A shift to genuinely multi-hazard, multi-risk disaster management is both practicable and worthwhile, but meaningful uptake needs coordinated progress on methods, usable tools, equity, local coproduction and early-career capacity. Without standardized concepts, stronger spatio-temporal evidence, and operationally tested decision-support systems, benefits will remain patchy and context-dependent.

Main Finding

The MYRIAD-EU project (2021–2025) demonstrates that shifting disaster risk management toward a genuinely multi-hazard, multi-risk paradigm is feasible and valuable, but requires coordinated advances across conceptual mainstreaming, evidence on spatio-temporal dynamics of hazard–exposure–vulnerability, scenario methods (current and future), usable decision-support tools, explicit equity integration, deep case-study coproduction, support for Multi-Hazard Early Warning Systems (MHEWS), and strengthened Early Career Researcher (ECR) leadership. The project synthesizes progress and remaining challenges and proposes concrete directions for continued research and practice.

Key Points

• A multi-hazard, multi-risk approach increases societal resilience but is complex and cross-disciplinary.
• Progress was made on a six-point research agenda (proposed in 2022); results and gaps were evaluated across MYRIAD-EU activities.
• Priority needs identified:
  • Mainstream and harmonize concepts, definitions, and terminologies to enable comparability and communication across disciplines and stakeholders.
  • Build stronger empirical evidence on how hazard, exposure, and vulnerability interact across space and time to shape aggregated multi-risks.
  • Develop methods to generate both present-day and future multi-hazard and multi-risk scenarios that integrate climate, socio-economic change, and cascading effects.
  • Increase availability of appropriate, solutions-oriented, and user-friendly tools for practitioners and decision-makers.
  • Explicitly integrate equity considerations and aim for equitable disaster risk reduction (DRR) and adaptation.
  • Deeply test and coproduce multi-risk knowledge via in-depth case studies with local stakeholders.
  • Support development and operationalization of Multi-Hazard Early Warning Systems (MHEWS).
  • Empower Early Career Researchers through leadership roles and capacity-building in project structures.
• The paper outlines concrete ways these priorities could be advanced in future research and practice.

Data & Methods

• Project scope: synthesis and reflection on interdisciplinary research and practice conducted across MYRIAD-EU (2021–2025).
• Methods used (as reported/synthesized):
  • Comparative synthesis of project outputs, lessons learned, and stakeholder feedback.
  • In-depth, place-based case studies co-produced with local stakeholders to test methods and tools.
  • Development and testing of scenario methods for multi-hazard and multi-risk assessments (present and future).
  • Tool development aimed at decision-relevant, usable interfaces; iterative testing with end users.
  • Engagement with practitioners and policymakers to evaluate needs for MHEWS and operational uptake.
  • Emphasis on participatory approaches to capture equity and local vulnerability dynamics.
• Typical data types leveraged (implicit in project activities):
  • Hazard records and models (historic catalogs, probabilistic hazard models).
  • Remote sensing and exposure datasets (assets, infrastructure, land use).
  • Socio-economic and demographic data for vulnerability and adaptive capacity.
  • Climate and socio-economic scenario projections for forward-looking analyses.
  • Stakeholder-derived qualitative data from coproduction processes.

Implications for AI Economics

• Research opportunities
  • Develop spatio-temporal AI/econometric models that jointly capture interacting hazards, dynamic exposure, and evolving vulnerability (e.g., panel/time-series spatial models, spatio-temporal deep learning with uncertainty quantification).
  • Build scenario generation pipelines combining climate projections, land-use change, and socio-economic trajectories; use generative models and structured simulators to produce plausible multi-hazard futures for economic evaluation.
  • Integrate causal inference methods to estimate adaptation/DRR intervention effectiveness and distributional impacts, enabling more credible policy counterfactuals.
• Data & model design
  • Prioritize models that handle heterogeneous data (remote sensing, administrative, survey), low-data regions (transfer learning, domain adaptation), and structural change over time.
  • Emphasize probabilistic forecasting and calibration (e.g., ensemble methods, Bayesian models) because economic decisions hinge on risk distributions, not point estimates.
  • Include explainability and human-in-the-loop elements so model outputs are actionable for practitioners and policymakers.
• Decision & valuation frameworks
  • Extend cost–benefit and cost–effectiveness frameworks to multi-hazard contexts, including cascading and correlated losses across sectors and time.
  • Incorporate equity-weighted welfare metrics and distributional social welfare functions when valuing interventions, reflecting explicit equity objectives highlighted by MYRIAD-EU.
  • Use decision-analytic measures (expected value of information, robust decision-making criteria) to prioritize investments in MHEWS, data collection, and adaptation.
• Tools & markets
  • There is demand for usable, solutions-oriented decision tools—opportunities for AI-driven platforms that bundle risk analytics, scenario exploration, and stakeholder-tailored visualizations.
  • Market potential for risk-data products and AI services that support municipal and national MHEWS and resilience planning.
• Evaluation & governance
  • Design evaluation metrics beyond predictive accuracy: calibration, sharpness, decision-relevance, fairness metrics (e.g., subgroup calibration, disparity in expected losses), and economic utility loss.
  • Address governance, privacy, and data-access issues—models must be interoperable, transparent, and aligned with stakeholder coproduction norms.
• Capacity & workforce implications
  • Investing in ECR leadership and interdisciplinary training will grow the pipeline of researchers who can combine AI, economics, and risk science—important for long-term innovation and implementation.
  • Funders and institutions should support collaborative, place-based projects that integrate modeling with stakeholder engagement to improve uptake and validity.
• Concrete research/practice actions for AI economists
  • Build open, benchmarked multi-hazard datasets with standardized metadata and labels to promote method comparison and transferability.
  • Prototype decision-support systems that embed probabilistic multi-hazard forecasts into economic optimization for DRR investment prioritization.
  • Develop and test equity-aware loss functions and allocation rules in resource-constrained adaptation settings.
  • Collaborate with practitioners to co-design model outputs and UX for operational MHEWS and resilience planning, then evaluate real-world decisions and outcomes.

Overall, MYRIAD-EU’s synthesis points to a rich agenda for AI economics: methodological innovation for dynamic, interacting risks; tools that translate probabilistic multi-hazard information into economic decisions; and attention to equity, coproduction, and capacity building to ensure models lead to fair, usable policy outcomes.