Irrigation monitoring from satellite at hyper-high resolution: Paving the way for remote-sensing-based agricultural water management support services

Abstract

Recent advances in satellite retrievals of key hydrological variables have fostered the development of approaches for tracking the irrigation footprint on water resources. Nevertheless, constraints due to the native spatial and temporal resolutions of remotely sensed data still limit the building of supporting systems for agricultural water management relying on Earth Observation. This work aims at filling this gap by applying well-established irrigation mapping and quantification techniques with multiresolution satellite data as input to reproduce irrigation dynamics at the unprecedented spatial sampling of 10 m. Results are validated across different scales of interest for water allocation managers, i.e., from the consortium to the single farm level. The irrigation quantification experiment, carried out through the SM-based (Soil-Moisture-based) inversion approach, provides satisfactory results especially in light of the scarcity of ancillary information for refining the estimates. Percentage errors aggregated at the consortium and the farm scales equal to −24 % and to −14 %, respectively, are found. Such results are achieved without considering losses due to irrigation efficiency, as this information is not explicitly available. The irrigation mapping experiment, carried out by leveraging the TSIMAP (Temporal Stability derived Irrigation MAPping) method, is validated at the farm scale only. An overall accuracy of 93 % is reached, corresponding to two agricultural fields misreproduced as non-irrigated out of the total number equal to twenty-eight. The outcomes of this study show the potential of hyper-high resolution implementations of the considered irrigation mapping and quantification techniques for supporting agricultural water managers, highlighting improvements needed to further meet potential users’ requirements.