STOCHASTIC ALTERNATIVES TO MODELS BASED ON THE DALTON'S APPROACH FOR ESTIMATING MONTHLY PAN EVAPORATION FROM RESERVOIR DAMS IN ALGERIA

Abstract

In this contribution, we assess the capabilities of monthly pan evaporation estimation using two methods: i) the first is a purely stochastic model designed by ARIMA (autoregressive with integrated moving average), which is based only on endogenous variables, and ii) the second is derived from Dalton's physical approach based on exogenous variables. The areas of application of the two methods are two sites of two dams, Ain Zada and Beni Haroun, with distinct climates of Algeria: semiarid and subhumid. The actual evaporations are measured directly from the pans for seven years. The first six years of observations are used for the calibration of the two models, and the last year is reserved for the test. The performance of the models is evaluated using three commonly used metrics, namely, the correlation coefficient (R), root mean square error (RMSE) and mean absolute percentage error (MAPE). According to the results of this research, it appears that ARIMA models perform better than models derived from Dalton’s approach with significantly improved predictive capabilities.