Machine Learning and Conventional Methods for Reference Evapotranspiration Estimation Using Limited-Climatic-Data Scenarios

Abstract

Reference evapotranspiration (ET0) is one important agrometeorological parameter for hydrological studies and climate risk zoning. ET0 calculation by the FAO Penman–Monteith method requires several input data. However, the availability of climate data has been a problem in many places around the world, so the study of scenarios with different combinations of climate data has become essential. The aim of this study was to evaluate the performance of artificial neural network (ANN), random forest (RF), support vector machine (SVM), and multiple linear regression (MLR) approaches to estimate monthly mean ET0 with different input data combinations and scenarios. Three scenarios were evaluated: at the state level, where all climatological stations were used (Scenario I–SI), and at the regional level, where the Minas Gerais state was divided according to the climatic classifications of Thornthwaite (Scenario II–SII) and Köppen (Scenario III–SIII). ANN and RF performed better in ET0 estimation among the models evaluated in the SI, SII, and SIII scenarios with the following data combinations: (i) latitude, longitude, altitude, month, mean, maximum and minimum temperature, and relative humidity and (ii) latitude, longitude, altitude, month, mean temperature, and relative humidity. SVM and MLR models are recommended for all scenarios in situations with limited climatic data where only air temperature and relative humidity data are available. The results and information presented in this study are important for the agricultural chain and water resources in Minas Gerais state.

Keywords

• artificial neural network
• random forest
• support vector machine
• multiple linear regression
• crop water requirements
• meteorological data