Preparing proper disaster prevention measures is important for decreasing the casualties and property losses resulting from floods. One of the most efficient measures in this regard is real-time flood forecasting using quantitative precipitation forecasts (QPFs) based on either short-term radar-based extrapolation or longer-term numerical weather prediction. As both methods have individual advantages and limitations, in this study we developed a new real-time blending technique to improve the accuracy of rainfall forecasts for hydrological applications. We tested the hydrological applicability of six QPFs used for urban flood forecasting in Seoul, South Korea: the McGill Algorithm for Prediction Nowcasting by Lagrangian Extrapolation (MAPLE), KOrea NOwcasting System (KONOS), Spatial-scale Decomposition method (SCDM), Unified Model Local Data Assimilation and Prediction System (UM LDAPS), and Advanced Storm-scale Analysis and Prediction System (ASAPS), as well as our proposed blended approach based on the assumption that the error of the previously predicted rainfall is similar to that of current predicted rainfall. We used the harmony search algorithm to optimize real-time weights that would minimize the root mean square error between predicted and observed rainfall for a 1 h lead time at 10 min intervals. We tested these models using the Storm Water Management Model (SWMM) and Grid-based Inundation Analysis Model (GIAM) to estimate urban flood discharge and inundation using rainfall from the QPFs as input. Although the blended QPF did not always have the highest correlation coefficient, its accuracy varied less than that of the other QPFs. In addition, its simulated water depth in pipe and spatial extent were most similar to observed inundated areas, demonstrating the value of this new approach for short-term flood forecasting.

Flood prediction is difficult in urban areas because only sparse gauge data and radar data of low accuracy are usually used to analyze flooding and inundation. Sub-basins of urban areas are extremely small, so rainfall data of high spatial resolution are required for analyzing complex drainage systems with high spatial variability. This study aimed to produce three types of quantitative precipitation estimation (QPE) products using rainfall data that was derived from 190 gauges, including the new high-density rain-gauge network operated by the SK Planet company, and the automated weather stations of the Korea Meteorological Administration, along with weather radar data. This study also simulated urban runoff for the Gangnam District of Seoul, South Korea, using the obtained QPE products to evaluate hydraulic and hydrologic impacts according to three rainfall fields. The accuracy of this approach was assessed in terms of the amount and spatial distribution of rainfall in an urban area. The QPE products provided highly accurate results and simulations of peak runoff and overflow phenomena. They also accurately described the spatial variability of the rainfall fields. Overall, the integration of high-density gauge data with radar data proved beneficial for quantitative rainfall estimation.

The mountainous Yeongdong region of South Korea contains mountains over 1 km. Owing to this topographic blockage, the region has a low-density rain-gauge network, and there is a low-altitude (~1.5 km) observation gap with the nearest large S-band radar. The Korean government installed an X-band dual-polarization radar in 2019 to improve rainfall observations and to prevent hydrological disasters in the Yeongdong region. The present study analyzed rainfall estimates using the newly installed X-band radar to evaluate its hydrological applicability. The rainfall was estimated using a distributed specific differential phase-based technique for a high-resolution 75 m grid. Comparison of the rainfall estimates of the X-band radar and the existing rainfall information showed that the X-band radar was less likely to underestimate rainfall compared to the S-band radar. The accuracy was particularly high within a 10 km observation radius. To evaluate the hydrological applicability of X-band radar rainfall estimates, this study developed a rain-based flood forecasting method—the flow nomograph—for the Samcheok-osib stream, which is vulnerable to heavy rain and resultant floods. This graph represents the flood risk level determined by hydrological–hydraulic modeling with various rainfall scenarios. Rainfall information (X-band radar, S-band radar, ground rain gauge) was applied as input to the flow nomograph to predict the flood level of the stream. Only the X-band radar could accurately predict the actual high-risk increase in the water level for all studied rainfall events.

To reduce hydrological disasters, it is necessary to operate rain gauge stations at locations where the spatio-temporal characteristics of rainfall can be reflected. Entropy has been widely used to evaluate the designs and uncertainties associated with rain gauge networks. In this study, the optimal rain gauge network in the Daegu and Gyeongbuk area, which requires the efficient use of water resources due to low annual precipitation and severe drought damage, was determined using conditional and joint entropy, and the selected network was quantitatively evaluated using the root mean square error (RMSE). To consider spatial distribution, prediction errors were generated using kriging. Four estimators used in entropy calculations were compared, and weighted entropy was calculated by weighting the precipitation. The optimal number of rain gauge stations was determined by calculating the RMSE reduction and the reduction ratio according to the number of selected rain gauge stations. Our findings show that the results of conditional entropy were better than those of joint entropy. The optimal rain gauge stations showed a tendency wherein peripheral rain gauge stations were selected first, with central stations being added afterward.

This study proposes a multi-task deep learning model for simultaneous prediction of time-series water levels and flood risk thresholds, aiming to enhance flood forecasting precision. Using AutoKeras, single-task and multi-task models were optimised to predict water levels 10–360 min ahead based on 720 min of prior data. The multi-task model consistently outperformed the single-task model across multiple evaluation metrics, including correlation coefficients, root mean squared error, Nash–Sutcliffe efficiency, and Kling–Gupta efficiency scores. Real-time prediction tests on actual rainfall events further validated the multi-task model's improved accuracy and applicability in operational flood forecasting. The study demonstrates significant progress in flood prediction methodologies, offering a more comprehensive approach to forecasting and categorising flood incidents.

This study examines the effectiveness of various deep learning algorithms in nowcasting using weather radar data from South Korea. Herein, the algorithms examined include RainNet, ConvLSTM2D U-Net, a U-Net-based recursive model, and a generative adversarial network. Moreover, this study used S-band radar data from the Ministry of Environment to assess the predictive performance of these models. Results show the efficacy of these algorithms in short-term rainfall prediction. Specifically, for a threshold of 0.1 mm/h, the recursive RainNet model achieved a critical success index (CSI) of 0.826, an F1 score of 0.781, and a mean absolute error (MAE) of 0.378. However, for a higher threshold of 5 mm/h, the model achieved an average CSI of 0.498, an F1 score of 0.577, and a MAE of 0.307. Furthermore, some models exhibited spatial smoothing issues with increasing rainfall-prediction times. The findings of this research hold promise for applications of societal importance, especially for preventing disasters due to extreme weather events.