Using EO to Forecast Crop Water Demand for Irrigation Scheduling

Photo of an irrigation system over farmland

Project Summary

This project aims to harness the large trove of satellite and agrometeorological datasets to forecast crop water demand at time scales relevant for irrigation scheduling in California, Utah, and New Mexico. Dr. Ghezzehei is taking a machine learning-enabled approach to develop tools for (a) mapping soil and crop water status by synthesizing ground-based soil and plant sensing with satellite EO (Landsat, HLS) and, (b) predicting actual and reference ET using historical and predicted weather and ET datasets (Melton et al., 2021). These approaches will build upon recent developments including predicting soil moisture of complex, heterogeneous landscapes using machine-learning and UAV data (Araya et al., 2021) and prediction of net-radiation using forecasted temperature (Yan and Ghezzehei, in prep).

Study Area:

California | New Mexico | Utah

Earth Observations Used

OpenET | Landsat | HLS

Lead Institution

Project Lead

Teamrat Ghezzehei

UC Merced