Disasters Collide: Scientists Map a New Path Beyond the Sendai Framework

When a magnitude 9.0 earthquake struck northeastern Japan in March 2011, it did not stop at shaking the ground. The quake unleashed a tsunami that overtopped protective barriers, and the flooding cascaded into the Fukushima nuclear disaster, overwhelming response systems and exposing vulnerabilities that no single-hazard plan had anticipated. That tragic chain of events has become the defining case study for what scientists call multi-hazard risk: the reality that floods, earthquakes, heatwaves, and landslides rarely behave as isolated problems. A new perspective paper published in the journal Geoscience Communication, led by Timothy Tiggeloven of Vrije Universiteit Amsterdam and more than forty co-authors, argues that the world’s flagship disaster risk agreement, the Sendai Framework for Disaster Risk Reduction 2015–2030, is falling short of its goals precisely because it has struggled to put this interconnected understanding into practice.

The paper synthesises discussions from the 3rd International Conference on Natural Hazards and Risks in a Changing World, held in Amsterdam on 12–13 June 2024 and organised by the MYRIAD-EU project, RISK-KAN, and the NatRiskChange research training group. Roughly 280 scientists and practitioners attended, spanning 14 specialised sessions. The organisers also surveyed 86 participants before the meeting, about 40 percent of attendees, asking them to name the biggest impediments to understanding and managing multi-hazard risks over the next five years. Respondents, two-thirds of whom were early career researchers, repeatedly pointed to funding constraints, data limitations, fragmented governance, and poor communication between science and practice. The team then conducted an inductive thematic analysis of session reports and a post-conference expert consultation, distilling the material into four perspective themes aligned with the gaps identified in the Sendai Framework’s 2023 Mid-Term Review.

That Mid-Term Review delivered a sobering verdict. Most of the framework’s seven global targets are unlikely to be met by 2030, with particular shortfalls in reducing the number of people affected by disasters, economic losses, and damage to critical infrastructure. Progress has been made on expanding early warning system coverage and international cooperation, but the review found a persistent failure to prevent the creation of new risks, as disaster impacts continue to rise in absolute terms despite growing investment in risk reduction. The authors of the new paper cluster the implementation gaps into three interlocking problems: persistent knowledge and data gaps, fragmented governance across scales and sectors, and insufficient resources and capacity. Crucially, these problems reinforce one another. Knowledge gaps justify siloed approaches, siloed governance suppresses demand for integrated knowledge, and resource constraints block both knowledge generation and governance reform.

The first perspective theme addresses assessments and tools for resilient decision-making. The authors highlight narrative-based scenario methods known as storylines, which describe plausible ways that risks might unfold under different social, environmental, and policy conditions. Because storylines can blend qualitative and quantitative data and be co-created with communities, they offer a way to communicate complex multi-risk interactions to non-specialists. The British Geological Survey maintains an online repository of storylines and storymaps designed for exactly this kind of public engagement. The paper also catalogues a growing ecosystem of decision-support platforms: RiskScape, an open-source engine for assessing risks from volcanic eruptions, earthquakes, and floods; CLIMADA for climate risk and adaptation modelling; the MYRIAD-EU dashboard for multi-hazard scenario generation; and DAPP-MR, which uses interactive visualisations of dynamic adaptation pathways to support multi-risk decisions under uncertainty. Platforms such as FloodAdapt incorporate citizen input to guide local interventions and promote equitable recovery.

Yet the authors caution that sophisticated tools often remain confined to expert users, limiting their uptake by decision-makers and communities. They recommend co-developing tools with end-users from the outset, establishing interoperability standards and open-access repositories, and investing in boundary organisations and knowledge brokers who translate between scientific, policy, and community contexts. Data governance emerges as a central obstacle. Satellite datasets from national space agencies arrive in multiple formats with inconsistent documentation, proprietary insurance risk models keep their methodologies opaque, and institutional competition for funding discourages data sharing across organisations. The paper notes that some valuable products, such as NASA’s high-definition VIIRS Black Marble nighttime light data, remain restricted to collaborators on funded projects. It also flags a paradox for critical infrastructure: detailed data are essential for modelling interdependencies, yet publishing them can expose vulnerabilities, particularly for energy grids facing hybrid threats.

The second theme confronts what the authors call complex risk landscapes, where systemic risks amplify through cascading impacts, compounding effects, contagion, and feedback loops. Recent events illustrate the stakes. The 2024 Swiss floods and landslides showed how interconnected vulnerabilities can trigger cascading failures, while research on East Africa’s alternating droughts and floods has traced how ethnic conflict, food insecurity, and health crises compound into economic disruption. A core methodological challenge is deciding where to draw system boundaries, whether to analyse a triggering event or whole-system dynamics, and how to balance quantitative and qualitative methods. Temporal complexity compounds the difficulty: climate hazards tend to shift gradually, while socioeconomic vulnerabilities can change abruptly after policy shifts or economic shocks. Standard vulnerability assessments typically fail to capture these dynamics, and longitudinal post-disaster surveys, which are essential for understanding recovery trajectories and persistent mental health effects, remain scarce.

To support integrated decision-making for systemic risks, the authors point to emerging frameworks such as the Risk-Tandem Framework, which structures stakeholder engagement and knowledge co-production to bridge gaps between risk science, policy, and practice. They also advocate agent-based modelling to simulate environmental and social interactions, causal network analysis to map risk drivers and tipping points, and Bayesian networks that incorporate stakeholder perceptions into assessments. A persistent institutional barrier is the siloed relationship between disaster risk reduction and climate change adaptation, which remain largely separate communities despite deep thematic overlap. The paper argues that frameworks like the Sendai Framework and the Paris Agreement’s Global Goal on Adaptation are beginning to bridge these divides, but that further scientific support is needed to make coordination operational rather than aspirational.

The third theme examines emerging technologies, and here the tone is both excited and wary. Artificial intelligence and machine learning can now process vast, multi-dimensional datasets capturing the interplay of multiple hazards, sometimes in real time. Long short-term memory networks support rapid-onset predictions for floods and landslides, gradient-boosting models such as XGBoost are used for drought and food security forecasting, and natural language processing tools mine policy documents and media to quantify hazard impacts. Earth observation combined with AI enables high-resolution exposure assessments, and digital twins, virtual replicas of cities and infrastructure synchronised with sensor and satellite data, allow stakeholders to simulate disaster scenarios in near real time. Crowdsourced efforts such as the Humanitarian OpenStreetMap Team and drone-based rapid assessments fill critical data gaps from the bottom up.

But the authors are emphatic that these technologies carry risks of their own. AI systems trained predominantly on data from high-income countries may perform poorly in data-scarce regions, creating warning gaps precisely where vulnerability is highest. Temporal biases underrepresent short-duration, high-magnitude events, a limitation made vivid by Hurricane Helene’s 2024 flooding in the southeastern United States, which struck small, steep, poorly observed mountain watercourses and led to major underestimation of high-end flood risk. Computational demands and costs can exclude researchers in resource-constrained settings, potentially widening global inequalities in disaster risk management capacity. The paper calls for responsible AI frameworks that prioritise interpretability, establish accountability when systems fail, integrate uncertainty quantification, and ensure that AI complements rather than replaces local and indigenous knowledge.

The final theme tackles multi-level governance, where the Mid-Term Review found persistent fragmentation: local authorities lack capacity, national frameworks fail to meet local needs, and global policies rarely translate into enforceable local action. The authors recommend statutory coordination mechanisms specifying mutual obligations between governance levels, partnership funding requiring co-financing across scales, and dedicated boundary-spanning roles connecting local to national actors. Nature-based solutions, from ecological security patterns at regional scales to stormwater management at the site level, exemplify how participatory, multi-stakeholder processes can deliver risk reduction alongside health, food, and water benefits. On early warning, the UN’s Early Warnings for All initiative offers a pathway toward all-of-society engagement, but few countries have anticipatory action plans or functioning multi-hazard early warning systems, and many depend on unsustainable external funding. Forecasting multi-hazard events remains especially hard when geophysical hazards with deep uncertainties, such as earthquake-triggered tsunamis or volcano-earthquake interactions, demand fundamentally different lead times and communication strategies than mature hydrometeorological systems.

As the 2030 deadline approaches and negotiations over a successor framework begin, the paper’s central message is that effective disaster risk reduction requires both incremental technical improvements and fundamental shifts in governance, data sharing, and inclusive engagement. The authors acknowledge structural obstacles, including power asymmetries embedded in institutions and disciplinary silos perpetuated by academic incentives. Still, they map their recommendations directly onto the Sendai Framework’s four priorities, from integrated methodologies for understanding risk to participatory early warning indicators for preparedness and recovery. Their survey respondents shared a common conviction: progress hinges on institutions capable of long-term investment in knowledge generation, and on bidirectional exchange between research and practice. With climate change intensifying extremes and societies growing more interconnected, the authors argue that the next global framework must treat multi-hazard thinking not as an optional refinement but as the foundation of disaster resilience, supporting the broader Sustainable Development Goals through enhanced societal resilience.

Subject of Research: Multi-hazard disaster risk assessment and management in the context of the UN Sendai Framework

Article Title: Multi-hazard risk assessment and management: pathways for the Sendai Framework and beyond

Article References: Tiggeloven, T., Raymond, C., de Ruiter, M. C., Sillmann, J., Thieken, A. H., Buijs, S. L., Ciurean, R., Cordier, E., Crummy, J. M., Cumiskey, L., De Polt, K., Duncan, M., Ferrario, D. M., Jäger, W. S., Koks, E. E., van Maanen, N., Murdock, H. J., Mysiak, J., Nirandjan, S., … Ward, P. J. (2026). Multi-hazard risk assessment and management: pathways for the Sendai Framework and beyond. Geoscience Communication, 9(2), 185-221. https://doi.org/10.5194/gc-9-185-2026

Image Credits: AI Generated

DOI: 10.5194/gc-9-185-2026

Keywords: multi-hazard risk, Sendai Framework, disaster risk reduction, systemic risk, early warning systems, artificial intelligence, earth observation, digital twins, climate adaptation, governance, storylines, vulnerability