Assessment of reference evapotranspiration using remote sensing and forecasting tools under semi-arid conditions

Reference evapotranspiration ( ET o ) assessment at regional scale is a challenge to agricultural and hydrological research requiring the characterization of vast areas to obtain accurate meteorological variables. To account for this, a new approach considering a modified version of the Makkink equation and based on a combination of remotely sensed solar radiation ( R s ) considering the MSG satellite, and numerical weather forecast of near surface air temperature ( T 2 m ), was developed. The new approach is referred as Makkink-Advection equation (MAK-Adv). Once R s and T 2 m were validated for the semiarid conditions registered in Southern Spain with very satisfactory results (RMSE = 1.6 MJ m −2 d −1 and 1.4 °C respectively, with linear regressions with slope = 0.99 and R 2 = 0.97), the MAK-Adv approach was calibrated and validated with ET o measurements from a weighing lysimeter under near-reference conditions in Southern Spain. The new approach has provided accurate daily ET o estimations compared with lysimeter data (RMSE = 0.50 mm d −1 , RE = 12.3% and MAE = 0.39 mm d −1 ), and remains accurate during summer time (RMSE = 0.62 mm d −1 and RE = 9%), the critical period when accurate ET o data are essential for a correct agricultural water management. Once calibrated and validated, and thanks to the use of remotely sensed data, the MAK-Adv approach allowed the ET o assessment at regional scale. For a semi-arid region located in southern Spain with an area of 87,300 km 2 , a detailed spatial ET o assessment was performed, determining ET o values for the totality of the area with a resolution of about 9 km 2 allowing to evaluate the spatial and temporal variability at basin/region scale. This new approach gets over the serious limitations for a correct ET o assessment when traditional methodologies are not able to provide these data for vast areas, even with the modern weather station networks. Thus, MAK-Adv approach is especially useful for those areas with missing or inaccurate meteorological in situ observations.