Evaluating the ability of satellite remote sensing to detect changes in actual evapotranspiration under contrasting field management practices in the Czech Republic

Soil management affects the field water balance, yet quantitative assessments of individual practices remain limited. In this study, the abilities of two remote sensing energy balance models, the Two-source Energy Balance Model (TSEB) and Mapping Evapotranspiration at High Resolution with Internalized Calibration (METRIC), were analyzed to detect changes in actual evapotranspiration (ET a ) using Landsat imagery. These changes were evaluated against contrasting crop management practices in two neighboring fields at the Polkovice site in the Czech Republic. Two eddy covariance (EC) systems provided high-precision field measurements from 2020 to 2023 to allow comparison with the model outputs. Interpolated model outputs captured the EC-measured ET a reduction during the slug infestation period, with similar decreases of 90 mm for METRIC, 92 mm for TSEB, and 87 mm measured by EC. However, the results failed to capture ET a differences between fields observed in the daily EC measurements under the cover crop and shallow tillage treatments. For the cover crop treatment, METRIC underestimated the EC-measured difference by 12 mm and the TSEB-estimated difference by 24 mm during August–October 2020. For shallow tillage, METRIC and TSEB models underestimated the difference by 16 mm and 26 mm, compared with the EC measurements during August and September 2023. These findings suggested that while both models could account for significant differences in ET a between fields, they may struggle to detect subtle management-induced variations, particularly when relying on coarse- to medium-temporal-resolution satellite data. Thus, new high-resolution missions and harmonized multi-platform LST datasets with improved resolution hold promise. • Both METRIC and TSEB captured major actual evapotranspiration (ET a ) field differences. • Models were less effective at detecting subtle changes, such as shallow tillage or cover crops. • The primary limitation of remotely sensed ET a is the low temporal resolution of Landsat imagery. • Models captured higher spatial ET a variability in the no-till field than in the conventional field.