Water Level Prediction of Small and Medium-Sized Rivers Using Artificial Neural Networks

Abstract

Typhoons and rainy seasons are typically a frequent occurrence in Korea’s summer climate (June to August); however, the frequency of local heavy rains has increased in recent years owing to climate change. These heavy rains are characterized by substantial rainfall within a short period of time, causing flash flooding. This in turn causes the water level of small and medium-sized rivers with a narrow river channel to increase rapidly, resulting in considerable damage. The rainfall-runoff model is a physical model that is used to predict the river level of most large rivers according to rainfall events, and damage is consequently minimized through flood forecasting and warning. However, data construction on small and medium-sized rivers is insufficient compared to that on large rivers, thereby leading to difficulties in accurate flood forecasting and warning through physical models. This study, therefore, attempts to predict the river level of small and medium-sized rivers using minimal information provided by the Ministry of Environment and artificial neural networks.