Abstract
STUDY REGION: Jucar River System (Spain) and Sicily Island (Italy). STUDY FOCUS: Penman-Monteith crop reference evapotranspiration (PM-ETo) is critical for irrigation planning and hydrological modeling. Its estimation typically requires dense agricultural weather networks with automated stations. Alternatively, reanalysis datasets like ERA5-Land and AgERA5 offer spatially comprehensive data, but their resolution is often insufficient. Spatial interpolation techniques are thus required to estimate PM-ETo at unsampled locations. This study applied the DRI (Dynamic Regression-Based Interpolation) algorithm to generate high-resolution (100 m) PM-ETo maps for both regions using three data sources: meteorological station records and ERA5-Land and AgERA5 reanalysis products. The performance of AgERA5 for PM-ETo estimation was also assessed. Additionally, PM-ETo interpolated maps from the three sources were compared. NEW HYDROLOGICAL INSIGHTS FOR THE REGION: AgERA5, a bias-corrected downscaling of ERA5, effectively removed bias in Sicily when compared to in situ data, but not in the Jucar system. Nonetheless, AgERA5 outperformed ERA5-Land in both regions for PM-ETo estimation. Following interpolation, the resulting maps retained the same biases identified in the original datasets and preserved the frequency distributions of ground-truth maps. This indicates that the interpolation method does not distort the underlying meteorological fields between stations. The proposed approach offers a valuable tool for practitioners and modelers, enabling the generation of high-resolution, accurate, and practical PM-ETo maps to support irrigation planning and hydrological applications.