth and irrigation events. In this study we use both conventional and less conventional indices (blue and red-edge bands) constructed from 10 and 20 m Sentinel 2 observations to detect irrigation events, infer crop emergence, and estimate fractional crop cover. Irrigation events, and notably pre-season irrigations critical for salt management, are also detected using 10-m Sentinel 1 GRD image data. Combined with local weather data, crop coefficients, and estimates of crop type, these data sets make possible the identification of all crop stages and enable estimation of water from day of planting to end-of-season. The methodology and results from this approach will be demonstrated for several important crops including lettuce, wheat, and alfalfa, for irrigated farm sites supplied by the Colorado River, near Yuma, Arizona. Validation of daily and seasonal ET estimates will be shown using eddy covariance data collected between 2017-2021.