Flooding from torrential rain occurs in a short amount of time, while drought lasts for a longer period; the former may inflict huge losses in terms of both life and property. For these reasons, considerable research has been performed in the field of flood control system development. A physical model is mainly used for flood forecasting and warning. However, physical rainfall-runoff models for the conventional flood forecasting process require extensive information and data, and include uncertainties that can accumulate errors during the modeling process. On the other hand, ANFIS, which is a data-driven model combining the neural network and fuzzy techniques, can decrease the amount of physical data required for the construction of a conventional model and easily construct and evaluate a flood forecasting model using only rainfall and water level data. However, data-driven models have the disadvantage that they do not provide mathematical and physical logic, so that there are no logical correlations between the input and output data of the model. This study analyzes the characteristics of a data-driven model, ANFIS, according to its functional options and input data, such as changes in the clustering radius and the training data length. In addition, the suitability of ANFIS is evaluated through comparison with the results of HEC-HMS, which is widely used for rainfall-runoff models. In this study, the neuro-fuzzy technique is applied to the Cheongmicheon Basin using the observed precipitation and stream level data from 2008 to 2011.

No abstract

No abstract

In the conventional flood forecasting process, a rainfall-runoff model is used to predict runoff at a specific location. However, the process of determining the required parameters for the model is sometimes very complicated and requires extensive information and data. In addition, considerable amount of uncertainties may be included during the parameter estimation processes. Errors can occur during the pre-processing and main processing stages of the modeling, and errors from each step accumulate into the model result. In this study, a neuro-fuzzy technique is used to minimize the amount of uncertainties included in a conventional flood forecasting model for more accurate forecasting of floods. The adaptive neuro-fuzzy inference system (ANFIS), which is a data-driven model that combines a neural network and the fuzzy technique, can decrease the amount of physical data required for constructing a conventional model. By using only rainfall and water level data, ANFIS can easily construct and evaluate a flood forecasting model. Furthermore, the model construction process is relatively simple, and reliable results can be efficiently obtained in a reasonably short time once the model is developed. The developed model is applied to the Tancheon basin in Korea. The water level at the Daegok Bridge, which is located downstream of Tancheon, is forecast by the neuro-fuzzy method. The applicability and suitability of the model are studied by comparing the result with the observed stream level data from 2007 to 2011 in the Tancheon basin area. Tancheon is a tributary of the Han River and begins from the city of Yongin in Gyeonggi-do. It has a total length of 35.6 km and an area of 302 km 2 . The water level data from t + 1 to t + 18 is estimated by ANFIS using 10-min interval data. The results showed that the average height error was 24.48% and the average RMSE was 0.367 m.

With exacerbating climate change, the current reservoir storage capacity in South Korea is insufficient to meet the future scheduled water demand. No study has yet evaluated the effects of applying the water supply adjustment standard (Standard) and activating the reservoir emergency storage in response to extreme drought. The main objective is to assess the effects of applying Standard and activating emergency storage in meeting the water demand under extreme drought at six multipurpose reservoirs (Andong, Gimcheon-Buhang, Gunwi, Hapcheon, Imha, and Milyang) in the Nakdong River Basin, South Korea. We built a reservoir simulation model (HEC-ResSim), determined the extreme drought scenarios, and emergency storage capacity. We evaluated three reservoir operation cases (general operation, regular Standard, and revised Standard) from 2011 to 2100. The results show that applying the Standard and activating the emergency storage are effective in meeting the future water demand during extreme drought. In conclusion, we need to secure 110 million cubic meters (MCM) (Hapcheon reservoir) and 8 MCM (Gunwi reservoir) of water to reduce the number of days in the emergency stage. This research serves as a fundamental study that can help establish Standard and emergency storage activation criteria for other multipurpose reservoirs in preparation for extreme drought.