Predicting the Effect of Temperature Changes on Reference Evapotranspiration by Means of Time Series Modeling (Case Study: Khorramabad Basin)

Global warming phenomenon has affected the hydrologic balance, especially in the arid and semi-arid regions of the world. Therefore, it seems necessary to study these effects to achieve better water resources management system. In this study, maximum and minimum temperature information at the period of ۱۹۹۲-۲۰۱۷ of Khorramabad synoptic station were assessed. the changes of these two characteristics and reference evapotranspiration of Khorramabad plain were investigated by time series analysis. MSE, RMSE and R۲ indices were used to validate the models. The results has shown that both maximum and minimum temperature series are static and abnormal, so for normalization, the square root for the minimum data and the squared conversion for the maximum temperature data were used. The ACF chart of both series reaches its local peak at time intervals of multiples of ۱۲, indicating a seasonal trend with a period of ۱۲ months. Finally, the ARIMA model (۰,۰,۴) (۰,۱,۱) for the minimum temperature and the ARIMA model (۰,۰,۱) (۰,۱,۱) for the maximum temperature were the best chosen models. The values of R۲, RMSE and MSE for the selected maximum temperature model were ۰.۹۷۱, ۱.۶۵۶ and ۰.۹۹۱, respectively, and for the minimum temperature model ۰.۹۶۵, ۱.۳۰۴ and ۰.۹۹۱, respectively, which indicates the acceptable accuracy of the proposed models. Forecasts indicate an increase in the minimum and maximum temperatures in the whole future period compared to the base period. The peak of this increase occurs in June, July and August for the minimum and maximum temperatures respectively for Tmin: ۲.۰۳, ۱.۵۴, ۱.۷۵, and for Tmax: ۱.۹۱, ۲.۰۳, ۱.۷۷ Celsius. In the next period, the reference evapotranspiration will increase on average compared to the base period, with most of this increase occurring in March, April, and May.