Objective In recent years, the global trend of climate change has become increasingly pronounced, accompanied by a rising frequency and intensity of extreme weather events, particularly torrential rains. Rain-induced secondary water-related disasters such as flash floods, debris flows, and urban waterlogging pose serious threats not only to human lives and property but also to cultural heritage sites, especially in urban areas with fragile topography and limited disaster resilience. Despite growing awareness of the importance of cultural heritage disaster risk management in heritage conservation, a lack of systematic theoretical frameworks and empirical research persists in risk prediction and the formulation of disaster mitigation strategies that emphasize prevention and rapid response. Addressing this gap, the present study focuses on cultural heritage sites in Hitoyoshi City, Kumamoto Prefecture, Japan, which experienced severe flood damage due to extreme rainfall. This study aims to utilize publicly available geospatial datasets in conjunction with hydrological flood inundation simulations to systematically examine the interrelationships among the spatial distribution of cultural heritage sites, topographical and environmental characteristics, and material properties. The objective is to uncover the compound factors contributing to flood vulnerability and to construct a scientifically rigorous and operationally-feasible risk identification framework that can enhance the disaster preparedness of cultural heritage assets under the escalating impacts of climate change.

Methods The research methodology integrated multiple data sources. First, publicly available geospatial datasets were collected, including high-resolution digital elevation model (DEM) and land use classification data. Second, based on hydrometeorological data from the Japan Meteorological Agency, the HEC-RAS hydrological model was employed to simulate the flood inundation extent caused by the heavy rainfall event in July 2020. Third, GIS-based spatial analysis was conducted by overlaying the locations of cultural heritage sites with the simulated flood extent to identify high-risk areas in Hitoyoshi City vulnerable to flood hazards. Fourth, through the integration of GIS and geospatial data from the Geospatial Information Authority of Japan, a comprehensive analysis was performed from the perspectives of heritage site distribution, surrounding environmental conditions, and structural/material characteristics of the buildings, to elucidate the underlying causes and compound factors of disaster vulnerability. Finally, based on field survey data, the applicability and accuracy of the geospatial datasets for flood risk analysis and cultural heritage protection were assessed, and the practical feasibility of applying these data for disaster mitigation planning was discussed.

Results This study performed a flood inundation simulation based on the extreme rainfall event of July 2020. The results revealed a high degree of spatial consistency with the official peak discharge of the Kuma River (approximately 8,000 m³/s), thereby confirming the reliability of the model's input parameters. Subsequent accuracy assessments demonstrated that the model conducted well in identifying flood-prone areas, providing a robust foundation for cultural heritage vulnerability assessments and disaster risk reduction planning. Among 23 flood-affected cultural heritage sites, 19 were within the inundation zone and 4 outside it, highlighting the limitations of using flood extent alone for risk prediction. Notably, 15 sites were densely concentrated in Kamiaoi-machi, an area with high heritage density and severe damage, indicating the need for targeted mitigation strategies. Based on the levee design elevation standard (approximately 100 m), topographic analysis classified the study area into five elevation categories. Approximately 87% of the damaged sites were situated in low-lying areas (100~110 m), indicating high exposure and vulnerability to flooding. However, 13% of the damaged sites were located at higher elevations, where low ground permeability and structural vulnerability also contributed to flood damage. Land use analysis revealed that about 78% of the affected sites were located within impervious built-up areas, which also had the largest inundation extent (308.6 hectares). These areas are prone to surface runoff accumulation under extreme rainfall conditions, leading to increased flood exposure and risk. Furthermore, the simulated flood depths showed strong correlation with field survey observations. In areas where depths exceeded one meter, cultural heritage sites frequently suffered significant inundation or structural damage, indicating the reliability of simulated depth values. Most damaged structures were wooden, whose inherent fragility made them particularly vulnerable to flood impacts. Field observations further confirmed that structural materials significantly influence flood resilience. In conclusion, cultural heritage flood risk is shaped not only by location within floodplains but also by the combined effects of low elevation, impervious land use, and fragile building materials. This study affirms that flood damage to heritage sites is a compound disaster, requiring multi-dimensional environmental analysis for effective risk identification and planning.

Conclusion This study integrates hydrological modeling with geospatial analysis to effectively identify high-risk disaster areas affecting urban cultural heritage. Through a comprehensive analysis of spatial distribution, topographical features, surrounding environments, and structural materials of affected heritage sites, the study identifies the compound factors contributing to disaster vulnerability. Based on these findings, targeted disaster mitigation strategies are proposed, and the theoretical and practical value of geospatial technologies in cultural heritage risk prediction and disaster management is further explored. In light of increasing disaster risks driven by climate change, particularly in local cities with limited financial and technical resources, this study proposes and validates a risk identification method using publicly available geospatial data. This approach offers both operational simplicity and high accuracy, demonstrating strong practical feasibility for enhancing disaster resilience in resource-constrained urban areas. The methodology and findings of this study may also serve as valuable references for flood-risk prediction, disaster prevention planning, and policy development related to cultural heritage protection in China.