Exploring the Potential of Temperature-Based Methods for Regionalization of Daily Reference Evapotranspiration in Two Spanish Regions

Climatological data for 101 stations in Catalonia and Navarre in northeastern Spain were used for comparing estimates of daily reference evapotranspiration (${\mathrm{ET}}_o$) by several temperature-based methods against daily estimates provided by the FAO-56 Penman–Monteith equation. Improved versions of both the Hargreaves–Samani (HS) and the Penman–Monteith temperature (PMT) equations were achieved. Specifically, the Hargreaves coefficient (in the case of the HS) and the empirical coefficient for the estimation of solar radiation (in the case of the PMT) were calibrated for each station and then regionalized by means of quadratic regressions based on the average daily temperature range. The results showed that both the calibrated and regionalized versions of the HS model performed better than the corresponding PMT model versions, with average mean absolute error (MAE) of $0.53\pm 0.16$ and $0.58\pm 0.18\mathrm{mm}·{\mathrm{d}\mathrm{ay}}^{-1}$, for the calibrated and regionalized HS, respectively, versus $0.63\pm 0.17$ and $0.64\pm 0.18\mathrm{mm}·{\mathrm{d}\mathrm{ay}}^{-1}$ for the calibrated and regionalized PMT, respectively. Furthermore, a regionalization approach for the Hargreaves coefficient based on a climate classification led to even better results for the regionalized HS model ($\mathrm{MAE}=0.56\pm 0.17\mathrm{mm}·{\mathrm{d}\mathrm{ay}}^{-1}$). This study highlighted the potential of the calibrated Hargreaves coefficient and its high correlation with the average daily temperature range, allowing reliable regionalization equations, especially over homogeneous climatic regions.