Gwangseob Kim
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Gwangseob Kim has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Bashir Adelodun
Arusha Climate and Environmental Research Centre, Aga Khan University, Arusha, Tanzania
2 shared publications - Hong-Tae Kim
Department of Anatomy, Daegu Catholic University School of Medicine, Daegu 42472, Republic of Korea
2 shared publications - Hyo-Jeong Kim
Department of Civil Engineering, Kyungpook National University, Daegu 41566, Republic of Korea
2 shared publications Hyungon Cho
2 shared publications- Hyungon Cho
Department of Civil Engineering, Kyungpook National University, Daegu 41566, Republic of Korea
2 shared publications - Sanghwa Oh
School of Architectural, Civil, Environmental and Energy Engineering, Kyungpook National University, Daegu 41566, Republic of Korea
2 shared publications
Latest Publications
- Open Access
Comparative Evaluation of Deep Learning Models for Urban Flood Forecasting
A rapid and accurate urban inundation prediction system is important for reducing the impacts of urban flooding. This study comparatively evaluates three deep learning models—a Transformer, CNN-LSTM, and GNN—as surrogate models that link rainfall time series with the spatial distribution of inundation. A dataset of 1000 rainfall–inundation pairs was constructed using the physics-based EPA-SWMM/FLO-2D framework under a range of synthetic and observed rainfall events with durations of 1 h, 2 h, and 3 h. The three deep learning models were trained to approximate the inundation responses generated by the physics-based simulations, and their predictive performance was evaluated using RMSE, NSE, IoU, and F1 score, together with computational time, threshold sensitivity, and robustness to different train–test partitions. For the 30 common evaluation events, the Transformer achieved an RMSE of 0.034 ± 0.011 m, NSE of 0.908 ± 0.018, IoU of 0.903 ± 0.020, and F1 score of 0.949 ± 0.016, showing the highest overall predictive performance among the three models under the conditions considered in this study. The training times of the three models were relatively similar, ranging from 3268.1 to 3407.5 s, while inference times ranged from 0.031 to 0.044 s, substantially shorter than the mean EPA-SWMM/FLO-2D simulation time of 293.1 s. Threshold-sensitivity analysis and repeated random train–test splits showed a similar relative performance pattern among the models, while 50-run Monte Carlo cross-validation of the Transformer indicated relatively consistent performance across repeated data partitions. These results demonstrate the potential of deep learning surrogate models for rapid urban inundation prediction and provide a comparative assessment of the suitability of different architectures for approximating physics-based inundation responses.
Open AccessComparative Evaluation of Deep Learning Models for Urban Flood Forecasting
A rapid and accurate urban inundation prediction system is important for reducing the impacts of urban flooding. This study comparatively evaluates three deep learning models—a Transformer, CNN-LSTM, and GNN—as surrogate models that link rainfall time series with the spatial distribution of inundation. A dataset of 1000 rainfall–inundation pairs was constructed using the physics-based EPA-SWMM/FLO-2D framework under a range of synthetic and observed rainfall events with durations of 1 h, 2 h, and 3 h. The three deep learning models were trained to approximate the inundation responses generated by the physics-based simulations, and their predictive performance was evaluated using RMSE, NSE, IoU, and F1 score, together with computational time, threshold sensitivity, and robustness to different train–test partitions. For the 30 common evaluation events, the Transformer achieved an RMSE of 0.034 ± 0.011 m, NSE of 0.908 ± 0.018, IoU of 0.903 ± 0.020, and F1 score of 0.949 ± 0.016, showing the highest overall predictive performance among the three models under the conditions considered in this study. The training times of the three models were relatively similar, ranging from 3268.1 to 3407.5 s, while inference times ranged from 0.031 to 0.044 s, substantially shorter than the mean EPA-SWMM/FLO-2D simulation time of 293.1 s. Threshold-sensitivity analysis and repeated random train–test splits showed a similar relative performance pattern among the models, while 50-run Monte Carlo cross-validation of the Transformer indicated relatively consistent performance across repeated data partitions. These results demonstrate the potential of deep learning surrogate models for rapid urban inundation prediction and provide a comparative assessment of the suitability of different architectures for approximating physics-based inundation responses.
- Open Access
- 28 February 2026
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 26 (1) , 77-86
Analysis of Urban Flood Characteristics under Variations in Storm Temporal Patterns
Because urban flood characteristics are governed not only by total rainfall amount but also by its temporal distribution, this study applied the HSD (Heavy Storm Distribution) method incorporating the Truncated Laplace distribution. Various rainfall temporal patterns—including both unimodal and bimodal distributions—were implemented for the study area to analyze the resulting urban inundation characteristics. For unimodal rainfall events, the overflow volume produced using the HSD-based temporal distribution was approximately 51% greater than that obtained using the conventional Huff method. For bimodal events, the reduced temporal concentration of rainfall led to an overall decrease in overflow volume compared with unimodal events; however, depending on the specific configuration of the bimodal pattern, overflow decreased by up to 32%, with an average reduction of approximately 13%. The overflow results, which reflected diverse temporal rainfall distribution characteristics depending on the location and relative magnitude of the peaks, demonstrated that analyses of urban inundation must incorporate the effects of multiple temporal rainfall patterns. These findings also indicated the necessity of flood-mitigation planning that accounts for variability in temporal rainfall distribution.
- 28 February 2026
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 26 (1) , 77-86
Open AccessAnalysis of Urban Flood Characteristics under Variations in Storm Temporal Patterns
Because urban flood characteristics are governed not only by total rainfall amount but also by its temporal distribution, this study applied the HSD (Heavy Storm Distribution) method incorporating the Truncated Laplace distribution. Various rainfall temporal patterns—including both unimodal and bimodal distributions—were implemented for the study area to analyze the resulting urban inundation characteristics. For unimodal rainfall events, the overflow volume produced using the HSD-based temporal distribution was approximately 51% greater than that obtained using the conventional Huff method. For bimodal events, the reduced temporal concentration of rainfall led to an overall decrease in overflow volume compared with unimodal events; however, depending on the specific configuration of the bimodal pattern, overflow decreased by up to 32%, with an average reduction of approximately 13%. The overflow results, which reflected diverse temporal rainfall distribution characteristics depending on the location and relative magnitude of the peaks, demonstrated that analyses of urban inundation must incorporate the effects of multiple temporal rainfall patterns. These findings also indicated the necessity of flood-mitigation planning that accounts for variability in temporal rainfall distribution.
- Open Access
Bayesian Estimation of Neyman–Scott Rectangular Pulse Model Parameters in Comparison with Other Parameter Estimation Methods
Neyman–Scott rectangular pulse is a stochastic rainfall model with five parameters. The impacts of initial values and optimization methods on the parameter estimation of the Neyman–Scott rectangular pulse model were investigated using both the method of moments and the method of maximum likelihood. The estimates using the method of moments were influenced by the optimization method and were sensitive to the initial values and the aggregation scale of the data. Thus, by using frequentist estimation methods, we cannot guarantee the unique values as estimates. The aim of this study is to find more reliable unique values as estimates using a Bayesian approach. In this approach, parameters are estimated from the posterior distribution, and model performance is assessed by comparing observed values with fitted values. Slice sampling within the Gibbs sampler algorithm demonstrates superior convergence and model fitting, yielding unique estimates for the model parameters. The main conclusion of this study is that Bayesian estimation methods outperform other estimation methods in terms of providing reliable and stable estimates that improve rainfall generation accuracy.
Open AccessBayesian Estimation of Neyman–Scott Rectangular Pulse Model Parameters in Comparison with Other Parameter Estimation Methods
Neyman–Scott rectangular pulse is a stochastic rainfall model with five parameters. The impacts of initial values and optimization methods on the parameter estimation of the Neyman–Scott rectangular pulse model were investigated using both the method of moments and the method of maximum likelihood. The estimates using the method of moments were influenced by the optimization method and were sensitive to the initial values and the aggregation scale of the data. Thus, by using frequentist estimation methods, we cannot guarantee the unique values as estimates. The aim of this study is to find more reliable unique values as estimates using a Bayesian approach. In this approach, parameters are estimated from the posterior distribution, and model performance is assessed by comparing observed values with fitted values. Slice sampling within the Gibbs sampler algorithm demonstrates superior convergence and model fitting, yielding unique estimates for the model parameters. The main conclusion of this study is that Bayesian estimation methods outperform other estimation methods in terms of providing reliable and stable estimates that improve rainfall generation accuracy.
- Open Access
- 22 November 2023
- Published by MDPI AG in Atmosphere
- Vol. 14 (12) , 1718
Assessment of the Spatiotemporal Changes in the Extreme Precipitation Climate Indices over the Chungcheong Region of South Korea during 1973–2020
This study analyzed the changes and trends in twelve extreme precipitation-based climate indices obtained using daily data from 10 synoptic stations in the Chungcheong region of South Korea during the 1973–2020 period. The climate indices were used to assess the trends in the extreme precipitation characteristics of duration, frequency, and intensity using the innovative trend analysis (ITA) method. The results of the ITA were further compared with two other non-parametric test methods such as Mann–Kendall (MK) and Spearman’s rho (SR). The results showed that most stations exhibited significant increasing trends in all the investigated climate indices at a 95% confidence level as indicated by the ITA method, with only a few stations indicating significant decreasing trends in R95p, R99p, Rx3day, and Rx5day. The sub-trend analysis further revealed the dominance of neutral behavior around the low-value cluster, especially for the extreme precipitation duration. At the same time, increasing trends dominate the high-value cluster at most stations. Meanwhile, only R10mm, R99p, and R95p exhibited monotonic trends in the Boeun and Seosan stations, respectively. Further, the ITA exhibited superior performance over the MK and SR methods by indicating the presence of more significant trends in the climate indices at most stations. The distribution of the extreme precipitation indices for duration, frequency, and intensity indicate the pronounced risk of flood conditions around the north–central and some parts of southern regions, while the western region indicates a potential drought risk, which could greatly impact the water resources and consequently agricultural activities in the study area. The results of this study provide essential information for addressing the climate-related problems of water resource management and agriculture in the study area and other related climatic regions.
- 22 November 2023
- Published by MDPI AG in Atmosphere
- Vol. 14 (12) , 1718
Open AccessAssessment of the Spatiotemporal Changes in the Extreme Precipitation Climate Indices over the Chungcheong Region of South Korea during 1973–2020
This study analyzed the changes and trends in twelve extreme precipitation-based climate indices obtained using daily data from 10 synoptic stations in the Chungcheong region of South Korea during the 1973–2020 period. The climate indices were used to assess the trends in the extreme precipitation characteristics of duration, frequency, and intensity using the innovative trend analysis (ITA) method. The results of the ITA were further compared with two other non-parametric test methods such as Mann–Kendall (MK) and Spearman’s rho (SR). The results showed that most stations exhibited significant increasing trends in all the investigated climate indices at a 95% confidence level as indicated by the ITA method, with only a few stations indicating significant decreasing trends in R95p, R99p, Rx3day, and Rx5day. The sub-trend analysis further revealed the dominance of neutral behavior around the low-value cluster, especially for the extreme precipitation duration. At the same time, increasing trends dominate the high-value cluster at most stations. Meanwhile, only R10mm, R99p, and R95p exhibited monotonic trends in the Boeun and Seosan stations, respectively. Further, the ITA exhibited superior performance over the MK and SR methods by indicating the presence of more significant trends in the climate indices at most stations. The distribution of the extreme precipitation indices for duration, frequency, and intensity indicate the pronounced risk of flood conditions around the north–central and some parts of southern regions, while the western region indicates a potential drought risk, which could greatly impact the water resources and consequently agricultural activities in the study area. The results of this study provide essential information for addressing the climate-related problems of water resource management and agriculture in the study area and other related climatic regions.
- Open Access
- 18 October 2023
- Published by MDPI AG in Preprints.org
Assessment of Spatiotemporal Changes in Precipitation Extreme Climate Indices over Chungcheong Region of South Korea during 1973–2020
This study analyzed the changes and trends in twelve extreme precipitation-based climate indices obtained using daily data from 10 synoptic stations in the Chungcheong region of South Korea during the 1973-2020 period. The climate indices were used to assess the trends in the extreme precipitation characteristics of duration, frequency, and intensity using the innovative trend analysis (ITA) method. The results of the ITA were further compared with two other non-parametric test methods such as Mann-Kendall (MK) and Spearman’s rho (SR). The results showed that the most stations exhibited significant increasing trends in all the investigated climate indices at a 95% confidence level as indicated by the ITA method, with only a few stations indicating significant decreasing trends in R95p, R99p, Rx3day, and Rx5day. The sub-trend analysis further revealed the dominance of neutral behavior around the low-value cluster, especially for the extreme precipitation duration. At the same time, increasing trends dominate the high-value cluster in most stations. Meanwhile, only R10mm, R95p, and R99p exhibited monotonic trends in the Boeun, Seosan, and Boeun stations, respectively. Further, the ITA exhibited superior performance over the MK and SR methods by indicating the presence of more significant trends in the climate indices at most stations. The distribution of extreme precipitation indices for duration, frequency, and intensity indicate the pronounced risk of flood condition around the north-central and some parts of southern regions, while the western region indicate a potential drought risk, which could greatly impact the water resources and consequently agricultural activities in the study area. The results of this study provide essential information for addressing the climate-related problems of water resources management and agriculture in the study area and other related climatic regions.
- 18 October 2023
- Published by MDPI AG in Preprints.org
Open AccessAssessment of Spatiotemporal Changes in Precipitation Extreme Climate Indices over Chungcheong Region of South Korea during 1973–2020
This study analyzed the changes and trends in twelve extreme precipitation-based climate indices obtained using daily data from 10 synoptic stations in the Chungcheong region of South Korea during the 1973-2020 period. The climate indices were used to assess the trends in the extreme precipitation characteristics of duration, frequency, and intensity using the innovative trend analysis (ITA) method. The results of the ITA were further compared with two other non-parametric test methods such as Mann-Kendall (MK) and Spearman’s rho (SR). The results showed that the most stations exhibited significant increasing trends in all the investigated climate indices at a 95% confidence level as indicated by the ITA method, with only a few stations indicating significant decreasing trends in R95p, R99p, Rx3day, and Rx5day. The sub-trend analysis further revealed the dominance of neutral behavior around the low-value cluster, especially for the extreme precipitation duration. At the same time, increasing trends dominate the high-value cluster in most stations. Meanwhile, only R10mm, R95p, and R99p exhibited monotonic trends in the Boeun, Seosan, and Boeun stations, respectively. Further, the ITA exhibited superior performance over the MK and SR methods by indicating the presence of more significant trends in the climate indices at most stations. The distribution of extreme precipitation indices for duration, frequency, and intensity indicate the pronounced risk of flood condition around the north-central and some parts of southern regions, while the western region indicate a potential drought risk, which could greatly impact the water resources and consequently agricultural activities in the study area. The results of this study provide essential information for addressing the climate-related problems of water resources management and agriculture in the study area and other related climatic regions.
- Open Access
- 31 December 2022
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 22 (6) , 79-88
Development of an Urban Flood Forecast Model Using Lumped Pipe Networks
Typical urban flood simulation is conducted using physical models such as the 1D storm water analysis model and 2D inundation analysis model. Although 2D inundation analysis can predict flow velocity, inundation depth, and inundation area throughout an inundated urban area, it is difficult for it to produce a near real-time urban flood forecast for a metropolitan area such as Seoul. In this study, a physical urban flood forecast model was developed using lumped pipe networks to produce a near real-time urban inundation forecast. The dense pipe networks within a drainage basin were simplified as a single conceptual lumped pipe that has drainage and storage functions, and new pipe networks were constructed using lumped pipe networks. The model was applied to the August 2018 storm events in Seoul and showed a prediction accuracy of 0.71. The results demonstrated that the model can obviate the limitations of the near real-time operation of existing physical flood forecasting models to yield useful information for urban flood response, though showing room for improvement.
- 31 December 2022
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 22 (6) , 79-88
Open AccessDevelopment of an Urban Flood Forecast Model Using Lumped Pipe Networks
Typical urban flood simulation is conducted using physical models such as the 1D storm water analysis model and 2D inundation analysis model. Although 2D inundation analysis can predict flow velocity, inundation depth, and inundation area throughout an inundated urban area, it is difficult for it to produce a near real-time urban flood forecast for a metropolitan area such as Seoul. In this study, a physical urban flood forecast model was developed using lumped pipe networks to produce a near real-time urban inundation forecast. The dense pipe networks within a drainage basin were simplified as a single conceptual lumped pipe that has drainage and storage functions, and new pipe networks were constructed using lumped pipe networks. The model was applied to the August 2018 storm events in Seoul and showed a prediction accuracy of 0.71. The results demonstrated that the model can obviate the limitations of the near real-time operation of existing physical flood forecasting models to yield useful information for urban flood response, though showing room for improvement.
- Open Access
- 28 February 2022
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 22 (1) , 227-239
Analysis of the Risk Management Strategy for Future Flood-Related Property Damage Due to Climate Change
- 28 February 2022
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 22 (1) , 227-239
Open AccessAnalysis of the Risk Management Strategy for Future Flood-Related Property Damage Due to Climate Change
- Open Access
Case Study of Urban Flood Inundation—Impact of Temporal Variability in Rainfall Events
This study aimed to calculate and analyze total overflows that accumulate in urban manholes in the target drainage basin of Samsung-dong, Seoul in heavy rainfall events with different temporal distribution characteristics, using the EPA’s Storm Water Management Model (EPA-SWMM model). Inundation behaviors were analyzed using the two-dimensional flood model (FLO-2D). The extreme rainfall events were produced using different exceedance probability Huff distributions for different durations and return periods, such as from 1 to 3 h and 10 years, 50 years, 80 years, 100 years, respectively. The inundation model was validated using the actual flood observations on 21 September 2010 in the Samsung-dong drainage basin. The total overflow amount showed considerable differences according to the different time distribution characteristics, such as the temporal location of the storm peak and the concentration level of the storm. Furthermore, the inundation behaviors were also related to the temporal characteristics of storms. The results illustrated that the consideration of the temporal distribution characteristics of extreme rainfall events is essential for an accurate understanding of the rainfall–runoff response and inundation behavior in urban drainage basins.
Open AccessCase Study of Urban Flood Inundation—Impact of Temporal Variability in Rainfall Events
This study aimed to calculate and analyze total overflows that accumulate in urban manholes in the target drainage basin of Samsung-dong, Seoul in heavy rainfall events with different temporal distribution characteristics, using the EPA’s Storm Water Management Model (EPA-SWMM model). Inundation behaviors were analyzed using the two-dimensional flood model (FLO-2D). The extreme rainfall events were produced using different exceedance probability Huff distributions for different durations and return periods, such as from 1 to 3 h and 10 years, 50 years, 80 years, 100 years, respectively. The inundation model was validated using the actual flood observations on 21 September 2010 in the Samsung-dong drainage basin. The total overflow amount showed considerable differences according to the different time distribution characteristics, such as the temporal location of the storm peak and the concentration level of the storm. Furthermore, the inundation behaviors were also related to the temporal characteristics of storms. The results illustrated that the consideration of the temporal distribution characteristics of extreme rainfall events is essential for an accurate understanding of the rainfall–runoff response and inundation behavior in urban drainage basins.
- Open Access
Impacts of Climate Change Scenarios on Non-Point Source Pollution in the Saemangeum Watershed, South Korea
Non-point source (NPS) pollution is a primary cause of water pollution in the Saemangeum watershed in South Korea. The changes in NPS pollutant loads in the Saemangeum watershed for an 81-year period (2019–2099) were simulated and analyzed by applying the soil and water assessment tool. Six climate model (BCC-CSM1–1, CanESM2, GFDL-ESM2G, HadGEM2-CC, INM-CM4, and MIROC-ESM) outputs using representative concentration pathway (RCP) scenarios (RCP 4.5 and RCP 8.5) were obtained from the South Korean Asia-Pacific Economic Cooperation (APEC) Climate Center. Simulated streamflow and water quality were evaluated using the Nash–Sutcliffe efficiency (NSE) index and coefficient of determination (R2). The model satisfactorily simulated streamflow with positive NSE values and R2 > 0.5. Based on two climate change scenarios (RCP 4.5 and RCP 8.5), gradual increases of 70.9 to 233.8 mm and 1.7 to 5.7 °C in annual precipitation and temperature, respectively, are likely for two time periods (2019–2059 and 2060–2099). Additionally, the expected future average annual and monthly streamflow, sediment, and total phosphorus showed changes of 5% to 43%, 3% to 40%, and −55% to 15%, respectively, whereas the expected future average annual and monthly total nitrogen showed decreases of −5% to −27%. Future NPS pollutant loads in the Saemangeum watershed should be managed according to different climate change scenarios.
Open AccessImpacts of Climate Change Scenarios on Non-Point Source Pollution in the Saemangeum Watershed, South Korea
Non-point source (NPS) pollution is a primary cause of water pollution in the Saemangeum watershed in South Korea. The changes in NPS pollutant loads in the Saemangeum watershed for an 81-year period (2019–2099) were simulated and analyzed by applying the soil and water assessment tool. Six climate model (BCC-CSM1–1, CanESM2, GFDL-ESM2G, HadGEM2-CC, INM-CM4, and MIROC-ESM) outputs using representative concentration pathway (RCP) scenarios (RCP 4.5 and RCP 8.5) were obtained from the South Korean Asia-Pacific Economic Cooperation (APEC) Climate Center. Simulated streamflow and water quality were evaluated using the Nash–Sutcliffe efficiency (NSE) index and coefficient of determination (R2). The model satisfactorily simulated streamflow with positive NSE values and R2 > 0.5. Based on two climate change scenarios (RCP 4.5 and RCP 8.5), gradual increases of 70.9 to 233.8 mm and 1.7 to 5.7 °C in annual precipitation and temperature, respectively, are likely for two time periods (2019–2059 and 2060–2099). Additionally, the expected future average annual and monthly streamflow, sediment, and total phosphorus showed changes of 5% to 43%, 3% to 40%, and −55% to 15%, respectively, whereas the expected future average annual and monthly total nitrogen showed decreases of −5% to −27%. Future NPS pollutant loads in the Saemangeum watershed should be managed according to different climate change scenarios.
- Open Access
- 31 December 2018
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 18 (7) , 613-620
Analysis of the Temporal Distribution of Rainfall Using the Heavy Storm Distribution Method Reflecting Concentrated Duration Characteristics in Busan
- 31 December 2018
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 18 (7) , 613-620
Open AccessAnalysis of the Temporal Distribution of Rainfall Using the Heavy Storm Distribution Method Reflecting Concentrated Duration Characteristics in Busan