Evaluation of multivariate linear regression for reference evapotranspiration modeling in different climates of Iran

Abstract The study aimed to evaluate the accuracy of empirical equations (Hargreaves-Samani; HS, Irmak; IR and Dalton; DT) multivariate linear regression models (MLR1–6) for estimating reference evapotranspiration (ET Ref ) in different climates Iran based on Köppen method including arid desert (Bw), semiarid (Bs), humid with mild winters (C), severe (D). For this purpose, climatic data 33 meteorological stations during 30 statistical years 1990–2019 were used a monthly time step. Based various (minimum maximum temperature, relative humidity, wind speed, solar radiation, extraterrestrial vapor pressure deficit), addition 6 three as MLR1, MLR2, HS (temperature-based), MLR3 (radiation-based), MLR4, MLR5 DT (mass transfer-based), MLR6 (combination-based) also estimate evapotranspiration. results these compared using root mean square error (RMSE), absolute (MAE), scatter index (SI), determination coefficient ( R 2 ), Nash-Sutcliffe efficiency (NSE) criteria FAO 56 Penman-Monteith target data. All MLR gave better than equations. showed that simplest model (MLR1) minimum temperature was more accurate lowest highest related equation RMSE 10.8–15.1 mm month ?1 22–28.3 , respectively. Also, among all evaluated equations, radiation-based such Bw Bs MAE = 8.01–11.2 had higher C D 13.44–14.48 . In general, ability excellent from SI < 0.2.