Multi-scale monitoring of irrigation volumes and evapotranspiration by assimilating satellite data into an energy-water balance model

The agricultural sector is the biggest and least efficient water user, accounting for around 70% of total water use in the Mediterranean region, which is already strongly impacted by climate change with prolonged drought periods, imposing limitation to irrigation water availability. The objective of this study was to develop a procedure for the monitoring irrigation water use and evapotranspiration across different agricultural districts in the Po Valley at 30 m of spatial resolution from 2022 to 2024 and over the entire agricultural area of the basin at 250 m from 2015 to 2024.

The analysis is based on the FEST-EWB model, that computes continuously in time both soil moisture (SM) and evapotranspiration based on the coupling of the energy and water balances. A specific model procedure has been implemented to consider the presence of flooded paddies. The model has been calibrated and validated over non-irrigated areas, against land surface temperature (LST) from downscaled MODIS data at 250m and LANDSAT data at 100 m; Sentinel 1 soil moisture data and local eddy covariance evapotranspiration measurements.

The model has been run using as input the past meteorological forcings (ECMWF ERA5-Land or ground network) and vegetation data from Sentinel2 at 30 m and MODIS at 250 m. Groundwater dynamic was considered from the available groundwater wells from the regional networks.

The actual irrigation volumes have been estimated through the calibrated model implementing three different irrigation strategy: the FAO approach based on SM crop stress thresholds (Allen et al., 1998), the separate and jointly assimilation of satellite LST and SM data to update the modeled fluxes and estimate the irrigation volumes. The different irrigation efficiencies have been considered when modelling the irrigation volumes.

Overall, the results suggested that the yearly total irrigation volumes modeled with the FAO approach are generally underestimated in respect with the observed water allocations. Higher agreement was found when satellite LST or SM are assimilated, but with differences across the years.