Estimation of the Reference Evapotranspiration Using Neural Networks Model and Limited Climatic Variables

The generalized regression neural networks model (GRNNM) embedded genetic algorithm (GA) is developed and applied to estimate the alfalfa reference evapotranspiration (ET_r) in rural regions of the north Gyeongsangbuk-do such as Yeongju, Bongwha, and Andong station respectively, South Korea. Since the observed data of ET_r using the lysimeters have not existed in this region, the reliable data of ET_r will be necessary to prevent crops damages as well as reduce the drought disaster mitigation. The developing processes of the GRNNM-GA consist of the two major parts such as the training and the validation performance respectively. From this study, the reliable data of ET_r is constructed and suggested the reference data for irrigation and drainage networks system. It is possible to construct the reference evapotranspiration estimation system (RETES) in rural regions, the north Gyeongsangbuk-do of South Korea.