A flood hazard rating prediction model was developed that is based on a long short-term memory (LSTM) neural network and random forest. The target area was Samseong District in Seoul, which has a history of severe flooding. The Storm Water Management Model was used to generate training data for the LSTM model to predict the total overflow as the rainfall input data. Two-dimensional numerical analysis was performed to calculate inundation and flow velocity maps for training the random forest, which was used to generate a map of the predicted flood hazard rating of grid units given the total accumulative overflow of the target area. To confirm the goodness of fit, the proposed model was used to predict a flood hazard rating map for a rainfall event observed on July 27, 2011. The prediction accuracy for the flood hazard rating of each grid was 99.86% when the debris factor was considered and 99.99% when the debris factor was not considered.

This study proposed a real-time flood extent prediction method to shorten the time it takes from the flood occurrence to an alert issuance. This method uses logistic regression to generate a flood probability discriminant for each grid constituting the study area, and then predicts the flood extent with the amount of runoff caused by rainfall. In order to generate the flood probability discriminant for each grid, a two-dimensional (2D) flood inundation model was verified by applying the Typhoon Chaba, which caused great damage to the study area in 2016. Then, 100 probability rainfall scenarios were created by combining the return period, duration, and time distribution using past observation rainfall data, and rainfall-runoff–inundation relation databases were built for each scenario by applying hydrodynamic and hydrological models. A flood probability discriminant based on logistic regression was generated for each grid by using whether the grid was flooded (1 or 0) for the runoff amount in the database. When the runoff amount is input to the generated discriminant, the flood probability on the target grid is calculated by the coefficients, so that the flood extent is quickly predicted. The proposed method predicted the flood extent in a few seconds in both cases and showed high accuracy with 83.6~98.4% and 74.4~99.1%, respectively, in the application of scenario rainfall and actual rainfall.

In the midst of climate change, the need for accurate predictions of dam inflow to reduce flood damage along with stable water supply from water resources is increasing. In this study, the process and method of selecting the optimal deep learning model using hydrologic data over the past 20 years to predict dam inflow were shown. The study area is Andong Dam and Imha Dam located upstream of the Nakdong River in South Korea. In order to select the optimal model for predicting the inflow of two dams, sixteen scenarios (2 × 2 × 4) are generated considering two dams, two climatic conditions, and four deep learning models. During the drought period, the RNN for Andong Dam and the LSTM for Imha Dam were selected as the optimal models for each dam, and the difference between observations was the smallest at 4% and 2%, respectively. In typhoon conditions, the GRU for Andong Dam and the RNN for Imha Dam were selected as optimal models. In the case of Typhoon Maemi, the GRU and the RNN showed a difference of 2% and 6% from the observed maximum inflow, respectively. The optimal recurrent neural network-based models selected in this study showed a closer prediction to the observed inflow than the SFM, which is currently used to predict the inflow of both dams. For the two dams, different optimal models were selected according to watershed characteristics and rainfall under drought and typhoon conditions. In addition, most of the deep learning models were more accurate than the SFM under various typhoon conditions, but the SFM showed better results under certain conditions. Therefore, for efficient dam operation and management, it is necessary to make a rational decision by comparing the inflow predictions of the SFM and deep learning models.

Hydraulic structures installed in rivers inevitably create a water level difference between upstream and downstream regions. The potential energy due to this difference in water level is converted into kinetic energy, causing high-velocity flow and hydraulic jumps in the river. As a result, problems such as scouring and sloping downstream may occur around the hydraulic structures. In this study, a FLOW-3D model was constructed to perform a numerical analysis of the Changnyeong-Haman weir in the Republic of Korea. The constructed model was verified based on surface velocity measurements from a field gate operation experiment. In the simulation results, the flow discharge differed from the measured value by 9–15 m³/s, from which the accuracy was evaluated to be 82–87%. The flow velocity was evaluated with an accuracy of 92% from a difference of 0.01 to 0.16 m/s. Following this verification, a flow analysis of the hydraulic structures was performed according to boundary conditions and operation conditions for numerous scenarios. Since 2018, the Changnyeong-Haman weir gate has been fully opened due to the implementation of Korea’s eco-environmental policy; therefore, in this study, the actual gate operation history data prior to 2018 was applied and evaluated. The evaluation conditions were a 50% open gate condition and the flow discharge of two cases with a large difference in water level. As a result of the analysis, the actual operating conditions showed that the velocity and the Froude number were lower than the optimal conditions, confirming that the selected design was appropriate. It was also found that in the bed protection section, the average flow velocity was high when the water level difference was large, whereas the bottom velocity was high when the gate opening was large. Ultimately, through the reviewed status survey data in this study, the downstream flow characteristics of hydraulic structures along with adequacy verification techniques, optimal design techniques such as procedures for design, and important considerations were derived. Based on the current results, the constructed FLOW-3D-based model can be applied to creating or updating flow analysis guidelines for future repair and reinforcement measures as well as hydraulic structure design.

In this study, the sensitivity of the parameters was analyzed using PEST (Parameter ESTimation) to improve the accuracy of the runoff and flooding analysis in urban areas. Using four parameters (watershed width, roughness coefficient of impervious and pervious areas, and Manning’s roughness coefficient for conduits) with high sensitivity, six scenarios were created according to the number of parameters considered, and a PEST-SWMM (Storm Water Management Model) combined simulation was performed. The scenarios were applied to the Seocho 3, 4, 5, Yeoksam, and Nonhyun drainage basins in which inundation damage occurred due to the heavy rain on 21 July 2013. The sensitivity of the four parameters was in the order of Manning’s roughness coefficient for conduits, the roughness coefficient of the impervious area, the watershed width, and the roughness coefficient of the pervious area. When the PEST–SWMM coupled analysis for each scenario was performed using the analyzed sensitivity results, the RMSE (Root Mean Square Error) decreased by up to 2.37 cm and the RPE (Relative Peak Error) decreased by 22.04% compared to the SWMM alone. When the accuracy of each scenario was analyzed, similar or better accuracy was obtained as far as the parameters were considered. However, the further consideration of less sensitive parameters tends to reduce the accuracy. In this study, it was found that a more efficient simulation in terms of accuracy and calculation time could be obtained when constructing scenarios by considering only highly sensitive parameters. Additionally, when combining two-dimensional (2D) flood analysis with other rainfall events, it can help study real-time flood forecasting in urban areas.

Accurate flood level prediction is crucial for mitigating flood damage caused by typhoons or localized heavy rainfall. However, predicting flood levels is challenging due to changes in river environments and external factors, such as dam or weir operations. To address these challenges, this study proposes a methodology for constructing an optimal combination of input data using basic hydrological information and predicting flood levels in real time through a deep learning model. The study focuses on identifying the best input data combination tailored to each river basin's characteristics, considering both natural runoff rivers and those influenced by dam discharges. The Long Short‐Term Memory (LSTM) model, known for its superior performance in time‐series forecasting, was employed. The results demonstrate high accuracy in flood level prediction, particularly within a 3‐h lead time.