Evaluating Class A Pan–Based Estimates of Daily Reference Evapotranspiration with Respect to Irrigation Scheduling on Sandy Soils in a Hot Arid Environment

The performance of Class A pan data and six commonly used pan coefficient ($k_p$) approaches were evaluated with respect to the reference evapotranspiration ($E{T}_0$) in a typical farm setup of a hot arid environment. To assess the possible impacts of different microclimates and periods of the year, data from two locations on the farm and two periods of the year were considered. The results show a low agreement between modeled and real $k_p$ with a coefficient of determination ($R^2$) ranging from 0.60 to no correlation. Reference evapotranspiration from pan evaporation ($E{T}_{0,\mathrm{pan}}$) was evaluated, and $E{T}_0$ estimates delivered better results with $R^2$ ranging from 0.94 to 0.90. However, comparing daily $E{T}_0$ based on hourly $E{T}_0$ values with $E{T}_0$ based on daily averages of weather data yielded significant differences in model performance. Furthermore, different periods of the year and different microclimatic conditions also affected model performance and influenced the ranking of the models. Due to the sensitivity of the $k_p$ coefficient models against these impacts, it was not possible to identify a robust “best” model that delivered nonbiased $E{T}_0$ estimations within an acceptable range of accuracy. These findings indicate that the drivers of pan evaporation are not sufficiently reflected in easy-to-measure weather data used by the $k_p$ coefficient models.