An Expansion of the Ungaged Pan Evaporation Using Neural Networks Model in Rural Regions, South Korea

The generalized regression neural networks model (GRNNM) embedded genetic algorithm(GA) is developed and applied to expand the ungaged or the missing pan evaporation (PE) in rural regions of the north Gyeongsangbuk-do such as Yeongju, Bongwha, and Andong stations respectively, South Korea. The Songriwon dam with small-sized reservoir, which is located near to Yeongju and Bongwha station, will be constructed for flood control and water supply and it will damage many kinds of specific and traditional crops such as Korean ginseng, the red pine mushrooms and honey-sweet apples and so on. Therefore, reliable evaporation data will be necessary to prevent crops damages and control water transfer as well as to reduce the drought mitigation. The developing processes of the GRNNM-GA consist of the three major parts such as the training, the testing, and the reproduction performance respectively. From this study, the reliable expansion system of PE is constructed and suggested the basic data for irrigation and drainage networks in rural regions, the north Gyeongsangbuk-do of South Korea.