Irrigation quantification through ensemble physical modelling

Agriculture is the largest freshwater user worldwide, averaging 70% of freshwater consumption. Despite this, irrigation represents the most uncertain flux of the water cycle, since it is rarely measured and is difficult to predict because of complex and site-specific water management patterns. In this work, a methodology to quantify irrigation incidence and volume is presented, integrating ground and satellite data within energy-water balance modelling. The proposed algorithm adopts a Montecarlo approach, simulating the impacts of multiple irrigation schedules over selected hydrological variables and extracting the most likely ones by comparing model results with different references. These include a benchmark simulation (fed with observed irrigation), in-situ measurements and satellite observations. The tested hydrological variables were surface soil moisture, deep soil moisture, evapotranspiration and land surface temperature. Of these, the former qualified as the most suited to the algorithm. The procedure was tested at field scale over a variety of Italian case studies, ranging from semi-arid to wet climates, on-demand to turn irrigation, homogeneous to heterogeneous agricultural landscapes, and including small vegetables, high-stemmed and arboreal crops. The results indicated three main conclusions: (1) the algorithm works best with fewer irrigation events (<10), as very frequent events (>2–3 per week) overcrowd the signal; (2) high-ET periods increase ensemble spread and improve the efficacy of the procedure, allowing to better distinguish between different schedules; (3) uncertainty in satellite retrievals of SSM, specifically over heterogeneous agricultural landscapes, negatively influences the algorithm accuracy by muddling the signal coming from the target field.