Predicting daily reference evapotranspiration rates in a humid region, comparison of seven various data-based predictor models

The reference crop evapotranspiration (ET0) is one of the major components of the hydrological cycle, and its prediction is of great importance in agricultural operations, especially irrigation, of field and horticultural crops. The present study aims to evaluate the performances of two stochastic and machine learning models in predicting ET0 for Mazandaran province, which is one of the most important centers of rice cultivation (as a high-water use plant) in Iran. The studied data belong to 5 synoptic stations in Mazandaran province. They include minimum, maximum, and mean air temperature, minimum, maximum, and mean relative humidity, wind speed, and sunshine duration. These data are received on a daily basis from the Iranian Meteorological Organization during the period 2003-2018. Then, these variables and the FAO-56 Penman-Monteith model are used to calculate daily ET0 rates. Moreover, stochastic models including autoregressive (AR), moving average (MA), autoregressive moving average (ARMA), and autoregressive integrated moving average (ARIMA), and machine learning models including least square support vector machine (LSSVM), adaptive neuro-fuzzy inference system (ANFIS), and generalized regression neural network (GRNN) are used to predict ET0. Predictor inputs include ET0 time lags selected by Autocorrelation Function (ACF) and partial ACF (PACF). The time series models of ARMA and ARIMA, and the machine learning model of LSSVM provide the most accurate predictions with the slight superiority of ARMA and ARIMA over LSSVM in most cases. As a result, it is found that stochastic models are superior to machine learning models due to their more accurate prediction and less complexity. The ARMA model (root mean square error = 0.623mm/day, Wilmott index = 0.962, and R-2 = 86.22%) shows the highest prediction accuracy. The current approach can be applied to predict irrigation water requirements and has research value under similar or different climatic conditions.