Research and Practice

Principal Investigator

Xuan Liu（BreathingCORE Limited）

Co-investigator from IRIDeS

Erick Mas（Disaster Geo-informatics Lab）

Summary of the Research

This project develops SimDisaster: LLM agents simulate tsunami evacuation at population scale with a Disaster World Model (DWM) tracking road passability, expected inundation depth, and safe-area capacity so agents reason before acting. Agent personas are first grounded in published post-disaster survey findings (delay reasons, mode choice, information-seeking behavior) and then refined through pattern-oriented calibration against multiple aggregate patterns from NTT DOCOMO Mobile Spatial Statistics for the 2024 Noto Peninsula earthquake—departure-time curves, zone-to-zone flows, and safe-area arrivals. Agents produce interpretable natural-language decisions throughout.

Principal Investigator

Hiroki Ishibashi（Nihon University）

Co-investigator from IRIDeS

Shunichi Koshimura（Disaster Geo-informatics Lab）

Summary of the Research

Earthquake damage to bridges can significantly degrade the road network functionality, which is a key performance indicator of resilience. This project develops a resilience-based disaster management framework for bridges under seismic and tsunami hazards to enhance disaster preparedness for megathrust earthquakes anticipated in Japan. The framework integrates probabilistic hazard analysis, bridge vulnerability assessments and network-level performance (i.e., resilience) evaluation. A bi-level optimization model is formulated in the framework. The upper-level problem determines optimal earthquake countermeasures for individual bridges to maximize network resilience. Under the countermeasures determined at the upper level, the lower-level problem optimizes post-earthquake bridge repair scheduling under resource constraints to maximize the resilience of the road network during the recovery process.

Principal Investigator

Shigeo Tatsuki（Doshisha Univercity）

Co-investigator from IRIDeS

Shunichi Koshimura（Disaster Geo-informatics Lab）

Summary of the Research

Based on the operational workflow for developing individualized evacuation plans (including evacuation drills) in Kijo Town, Miyazaki Prefecture, this study reconstructs the process into a Social Digital Twin that incorporates state variables, temporal dimensions, feedback mechanisms, and simulation functions, enabling alignment with real-world operations.
Furthermore, by linking this model with an individualized evacuation plan database, the study implements a personalized risk communication system that delivers real-time reminders of required actions according to alert levels, while also confirming receipt and subsequent actions for each individual.

Principal Investigator

Shohei Nagata（IRIDeS, Tohoku University）

Summary of the Research

This study examines the applicability of social impact assessment methods using human mobility data for the international deployment of disaster digital twins. Building on the research team’s previous work in Japan, it investigates whether approaches developed to monitor evacuation behavior and post-disaster recovery processes can be applied in other national contexts. The study consists of three main components: a workshop with Japanese and U.S. researchers and stakeholders, a comparison of human mobility datasets from the two countries, and analyses of human mobility during actual disasters in the U.S. Through these efforts, the study will clarify the conditions, scope, and limitations of mobility-based social impact assessment in the U.S. context and provide foundational knowledge to support the future international development of disaster digital twins.

Principal Investigator

Bruno Adriano (IRIDeS,Tohoku University)

Summary of the Research

Building a highly accurate 3D "Digital Twin" to predict tsunami impacts requires detailed underwater maps (bathymetry). Currently, acquiring this data using traditional ship surveys is slow and expensive. This project will develop a groundbreaking, cost-effective method to estimate coastal water depths directly from space by combining AI with optical and radar (SAR) satellite imagery. Focusing on the high-risk area of Kochi City, we will build Japan's first comprehensive AI training datasets for shallow- and intermediate-depth mapping. This rapid, scalable approach will revolutionize how coastal communities worldwide update hazard models and prepare for future tsunamis.

Principal Investigator

Shosuke Sato（IRIDeS, Tohoku University）

Summary of the Research

This research advances the high-fidelity "Disaster Digital Twin" developed under the JST PRESTO program by integrating dynamic evacuation logs of actual survivors. The project focuses on creating multiple interactive scenarios that reproduce realistic evacuation obstacles, such as safety checks and decision-making delays. Through empirical experiments with 50 participants, we will quantitatively evaluate the system's impact on changing the evacuation mindset of non-victims, aiming to establish a highly effective, evidence-based disaster education model.

Principal Investigator

Ray Y. Chuang（National Taiwan University）

Co-investigator from IRIDeS

Erick Mas（Disaster Geo-informatics Lab）

Summary of the Research

This research develops a disaster digital twin framework to reconstruct a landslide dam failure event in Taiwan and associated population responses. By integrating environmental data and mobile phone–based mobility data, the study represents both hazard processes and human dynamics within a unified system. An agent-based model (ABM) is used to simulate evacuation behavior, where individuals respond to changing conditions. The results are used to build a dynamic platform that allows continuous updates of hazard and population changes. This case-study framework aims to compare observed and simulated evacuation responses to support scenario-based preparedness analysis.

Principal Investigator

Shinji Yamashita（NIPPON KOEI CO.,LTD.）

Co-investigator from IRIDeS

Erick Mas（Disaster Geo-informatics Lab）

Summary of the Research

Building on insights obtained from our previous validation of an evacuee estimation method, this study aims to improve the validity of scenario and parameter settings in flood evacuation simulations. To this end, we will develop a macroscopic mobility (flow) prediction model using neural networks that incorporates disaster information, damage history, and land-related data. The resulting predictions will be integrated into the parameterization of an agent-based (microscopic) evacuation simulation to enhance both accuracy and validity, with the ultimate goal of supporting evacuation planning and the evaluation of mitigation measures.

Principal Investigator

Misao Hashimoto（Gifu University）

Co-investigator from IRIDeS

Akihiro Shibayama（Disaster Culture and Archive Studies）

Summary of the Research

Disaster prevention researchers from Gifu University, Co-Innovation University, Tohoku University,Shinshu University, Kumamoto University, Senshu University and the National Research Institute for Earth Science and Disaster Prevention will bring together their accumulated knowledge of regional disaster archives and conduct research on the construction of a platform that enables cooperation and collaboration to facilitate system construction, operation and maintenance. The utilisation of the archives constructed will also be discussed and be implemented.

Principal Investigator

Hiroyuki Miura（Hiroshima University）

Co-investigator from IRIDeS

Osamu Murao（International Strategy for Disaster Mitigation Lab）

Summary of the Research

This study will develop indicators of disaster recovery using time-series analysis that leverages satellite remote sensing data and machine learning for developing countries where spatial data infrastructure is insufficient. We will conduct an analysis of the areas affected by the 2018 Sulawesi earthquake in Indonesia to verify the validity and usefulness of the proposed indicators.

Principal Investigator

Amalia Nafisah Rahmani Irawan（Tohoku University）

Co-investigator from IRIDeS

Daisuke Komori（Water Disaster Risk Research Lab）

Summary of the Research

This study develops and validates a flood hazard and risk S-curve framework for two contrasting river catchments in West Java, Indonesia, the Citarum and Cimanuk basins. Relying solely on globally available open datasets: ERA5 precipitation, Global Surface Water (GSW), and WorldPop population data, integrated with the Rainfall-Runoff-Inundation (RRI) model, the framework quantifies at 500 m grid scale how effectively DRR investments have reduced flood risk relative to fundamental hazard levels over 1986 – 2025 as main study period. The contrasting DRR histories of the two basins; Citarum's long-established multi-reservoir system versus Cimanuk's Jatigede reservoir operational only since 2015, provide a natural comparative setting for validating the method's sensitivity to detect real differences in risk reduction effectiveness.

Principal Investigator

Aiko Sakurai（Kobe University

Co-investigator from IRIDeS

Takeshi Sato（Disaster Education Research and Implementation Lab）

Summary of the Research

In Ishinomaki City, which marks the 15th anniversary of the Great East Japan Earthquake, this study conducts a tripartite survey involving students, teachers, and parents in collaboration with the municipal Board of Education.
This survey is uniquely enabled by the research team’s continuous engagement in both field-based support and academic research over the past 15 years since the disaster.
By examining the structure of tsunami risk understanding and evacuation decision-making across developmental stages, the study empirically demonstrates how disaster education is translated into a continuum from experience to understanding, judgment, and action, thereby contributing to the foundation for enhancing disaster resilience.

Principal Investigator

Julia Gerster（IRIDeS, Tohoku University）

Summary of the Research

In FY24, we created disaster kamishibai based on interviews with non-Japanese disaster survivors and presented them at an international symposium on diversity and disaster risk reduction. In FY25, we summarized the historical reasons for the success of kamishibai in disaster education and our findings that these performances enhance mental well-being among performers. With extended funding, we aim to further examine the effects of kamishibai on audience understanding of diverse needs in disaster contexts. This next phase will focus on audience surveys, performer interviews, the preparation of co-authored publications grounded in these findings, and a final dissemination workshop.

Principal Investigator

Elizabeth Maly（IRIDeS, Tohoku University）

Summary of the Research

Facing multi-hazard risks, both Taiwan and Japan have experienced water-related disasters and developed various methods of recording disaster experience and passing down disaster memories. Based on analaysis of disaster and risk narratives conveyed by museums, memorial sites, and community practices, the project explores the transformation of disaster memory into learning. Collaborative research and participatory workshops aim to advance disaster communication, learning and adaptation in hazard-prone environments, and strengthening human and social-ecological resilience. Through the development of a disaster communication game, this project aims to strengthen resilience in everyday life and better prepare for future risks in Japan and Taiwan.

Principal Investigator

Tomoya Kobayashi（Tohoku University）

Co-investigator from IRIDeS

Koichi Chida（Radiological Disasters and Medical Science Lab）

Summary of the Research

Rapid and accurate identification of disaster victims is essential for enhancing family well-being andoverall disaster resilience. Clarifying the causes of death among disaster victims is also crucial forpreventing disaster-related mortality.The objective of this research is to evaluate the utilization of postmortem imaging (Autopsy imaging:Ai) for personal identification and cause-of-death investigation in mass disaster situations.

Principal Investigator

Taro Kataoka（Hirosaki University）

Co-investigator from IRIDeS

Atsushi Kawauchi（Uehiro Disaster Risk Reduction Research Division）

Summary of the Research

This research proposes a specific model for a network-based protection system to safeguard and transmit local cultural heritage in depopulated areas. Drawing on case studies from depopulated regions in Aomori and Iwate Prefectures, among others, it identifies key issues and examines how collaboration among history, cultural heritage studies, and conservation science can help protect local cultural heritage from disasters.

Principal Investigator

Shin-ya Horiai（Hachinohe Institute of Technology）

Co-investigator from IRIDeS

Makoto Okumura（Regional Resilience Planning Lab

Summary of the Research

The city of Hachinohe City is selected as the study area, and population distribution patterns are estimated using mobile spatial statistics data. Based on the estimated results, evacuation direction guidance is derived from an optimization-based evacuation model on the existing road network, and its feasibility is evaluated through an agent-based tsunami evacuation traffic simulation. Furthermore, changes in bottlenecks according to actual residential patterns and the effectiveness of corresponding countermeasures are examined.

Principal Investigator

Anna Shinka (IRIDeS,Tohoku University)

Summary of the Research

This study aims to clarify the cognitive processes and individual differences involved in the understanding and use of hazard maps, and to establish design principles for spatial information presentation and education that enable actionable interpretation by all. By combining brain activity measurement, eye tracking, and questionnaires, the study analyzes how expertise and cognitive traits relate to map reading, and is expected to yield insights that contribute to inclusive disaster risk reduction.

Principal Investigator

Fuko Nakai（The University of Tokyo）

Co-investigator from IRIDeS

Kanako Iuchi（Regional Resilience Planning Lab）

Summary of the Research

This study investigates post-disaster life recovery processes following the 2024 Noto Peninsula Earthquake by integrating mobile positioning data with retrospective travel diaries collected through individual interviews. Through the contextual enrichment of mobile positioning data based on actual life recovery experiences, the research identifies critical limitations in mobile positioning data, including data gaps and biases. The study further develops new mobility indicators and analytical methods designed to more accurately reflect the nuances of daily life during recovery. Ultimately, this research provides a methodological framework for utilizing big data to capture the lived experiences of disaster survivors, to fill the gap between digital information and the ground-level reality.

Principal Investigator

Takashi Oda（The University of Tokyo）

Co-investigator from IRIDeS

Takeshi Sato（Disaster Education Research and Implementation Lab）

Summary of the Research

This study aims to develop a sustainable model of inclusive disaster risk reduction that supports the safety and independence of children with disabilities, based on two strands of practice: seven years of continuous implementation and evaluation of disaster education in a special needs school in Tochigi Prefecture, and disaster management practices to enhance the safety of school bus commuting in Miyagi Prefecture.

Principal Investigator

Hyejeong Park（IRIDeS, Tohoku University）

Summary of the Research

This research qualitatively investigates the risks of Natural Hazard-Triggered Technological (Natech) risks within the Wildland-Urban Interface (WUI) areas, using the 2025 WUI fires in Ofunato, Japan, and Yeongnam, South Korea. When WUI fires physically intersect with critical infrastructure, institutional misalignments often arise between civil evacuation and infrastructure-containment protocols. To address this structural vulnerability, the study employs a symmetrical comparative methodology, examining official government documents and existing spatial hazards to establish physical Natech risk zones. Finally, it assesses cross-sector protocols to identify regulatory gaps and structural contradictions within these high-risk boundaries. The primary outcome is a cross-national WUI-Natech policy framework. This will provide local municipalities with actionable, evidence-based recommendations to align infrastructure governance with cascading hazard risks.

Principal Investigator

Hirokazu Shimauchi（Future University Hakodate ）

Co-investigator from IRIDeS

Reika Nomura（Computational Safety Engineering Lab）

Summary of the Research

This project investigates a real-time tsunami wavefield prediction framework that integrates latent representation learning of offshore tsunami observation data with Physics-Informed Neural Networks (PINNs). Preliminary experiments on about 500 simulated tsunami scenarios in a restricted region suggest that the approach can interpolate among relatively similar scenarios to some extent, while also revealing reduced accuracy in coastal areas. Building on this foundation, the project will construct a dataset of approximately 12,000 tsunami simulation scenarios for the Pacific coast of Tohoku and examine the applicability of the framework to a broader range of tsunami-generation conditions. Particular attention will be paid to the relation between latent representation quality and prediction accuracy, and to sampling and loss design in coastal regions. Through these analyses, the study seeks to assess the feasibility and limitations of physics-aware machine learning for real-time tsunami prediction.

Principal Investigator

Shoko Araki（Iwate University）

Co-investigator from IRIDeS

Akihiro Shibayama（Disaster Culture and Archive Studies）

Summary of the Research

This study examines the 2025 Ofunato City forest fire to clarify the actual conditions of housing and livelihood reconstruction from the emergency response phase through the recovery and reconstruction processes, focusing on the interrelationship between residential location choices and financial support. In particular, in WUI (Wildland–Urban Interface) areas where forests and urban areas are in close proximity, residential locations and livelihoods (such as fishing and agriculture) are closely intertwined. Consequently, decisions regarding housing reconstruction are not merely a matter of site selection but are characterized by constraints related to the continuation of livelihoods and local institutional frameworks. Based on questionnaire surveys and interviews, this study analyzes (1) the actual conditions of emergency housing selection and residential site selection, (2) the actual application of insurance, mutual aid, and public support, and (3) the structural relationship between the two. Through this analysis, the study aims to clarify the decision-making process for rebuilding lives following small-scale disasters and the impact of those decisions on livelihoods.

Principal Investigator

Nilo Lemuel J. Dolojan（IRIDeS, Tohoku University）

Summary of the Research

This study will propose a new framework to quantify the attribution of landslide hazard and risk to climate change, called "impact-based attribution". This will be achieved by the fusion of physics-based slope stability simulations and large ensemble simulations of heavy rainfall under historical, current, and projected climate conditions. The simulation results will quantify how climate change-induced changes in rainfall intensity, duration, and distribution affect slope stability and landslide risks. The proposed research will transform conventional climate change attribution studies of extreme weather events to practical "impact-based" assessments of climate change on disaster events.

Principal Investigator

Daisuke Komori（IRIDeS, Tohoku University）

Summary of the Research

With regard to low-frequency, catastrophic disasters, geological investigations̶such as analyses of tsunami deposits and ice core samples̶and collaborative research with historians on historical documents have revealed the history of natural disasters spanning hundreds of years; however, there are no case studies of low-frequency, large-scale disasters on a centuries-long timescale specifically concerning flood disasters. Therefore, focusing on the Chao Phraya River in Thailand, we aim to estimate the scale and frequency of past floods based on changes in the elevation of temples built over several centuries in flood-prone areas of slow-flowing rivers, and to undertake the challenge of reproducing the worst-ever flood on record and assessing the risk of low-frequency, large-scale disasters on a centuries-long timescale. As slow-flowing rivers on the continent have low flood flow velocities and uniform flood patterns, it is presumed that, before the advent of modern technology, temple foundations were empirically designed based on the natural disaster risks and disaster experiences of the time. Therefore, we will identify flood-prone areas and the temples established there long ago using historical inundation maps, and model the relationship between the probable flood depth at these temples̶determined through field surveys and numerical modelling̶and changes in the height of the temple embankments. This research, which is rare even on a global scale, is expected to serve as a useful tool for contributing to Sustainable Development Goal (SDG) 11: Sustainable Cities and Communities.

Principal Investigator

Ryohei Yamashita（Ishikawa Prefectural University）

Co-investigator from IRIDeS

Yuta Hara（2030 Global DRR Agenda Office）

Summary of the Research

Research fatigue following catastrophic disasters and the inefficiency of reconstruction efforts—stemming from insufficient coordination among the public, private, and academic sectors—remain persistent challenges . In addition to conventional survey methods, modern information tools and media may exert both positive and negative influences on the intensification or mitigation of research activities . Using the Noto Peninsula Earthquake as a case study, this research aims to grasp the reality of these issues, decode their underlying mechanisms, and return the obtained insights to the field.

Principal Investigator

Judit Kroo（University of Melbourne ）

Co-investigator from IRIDeS

Julia Gerster（Disaster Memory Studies Lab）

Summary of the Research

This research investigates how post-disaster communities in Tohoku and the Noto Peninsula transform practices of remembering and forgetting into resources for future-making. Building on the team's previous project examining the dialectic between memory and forgetting after the 2011 and 2024 disasters, this extension develops a new theoretical framework examining how the lingering presence of disaster's material and affective traces (“haunting”) becomes a site where communities negotiate possible futures. Through ethnographic fieldwork, interviews, and collaborative analysis, the project will produce co-authored publications, and a collaborative workshop on how to balance opposing opinions on preserving disaster memories for sustainable and resilient community recovery.

Principal Investigator

Kenta Tozato（Hachinohe Institute of Technology）

Co-investigator from IRIDeS

Shuji Moriguchi（Computational Safety Engineering Lab）

Summary of the Research

This study aims to conduct numerical simulations of slope hazards caused by heavy rainfall in Hachinohe City, and to evaluate the risk of slope hazards over a wide area. Using the analysis results, we will construct a framework that enables real-time prediction of slope disaster hazard from rainfall information.

Principal Investigator

Motohiro Tsuboi（Japanese Red Cross Saitama Hospital）

Co-investigator from IRIDeS

Shinichi Egawa（International Cooperation for Disaster Medicine Lab）

Summary of the Research

This study aims to elucidate the systemic challenges underlying indirect disaster-related deaths in Japan by analyzing anonymized case records of individuals officially recognized as indirect disasterrelated deaths, obtained through information disclosure requests. In addition to cross-sectional analyses of variables such as time of death, cause of death, age, and sex, the study focuses on evacuation living conditions to examine factors and potential countermeasures related to indirect disaster-related deaths through both quantitative and qualitative approaches. Based on the findings, the study seeks to contribute to the visualization and reduction of indirect disaster-related health impacts by formulating recommendations and building consensus on the appropriate documentation of disaster relevance in death certificates and the definition of indirect disaster-related deaths.

Principal Investigator

Yuta Hara（IRIDeS, Tohoku University）

Summary of the Research

This study focuses on Inner Mongolia, China, and aims to conduct exploratory research to understand “sedentarization-based knowledge” in agro-pastoral communities of Han and Mongolian residents, as well as integrated resource management in the Yellow River basin. By examining afforestation and resettlement policies alongside the issue of soil salinization from a basin-wide perspective, the project seeks to reconcile livelihood improvement with the sustainable preservation of agro-pastoral cultures, while establishing a foundation for international collaborative research with the Institute of Grassland Research (IGR) of the Chinese Academy of Agricultural Sciences (CAAS).

Principal Investigator

Wei Yuan（IRIDeS, Tohoku University）

Summary of the Research

This research proposes a curriculum learning framework to improve human mobility resilience analysis under extreme weather events. It first characterizes weather shocks using snow depth and related hazard indicators. It then examines how mobility changes both across grids (inter-grid) and within grids (intra-grid) to capture disruption patterns at multiple spatial scales. Based on these weather–mobility interactions, the study develops a spatiotemporal prediction framework that progressively learns from easy to difficult patterns. Using the Niigata snow disaster as an illustrative case, the project aims to build a generalizable framework for mobility disruption assessment and disaster response support.

Principal Investigator

Kumiko Nakano（Tohoku University）

Co-investigator from IRIDeS

Yuta Hara（2030 Global DRR Agenda Office）

Summary of the Research

This study examine the environmental and ecological impacts of permafrost thaw as a slow-onset climate-related hazard in Arctic Alaska. It aims to conceptualize and identify the impacts of permafrost thaw, such as land deformation, inundation, related housing damage, and potential health risks from the release of water and microorganisms stored in permafrost, as a form of slow-onset disaster. 　By integrating geoscientific, epidemiological, and humanistic knowledge, the project aims to assess and respond to these risks through a participatory approach. Conducted as a Japan–U.S. collaborative effort, the study seeks to develop community-based risk communication strategies that enhance the disaster resilience and preparedness of Indigenous communities in Arctic Alaska. The project integrates spatial analysis and microbial ecology. Using Gambell, Alaska as a case study, Geographic Information System (GIS)–based analyses were conducted to evaluate land vulnerability and identify potential relocation suitability areas by integrating topographic and environmental spatial data. In parallel, permafrost samples were analyzed using 16S rRNA gene sequencing to characterize microbial community diversity and composition under thaw conditions. Future expansion of the study will include northern communities such as Wainwright, where thaw impacts are expected to be more pronounced. Ultimately, the project seeks to develop an integrated framework linking slow-onset environmental change, spatial vulnerability, and microbial ecological dynamics, contributing to adaptation planning and long-term risk assessment in Arctic regions.

Principal Investigator

Rieko Takahashi（IRIDeS, Tohoku University）

Summary of the Research

We will conduct a nationwide survey of individual evacuation plans at special needs schools, analysing best practices in reasonable accommodation and location trends using GIS. By providing ongoing guidance to education authorities based on scientific data, we aim to establish a ‘disaster management model’ that ensures the safety of pupils is equally guaranteed in every region of the country, regardless of local environmental conditions, and to develop a highly effective system.

Principal Investigator

Rieko Takahashi（IRIDeS, Tohoku University）

Summary of the Research

One of the social challenges surrounding children and adults requiring medical care during disasters is that their needs are sometimes ‘overlooked and excluded’. To address this, it is necessary to secure human resources among support staff and to foster the acquisition and development of multifaceted perspectives. In this study, we will design and evaluate the effectiveness of an educational intervention aimed at ‘expanding the cognitive framework’ for new support staff, with the aim of developing a training package that can be rolled out to other regions.

Principal Investigator

Yo Fukushima（IRIDeS, Tohoku University）

Summary of the Research

By applying insights from public health, this study aims to establish a practical foundation for "Bosai Communication Studies" that promotes evidence-based behavioral change. Centered on a collaborative platform between academia and practitioners, we will develop a system for sharing case studies, facilitate interdisciplinary workshops, and co-design community-based intervention prototypes. This research validates a co-creation model that circulates practical wisdom from the field and academic insights.

Principal Investigator

Sebastien P. Boret（IRIDeS, Tohoku University）

Summary of the Research

This research draws on three years of research (2023–2025) examining the state of inclusive Disaster Risk Reduction (DRR) among persons with disabilities (PwDs) since the 2004 Tsunami in Indonesia. Our initial findings show that post-tsunami global and national policies frequently fail to translate into accessible infrastructure for Acehnese persons with disabilities. Building on this empirical insight, the study develops a novel framework that conceptualises disability as a form of social capital, analyses how societal capacity shapes inclusion, and proposes a multi-layered pathway to integrate their specialised knowledge into formal systems, thereby strengthening disaster resilience across at the governance and grassroots levels.

Principal Investigator

Sebastien P. Boret（IRIDeS, Tohoku University）

Summary of the Research

This pilot study examines how inclusive artificial intelligence (AI) can help address care gaps for persons with disabilities (PwDs) and older people during evacuation. Focusing on the Great East Japan Earthquake, it investigates both the potential and the challenges of deploying AI in real-world evacuation settings. The study draws on field surveys of PwDs and older adults, as well as interviews with practitioners involved in disaster response and care. By identifying critical gaps in existing support systems, the study establishes a baseline for inclusive, technology-enabled disaster risk reduction and informs transdisciplinary collaboration toward more resilient and accessible evacuation strategies.

Principal Investigator

Natsuko Chubachi（IRIDeS, Tohoku University）

Through semi-structured interviews with citizens and Disaster Risk Reduction (DRR) officials, we will identify elements and contexts that encourage or hinder citizens' everyday DRR behavior and awareness. In particular, this research will focus on citizens' behavior and perception regarding the seismic reinforcement of private houses and the construction of a DRR social network in their community. Those two are critical for DRR but challenging for many people to implement. The findings will help to identify and resolve weaknesses in everyday DRR.

Principal Investigator

Chiaki Oguchi（Institute of Science Tokyo）

Co-investigator from IRIDeS

Takayuki Takahashi（Inland Earthquake and Volcano Lab）

Summary of the Research

This research will focus on the alluvial lowlands of the Taya area, located in the middle reaches of the Kashio River in Sakae Ward, Yokohama City. By conducting intensive boring surveys and radiocarbon dating, and reconstructing the paleotopography and paleoenvironment with high precision from historical ground elevation and flood inundation levels, we aim to make clear the relationship between the location of Taya caves, which have existed since the Kamakura period (14th century), and flood risk avoidance, thereby contributing to the evaluation of modern flood control methods for small and medium-sized rivers.

Principal Investigator

Naoto Nihonmatsu（University of Toyama）

Co-investigator from IRIDeS

Ryo Saito（Cognitive Sciences Lab）

Summary of the Research

The purpose of this study is to clarify the mechanisms underlying the factors that promote or hinder satisfactory interaction within the family risk communication process. Through semi-structured interviews with married couples, we will collect accounts of their decision-making processes during the nuclear accident. In particular, we will focus on and examine the processes by which role perceptions, differences in evaluation criteria, and evaluative dissonance arise between spouses.

Principal Investigator

Takeshi Sato（IRIDeS, Tohoku University）

Summary of the Research

In order to promote the Sendai Framework for Disaster Risk Reduction in the education sector, and while the Comprehensive School Safety Framework (GADRRES) is being developed, this project will promote the mainstreaming of disaster risk reduction in the education sector by conducting pioneering research on support coordination functions and support receiving plans for schools in disaster-stricken areas, based on surveys and analyses of existing support frameworks such as DMAT, DHEAT, and GADM.

Principal Investigator

Kumpei Tsuji（Tohoku University）

Co-investigator from IRIDeS

Reika Nomura（Computational Safety Engineering Lab）

Summary of the Research

This study aims to clarify the mechanical mechanisms underlying the disaster mitigation functions of forests by developing a numerical framework that integrates tree morphology generation and fluid–structure interaction analysis. A tree morphology generation method based on the L-system and the Space Colonization Algorithm is employed to automatically generate three-dimensional tree models with branching and tapering structures. These models are incorporated into fluid simulations to analyze the interaction between trees and surrounding flow fields.
The developed framework enables fluid–structure interaction analyses that account for both the complex morphology of trees and their anisotropic material behavior. Detailed simulations are first conducted for single-tree models to investigate how tree morphology and material properties influence flow structures and hydrodynamic forces. The framework is then extended to forest-scale simulations using beam-element finite element models, allowing efficient representation of large numbers of trees. Through systematic numerical experiments, key indicators such as flow reduction, drag characteristics, and energy dissipation are evaluated to quantify the flow attenuation effects of forests. The results are expected to provide a quantitative understanding of the mechanical mechanisms through which forest structures contribute to disaster mitigation.

Principal Investigator

Itaru Morita（Nippon Koei Co.,Ltd.）

Co-investigator from IRIDeS

Erick Mas（Disaster Geo-informatics Lab）

Summary of the Research

This research aims to quantify changes in population dynamics and economic activity during disasters using human flow data, and to construct a general-purpose early post-disaster estimation (preliminary estimation) model applicable to multiple disasters. Through this, we aim to establish evaluation technologies that contribute to the resilience of regional industries and other sectors.