Vector Autoregression Model-Based Forecasting of Reference Evapotranspiration in Malaysia

Evapotranspiration is one of the hydrological cycle’s most important elements in water management across economic sectors. Critical applications agriculture domain include irrigation practice improvement and efficiency, as well resource preservation. The main objective this research to forecast reference evapotranspiration using vector autoregression (VAR) model investigate meteorological variables’ causal relationship with a statistical approach. acquired 20-year, 1-year, 2-month climate datasets from Penang, Malaysia, were split into 80% training data 20% validation data. Public weather are used train initial VAR model. A Raspberry Pi IoT device connected DHT11 temperature sensor was outfitted at designated experimental crop site. In situ acquisition done sensors measure ambient humidity. collected humidity conjunction calculate forecast. results demonstrated that 20-year dataset showed better performance consistent forecasting general evapotranspiration, derived root mean square error (RMSE) correlation coefficient (CORR) 1.1663 −0.0048, respectively. As for 1-year model, RMSE CORR recorded 1.571 −0.3932, However, both positive negative due seasonal effects Penang. ranged between 0.5297 2.3562 2020, 0.8022 1.8539 2019, 2.0921 2018. CORR, it −0.5803 0.2825 −0.3817 0.2714 conclusion, tested estimating (ET0) based on smaller RMSEs demonstrates predicting true values, producing variations