A two-soil-layer agro-hydrological model for irrigation requirements: Evaluation and trends in the Po River basin

Increasing hydrological and agricultural droughts in the Po River Basin in northern Italy, one of the most intensively irrigated agricultural regions in Europe, have raised growing concerns about water resources management and highlighted the need for accurate estimates of irrigation requirements (IR). This study presents AGRHYD, a two-compartment (topsoil–subsoil) soil-water balance model built on the FAO 56 dual crop coefficient approach to estimate crop actual evapotranspiration (ETc act), crop evapotranspiration (ETc) and IR at point and regional scales. The explicit topsoil compartment, responsible for soil evaporation, facilitates future reconciliation and integration with satellite-based soil moisture retrievals. Since IR is a theoretical construct and cannot be directly measured, the model’s daily ETc act estimates for wheat and vine, two of the typical rainfed crops in the region, were evaluated against two independent satellite-based evapotranspiration products, WaPOR and GLEAM yielding Pearson’s correlation coefficient (r) up to 0.88 and an RMSE as low as 0.91 mm day⁻¹ for wheat, and r up to 0.91 and an RMSE as low as 0.58 mm day⁻¹ for vine, with satisfactory reproduction of ETc act temporal dynamics at daily, seasonal, and annual scales. IR was subsequently estimated for maize and forages, two of the main irrigated crops in the region, over the period 1981–2022, revealing pronounced spatial heterogeneity with the highest demands concentrated in eastern Piedmont and Emilia-Romagna. Long-term trend analysis identified statistically significant increasing trends in annual IR across the basin with mean rates of 0.88 mm year⁻¹ for maize and 1.17 mm year⁻¹ for forages, driven primarily by trends in June and July. Concurrent declining precipitation and increasing crop reference evapotranspiration (ET₀) across the basin during these months broadly provide a physically consistent hydroclimatic explanation. A sensitivity analysis confirmed that the key trend findings are robust to uncertainty in crop calendar timing. The resulting IR dataset and long-term trend diagnostics constitute a valuable resource for irrigation planning and water allocation in the Po River Basin, while the model's reliable performance and modest data requirements make it transferable to other irrigated regions.