Mutual relationships between evapotranspiration and soil water storage in a small agricultural catchment and their consistency from a statistical viewpoint

Accurate measurements and estimates of evapotranspiration (ET) and soil moisture are essential for efficient crop management and understanding of hydrological processes in agricultural catchments. In this study, we used satellite imagery for a small catchment (Nučice, Czech Republic) to retrieve vegetative indices (VIs, including NDVI, SAVI and EVI), hence to analyse their spatial and temporal variability. Furthermore, we investigated the relationship between vegetative indices (VIs), measured evapotranspiration (ET), and soil water storage (SWS). Also, for each variable, we aggregated weekly data at the seasonal temporal resolution, being able to study the trends of each variable’s statistical moments. Moreover, we employed the Normalized Nash-Sutcliffe Efficiency (NNSE) index to study the error and the bias between normalized variables within the same seasonality. We found linear relationships between VIs, ET, and SWS when they exceeded a certain threshold. We were able to estimate the ET by exploiting its linear relationships with VIs and SWS, thus bridging the measurement gap. Our results suggest ET prediction based on VIs can be used during the growing season but may give inaccurate results after harvest (when VIs have low values). SWS can provide a reasonable estimate of the ET when no vegetation is present. Furthermore, the good correspondence between the seasonal NNSE indices and the trends of statistical moments of ET, VIs, and SWS suggest that subsurface processes might be inferred from seasonal vegetation cover. Therefore, this allows us to anticipate the likelihood of seasonal correlations across surrogate variables, further studying the spatial variations of SWS throughout the catchment in connection to ET and VIs.

This research has been supported by the Grant Agency of the Czech Technical University in Prague, Grant No. SGS20/156/OHK1/3T/11 and TUdi project, EU Horizon 2020 Grant Agreement No 101000224.