5:15 PM - 7:15 PM
[HDS10-P04] An Attempt to Extract High-Risk Areas Utilizing Probabilistic Tsunami Hazard Information
Keywords:Tsunami, Risk
Introduction
The National Research Institute for Earth Science and Disaster Resilience has published probabilistic hazard information regarding maximum tsunami height along coastlines as part of its probabilistic tsunami hazard assessment results (e.g., Fujiwara et al. (2020)). Additionally, Nakamura et al. (2025) have conducted calculations for land areas targeted at the Nankai Trough earthquake, progressing the development of tsunami hazard information that addresses not only coastal areas but also inland inundation. In recent years, tsunami hazard information based on numerous tsunami propagation and inundation calculations, which reflect significant uncertainty regarding tsunami risk, has started to be publicly released.
On the other hand, research on methods for utilizing this hazard information is still in progress. Kito et al. (2024) proposed a levee risk assessment using conditional exceedance probabilities at the crest height and cost-benefit analysis methods for levee heightening. However, efforts to link tsunami hazard information with exposure levels in surrounding areas for regional risk assessment on a broader scale have not been sufficiently carried out. This study attempts to extract high-risk areas by linking tsunami hazard information targeted at the Nankai Trough earthquake from Fujiwara et al. (2020) and Nakamura et al. (2025) with exposure levels in land areas.
Tsunami Hazard and Exposure
The tsunami hazard information utilized includes the hazard data from Fujiwara et al. (2020) focused on maximum water level increases along coastlines and the hazard information from Nakamura et al. (2025) targeting inundation depths. For the purpose of data simplification, the inundation depths were set to a 50m mesh unit. Additionally, population data at a 250m mesh unit from the 2020 National Census was employed to represent exposure levels.
Extraction of High-Risk Area
(1) Population within Flooded Areas
By overlaying the maximum inundation depths from Nakamura et al. (2025) with the population data at the 250m mesh unit, the population within the flooded areas was extracted. The extracted population was aggregated at the municipal level, organizing it based on the total population within the flooded area and the population percentage represented in the municipalities. If there was at least one computational grid within the 250m mesh that was inundated, the entire population of that 250m mesh was extracted as the population within the flooded area.
(2) Exposed Population Not Protected by Levees
First, the "scoring of levees" was performed using the method from Kito et al. (2024). The score of the levee is defined as the conditional non-exceedance probability, given that the maximum water level increase equals the levee crest height. The levee crest height is referenced from the Cabinet Office (2013), while the maximum water level increase at the levee uses tsunami hazard information from Fujiwara et al. (2020) along the coastline. Next, the 250m meshes on land were related to the levees, associating each mesh with the nearest levee in a straight line. Finally, by extracting the population in the 250m meshes with scores below a certain threshold for the nearest levee, the population not protected by levees was identified. Furthermore, the extracted population was aggregated at the municipal level, and the results were displayed based on both the total population and population percentage for each municipality (Figure 1).
Summary
By overlaying tsunami hazard information with exposed populations, an attempt was made to extract high-risk areas. In this effort, not only the population that could be inundated was extracted, but also the exposed population that could not be protected by levees was identified using conditional hazard related to levee crest heights. Utilizing the population that cannot be protected by levees can contribute to efficient risk management, such as prioritizing levee improvements and other related measures.
This study was conducted as part of the research project "Research and Development on Natural Disaster Hazards and Risks" by the National Research Institute for Earth Science and Disaster Resilience.
The National Research Institute for Earth Science and Disaster Resilience has published probabilistic hazard information regarding maximum tsunami height along coastlines as part of its probabilistic tsunami hazard assessment results (e.g., Fujiwara et al. (2020)). Additionally, Nakamura et al. (2025) have conducted calculations for land areas targeted at the Nankai Trough earthquake, progressing the development of tsunami hazard information that addresses not only coastal areas but also inland inundation. In recent years, tsunami hazard information based on numerous tsunami propagation and inundation calculations, which reflect significant uncertainty regarding tsunami risk, has started to be publicly released.
On the other hand, research on methods for utilizing this hazard information is still in progress. Kito et al. (2024) proposed a levee risk assessment using conditional exceedance probabilities at the crest height and cost-benefit analysis methods for levee heightening. However, efforts to link tsunami hazard information with exposure levels in surrounding areas for regional risk assessment on a broader scale have not been sufficiently carried out. This study attempts to extract high-risk areas by linking tsunami hazard information targeted at the Nankai Trough earthquake from Fujiwara et al. (2020) and Nakamura et al. (2025) with exposure levels in land areas.
Tsunami Hazard and Exposure
The tsunami hazard information utilized includes the hazard data from Fujiwara et al. (2020) focused on maximum water level increases along coastlines and the hazard information from Nakamura et al. (2025) targeting inundation depths. For the purpose of data simplification, the inundation depths were set to a 50m mesh unit. Additionally, population data at a 250m mesh unit from the 2020 National Census was employed to represent exposure levels.
Extraction of High-Risk Area
(1) Population within Flooded Areas
By overlaying the maximum inundation depths from Nakamura et al. (2025) with the population data at the 250m mesh unit, the population within the flooded areas was extracted. The extracted population was aggregated at the municipal level, organizing it based on the total population within the flooded area and the population percentage represented in the municipalities. If there was at least one computational grid within the 250m mesh that was inundated, the entire population of that 250m mesh was extracted as the population within the flooded area.
(2) Exposed Population Not Protected by Levees
First, the "scoring of levees" was performed using the method from Kito et al. (2024). The score of the levee is defined as the conditional non-exceedance probability, given that the maximum water level increase equals the levee crest height. The levee crest height is referenced from the Cabinet Office (2013), while the maximum water level increase at the levee uses tsunami hazard information from Fujiwara et al. (2020) along the coastline. Next, the 250m meshes on land were related to the levees, associating each mesh with the nearest levee in a straight line. Finally, by extracting the population in the 250m meshes with scores below a certain threshold for the nearest levee, the population not protected by levees was identified. Furthermore, the extracted population was aggregated at the municipal level, and the results were displayed based on both the total population and population percentage for each municipality (Figure 1).
Summary
By overlaying tsunami hazard information with exposed populations, an attempt was made to extract high-risk areas. In this effort, not only the population that could be inundated was extracted, but also the exposed population that could not be protected by levees was identified using conditional hazard related to levee crest heights. Utilizing the population that cannot be protected by levees can contribute to efficient risk management, such as prioritizing levee improvements and other related measures.
This study was conducted as part of the research project "Research and Development on Natural Disaster Hazards and Risks" by the National Research Institute for Earth Science and Disaster Resilience.