Use of ECOSTRESS data for measurements of the surface water temperature: Significance of data filtering in accuracy assessment

The ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) program was launched in 2018 and provided the scientific community with extremely valuable surface temperature data distributed with a fine resolution of 70 m. This global data can be used in environmental, engineering and many other fields to measure surface temperature in hard-to-reach places or add data to discrete temperature measurements from the field. One of the aspects of data use by the broad scientific community is a clear methodology and understanding of potential error between ECOSTRESS sensor data measured by radiometer and surface temperature measured by conventional temperature sensors.This study uses three years of ECOSTRESS data (2018–2021) and surface water temperature data collected by the New York City Department of Environmental Protection on water supply reservoirs over the same time period. Data between two sources were matched spatially and temporarily. We developed a methodology of data filtering using temporal constraints and ECOSTRESS auxiliary information on clouds and signal quality indicators (flags) supplied in NASA provided documentation.The results of the study showed that the successful use of ECOSTRESS measurements requires data filtering using flags and temporal constraints. This might cause considerable reduction of original ECOSTRESS data and the number of measurements available for analysis. In our experimental settings, use of flags and temporal constraints reduced the original dataset with 442 records down to only six records; for this dataset the average temperature difference between ECOSTRESS data and measured data was found to be 2.6 °C. This number might vary according to local environmental conditions affecting surface water temperature. The proposed methodology signifies its importance in assessing the accuracy in estimating surface water temperature.