Eunjin Kang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Eunjin Kang has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Jungho Im
Department of Civil, Urban, Earth, and Environmental Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan, South Korea
6 shared publications - Cheolhee Yoo
Smart City Convergence Major, Pusan National University, Busan, South Korea
4 shared publications - Deoksu Kim
Coastal Disaster & Safety Research Department, Korea Institute of Ocean Science & Technology, Busan, Republic of Korea
3 shared publications - Jang-Geun Choi
Center for Ocean Engineering University of New Hampshire Durham NH USA
3 shared publications Deoksu Kim
2 shared publications- Hojin Kim
Marine Natural Disaster Research Department Korea Institute of Ocean Science and Technology Busan Republic of Korea
2 shared publications
Latest Publications
- Open Access
- 9 September 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Oceans
- Vol. 131 (9)
An Event‐Based Framework Reveals Regionally Distinct Coastal Cold‐Water Regimes Along the East Coast of Korea
Coastal cold‐water events (CWEs) reduce nearshore sea surface temperature (SST) and can cause ecological and socioeconomic impacts; however, objective detection and process‐based diagnostics remain limited in marginal seas influenced by western boundary currents. We developed a station‐based framework that identifies CWEs by isolating event‐scale SST variability after removing diurnal and seasonal signals, and applied it to hourly SST observations from 15 coastal stations along the east coast of Korea during 2004–2023. Spatial coherence revealed three distinct regimes (southern, central, and northern). CWEs in the northern and central regimes occurred earlier in the warm season, intensified more rapidly, and exhibited pronounced cooling‐recovery asymmetry, whereas CWEs in the southern regime occurred later, persisted longer, and evolved more gradually. Lagged composites of cross‐shore Ekman transport and CWE‐targeted SST showed that southern CWE development was consistent with sustained wind forcing, whereas intense cooling in the central and northern regimes occurred despite weaker, shorter wind impulses, implicating oceanic preconditioning. We infer that boundary‐current preconditioning sharpens the coastal thermocline, enabling moderate wind to uplift cold subsurface water. Cooling durations matched the theoretical two‐layer upwelling timescale in the central and northern regimes but not in the south, where the low Burger number places upwelling under bottom boundary layer control, so cooling is gated by sustained wind. Applying the metric to a global ocean reanalysis revealed state‐dependent skill, with systematic warm bias and larger errors during CWEs. The framework provides a transferable basis for CWE diagnostics, event‐focused model evaluation, and coastal‐hazard prediction.
- 9 September 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Oceans
- Vol. 131 (9)
Open AccessAn Event‐Based Framework Reveals Regionally Distinct Coastal Cold‐Water Regimes Along the East Coast of Korea
Coastal cold‐water events (CWEs) reduce nearshore sea surface temperature (SST) and can cause ecological and socioeconomic impacts; however, objective detection and process‐based diagnostics remain limited in marginal seas influenced by western boundary currents. We developed a station‐based framework that identifies CWEs by isolating event‐scale SST variability after removing diurnal and seasonal signals, and applied it to hourly SST observations from 15 coastal stations along the east coast of Korea during 2004–2023. Spatial coherence revealed three distinct regimes (southern, central, and northern). CWEs in the northern and central regimes occurred earlier in the warm season, intensified more rapidly, and exhibited pronounced cooling‐recovery asymmetry, whereas CWEs in the southern regime occurred later, persisted longer, and evolved more gradually. Lagged composites of cross‐shore Ekman transport and CWE‐targeted SST showed that southern CWE development was consistent with sustained wind forcing, whereas intense cooling in the central and northern regimes occurred despite weaker, shorter wind impulses, implicating oceanic preconditioning. We infer that boundary‐current preconditioning sharpens the coastal thermocline, enabling moderate wind to uplift cold subsurface water. Cooling durations matched the theoretical two‐layer upwelling timescale in the central and northern regimes but not in the south, where the low Burger number places upwelling under bottom boundary layer control, so cooling is gated by sustained wind. Applying the metric to a global ocean reanalysis revealed state‐dependent skill, with systematic warm bias and larger errors during CWEs. The framework provides a transferable basis for CWE diagnostics, event‐focused model evaluation, and coastal‐hazard prediction.
- Open Access
- 11 May 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (20) , 14773-14774
Correction to “Quantifying Multi-pollutant Co-exposure via Deep Learning-Based Simultaneous Prediction Using Geostationary Satellite Data”
- 11 May 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (20) , 14773-14774
Open AccessCorrection to “Quantifying Multi-pollutant Co-exposure via Deep Learning-Based Simultaneous Prediction Using Geostationary Satellite Data”
- Open Access
- 20 March 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (12) , 9319-9332
Quantifying Multi-pollutant Co-exposure via Deep Learning-Based Simultaneous Prediction Using Geostationary Satellite Data
Simultaneous prediction of multiple air pollutants is essential for quantifying human co-exposure and evaluating the health impacts of pollutant mixtures. However, spatial and temporal gaps in geostationary satellite observations, chemical transport models, and ground-based monitoring networks hinder accurate hourly assessments of multi-pollutant dynamics. Here, we present Deep Learning for Multiple Air Pollutant analysis (DeepMAP), a deep learning framework that simultaneously predicts six major air pollutants─PM10, PM2.5, O3, NO2, CO, and SO2─at hourly resolution. DeepMAP demonstrated robust performance across multiple pollutants and generalized well to unseen regions. The framework accurately captured dynamic high-concentration co-pollution episodes during March 2021, with normalized RMSE values below 0.36 for all pollutants. DeepMAP revealed that PM10-PM2.5 co-exceedance was the most frequent across East Asia (91 days/year), followed by PM10-PM2.5-NO2 (42), PM2.5-O3 (18), and PM10-PM2.5-O3 (12). Hotspots for PM10-PM2.5-NO2–O3 co-exceedance were identified over the North China Plain, East China, and South Korea, where the regional annual totals reached 24, 19, and 15 days, respectively. A novel co-exposure index further identified three distinct hotspot regions where the contribution of NO2 was approximately twice that observed elsewhere. Our findings provide a high-resolution, data-driven framework for characterizing multi-pollutant co-exposure and identifying regional priorities for air quality management and public health protection.
- 20 March 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (12) , 9319-9332
Open AccessQuantifying Multi-pollutant Co-exposure via Deep Learning-Based Simultaneous Prediction Using Geostationary Satellite Data
Simultaneous prediction of multiple air pollutants is essential for quantifying human co-exposure and evaluating the health impacts of pollutant mixtures. However, spatial and temporal gaps in geostationary satellite observations, chemical transport models, and ground-based monitoring networks hinder accurate hourly assessments of multi-pollutant dynamics. Here, we present Deep Learning for Multiple Air Pollutant analysis (DeepMAP), a deep learning framework that simultaneously predicts six major air pollutants─PM10, PM2.5, O3, NO2, CO, and SO2─at hourly resolution. DeepMAP demonstrated robust performance across multiple pollutants and generalized well to unseen regions. The framework accurately captured dynamic high-concentration co-pollution episodes during March 2021, with normalized RMSE values below 0.36 for all pollutants. DeepMAP revealed that PM10-PM2.5 co-exceedance was the most frequent across East Asia (91 days/year), followed by PM10-PM2.5-NO2 (42), PM2.5-O3 (18), and PM10-PM2.5-O3 (12). Hotspots for PM10-PM2.5-NO2–O3 co-exceedance were identified over the North China Plain, East China, and South Korea, where the regional annual totals reached 24, 19, and 15 days, respectively. A novel co-exposure index further identified three distinct hotspot regions where the contribution of NO2 was approximately twice that observed elsewhere. Our findings provide a high-resolution, data-driven framework for characterizing multi-pollutant co-exposure and identifying regional priorities for air quality management and public health protection.
- Open Access
- 14 March 2026
- Published by Copernicus GmbH
Bridging temporal gaps: AI-based temporal downscaling of biweekly NH3 to daily scale with spatial transferability
Ammonia (NH₃) is an important atmospheric pollutant with environmental and public health impacts. In recent decades, NH₃ concentrations have increased due to intensified agricultural activities and industrial development, underscoring the need for high-resolution monitoring. However, sparse biweekly ground-based observations from the Ammonia Monitoring Network (AMoN) remain a major limitation for comprehensive spatiotemporal analysis. The United States (US) is a region where NH₃ monitoring is particularly important due to its extensive agricultural activities. In this study, we developed machine learning–based frameworks, including a deep neural network (DNN), random forest, and light gradient boosting machine, to estimate nationwide biweekly NH₃ concentrations and temporally downscale them to daily values across the contiguous US from 2017 to 2022. The models incorporate satellite-derived NH₃ column measurements, meteorological variables, land cover characteristics, livestock density, and AMoN ground-based observations. Among the tested approaches, the DNN demonstrated the strongest performance under both spatial cross-validation and independent testing, achieving a correlation coefficient of 0.79, a root mean square error of 0.98 µg m⁻³, and an index of agreement of 0.83. The model effectively reproduced fine-scale spatial variability in daily NH₃ concentrations at a 9 km resolution. Shapley additive explanations further revealed that temporally varying predictors—such as day of year and meteorological conditions—played a dominant role, alongside land cover and cattle density, supporting robust temporal downscaling from biweekly to daily scales. To assess spatial transferability, the framework was additionally applied to ground-based monitoring stations in the United Kingdom, where daily NH₃ observations are available, using leave-one-station-out and leave-one-year-out cross-validation schemes. Overall, our results demonstrate the potential of machine learning approaches to bridge temporal gaps in NH₃ observations and to generate high-resolution daily concentration estimates.
- 14 March 2026
- Published by Copernicus GmbH
Open AccessBridging temporal gaps: AI-based temporal downscaling of biweekly NH3 to daily scale with spatial transferability
Ammonia (NH₃) is an important atmospheric pollutant with environmental and public health impacts. In recent decades, NH₃ concentrations have increased due to intensified agricultural activities and industrial development, underscoring the need for high-resolution monitoring. However, sparse biweekly ground-based observations from the Ammonia Monitoring Network (AMoN) remain a major limitation for comprehensive spatiotemporal analysis. The United States (US) is a region where NH₃ monitoring is particularly important due to its extensive agricultural activities. In this study, we developed machine learning–based frameworks, including a deep neural network (DNN), random forest, and light gradient boosting machine, to estimate nationwide biweekly NH₃ concentrations and temporally downscale them to daily values across the contiguous US from 2017 to 2022. The models incorporate satellite-derived NH₃ column measurements, meteorological variables, land cover characteristics, livestock density, and AMoN ground-based observations. Among the tested approaches, the DNN demonstrated the strongest performance under both spatial cross-validation and independent testing, achieving a correlation coefficient of 0.79, a root mean square error of 0.98 µg m⁻³, and an index of agreement of 0.83. The model effectively reproduced fine-scale spatial variability in daily NH₃ concentrations at a 9 km resolution. Shapley additive explanations further revealed that temporally varying predictors—such as day of year and meteorological conditions—played a dominant role, alongside land cover and cattle density, supporting robust temporal downscaling from biweekly to daily scales. To assess spatial transferability, the framework was additionally applied to ground-based monitoring stations in the United Kingdom, where daily NH₃ observations are available, using leave-one-station-out and leave-one-year-out cross-validation schemes. Overall, our results demonstrate the potential of machine learning approaches to bridge temporal gaps in NH₃ observations and to generate high-resolution daily concentration estimates.
- 7 July 2025
- Published by Elsevier in Journal of Hazardous Materials
- Vol. 496, 139166
Bridging temporal gaps: AI-based temporal downscaling of biweekly NH3 to daily scale with spatial transferability
- 7 July 2025
- Published by Elsevier in Journal of Hazardous Materials
- Vol. 496, 139166
Bridging temporal gaps: AI-based temporal downscaling of biweekly NH3 to daily scale with spatial transferability
- 30 April 2025
- Published by Korean Society for Atmospheric Environment in Journal of Korean Society for Atmospheric Environment
- Vol. 41 (2) , 179-198
Satellite-based Air Quality Monitoring Using Artificial Intelligence: Research Trends and Future Perspectives
- 30 April 2025
- Published by Korean Society for Atmospheric Environment in Journal of Korean Society for Atmospheric Environment
- Vol. 41 (2) , 179-198
Satellite-based Air Quality Monitoring Using Artificial Intelligence: Research Trends and Future Perspectives
- Open Access
- 18 March 2025
- Published by Copernicus GmbH
Quantifying the Role of Atmospheric Pollutant Deposition in Nutrient Flux of the Yellow Sea
The impact of persistent anthropogenic activities on ecosystems remains a critical focus in the context of sustainability. Pollutants deposited into the ocean, alongside nutrients introduced via river discharge, are recognized as major sources influencing marine primary productivity. In particular, atmospheric pollutant emissions in East Asia have been continuously rising, and the resulting deposition of these pollutants into the ocean is estimated to be substantial. Despite their significance, quantitative assessments of the ecological and biogeochemical impacts remain largely unexplored. The excessive nitrogen deposition has the potential to accelerate eutrophication processes, leading to harmful algal blooms and subsequent oxygen depletion in the Yellow Sea. These phenomena are likely to significantly disrupt local biodiversity and marine food webs, posing challenges to sustainable ecosystem management. Furthermore, the reduction in atmospheric pollutants during the COVID-19 lockdown in 2020 has triggered discussions regarding its effects on the productivity of the Yellow Sea, reflecting the complex interactions between atmospheric deposition and marine ecosystems. Therefore, this study quantitatively evaluates the nitrogen flux introduced through atmospheric deposition and its influence on the marine ecosystem of the Yellow Sea using a numerical approach based on the Regional Ocean Modeling System (ROMS). Atmospheric pollutant data reproduced through deep learning served as input, enabling experiments ranging from one-dimensional vertical models to full-scale simulations of the entire Yellow Sea. The study was conducted using the NPZD (Nutrient-Phytoplankton-Zooplankton-Detritus) model framework.
- 18 March 2025
- Published by Copernicus GmbH
Open AccessQuantifying the Role of Atmospheric Pollutant Deposition in Nutrient Flux of the Yellow Sea
The impact of persistent anthropogenic activities on ecosystems remains a critical focus in the context of sustainability. Pollutants deposited into the ocean, alongside nutrients introduced via river discharge, are recognized as major sources influencing marine primary productivity. In particular, atmospheric pollutant emissions in East Asia have been continuously rising, and the resulting deposition of these pollutants into the ocean is estimated to be substantial. Despite their significance, quantitative assessments of the ecological and biogeochemical impacts remain largely unexplored. The excessive nitrogen deposition has the potential to accelerate eutrophication processes, leading to harmful algal blooms and subsequent oxygen depletion in the Yellow Sea. These phenomena are likely to significantly disrupt local biodiversity and marine food webs, posing challenges to sustainable ecosystem management. Furthermore, the reduction in atmospheric pollutants during the COVID-19 lockdown in 2020 has triggered discussions regarding its effects on the productivity of the Yellow Sea, reflecting the complex interactions between atmospheric deposition and marine ecosystems. Therefore, this study quantitatively evaluates the nitrogen flux introduced through atmospheric deposition and its influence on the marine ecosystem of the Yellow Sea using a numerical approach based on the Regional Ocean Modeling System (ROMS). Atmospheric pollutant data reproduced through deep learning served as input, enabling experiments ranging from one-dimensional vertical models to full-scale simulations of the entire Yellow Sea. The study was conducted using the NPZD (Nutrient-Phytoplankton-Zooplankton-Detritus) model framework.
- Open Access
- 5 December 2024
- Published by Taylor & Francis in GIScience & Remote Sensing
- Vol. 61 (1)
An explainable AI framework for spatiotemporal risk factor analysis in public health: a case study of cardiovascular mortality in South Korea
- 5 December 2024
- Published by Taylor & Francis in GIScience & Remote Sensing
- Vol. 61 (1)
Open AccessAn explainable AI framework for spatiotemporal risk factor analysis in public health: a case study of cardiovascular mortality in South Korea
- 8 July 2023
- Published by Elsevier in Atmospheric Environment
- Vol. 309
Direct aerosol optical depth retrievals using MODIS reflectance data and machine learning over East Asia
- 8 July 2023
- Published by Elsevier in Atmospheric Environment
- Vol. 309
Direct aerosol optical depth retrievals using MODIS reflectance data and machine learning over East Asia