Time-varying and spatially variable potential evapotranspiration (PET, also called evaporative demand) estimates are critical for accurate accounting of Evapotranspiration (ET) in urban and agriculture-intensive regions of the Colorado River Basin, where real-time estimates and short-range predictions of evaporative demand are needed for stakeholder water scheduling. Remotely sensed estimates of PET present an attractive alternative source of input to the current historical PET (i.e., the NOAA-ATLAS climatological pan data) and other ground-based observations due to their higher spatial resolution, independence from ground-based networks, and ability to capture the current dynamics of watershed behavior.
This project is based on the hypothesis that the observational capacities of satellite systems have advanced significantly in the past decade and can now provide more accurate and timely estimates of potential evapotranspiration and other land-surface parameters. The integration of these high spatial and temporal resolution products will improve hydrologic modeling and prediction of snowmelt and streamflow at RFC forecast locations.
The researchers will focus efforts on modifying a Moderate Resolution
Imaging Spectroradiometer (MODIS)-based daily PET time series (previously developed by UCLA: UCLA-MODIS) for use in operational hydrologic modeling systems. The MODIS-based PET currently provides a daily, 250-m product that represents near real-time PET, a key variable in watershed budgets and streamflow predictions. Given the array of available high-temporal resolution GOES-R products, researchers from UCLA and the NWS Colorado Basin River Forecast Center (CBRFC) will adapt the UCLA-MODIS PET algorithm with new (and any proposed) GOES-R satellite products. The developed GOES-derived PET will be evaluated against various products used at the CBRFC, including the historical NOAA-ATLAS (climatological pan) regional values and the NLDAS-derived PET being developed by CBRFC scientists. Evaluation will also include adaptation of the GOES-MODIS PET satellite-based product for integration into the operational rainfall-runoff model (SAC-SMA / SNOW-17) and testing forecast skill at a suite of operational basins in the CBRFC that represent the hydroclimatic variability in the western US. The expected result is the integration of near realtime GOES-R satellite-based products into operational hydrologic forecasting, with advanced products that are better able to capture the observed heterogeneity of mountainous terrain and ET processes at a range of elevations and land cover types.