1. Introduction
The rise in sea level and the increase in sea surface temperature caused by climate change are amplifying typhoon intensity and scale, thereby exacerbating disaster risks in coastal regions. In particular, such meteorological changes influence the track and strength of typhoons approaching the Korean Peninsula, acting as key factors that heighten both the likelihood of storm surge occurrence and the scale of potential damage. According to the Korea Meteorological Administration (KMA, 2022) typhoon statistics, a total of 26 typhoons formed in 2022, a figure comparable to the 30-year average of 25.1. Among them, however, only two affected the Korean Peninsula, fewer than the average of 3.4. Nevertheless, the average number of typhoons impacting the Korean Peninsula has shown an upward trend, reaching four in the past ten years. Notably, the two typhoons that occurred in 2022 intensified into super typhoons (maximum wind speed exceeding 54 m/s) and maintained very strong intensity (44–53 m/s) even upon approaching the Korean Peninsula.
According to Park (2022), although the overall frequency of typhoon occurrence in the northwestern Pacific has declined over the past 40 years since 1980, the proportion of typhoons affecting the Korean Peninsula has gradually increased. This trend indicates that typhoon tracks are shifting under the influence of climate change. Wu et al. (2005), in their analysis of typhoon tracks affecting East Asia from 1965 to 2003, reported that the westward expansion of the Western Pacific subtropical high has shifted typhoon tracks further westward, thereby increasing the risk posed to East Asia. Similarly, Kossin et al. (2013), using satellite data, demonstrated that the genesis locations of tropical depressions have gradually migrated from the equatorial region, implying that mid-latitude regions such as the Korean Peninsula may experience increased typhoon impacts in the future.
Tropical depressions occur only a limited number of times and last for relatively short periods, which makes it difficult to fully capture the characteristics and risks of storm surges using observational data alone. In addition, climate change and sea level rise are altering the genesis regions, tracks, and intensities of typhoons, meaning that risk assessments based solely on past statistics are limited to provide a reliable picture of future hazards. To overcome these limitations, recent research has focused on generating large numbers of synthetic typhoons and linking them with numerical or artificial intelligence models for storm surge prediction. Eum et al. (2020) used the Tropical Cyclone Risk Model (TCRM), a synthetic typhoon generation algorithm, to produce thousands of synthetic typhoons and then trained a deep learning model to estimate storm surge heights. This model showed root mean square errors of just 0.09–0.30 m and correlation coefficients of 0.65–0.94, suggesting the potential for developing real-time operational forecasting systems. Ruiz-Salcines et al. (2021) constructed synthetic typhoon scenarios to examine the evolution of inundation risks under sea level rise and typhoon intensification. Thus, they warned that a Hurricane Patricia-scale event, previously thought to occur only once in 4,000 years, could recur roughly every 198 years, dramatically increasing both the intensity and scope of flooding in Mexico’s Colima region. Yin et al. (2021) went a step further, generating more than 5,000 synthetic typhoons to assess flood risk in coastal regions of Shanghai, China. Their results showed that sea level rise could double the likelihood of inundation by 2050, increase it twentyfold by 2100, and expand the affected area by more than 1,360%. More recently, Woods et al. (2023) used observed typhoon data from 1980 to 2017 to build climate change scenarios up to 2050, and then estimated storm surge heights along the coasts of southern China and Southeast Asia. They found that due to changes in typhoon tracks and intensities, the length of coastline exposed to storm surges greater than 2.5 m could more than double. They also warned that regions that historically experienced no damage could also become high-risk zones in the future. In South Korea, Lee et al. (2024) applied 735 synthetic typhoon scenarios to estimate the probability distribution of flood depths in Busan and to quantify flood vulnerability. Yang et al. (2024) used scenarios synthesized from historical typhoon records to evaluate storm surge hazards under sea level rise for 50-, 100-, and 150-year return periods. They reported that a 1-m rise in mean sea level could increase storm surge risk by 25–60%. Seo et al. (2024) selected 235 synthetic typhoons from a pool of approximately 3,000 generated with TCRM to estimate storm surge heights at 31 Korean trade ports and analyzed extreme value statistics by return period. By conducting Advanced CIRCulation (ADCIRC) simulations, they showed that typhoons traveling along the Yellow Sea and South Sea paths have major impacts on ports in the southeast and southwest, with local topography playing a key role in determining the scale of damage. Jin et al. (2024) built virtual scenarios based on historical typhoons such as Maemi (0314), Sarah (5914), and Hinnamnor (2211), which struck Korea’s southeastern coast, considering different landfall timings. Using a high-resolution ADCIRC model, they analyzed storm surge heights and overtopping volumes for Jeju Island, the southwest coast, and the southeast coast. Their findings emphasized that even for the same typhoon, the timing of landfall—whether at minimum central pressure or peak wind speed —can make a substantial difference in surge height and wave overtopping. They also highlighted the heightened flood vulnerability of semi-enclosed bays along the southeast coast. Kim et al. (2025) used ADCIRC to numerically evaluate how storm surges vary depending on typhoon tracks approaching the Korean Peninsula from the Yellow Sea, South Sea, or East Sea. In addition to typhoon intensity, they also quantitatively examined the uncertainties in track and coastal topography influencing both the magnitude and duration of storm surges.
In this study, 33 synthetic typhoons generated through probabilistic methods were selected to quantitatively estimate the Maximum Storm Surge Height (MSSH) and the Time-Integrated Storm Surge Height (TISSH) in the coastal waters surrounding the Hanbit Nuclear Power Plant (HNPP) site on the west coast of Korea. Particular attention is given to the point at which the typhoon center passes closest to the site, with storm surge characteristics and potential risk factors assessed based on key meteorological parameters—central pressure, maximum wind speed, radius of maximum winds—and the minimum distance to the site. By adopting this approach, the study goes beyond the limitations of earlier research, which largely focused only on peak surge heights. Instead, it considers both the magnitude of sea level rise and the cumulative effect over time, allowing for a more comprehensive quantification of how typhoon intensity and track influence storm surge responses.
2. Numerical Models and Scenario Generation
2.1 Tropical Cyclone Risk Model (TCRM)
TCRM, developed by Geoscience Australia, is a probabilistic framework designed to simulate tropical depression tracks and wind fields, as well as to assess cyclone-induced hazards. The model uses an autoregressive model to overcome the limitations of relatively short observational records, generating thousands of synthetic typhoons based on the statistical characteristics of historical events. TCRM is organized into five modules: data processing, track generation, wind field calculation, statistical analysis, and risk assessment.
In the track generation module, synthetic tracks are produced using the method proposed by Hall and Jewson (2007), which assumes that historical typhoon tracks follow a normal distribution and generates new tracks on a probabilistic basis. The statistical module estimates the probability distribution parameters for central pressure variations and generates a time series of central pressure using the Monte Carlo method, as follows:
Here, Δt is the time interval;
μ p i σ p i σ p i C o v ( X t , X t - 1 ) / σ X 2 Φ p i 1 - α p i 2
Another key parameter in defining cyclone structure is the radius of maximum wind (Rmax). In this study, it is calculated using the empirical formula proposed by Kim and Suh (2016):
Through these processes, TCRM provides the necessary meteorological parameters for generating synthetic typhoon scenarios, which are subsequently used as input data for storm surge modeling.
2.2 Advanced Circulation Model (ADCIRC)
ADCIRC (Luettich et al., 1992) is a numerical hydrodynamic model based on the finite element method, constructed on an unstructured triangular mesh. The model is particularly well-suited for long-term simulations of large-scale, long-period water surface fluctuations such as tides, storm surges, and tsunamis. In this study, the two-dimensional depth-integrated (2DDI) version of ADCIRC was applied. The governing equations are expressed as follows:
Here, t denotes time, ζ is the water surface elevation relative to mean sea level; R is the radius of the Earth; λ and φ are longitude and latitude; UH and VH represent the volume flux per unit width; f is the Coriolis parameter; g is gravitational acceleration; α is the elastic loading coefficient of the Earth; η is the equilibrium tidal potential; ρ0 is water density; ps is surface atmospheric pressure; and H is the total water depth, defined as the sum of bathymetric depth h and surface displacement ζ. U and V denote depth-averaged velocity components. τs λ and τs φ represent surface wind stresses, while τbλ and τbφ denote bottom friction stresses. Mλ and Mφ are the gradients of depth-averaged lateral stresses; Dλ and Dφ are turbulent diffusion terms; and Bλ and Bφ correspond to baroclinic pressure gradient terms (Dietrich, 2010).
The external forcing applied to ADCIRC—pressure field P (r,θ) and wind field V (r )—is derived from the parametric equations of Holland (1980) using typhoon parameters generated from TCRM:
Here, r is the radial distance from the typhoon center; θ is the azimuth; Pc is the central pressure; Pn is the ambient pressure outside the cyclone’s influence, ρa is the air density, and B is the Holland shape parameter (1–2.5).
2.3 Overview of Synthetic Typhoon Scenarios
As shown in Fig. 1, approximately 3,000 synthetic typhoons were generated using TCRM based on historical typhoon records from 1848 to 2020. From these, 235 typhoons were initially selected according to the following criteria: (i) northward propagation beyond 30°N and (ii) entry into the longitude range of 120°–135°E, thereby posing potential influence on the Korean Peninsula. Among these, 33 synthetic typhoons with westward trajectories toward the Yellow Sea were ultimately chosen based on minimum central pressure. These selected events were then applied as input to the ADCIRC model for storm surge simulations. In addition, the tracks of two historical typhoons that significantly affected the Yellow Sea—Muifa (1109) and Bolaven (1215)—were also plotted in Fig. 1 (JTWC, n.d.) and employed to validate and assess the performance of the ADCIRC model.
The synthetic typhoons entering the Yellow Sea primarily followed a northward track passing west of Jeju Island before traversing the sea basin. The storm surges induced by these events were subsequently estimated. The meteorological characteristics of the 33 typhoons are as follows: minimum central pressures ranged from 915 to 985 hPa, maximum sustained wind speeds ranged from 26.52 to 68.34 m/s, and the minimum approach distances to the HNPP site ranged from 0.517 to 301.46 km.
2.4 Computational Domain and Grid System
The computational domain, shown in Fig. 2, encompasses not only the East Sea, South Sea, and Yellow Sea surrounding the Korean Peninsula but also portions of the northwestern Pacific, including Japan (excluding Hokkaido), Taiwan, and the Ogasawara Islands. To improve computational efficiency and accuracy, a coarse regional mesh was first constructed and then refined with high-resolution grids in the coastal and nearshore zones. The fine-resolution grid was developed using the most up-to-date nautical charts and bathymetric survey data.
In the nearshore areas, the horizontal grid spacing was set to approximately 10 m, with grid resolution progressively coarsening toward the open ocean. The final computational mesh consisted of 398,909 elements and 216,241 nodes, providing a robust, high-resolution framework for storm surge analysis in coastal regions. To minimize spurious reflections at the boundaries, radiation conditions were applied to the open boundaries.
2.5 Validation of the ADCIRC Model
To evaluate the accuracy and reliability of the storm surge simulations, model results were compared with observed storm surge heights generated by historical Typhoons Muifa (1109) and Bolaven (1215). Both events, as shown in Fig. 1, entered the Yellow Sea and propagated northward along the coastline, producing pronounced storm surges along the west coast of Korea. In particular, significant surges were recorded at Incheon and Boryeong tide gauges. Validation was conducted against 16 tide gauge stations in total: 7 along the west coast, 7 along the south coast, and 2 in Jeju Island (see Fig. 3).
For Typhoon Muifa (1109) in Fig. 4(a), the coefficient of determination (R2) between simulated and observed values was 0.742, indicating a strong correlation. At the Mokpo tide gauge, located near the HNPP site, the observed surge of 0.377 m was closely reproduced by the model with a simulated value of 0.363 m, corresponding to a 96.3% agreement. In the case of Typhoon Bolaven (1215) in Fig. 4(b), the model slightly overestimated the surge at Mokpo by 0.173 m compared to the observed 0.525 m. Nevertheless, the overall accuracy was higher than that of Muifa, with an R2 value of 0.755.
These results demonstrate that the ADCIRC model adequately reproduces both the spatial distribution and magnitude of storm surge responses, reflecting the combined effects of regional bathymetry and typhoon tracks. Although certain locations exhibited over- or underestimation, these discrepancies primarily arise from insufficient grid resolution around the complex island-dense Yellow Sea coastline and from decreased representativeness at sites far from the storm tracks. Despite these limitations, the model’s overall agreement with observational data confirms that the ADCIRC framework employed in this study provides a reliable tool for simulating storm surge responses under varying typhoon intensities and trajectories.
3. Numerical Results
3.1 Spatial Distribution of Maximum Storm Surge Height
Fig. 5 presents the simulated spatial distributions of the maximum storm surge height (MSSH) for synthetic typhoons No. 33, No. 15, and No. 9. Their meteorological characteristics—central pressure (pc), maximum sustained wind speed (Vmax), radius of maximum wind (Rmax) —as well as the minimum approach distance to the HNPP site (Dmin) are summarized in Table 1. Here, pc, Vmax, and Rmax represent values corresponding to the time when the typhoon center was located at Dmin from the site.
Although typhoons No. 33 (Fig. 5(a)) and No. 15 (Fig. 5(b)) exhibit comparable approach distances (117.48 km and 127.76 km, respectively), their meteorological properties differ substantially. Typhoon No. 15, with a lower pc and larger Vmax and Rmax, induced broader and more elevated storm surge levels across the Yellow Sea, as shown in Fig. 5(b). In contrast, Typhoon No. 33, despite its weaker intensity, produced a pronounced local surge near Boryeong (Fig. 5(a)) as it tracked closer to the coastline during its northward movement. This observation demonstrates that even relatively weak typhoons can generate considerable localized surges if they pass close to the shoreline.
Typhoon No. 9 (Fig. 5(c)), with meteorological parameters comparable to No. 33, passed approximately 300 km from the HNPP site. Consequently, no significant surges developed across most of the southwestern Yellow Sea or along the adjacent coastlines. This trend underscores that in addition to typhoon intensity, the approach distance and track relative to the target coastline are critical determinants of the spatial extent of storm surge impacts.
These results reaffirm the general relationship that storm surge magnitudes tend to increase with lower central pressure, higher maximum wind speeds, and larger wind radii. Furthermore, the proximity between the typhoon center and the target coast exerts a decisive influence on surge occurrence and spatial distribution.
3.2 Storm Surge Waveform
Fig. 6 illustrates the storm surge waveforms at the HNPP site generated by synthetic typhoons No. 6, No. 8, No. 10, and No. 19. As summarized in Table 1, typhoons No. 6 and No. 19 followed tracks in close proximity to the site, with approach distances (Dmin) of 0.46 km and 4.52 km, respectively. Among them, Typhoon No. 19 was the more intense event, characterized by a lower pc, higher Vmax, and larger Rmax. In contrast, typhoons No. 8 and No. 10 exhibited similar meteorological intensities, but their approach distances differed significantly (188.2 km and 84.39 km, respectively). The waveforms in Fig. 6 were aligned by resetting the time axis relative to the occurrence of the MSSH, which was defined as the reference time (tp).
When typhoons approached the HNPP site directly—as in the cases of No. 6 and No. 19—the storm surge hydrographs rose steeply within a short time window, with No. 19 producing the highest MSSH overall. In contrast, more distant events such as No. 8 and No. 10 generated lower peak water levels but exhibited longer surge durations. Interestingly, Typhoon No. 6, despite being stronger than No. 8 and passing at a closer distance, produced a lower MSSH. This can be attributed to the complex bathymetry and island chains located southwest of HNPP, which acted as natural barriers and attenuated water level amplification.
Meanwhile, Jin et al. (2024) and Kim et al. (2025) have emphasized that storm surge hazard assessments should not rely solely on the peak water level at a single time step. Instead, the overall volume under the surge hydrograph is also a critical indicator. Accordingly, this study adopted the volumetric approach proposed by Lee et al. (2022; 2023) to compute the time-integrated storm surge height (TISSH) from the hydrographs shown in Fig. 6, using Eq. (10). This index reflects both the intensity and persistence of surges, making it a suitable metric for evaluating coastal overtopping and inundation risks.
Here, η(t ) represents the storm surge height at time t, while t1 and t2 denote the onset and termination times of significant surge elevation, respectively.
3.3 Effects of Typhoon Intensity
Figs. 7 and 8 present scatter plots of the maximum storm surge height (MSSH) at the HNPP site and the time-integrated storm surge height (TISSH), computed using Eq. (10), for the 33 synthetic typhoons entering the Yellow Sea. In each figure, panel (a) shows the relationship with central pressure (pc), panel (b) with maximum wind speed (Vmax), and panel (c) with radius of maximum wind (Rmax). The meteorological characteristics and approach distances (Dmin) of the typhoons marked in the figures can be referenced in Table 1.
Because storm surges arise from the combined effects of sea level rise induced by low atmospheric pressure and surface wind stress from strong winds, the meteorological properties of typhoons directly influence the magnitude of surges. As shown in Fig. 7, MSSH exhibits a clear correlation: lower central pressure, higher maximum wind speed, and larger wind radii correspond to greater storm surge heights. A similar pattern is evident in Fig. 8 for TISSH, indicating that larger typhoons produce greater surge volumes over time.
Among the events highlighted in Fig. 7, typhoons No. 19 and No. 28 had comparable intensities and both produced high MSSH. However, despite passing very close to the HNPP site (Dmin = 4.52 km), Typhoon No. 19 generated a lower surge height than Typhoon No. 28 (Dmin = 148.11 km). This can be attributed to the numerous islands southwest of HNPP, which absorbed wave energy and reduced local surge amplification. A similar mitigating effect is evident for weaker Typhoon No. 13, which resulted in very low MSSH.
In contrast, Fig. 8 shows that Typhoons No. 28 and No. 13 yielded relatively high TISSH values. Although their approach distances were large (148.11 km and 127.17 km, respectively), both generated broad spatial distributions of storm surges with prolonged waveforms, consistent with the earlier analysis in Fig. 6. On the other hand, typhoons No. 6 and No. 19, despite their very close passes (0.46 km and 4.52 km), produced narrow surge waveforms and consequently low TISSH. Other weaker typhoons, such as No. 2 and No. 13, consistently showed small TISSH overall.
The combined results of Figs. 7 and 8 confirm the general trend that stronger typhoons produce higher MSSH and TISSH. However, they also reveal that outcomes can differ substantially under similar intensity conditions depending on the storm track and coastal geomorphology. For example, typhoons passing very close to the site often induced high but short-lived surges, resulting in relatively small cumulative effects, whereas typhoons passing at intermediate distances produced longer-duration surges, yielding higher TISSH despite lower peak water levels. This underscores the importance of considering both typhoon track and local coastal conditions in addition to intensity when evaluating storm surge risks.
These results reaffirm that meteorological factors such as central pressure, maximum wind speed, and radius of maximum wind are key determinants of storm surge magnitude. At the same time, they demonstrate the limitations of assessing surge risk based solely on typhoon intensity. A comprehensive evaluation must also account for the proximity of the typhoon track to the target region and the coastal characteristics of the affected area.
3.4 Effects of Closest Approach Distance
Figure 9 presents scatter plots showing the influence of the closest approach distance (Dmin) of typhoons on (a) MSSH and (b) TISSH at the HNPP site. The meteorological characteristics of each typhoon are summarized in Table 1.
As shown in Fig. 9(a), MSSH generally increases as Dmin decreases. This reflects the fact that shorter distances between the typhoon center and HNPP reduce energy dissipation during surge propagation, amplifying the direct coastal impact. For example, typhoons No. 19 and No. 28 exhibited similar intensities, yet No. 19, with a much smaller Dmin of 4.52 km, generated a lower MSSH than No. 28. As discussed earlier, this arises from the shielding effect of the islands located southwest of HNPP, which dampened the surge elevation. In contrast, the weak Typhoon No. 13 produced very low MSSH regardless of its approach distance, underscoring that surge height cannot be explained solely by proximity.
Meanwhile, Fig. 9(b) depicts the distribution of TISSH with respect to Dmin. In general, as the typhoon track shifts farther from HNPP, the propagation distance of surge waves increases, leading to lower peak water levels but prolonged water-level rise, resulting in larger TISSH up to a certain distance. Afterward, as Dmin increases, the storm’s influence diminishes, leading to a decreased trend in TISSH. This indicates that strong typhoons passing within an intermediate distance can generate larger TISSH values. For instance, Typhoon No. 15 (Dmin = 127.17 km) produced greater TISSH than Typhoon No. 28 (Dmin = 148.11 km), despite having higher central pressure and weaker intensity. In contrast, very weak typhoons such as No. 13 generated consistently low TISSH, highlighting the critical role of both storm intensity and coastal conditions in addition to proximity.
3.5 Assessment of Potential Storm Surge Risks
The numerical results demonstrate that the scale of storm surges at the HNPP site is governed by the combined effects of typhoon meteorological factors, closest approach distance, and interactions with the complex island topography. Notably, intermediate-distance typhoons often produced larger TISSH than those passing in close proximity. This can be explained by the long-duration accumulation of surge waves generated by strong but more distant typhoons, which maintained moderately elevated water levels over extended periods. This finding underscores the limitations of using peak MSSH alone as an indicator, as it risks underestimating the persistence of overtopping and inundation. In contrast, volume-based indices such as TISSH provide a more reliable measure of disaster intensity. Therefore, by integrating the water level relative to the HNPP site’s ground elevation, one could directly estimate overtopping duration, overtopping volume, inundation extent, and inundation depth, thereby producing practical storm surge risk indicators.
Geographically, the HNPP site benefits from natural terrain features that provide partial buffering against storm surges. The dissipation, attenuation, and blocking effects of the archipelagic setting are especially effective against typhoons passing in close proximity. This geographical advantage suggests that HNPP enjoys relatively favorable coastal defense conditions compared to many other parts of the west coast.
These findings confirm that maximum surge height should be complemented by cumulative measures when assessing storm surge hazards. They also highlight the decisive influence of coastal morphology, such as island clusters, on surge behavior. Future analyses should therefore adopt methodologies capable of explicitly reflecting coastal conditions. Examples include high-resolution computational grids in numerical models, parameterizations that account for regional topographic features, and coastal vulnerability indices designed to quantify geomorphic influences. Such approaches would enable a comprehensive hazard assessment that integrates typhoon intensity, storm track, and coastal conditions.
4. Conclusions
This study evaluated the potential storm surge hazards at the HNPP site on the western coast of Korea using 33 synthetic typhoons generated with the TCRM framework. The analysis examined the quantitative response of MSSH and TISSH to typhoon meteorological characteristics (central pressure, maximum wind speed, and radius of maximum wind) and closest approach distance.
Storm surge simulations conducted with ADCIRC yielded the following key findings: First, MSSH increased with lower central pressure, higher maximum wind speed, and larger wind radii, reaffirming typhoon intensity as a primary factor. Second, typhoons passing very close to HNPP did not always generate the highest surge heights, largely due to the complex influence of the surrounding archipelago. Third, TISSH analysis, which incorporated the duration and cumulative volume of elevated sea levels, revealed that strong typhoons at intermediate distances often produced larger cumulative effects than closer tracks. These results demonstrate the limitations of assessing risk based solely on peak surge height and underscore the importance of adopting volumetric indices.
Additionally, the geographical setting of Yeonggwang, where HNPP is located, offers a degree of natural protection due to the surrounding archipelago. This buffering effect was particularly evident for typhoons passing nearby, confirming the MSSH and HNPP’s relatively favorable position in terms of coastal hazard mitigation.
Future research should integrate additional factors such as storm waves (Seo et al., 2023; Hwang et al., 2024), tide-surge interactions (Park et al., 2010; Yang et al., 2019), and sea-level rise scenarios (IPCC, 2021) to evaluate compound hazards and strengthen long-term safety assessments for nuclear power plant sites. Furthermore, high-resolution inundation modeling using three-dimensional numerical models (Hur et al., 2019; Hwang et al., 2022) will be necessary to quantify overtopping depth and inundation extent. Such analyses will enhance the practical evaluation of site-specific vulnerabilities and support the development of comprehensive disaster mitigation strategies that account for future climate conditions.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Print
