Post-fire evapotranspiration estimates in ground truth limited environments

Wildfires are catastrophic events with increasing incidence and severity, especially in Mediterranean climates like California. Wildfires can drastically change the hydrologic components of a watershed and require consistent and reliable data to monitor recovery. With the upcoming retirement of the Moderate Resolution Imaging Spectroradiometer (MODIS), a reliable evapotranspiration (ET) product at the subcatchment scale is necessary to monitor longitudinal ET, particularly post-fire. We evaluate NASA JPL’s ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS), MODIS, and the North American Land Data Assimilation System phase 2 (NLDAS-2) to assess performance in longitudinal evapotranspiration (ET) estimates and post-fire ET change detection using correlation analysis and linear regression for twelve subwatersheds located within the August Complex Fire (August 2020) in Trinity County, the Delta Fire (September 2018) in Shasta and Trinity Counties, and the Creek Fire (September 2020) in Fresno County for a four-year period. We found that MODIS and ECOSTRESS showed the strongest correlation in daily ET estimates (mean r = 0.56), while correlations between ECOSTRESS-NLDAS-2 (r = 0.06) and MODIS-NLDAS-2 (r = 0.30) were notably weaker. However, ECOSTRESS data availability constrained post-fire change detection and annual ET sum assessments. Its higher spatial resolution and correlation with MODIS have potential for finer-scale model parameterization of post-fire landcover changes, particularly vegetation loss in place of MODIS. NLDAS-2 can be useful in monthly and annual patterns at larger scales. By comparing ECOSTRESS with established platforms of ET estimation, we validate ECOSTRESS as a tool for post-fire ET monitoring, especially in areas where ground truth data is unavailable. These findings can allow for more informed and effective land management decisions, particularly in wildfire and vegetation monitoring.