Evaluating sensitivity of PMF-56 and Hargeaves reference evapotranspiration models to maximum and minimum air temperatures in humid regions of Iran

The main aim of this study was to evaluate sensitivity of ET0 values estimated by Penman-Monteith FAO 56 (PMF-56) and Hargreaves (HG) models to the mean maximum (Tmax) and mean minimum air temperatures (Tmin) in humid regions of northern Iran. For this purpose, a 50-year meteorological quality controlled data from 6 synoptic stations were applied and their daily ET0 values were estimated by PMF-56 and HG methods for the same period. The average Tmax and Tmin values were then altered from 2 to 10 percents under various scenarios and the corresponding ET0 values were recalculated and compared with the initial ET0 values. The results showed that the maximum relative increase of ET0 values is observed in summer months by 10% increase in Tmax. In general, the sensitivity of both ET0 models to the maximum temperature was higher than minimum temperature. In contrast, a 10% increase in Tmin during summer months caused up to 7% drop in ET0. It is shown that in small temperature variations (less than 8%), the response of the ET0 models is linear. However, higher temperature variations (more than 10%) caused non-linear results. The comparison of HG and PMF-56 results revealed that the sensitivity of PMF-56 aerodynamic model to the Tmin and Tmax flactuations was more realistic than the HG model, because of considering outgoing long-wave radiation and vapor pressure deficit in PMF-56 model. The results indicated that the sensitivitiy of the HG and PMF-56 models to the temperature parameters depend on the geographical location and the season. When the effect of teleconnection events (ENSO, NAO, MJO) on ET0 variability is studied, the application of Hargreaves model is not recommended.