Prediction of daily crop reference evapotranspiration (ET o) values through a least-squares support vector machine model

Abstract

Real-time prediction of daily reference crop évapotranspiration (ET o) is the basis for estimating crop évapotranspiration and for computing crop irrigation requirements. In recent years, least-squares support vector machines (LSSVMs) have been applied for forecasting in many fields of engineering. In this paper, LSSVMs are applied to forecast ET o using public weather forecasting data (minimum and maximum temperature, average relative humidity, wind scale and weather conditions). LSSVM-estimated ET o is compared with Penman-Monteith (PM)-estimated ET 0 using measured meteorological data. Based on a comparison between LSSVM and PM over a 2 month period, the results show that the root mean square error and mean absolute error are less than 0.5 and 0.4 mm d -1 respectively, and that the model efficiency is greater than 90%. This indicates that ET 0 can be successfully estimated using public weather forecasts through the LSSVM approach. © IWA Publishing 2011.