Abstract
This study develops hourly water level forecasting models with lead-times of 1 to 3 h using an artificial neural network (ANN) for Anyangcheon stream, one of the major tributaries of the Han River, South Korea. To consider the backwater effect from this river, an enhanced tributary water level forecasting model is proposed by adding multiple water level data on the main river as input variables into the conventional ANN structure which often uses rainfall and upstream water level data. Four types of ANN models per each lead-time are built with increasing complexity of the input vector, and their performances are compared. The results indicate that the inclusion of multiple water level data on the main river to the network provides water level forecasts with greater accuracy at the Ogeumgyo gauging station of interest. The final best ANN models for water level forecasts with lead-times of 1 to 2 h show good performance with root mean square errors (RMSE) below 0.06 m and 0.12 m, respectively. However, the final best ANN model for forecasting 3 h ahead was unsatisfactory, showing underestimation at many rising parts of the hydrograph.
Highlights
Flood forecasting in rivers is an essential non-structural countermeasure against flood disasters [1].For this purpose, flood forecasting is usually performed using process-based models, which require basic data collection and periodic data updating
Since urban floods frequently occur in tributaries of the Han River, South Korea, in which flood water level rises due to the backwater effect, accurate and efficient methodology to predict water levels is very important in producing countermeasures against flood disasters
(1) artificial neural network (ANN) models were constructed for forecasting water levels with lead-times of 1–3h by using the input data, areal average rainfall, water level data at the station of interest, water level data at an upstream gauging station, and water level data at multiple stations located near the confluence of the tributary, Anyangcheon stream, and the main river, the Han River
Summary
Flood forecasting in rivers is an essential non-structural countermeasure against flood disasters [1]. Adding effect from thelevel main stream into theisconventional rivers that include this In the this backwater study, an enhanced water prediction model proposed by training process of model in which rainfall and upstream water level data are used. This study demonstrates is included in the model construction To this end, models for hourly water level that forecasts can be improved if water level data in the main river is included in the ANN model forecasting areTodeveloped and their performance investigated for Anyangcheon stream,and a major construction. Models for hourly water level that forecasts can be improved if water level data in the main river is included in the ANN model forecasting areTodeveloped and their performance investigated for Anyangcheon stream,and a major construction This end, ANN models for hourlyiswater level forecasting are developed their tributary of the. Performance is investigated for Anyangcheon stream, a major tributary of the Han River, South Korea
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
