Combining Remote Sensing and Weather Forecast for Crop Evapotranspiration Estimation Under Different Irrigation Strategies

Accurate a priori estimation of crop water demand in terms of crop evapotranspiration (ETc) plays a key role in precision agriculture. This study assesses the effect of different irrigation strategies (full irrigation versus deficit irrigation, DI - including partial root-zone drying and regulated DI) on orange trees crop ETc obtained by combining remote sensing (RS) and weather forecast in Southern Italy. Spatially distributed daily ETc estimates were calculated using the ArcDualKc model that, based on FAO-56 dual crop coefficient (Kc), computes the basal crop coefficient (Kcb) and the soil evaporation coefficient (Ke). Sentinel-2 satellite images and forecast weather data provided by the limited-area prediction model of the Consortium for Small-scale Modelling (COSMO) were used as inputs for ETc estimates. Results showed slight differences between the full and DI treatments, with Kcb values of 0.64 (±0.07) for full irrigation and Kcb values varying from 0.60 to 0.63 (±0.07) in water deficit conditions. No substantial differences were observed in terms of Ke among the irrigation treatments, with Ke equal to 0.06 (±0.06). The reductions observed in Kcb in the DI treatments resulted in lower ETc (mm d‑1) rates with respect to full irrigation (up to 6%), with average values of 4.83 (±1.66) versus 4.56-4.74 (±1.61), respectively. To conclude, the adopted approach provides a valuable way for identifying the effects of different irrigation strategies at the field scale. However, the results might be influenced by the heterogeneity of Sentinel-2 pixels, which contain both vegetated and bare soil surfaces, highlighting the eventual need of using high-resolution images, such as those provided by UAV systems.