Assessing the accuracy of OpenET satellite-based evapotranspiration data to support water resource and land management applications

Authors: VOLK, J - Desert Research Institute; HUNTINGTON, J - Desert Research Institute; MELTON, F - California State University; ALLEN, R - University Of Idaho; Anderson, Martha; FISHER, J - California Institute Of Technology; KILIC, A - University Of Nebraska; RUHOFF, A - Universidade Federal Do Rio Grande Do Norte; SENAY, G - Us Geological Survey (USGS); MINOR, B - Desert Research Institute; MORTON, C - Desert Research Institute; OTT, T - Desert Research Institute; CARRARA, W - California State University; DOHERTY, CONOR - National Aeronautics And Space Administration (NASA); DUNKERLY, CHRISTIAN - Desert Research Institute; FRIEDRICHS, MACKENZIE - Us Geological Survey (USGS); GUZMAN, ALBERTO - California State University; HAIN, CHRISTOPHER - National Aeronautics And Space Administration (NASA); HALVERSON, GREGORY - National Aeronautics And Space Administration (NASA); JOHNSON, LEE - California State University; KANG, YANGHUI - University Of California Berkeley; Knipper, Kyle; ORTEGA-SALAZAR, SAMUEL - University Of Nebraska; PEARSON, CHRISTOPHER - Desert Research Institute; PARRISH, GABRIEL - Us Geological Survey (USGS); PURDY, A - California State University; REVELLE, PETER - University Of Nebraska; WANG, TIANXIN - University Of California Berkeley; YANG, YUN - Mississippi State University; LAIPELT, LEONARDO - Universidade Federal Do Rio Grande Do Norte; COMINI DE ANDRADE, BRUNO - Universidade Federal Do Rio Grande Do Norte

Submitted to: Nature Water
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/30/2023
Publication Date: 1/15/2024
Citation: Volk, J., Huntington, J., Melton, F., Allen, R.G., Anderson, M.C., Fisher, J., Kilic, A., Ruhoff, A., Senay, G.B., Minor, B., Morton, C., Ott, T., Carrara, W., Doherty, C., Dunkerly, C., Friedrichs, M., Guzman, A., Hain, C., Halverson, G., Johnson, L., Kang, Y., Knipper, K.R., Ortega-Salazar, S., Pearson, C., Parrish, G.E., Purdy, A.J., Revelle, P.M., Wang, T., Yang, Y., Laipelt, L., Comini De Andrade, B. 2024. Assessing the accuracy of OpenET satellite-based evapotranspiration data to support water resource and land management applications. Nature Water. 2:193-205. https://doi.org/10.1038/s44221-023-00181-7.
DOI: https://doi.org/10.1038/s44221-023-00181-7

Interpretive Summary: To advance evidence-based water management strategies in water-limited regions, a crucial piece of information is how much water is needed and how much is being used – by region, by season, and by land-use. The OpenET platform was designed to provide this information, routinely and transparently, across the western United States. The platform hosts six different remote sensing-based models of evapotranspiration (ET), all running as an ensemble at 30-m spatial resolution and daily timesteps, to provide information at spatial and temporal scales relevant to water management at the field to basin scales. In this paper, the accuracy of the models, and the ensemble average ET estimates, are evaluated in comparison with observations from 152 flux sites across the United States. Resultant errors (15.7mm/month) are consistent with accuracy thresholds specified by stakeholders involved in the project. The ensemble average estimate in most cases performed better than any individual model, demonstrating the value of running multiple modeling approaches. This assessment will provide added confidence to organizations, farmers, scientists, and other potential users of OpenET due to the high rigor and transparency of methods that were employed.

Technical Abstract: Remotely-sensed evapotranspiration (ET) data has strong potential to contribute to data-driven approaches for sustainable water management, but practitioners require robust and rigorous accuracy assessments. OpenET was developed to increase access to field-scale (30 m) remotely-sensed ET data for the Contiguous United States (CONUS). We compared OpenET data against data from 152 in situ stations, primarily eddy covariance flux towers, deployed across CONUS. Average monthly mean absolute error for cropland flux sites for the ensemble was 15.7 mm/month (17%), r2 was 0.9, and the mean bias error was -5.6 mm/month (6%) of the mean observed ET. Results for semi-arid to arid shrublands and forested sites demonstrated higher inter-model variability and lower accuracy relative to measurements from cropland sites. High accuracy and multi-model convergence across croplands demonstrates the utility of an ensemble approach, and should give confidence to users of OpenET for agricultural and water resource applications.