Background and objective: As climate change intensifies extreme rainfall events, the global risk of landslides continues to rise. In Korea, existing landslide hazard maps and warning systems often prove inadequate for accurately predicting large-scale or complex slope failures. Consequently, empirical models like LAHARZ, which enable the rapid assessment of debris-flow volume-extent relationships, are drawing increased attention. This study examines whether LAHARZ can be effectively applied in Yecheon-gun a non-volcanic region to efficiently estimate landslide damage extents and strengthen local disaster preparedness.

Methods: We collected debris-flow occurrence data from a July 2023 heavy rainfall event and derived a 5 m-resolution DEM from 1:5,000 topographic maps. Using satellite imagery, we delineated observed debris-flow areas in three sub-regions (Baekseok-ri, Geumgok-ri, Beolbang-ri) and input these datasets into LAHARZ to generate estimated damage extents, which were then compared with actual damage footprints.

Results: In the simpler terrain of Baekseok-ri, the model’s estimated damage extents closely matched observed debrisflow footprints, demonstrating robust performance. However, at multi-source sites (Beolbang-ri and Geumgok-ri), the estimated inundation ranges were either over- or underestimated. These discrepancies underscore the challenge of accurately capturing flow confluence when relying solely on volume-driven empirical formulas.

Conclusion: LAHARZ proves effective for rapid damage-extent estimation in single-source debris-flow events, yet it may fail to capture key physical interactions in more complex terrains where multiple initiation points merge. Incorporating high-resolution DEM data, physical modeling techniques, and field measurements is recommended to enhance the model’s accuracy and applicability in non-volcanic regions such as Yecheon County.

Key words: debris flow modeling, runout modeling, landslide hazard mapping, gis-based approach, laharz

Introduction

Landslides cause thousands of incidents globally each year, resulting in severe human and economic damage (Petley, 2012). In Korea, intensified rainfall events and more frequent heavy downpours attributed to climate change are further elevating landslide risks (IPCC, 2021). This problem is not merely tied to topographic characteristics; it also closely interacts with ongoing climatic shifts, continually threatening local communities.

Currently, domestic landslide risk management in Korea relies primarily on the Korea Forest Service’s hazard maps, landslide warnings, and erosion control projects (KFS, 2024c). However, for unexpected large-scale events or areas with complex terrain, these systems exhibit limited predictive capability, revealing vulnerabilities in proactive prevention. A notable example is the large-scale landslide that struck Yecheon-gun, Gyeongsangbuk-do, in July 2023. This event was not anticipated by existing hazard maps and thus caused extensive damage (KFS, 2024a). Such incidents underscore the urgent need for more scientific and systematic methods of landslide risk assessment (Kritikos and Davies, 2015; Chen et al., 2016).

In recent years, numerous studies have focused on quantifying the mobility of debris flows often referred to as “flow-like landslides” and producing automated hazard maps in GIS environments (Griswold and Iverson, 2008; Schilling, 2014). Many case studies show that erosion or deposition processes during debris-flow movement can significantly affect the inundation range and overall impact (Iverson, 2006; Yu et al., 2012; Mergili et al., 2012; Gregoretti et al., 2019). Consequently, overcoming the limitations of current risk assessment systems calls for advanced methods capable of evaluating debris-flow hazards under diverse climate and terrain conditions (Horton et al., 2013; Mergili et al., 2012). Moreover, as extreme rainfall events grow more frequent with climate change, large-scale debris flows often exceed the applicability range of conventional hazard maps (Ding et al., 2023; Guthrie et al., 2021), highlighting the pressing need for new models that capture these dynamic processes. In Korea, some recent research in the Gokseong area has integrated multi-remote sensing data to compare observed landslide damage and topographic changes with model estimations (Choi et al., 2024), emphasizing the importance of field data in enhancing model accuracy.

Models for estimating landslide runout generally fall into two categories: empirical and physical. Empirical models rely on statistical relationships between past debrisflow volumes and inundation areas, allowing relatively simple and rapid analyses of potential impact. A typical example is the LAHARZ model, which employs a power- law relationship between debris-flow volume and inundation distance (Schilling, 1998; Iverson et al., 1998). Although empirical models excel at quick hazard mapping over large regions, they have inherent limits in resolving finer-scale processes such as deposition or fluid viscosity. Physical models, conversely, are based on fluid dynamics and continuum mechanics, offering more detailed simulations of flow behavior but at considerable computational cost (Corominas et al., 2014). Thus, empirical models are advantageous for broad-area screening and swift decision- making, while physical models are better suited for focused analyses near specific structures or small-scale features.

Among empirical approaches, the LAHARZ model stands out for its ability to generate large-scale debris-flow hazard maps with minimal input data, making it particularly valuable when rapid assessments and broad coverage are required (Griswold and Iverson, 2008). Its empirical power-law formula allows practitioners to approximate runout extents using only flow volume estimations and a digital elevation model, eliminating the need for complex fluid dynamics simulations. This advantage becomes especially crucial for emergency response in data-sparse regions or when time-sensitive decisions are needed.

Based on this background, this study applies the LAHARZ model to a past debris-flow event in Yecheon-gun in order to estimate the resulting landslide runout and propose evidence- based risk management strategies. By deploying LAHARZ in a non-volcanic setting and comparing the model outputs with actual runout footprints, the research aims to refine risk zoning and improve local disaster preparedness. We anticipate that our findings will serve as foundational data for landslide risk management and proactive countermeasures in other regions with similar terrain characteristics.

In conclusion, this study presents a concrete example of addressing rising landslide risks driven by climate change. By demonstrating the practicality of data-driven modeling and quantitative estimation techniques, LAHARZ-based landslide runout simulations provide valuable reference material for enhancing both local community safety and sustainable land management.

Research Methods

Study area

Study area

During the heavy rainfall from July 9 to 19, 2023, Yecheon-gun in Gyeongsangbuk-do experienced multiple, simultaneous landslides (debris flows). In some areas, these events were not anticipated by existing landslide risk maps, raising concerns about limitations in current fore casting systems. Based on this background, the present study employs the LAHARZ model to analyze debris-flow runout in three locations within Yecheon-gun, Baekseok-ri in Hyoja-myeon, Geumgok-ri in Eunpung-myeon, and Beolbang-ri in Gamcheon-myeon where human casualties and property damage were reported (Fig. 1).

According to the Yecheon-gun Office (2024), Yecheon-gun is located in southeastern South Korea. Its northern region, influenced by the Sobaek Mountain Range, is predominantly mountainous, whereas the central and southern parts consist of lowlands and basins, forming a north-high-to-south-low topographical gradient. The Nakdong and Naeseongcheon Rivers also traverse the county, which spans about 661 km², 55.4% of which is forested.

Over the past five years, the average July rainfall in Yecheon-gun was around 100 mm. However, in July 2023, more than 600 mm of precipitation was recorded (YCDMIS, 2024), marking a substantial increase compared to previous years. Such a surge in rainfall suggests that localized downpours were the primary trigger for landslides. Indeed, data from the Korea Forest Service, spanning 2014 to 2023, indicate that 32 landslides occurred in Yecheon-gun in July 2023 alone, with some large-scale events transpiring beyond existing hazard-map zones and exacerbating damage to local communities.

Accordingly, this study applies the LAHARZ model to estimate debris-flow runout at the three selected sites Baekseok-ri, Geumgok-ri, and Beolbang-ri and re-evaluates high-risk zones based on the results. Through this approach, the study aims to propose strategies for improving landslide risk management in Yecheon-gun.

Framework

Framework

This research adopts a four-step procedure to estimate debris-flow runout using the LAHARZ model and to delineate high-risk zones (Fig. 2). In Step 1, areas in Yecheon-gun with reported human casualties and property damage were prioritized, guiding the selection of study sites. A 5 m × 5 m-resolution DEM was then generated from 1:5,000- scale topographic maps, and debris-flow event data from 2023 were compiled to form the base dataset. In Step 2, the debris-flow point data and DEM were converted within ArcGIS into a format compatible with LAHARZ. Debrisflow volumes and empirical coefficients were then determined, and the primary input parameters for model operation were configured.

In Step 3, the LAHARZ model was run to estimate debris- flow volumes using the Planimetric Area-Volume relationship. The flow paths and inundation ranges (or run-out extents) were subsequently calculated, and a GIS environment was used to visualize the results and identify high-risk areas. Finally, in Step 4, the modeled extents were compared with observed landslide footprints to evaluate accuracy, discuss applicability, and highlight limitations. By following these four steps, the study quantitatively assesses debris-flow runout and validates the effectiveness of the LAHARZ model in Yecheon-gun.

Data and materials

Data and materials

This study used reported landslide runout extents and initiation points from the heavy rainfall of July 2023 as key inputs to the LAHARZ model. Initially, we reviewed the 2023 landslide data provided by the Korea Forest Service (KFS, 2024). However, the KFS dataset either omitted certain areas such as Beolbang-ri or showed smaller damage extents than what could be observed from Google Earth imagery, making it difficult to pinpoint the actual runout footprints based on KFS data alone. Therefore, satellite imagery dated October 31, 2024, from Google Earth (2024) was consulted to remeasure the impacted areas. According to this imagery, Baekseok-ri (Hyoja-myeon) exhibited approximately 34,000 m² of runout from a single initiation point; Beolbang-ri (Gamcheon-myeon) recorded about 45,000 m² across two initiation points; and Geumgok-ri (Eunpung-myeon) totaled around 90,000 m² at three combined points. This indicates that the actual affected areas in Baekseok-ri and Geumgok-ri exceeded the figures reported by KFS, while Beolbang-ri did not appear in the KFS data at all. Consequently, the Google Earth-derived measurements serve as more complete baseline information for evaluating the LAHARZ model outputs against the observed zones of inundation or deposition.

To further investigate the terrain characteristics of the study sites, we obtained 1:5,000-scale topographic maps from the National Geographic Information Institute (NGII) and used them to create a 5 m × 5 m-resolution DEM. This DEM underwent preprocessing in ArcGIS such as coordinate system alignment and resolution adjustments before being converted into a format compatible with the LAHARZ model. By integrating the landslide initiation-point data with the DEM, we established a foundational dataset crucial for estimating flow volume and analyzing debris-flow pathways, both of which are essential for operating LAHARZ effectively (Table 1).

LAHARZ modeling

LAHARZ modeling

LAHARZ is a GIS-based empirical model developed by the U. S. Geological Survey (USGS) for the rapid estimation of lahar (volcanic mudflow) hazard zones (Schilling, 1998; Schilling, 2014). Although it was originally designed for volcanic terrain, subsequent work by Griswold and Iverson (2008) has demonstrated its applicability to non-volcanic debris flows and other flow-like landslides, thereby expanding its utility for generalized landslide hazard prediction. More recently, a Python- and GIS-based adaptation called Laharz_py has emerged, enabling automated debris-flow path and inundation (deposition) estimates based solely on a DEM and flow volume data(V).

A key advantage of LAHARZ lies in its straightforward power-law relationships, which allow hazard delineation across large areas from minimal input data. As it neither requires explicit modeling of fluid dynamics nor extensive geotechnical parameters, LAHARZ can be run quickly, making it particularly useful for broad-scale risk screening and time-sensitive decision-making. This makes the model attractive not only in volcanic contexts but also for diverse non-volcanic settings where rapid or wide-ranging assessments are essential.

LAHARZ proposes three principal approaches for debris- flow analysis: (1) Analyzing proximal hazard zones to derive cross-sectional area, (2) Using historical case data to estimate planimetric area, and (3) Drawing on other physical or empirical models to estimate flow volume. In this study, we adopt the second approach employing empirically derived planimetric-area relationships so that LAHARZ outputs can be compared directly with observed debris-flow runout extents. A key characteristic of LAHARZ lies in its empirical regression linking debris- flow volume (V) to cross-sectional area (A) and planimetric area (B) (Fig. 3). By examining multiple debris-flow case studies, Iverson et al. (1998) identified a power-law relationship in which χ₂ acts as an empirical coefficient bridging volume and the respective area terms.

These straightforward equations allow cross-sectional and planimetric areas to be back-calculated from volume alone. Conversely, if the planimetric area (for example, a debris-flow footprint observed on satellite imagery) is known, the corresponding volume (V) can be derived using the equation presented in Fig. 4. This regression-based approach enables rapid, large-scale estimation of landslide runout without resorting to more complex physical simulations, making it suitable for various flow-like landslides, including those in non-volcanic settings.

In this study, LAHARZ was applied to debris-flow events in Yecheon-gun via a planimetric-based analysis to test its viability in non-volcanic terrain. Specifically, debris-flow runout footprints (inundation areas) identified through satellite imagery served as planimetric inputs, while the empirical coefficient χ₂ was systematically varied (10, 20, 30, 40, and 50) to estimate flow volumes using Equation (3) (Table 2). Since the LAHARZ regression model was largely derived from international debris-flow cases, we hypothesized that applying a single coefficient to Korean non-volcanic debris flows might introduce inaccuracies. Accordingly, multiple χ₂-value scenarios were examined to encompass a range of volume estimates reflecting potential variations in geology, scale, and particle composition. Under each scenario, LAHARZ was then used to simulate debris-flow paths and inundation areas, and the outputs were compared with observed runout footprints to assess the suitability of each volume estimate.

Results and Discussion

This study applied the LAHARZ model to simulate landslide runout in Baekseok-ri, Geumgok-ri, and Beolbang-ri in Yecheon-gun, then compared the outputs with actual observations. The findings revealed distinct characteristics in each area.

First, in Baekseok-ri, where a single source point triggered the debris flow, the modeled runout around the village closely matched the observed data, suggesting that the empirical model captured flow behavior with reasonable accuracy. Among the various χ₂-value scenarios, χ₂= 10 provided the closest match to the actual runout footprint. However, because only a 5 m-resolution DEM derived from a 1:5,000 topographic map was used, small-scale features such as village structures, drainage channels, and agricultural contours were underrepresented. As a result, some flow paths were predicted to be shorter than those observed in reality (Fig. 5).

Next, in Beolbang-ri, two simultaneous landslides created a more complex flow pattern. According to the simulation results, for the upper source area on the map, χ₂= 10 generated a flow path most similar to the observed pattern, yet it overestimated the overall runout extent. In contrast, for the left source area, χ₂= 50 best approximated the actual runout footprint. These discrepancies underscore the difficulty of accurately modeling velocity changes and deposition processes caused by debris-flow confluence and divergence using only simple empirical equations (Fig. 6).

Finally, Geumgok-ri, which had debris flows from three source points merging into one, presented the most challenging scenario: no single set of parameter values adequately captured both the flow paths and the final runout footprint (Fig. 7). This shortfall likely arises from the power-law relationship between volume and area, which does not fully account for velocity changes or collision effects at confluence points, leading to overestimation in some areas and underestimation in others.

Overall, the LAHARZ model reproduced the inundation extent relatively well in simpler settings like Baekseok-ri (a single source), with χ₂= 10 emerging as the best fit. However, in more complex settings with multiple simultaneous debris flows such as Beolbang-ri and Geumgok-ri overestimation and underestimation occurred concurrently, highlighting the model’s inherent limitations in representing the physical interactions of converging debris flows.

In this study, the LAHARZ model demonstrated varying degrees of accuracy across different sites in Yecheon-gun. For single-source debris flows, as exemplified by Baekseok-ri, χ₂= 10 yielded the most accurate inundation extent, implying that a straightforward empirical equation can sufficiently simulate debris-flow travel distances and deposition patterns. However, for multi-source settings like Beolbang-ri and Geumgok-ri, the model’s outputs showed both overestimation and underestimation. Overestimation often occurred when debris flows from multiple initiation points converged, effectively combining volumes and increasing runout distances beyond actual observations. Underestimation was also observed where terrain factors or partial channel blockages caused actual debris-flow paths to diverge or deposit earlier than the model anticipated. These discrepancies highlight how (V^2/3)-type volume-area regressions can oversimplify complex flow dynamics. In multi-source events, velocity changes, collisions at confluence points, and intermittent deposition can significantly alter the final runout footprint. Furthermore, the 5 m-resolution DEM used in this study did not fully capture small drainage features, agricultural terraces, or local obstacles, which are critical for accurately modeling velocity dispersion and deposition thickness. Such limitations in topographic resolution and reliance on a single empirical coefficient (χ₂) mean that LAHARZ may both over-estimate and underestimate debris-flow extents under more intricate terrain conditions (Griswold and Iverson, 2008). Despite these shortcomings, the Baekseok-ri case reaffirms that LAHARZ remains highly efficient for rapid hazard- zone mapping of single-source debris flows, or at least for initial broad-scale assessments in non-volcanic terrain like Yecheon-gun. The model succeeded in reproducing the overall deposition pattern in simpler contexts, indicating that when large-scale or multi-source slope failures are not involved, LAHARZ can serve as a practical first-screening tool.

Conclusion

This research applied the LAHARZ model to debrisflow events in Baekseok-ri, Beolbang-ri, and Geumgok-ri (Yecheon-gun) and compared the modeled runout extents with observed data. The findings suggest that, although prior studies identified large-scale or multi-source debris flows as a limitation for empirical models, the single- source setting in Baekseok-ri confirms that LAHARZ can deliver reasonably accurate hazard delineations even in non-volcanic terrain. Hence, some of the previous study limitations, such as rapid mapping of simpler debris flows, have been effectively mitigated. However, the multi- source issue remains unresolved, as evidenced by the over- and underestimations in Beolbang-ri and Geumgok-ri. In multi-source contexts, volume-area formulas alone fail to capture dynamic processes such as velocity changes at confluence points, collisions between flows, and partial deposition. These factors lead to either excessive spread (overestimation) or premature deposition (underestimation), explaining the significant regional differences observed among the three study sites.

Recommended improvements involve incorporating physical modeling and in situ measurements to refine the volume-area relationship, especially where flows merge. Using higher-resolution DEMs or similarly detailed elevation data can further reduce discrepancies, and calibrating χ₂ for Korean geological contexts (for instance, considering soil composition and weathering profiles) may enhance predictive accuracy. Going forward, Korean-specific guidelines for the LAHARZ model could be established, acknowledging the country’s steep forested terrains and frequent heavy downpours. Site-specific calibration or hybrid approaches combining empirical and simplified physical models may provide a more robust framework for identifying landslide risk areas across the country. In particular, rapid or wide-ranging assessments can still rely on LAHARZ for preliminary zoning, but high-risk or multi-source areas may require additional modeling layers or higher-fidelity data.

Overall, while LAHARZ may not fully resolve the multi- source slope failure limitation highlighted in previous studies, it does offer a rapid, cost-effective solution for single-source or simpler debris flows a meaningful improvement in large-scale hazard mapping. By integrating higher-resolution terrain data, basic physical parameters, and contextual calibration, the model can evolve into a more reliable tool for complex terrains in Korea. This detailed case study from Yecheon-gun underscores the ongoing need to refine empirical models and incorporate complementary approaches, ultimately enhancing landslide- risk management in regions prone to sudden debris flows under changing climatic conditions.

Notes

This research was supported by the Cheongju University Research Scholarship Grants in 2023-2024.

REFERENCES

References

Chen, H.X., S. Zhang, M. Peng, L.M. Zhang. 2016. A physically-based multi-hazard risk assessment platform for regional rainfall-induced slope failures and debris flows. Engineering Geology. 203:15-29. https://doi.org/10.1016/j.enggeo.2015.12.009
[Article]

Choi, S.K., R.A. Ramirez, H.H. Lim, T.H. Kwon. 2024. Multi-source remote sensing-based landslide investigation: the case of the August 7, 2020, Gokseong landslide in South Korea. Scientific Reports. 14(1):12048.https://doi.org/10.1038/s41598-024-59008-4
[Article] [PubMed] [PMC]

Corominas, J., C. Van Westen, P. Frattini, L. Cascini, J.-P. Malet, S. Fotopoulou, F. Catani, M. Van Den Eeckhaut, O. Mavrouli, F. Agliardi, K. Pitilakis, M.G. Winter, M. Pastor, S. Ferlisi, V. Tofani, J. Hervás, J.T. Smith. 2014. Recommendations for the quantitative analysis of landslide risk. Bulletin of Engineering Geology and the Environment. 73:209-263. https://doi.org/10.1007/s10064-013-0538-8
[Article]

Ding, X.Y., W.J. Hu, F. Liu, X. Yang. 2023. Risk assessment of debris flow disaster in mountainous area of northern Yunnan province based on FLO-2D under the influence of extreme rainfall. Frontiers in Environmental Science. 11:1252206.https://doi.org/10.3389/fenvs.2023.1252206
[Article]

Google. 2024 Google Earth [Satellite imagery] Retrieved from https://earth.google.com.

Gregoretti, C., L.M. Stancanelli, M. Bernard, M. Boreggio, M. Degetto, S. Lanzoni. 2019. Relevance of erosion processes when modelling in-channel gravel debris flows for efficient hazard assessment. Journal of Hydrology. 568:575-591. https://doi.org/10.1016/j.jhydrol.2018.10.001
[Article]

Griswold, J.P., R.M. Iverson. 2008. Mobility statistics and automated hazard mapping for debris flows and rock avalanches. US Geological Survey No.2007–5276
[Article]

Guthrie, R., A. Befus. 2021. DebrisFlow Predictor: an agent-based runout program for shallow landslides. Natural Hazards and Earth System Sciences. 21(3):1029-1049. https://doi.org/10.5194/nhess-21-1029-2021
[Article]

Horton, P., M. Jaboyedoff, B.E.A. Rudaz, M. Zimmermann. 2013. Flow-R, a model for susceptibility mapping of debris flows and other gravitational hazards at a regional scale. Natural Hazards and Earth System Sciences. 13(4):869-885. https://doi.org/10.5194/nhess-13-869-2013
[Article]

IPCC. 2021 Summary for policymakers. Climate Change 2021: The Physical science basis, contribution of working group i to the sixth assessment report of the intergovernmental panel on climate change Retrieved from https://www.ipcc.ch/report/ar6/wg1/.
[Article]

Iverson, R.M. 2006. Forecasting runout of rock and debris avalanches. Landslides from massive rock slope failure Dordrecht: Springer Netherlands. (pp. 97-209). https://doi.org/10.5194/nhess-13-869-2013
[Article]

Iverson, R.M., S.P. Schilling, J.W. Vallance. 1998. Objective delineation of areas at risk from inundation by lahars. Geological Society of America Bulletin. 110(8):972-984.

Korea Forest Service (KFS). 2024a Historical Landslide Information Retrieved from https://www.data.go.kr/data/15074816/openapi.do.

Korea Forest Service (KFS). 2024b landslide hazard map Retrieved from https://map.forest.go.kr/forest/?systype=mapSearch&searchOption=landslide&longitude=14180192.1832882&latitude=4350576.8953995&scale=6.

Korea Forest Service (KFS). 2024c Landslide Information System Retrieved from https://sansatai.forest.go.kr/gis/main.do;jsessionid=qei7U_VUp1KqVxRmqKQIP6iHDHtuNe58pexVKidl.ldm10#mhms0.

Kritikos, T., T. Davies. 2015. Assessment of rainfall-generated shallow landslide/debris-flow susceptibility and runout using a GIS-based approach: application to western Southern Alps of New Zealand. Landslides. 12:1051-1075. https://doi.org/10.1007/s10346-014-0533-6
[Article]

Mergili, M., K. Schratz, A. Ostermann, W. Fellin. 2012. Physically-based modelling of granular flows with Open Source GIS. Natural Hazards and Earth System Sciences. 12(1):187-200. https://doi.org/10.5194/nhess-12-187-2012
[Article]

Petley, D. 2012. Global patterns of loss of life from landslides. Geology. 40(10):927-930. https://doi.org/10.1130/G33217.1
[Article]

Schilling, S.P. 1998. LAHARZ: GIS programs for automated mapping of lahar-inundation hazard zones. US Geological Survey, No. 98–638
[Article]

Schilling, S.P. 2014. Laharz_py: GIS tools for automated mapping of lahar inundation hazard zones. US Geological Survey, No. 2014–1073
[Article]

Schneider, D., C. Huggel, W. Haeberli, R. Kaitna. 2011. Unraveling driving factors for large rock-ice avalanche mobility. Earth Surface Processes and Landforms. 36(14):1948-1966. https://doi.org/10.1002/esp.2218
[Article]

Yecheon-gun Disaster Prevention Information System (YCDMIS). 2024 Local Rainfall Report Retrieved from https://www.data.go.kr/data/15074817/openapi.do.

Yecheon-gun Office. 2024 Location and Climate Retrieved from https://www.ycg.kr/open.content/ko/organization/basis.status/location.weather/.

Yu, G.A., W. Yao, H.Q. Huang, Z. Liu. 2021. Debris flows originating in the mountain cryosphere under a changing climate: A review. Progress in Physical Geography: Earth and Environment. 45(3):339-374. https://doi.org/10.1177/030913332096170
[Article]