This session provides a platform to demonstrate the latest technological advancements and innovations in multi-hazard risk assessment across various sectors and regions. It will feature presentations and discussions highlighting the implementation of cutting-edge technologies into useful applications to advance systemic disaster risk management and climate adaptation planning and ultimately contributing to Sustainable Development goals.
We particularly encourage submissions of research, case studies, and practical applications that showcase how these technologies can provide valuable insights into the complexities of multi-risk dynamics, optimise decision-making, and enhance resilience-building efforts.We also welcome critical discussions of implementation challenges, barriers, and lessons learned from both successful and unsuccessful deployment experiences.
Potential research topics include, but are not limited to:
- Examples of collaborative research efforts addressing stakeholder needs for multi-hazard tools and approaches
- AI/ML applications and digital twins for multi-hazard, multi-sector risk management
- Novel data collection technologies including LLMs, remote sensing for vulnerability and exposure mapping, and forensic assessment of past multi-risk events
- Resilience stress-testing for multi-hazard and high-impact low-probability events
- Innovative approaches in communications, knowledge-sharing, and capacity building across multi-hazard risk assessments and early warning systems
- Best practices for transferring innovations across different contexts and hazards
- Decision-support tools, open source software and novel risk assessment methods co-developed with stakeholders to enhance the preparedness of first responders and decision-makers to multi-risk
Orals: Mon, 4 May, 16:15–18:00 | Room 1.15/16
The growing interconnections between natural hazards, socio-economic systems, and vulnerabilities are increasing the frequency and impact of multi-hazard and compound risk events. Addressing this complexity requires innovative data-driven approaches that can integrate heterogeneous information and support both risk mitigation and preparedness across scales.
This talk showcases how Machine Learning (ML) and Artificial Intelligence, including Large Language Models (LLMs), can be effectively implemented in disaster risk management (DRM), with a focus on applications supporting the EU preparedness agenda at both European and global levels. First, I will present ML-based approaches for impact-oriented multi-hazard risk assessment, highlighting ensemble models developed to quantify compound hazard effects on flood losses at the subnational scale across Europe. I will then discuss ML applications for crisis anticipation, including forecasting food insecurity and conflict-induced human displacement, demonstrating how predictive models can support early warning and preparedness planning.
In a second part, the talk will illustrate how LLM-based methods can enhance data availability and knowledge integration for multi-hazard risk analysis. This includes automated geocoding of disaster locations from unstructured text to enable accurate subnational risk modelling, as well as the use of LLMs with Retrieval-Augmented Generation to extract factual crisis storylines and construct knowledge graphs from news and reports, supporting the analysis of cascading impacts and risk drivers.
Together, these examples demonstrate how AI-driven technologies can move beyond methodological innovation to deliver operational tools and evidence that directly support disaster risk reduction, preparedness, and decision-making, contributing to more resilient societies and informed policy-making that can adapt to evolving risk landscapes.
How to cite: Ronco, M.: Integrating AI and ML for Enhanced Multi-Hazard Risk Management, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-22846, https://doi.org/10.5194/egusphere-egu26-22846, 2026.
Understanding tropical cyclone (TC) risk is crucial for societal resilience and aligns with the United Nations Sustainable Development Goals. Although analyzing and ranking historical TCs help assess their associated risk, an optimal method to combine multiple risk factors into a single measure is still unclear. This makes it challenging for disaster risk practitioners to objectively assess the overall risk from historical TCs. We address this gap by employing Pareto optimality—a novel approach to objectively evaluate and rank the meteorological, hazard, and impact aspects of historical TCs in Japan from 1979 to 2019. Our findings demonstrate that Pareto-based ranking effectively identifies TCs that reflect complex and balanced trade-offs across competing risk metrics, preventing any single factor from dominating and providing a comprehensive view of overall risk. For example, the top three financially damaging TCs—Hagibis (2019), Bess (1982), and Mireille (1991)—are ranked alongside Tip (1979), Judy (1982), Bart (1999), Chaba (2004), and Tokage (2004) as the most impactful. Notably, Tip, Chaba, and Tokage do not appear among the top ten financially damaging TCs; however, they consistently rank high across multiple impact metrics, including fatalities, injuries, inundations, and house destructions. Similarly, ranking based on meteorological and multihazard intensity places TCs such as Songda (2004), Tokage (2004), and Nabi (2005) in the higher-ranked cluster due to their combined potential for wind and rainfall hazards—factors that could be overlooked if the focus is solely on meteorological intensity. We also highlight Kanto and Tokai regions as economic vulnerability hotspots, while Kinki, Kyushu, and Hokuriku regions are more affected by fatalities, injuries, and house destructions, underscoring the varied regional TC risk. The multidimensional ranking approach in this study addresses the complex nature of TC risk in Japan and offers a framework that can be adapted and applied to other vulnerable regions worldwide. Understanding these complex interactions between meteorological hazards, societal exposures, and vulnerability helps policymakers and disaster management agencies to develop more targeted and effective strategies for reducing TC-related risk.
How to cite: Islam, Md. R. and Sawada, Y.: Ranking the Unranked Disasters: A Multi-Dimensional Lens for Smarter Tropical Cyclone Risk Decisions, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17843, https://doi.org/10.5194/egusphere-egu26-17843, 2026.
Interconnected hazards and the resulting multi-hazard risks pose an increasing challenge for science, policy and practice. While recent research has advanced understanding of the physical processes linking hazards, place-based perspectives on how multi-hazard risk is experienced, governed and reduced remain comparatively limited. This contribution examines key themes for improving local-level understanding and management of multi-hazard risk across diverse rural and urban contexts.
Drawing on qualitative empirical evidence from multiple settings, the contribution synthesizes insights into what “managing multiple hazards” entails in practice. Cross-sectoral implications are considered across urban planning, maternal and child health systems, and the design and use of multi-hazard early warning systems. In addition, forensic analyses of selected multi-hazard events are used to identify recurring patterns in risk accumulation, institutional response and uneven impacts.
The contribution indicates that multi-hazard risk management is strongly shaped by place-specific configurations of vulnerability, infrastructure and governance, which influence trade-offs between sectors and over time. It argues for a shift from predominantly hazard-linkage framings towards community-centred, place-based approaches that better capture complexity and support context-sensitive solutions for populations at risk.
How to cite: Šakić Trogrlić, R. and de Ruiter, M.: Societal implications of multi-hazards: towards place-based understanding and management, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10335, https://doi.org/10.5194/egusphere-egu26-10335, 2026.
The increasing frequency of extreme climate events necessitates a transition from siloed Disaster Risk Management (DRM) and Climate Change Adaptation (CCA) towards seamlessly integrated, interoperable, and resilient systems. The DIRECTED project addresses this need through two primary pillars of innovation: the Risk-Tandem Framework for improved governance and the Data Fabric platform providing integrated information services. Both innovations are being co-developed with stakeholders in European Real World Labs (RWLs) in Denmark, Italy, Austria, Hungary and Germany.
This presentation highlights the DIRECTED project's innovation process, which combines technical solutions and knowledge co-production processes to support risk governance and strengthen climate resilience. Together, these innovations provide a transferable methodology for creating interoperable disaster risk management solutions that directly benefit first responders, local governments, and policymakers. We reflect on barriers and enablers in our approaches, as well as regional differences in the work with stakeholders in RWLs.
The project’s central governance innovation is the Risk-Tandem Framework, an iterative knowledge co-production process that bridges the gap between scientific risk management frameworks, modelling and practitioner needs. This framework has successfully translated complex stakeholder requirements into "user stories", leading to the co-design of tailored solutions such as cross-institutional emergency meetings, regional climate festivals and integrated emergency exercises.
On the technical front, the Data Fabric has been developed by DIRECTED to achieve interoperability among data, models and information products. It is a modular architecture combining existing with tailored additional open source components. The Data Fabric, which adheres to open standards based on international Open Geospatial Consortium (OGC) specifications, allows for seamless "data-to-model", "model-to-model" and “model-to-information” workflows. The platform combines several hazard and risk models, including CLIMADA, RIM2D, Danube Model, and SaferPlaces, with real-time forecast data from national weather services, e.g., DMI, GeoSphere, DWD, and HERA. The modular, cloud based implementation and open-source licensing invite community contributions and individualised set-ups of the Data Fabric. To support our users in the learning and sustained use of these innovations, guided workshops, e-learning modules, unified taxonomies, artistic communication outputs and virtual reality training have been developed with our stakeholders.
How to cite: Steinhausen, M., Irvine, T., Schweizer, P.-J., Bagli, S., Bharwani, S., Apel, H., Gräler, B., Cumiskey, L., Drews, M., Hochrainer-Stigler, S., Conradt, T., and Schröter, K.: Co-developing interoperable disaster risk management solutions for climate resilience: Innovations in tools, governance, and communication from the DIRECTED Project, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19997, https://doi.org/10.5194/egusphere-egu26-19997, 2026.
Systemic risk has emerged as a key feature of modern society, reflecting the growing complexity and interconnectivity of socio-ecological-technical systems. The concept is widely understood as the probability that disturbances cascade within a system or across interconnected systems, generating disproportionate, system-wide disruption (e.g., Kaufman and Scott, 2003; Sillmann et al., 2022; Gambhir et al., 2025). An alternative perspective frames systemic risk through the manifestation of systemic vulnerability, which is defined as the enduring, cross-scalar core of vulnerability that persists over time despite societal and technological advancements, mitigation efforts, or changes in hazard regimes (Armaș et al., 2025). Although they address systemic risk from different but complementary angles, these perspectives remain largely disconnected in both theory and application.
This study advances an integrative analytical framework aiming to clarify how systemic risk emerges at the intersection of these two presented perspectives, also showing that interdependence, nonlinearity, and feedback processes fundamentally shape impact dynamics. The primary focus of analysis is the interaction type, which functions as the fundamental unit for diagnosing cascading and compounding dynamics. Interaction types are organised along three orthogonal dimensions: mechanism, topology, and timing.
We also argue that understanding systemic vulnerability is essential for diagnosing systemic risk. The Systemic Vulnerability Model illustrates how such vulnerabilities reinforce impacts, fuel feedback loops, constrain recovery, and shape the likelihood of systemic collapse. Complementing this, Self-Organised Criticality (SOC) provides a theoretical underpinning that explains why, in highly connected systems, the accumulation of systemic vulnerability lowers certain system thresholds and leads to critical states. In these states, minor perturbations can trigger disproportionately large, system-wide failures, producing heavy-tailed loss distributions that challenge linear assumptions about hazard magnitude and impact.
The proposed analytical framework is intended as a conceptual and diagnostic tool rather than a predictive model. We do not propose a new definition of systemic risk but address the research gap on harmonising the currently disjoint discourse on systemic risk, supporting clear foundations for future studies on this topic. To continue this work, we aim to operationalise and empirically evaluate this analytical framework across diverse domains.
How to cite: Armaș, I. and Albulescu, A.-C.: Decoding systemic risk: An orthogonal interaction framework integrating systemic vulnerability and system-wide disruption, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7174, https://doi.org/10.5194/egusphere-egu26-7174, 2026.
Coastal and deltaic regions are increasingly exposed to compound hydro-climatological extremes, particularly the interaction of coastal and riverine flooding with extreme temperature events such as heatwaves and heat stress. These hazards interact across spatial and temporal scales, generating complex multi-hazard events that challenge conventional single-hazard risk-reduction and adaptation strategies. Despite growing attention in recent years, quantifying such multi-hazard interactions remains challenging due to limited long-term extreme-event data and incomplete understanding of the physical processes linking different hazards. This study addresses these gaps by quantifying and characterising compound and consecutive flood–temperature extremes across five coastal or deltaic regions in Bangladesh, India, Ghana, the United Kingdom, and the Netherlands. These case study regions are subject to multiple hydro-climatological extreme events. Long-term observational time series of tidal level, river water level, air temperature, and relative humidity were analysed for each case study. Coastal and riverine flood events were identified using the 90th percentile of tidal and river water levels, respectively, while extreme heat and heat stress events were defined using the 95th percentile of air temperature and wet bulb globe temperature. Interactions among hazards were examined using Kendall’s tau correlation to assess dependency structures, cross-correlation functions to identify precursor relationships and optimal time lags, and a non-parametric copula framework to estimate joint probabilities of hazards occurring in close succession. Results reveal distinct multi-hazard profiles for each region, including characteristic time lags between interacting hazards on an annual timescale. Coastal and riverine flooding exhibited strong multivariate dependence in most of the deltaic regions studied, with optimal time lags generally shorter than three days, indicating a high susceptibility to compound flooding. Similarly, all regions showed strong co-occurrence of extreme heat and heat stress events. Notably, heterogeneous temporally compounding events were observed between Global North and Global South regions. Temporally compounding events involving mixed combinations of flooding and temperature extremes (e.g., river flooding followed by extreme heat or coastal flooding followed by heat stress) were evident in coastal Bangladesh, whereas the United Kingdom and the Netherlands were primarily affected by compound flooding and compound heat events separately. The findings of this study advance the understanding of complex multi-hazard dynamics in vulnerable coastal and deltaic environments and provides evidence to support climate-resilient and adaptive management strategies.
How to cite: Adnan, M. S. G., Kebede, A. S., Addo, K. A., Dewan, A., Chakrabortty, R., White, C. J., and Ward, P. J.: Quantifying compound hydro-climatological extremes in coastal and deltaic regions , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3686, https://doi.org/10.5194/egusphere-egu26-3686, 2026.
Early Warning Systems (EWS) for hydro-meteorological hazards rely increasingly on integrated modelling frameworks capable of capturing complex surface and subsurface processes. Here we introduce STORMEW (Spatio-TempOral multi-hazaRd Model for Early Warning), a modular platform for multi-hazard spatio-temporal forecasting, analysis, and monitoring. In this study, STORMEW is configured to integrate SNOW-17, the CREST hydrological model, and a hybrid infinite-slope/Random Forest landslide model. The framework was applied in the Upper Garonne River Basin (Pyrenees, Spain) and forced with bias-corrected GEFS forecasts over a 10-year evaluation period (2010–2020).
GEFS-driven simulations effectively reproduced daily discharge (KGE: 0.53) and correctly identified the landslide initiation areas of the June 2013 event (ACC: 0.73). Building on these performance results, we implemented a consistent multi-hazard framework at the subbasin scale. Landslide hazard was derived from daily probability-of-failure (PoF) maps using the Percentage of Unstable Area (PUA), where cells with PoF > 0.5 were considered unstable and expert-defined PUA thresholds were used to classify four hazard categories. Flood hazard was likewise organised into four levels using subbasin-specific return periods of 2.5, 10, and 50 years computed from simulated daily discharge.
Under these criteria, the resulting classification showed minimal false alarms over the 2010–2020 period, correctly captured the June 2013 warning conditions, and discriminated high-flow events with no associated hazard (e.g., June 2018). Currently, the STORMEW system is implemented as a fully automated workflow that generates real-time flood and landslide warnings. To facilitate the interpretation of these outputs, a prototype web-based dashboard is being developed to visualize hazard dynamics in an operational context. Overall, this study demonstrates the capability of running a forecast-driven, multi-hazard EWS that links snow dynamics, hydrology, floods, and landslides for real-time early warning operations. Future work will explore the application of STORMEW in basins with differing climatic and hydrological conditions.
How to cite: Asurza Véliz, F. A., Hürlimann, M., Medina, V., Nieto, F., and Jaquet, E.: A Multi-Hazard Forecast-Driven Early Warning System for Hydrometeorological Hazards. Application to the Upper Garonne River Basin, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3181, https://doi.org/10.5194/egusphere-egu26-3181, 2026.
Challenges from cascading and compounding multi-hazard events are increasing, reinforcing the need to integrate emerging technologies into current risk assessment methodologies to disentangle the complexity of multi-risk events and support adaptation strategies. This work explored the integration of Earth Observation (EO) and machine learning to advance our understanding of multi-risk drought and heatwave events, with a case study in the downstream Adige River basin (Northern Italy). The study aimed at understanding the root causes of multi-risk drought and heatwave impacts and their spatiotemporal dynamics. To do so, a novel multi-risk assessment approach was adapted from the Forensic Investigation of Disasters (FORIN) framework and focused on the investigation of the upstream-downstream impact dynamics, with a testbed on irrigated agriculture. Two summer cropping periods with contrasting drought conditions (2022 and 2023, more and less dry respectively) were analysed, allowing for a spatiotemporal comparative investigation that sought to understand how differential traits (e.g., hazards, vulnerabilities) in similar settings (e.g., same area, different time) could explain the observed impacts. First, hydrometeorological hazards (SPEI 90-, 180-, and 365-days, temperature anomalies, river discharge, EO based soil moisture), exposure (spatiotemporal mapping of maize presence with in situ and EO data), and vegetation conditions were characterized. The latter were considered as a proxy of impact and calculated using Sentinel-2 derived indices (NDVI, NDMI) combined with a Principal Component Analysis into a composite stress index, then aggregated at the crop-field level (~20,000 fields in total) for each satellite image. Subsequently, unsupervised machine learning (HDBSCAN – Hierarchical Density-Based Spatial Clustering of Applications with Noise) was applied on the composite stress index for each date to identify field clusters and homogeneous areas having consistent vegetation conditions. Finally, clusters were further categorized into impacted or not impacted based on empirical stress thresholds rooted in NDVI and NDMI, to produce a final susceptibility map that represented the spatial frequency of stress occurrence. Results revealed a clear upstream–downstream stress gradient along the river, well summarized by the susceptibility map. This trend was mostly evident and statistically significant in 2022, and proved in line with upstream-downstream river discharge differences and provincial level yield data. Given the comparable hazard conditions along the case study, this suggested that additional factors might have had a strong influence on driving impacts, such as irrigation water management along the river. Additionally, correlation analysis revealed weak relationships between the composite stress index and the other variables (e.g., hazards, soil properties, field position) within the same year, suggesting the presence of complex socio-ecological aspects and physical vulnerabilities (e.g., water management, groundwater availability) that shaped vegetation stress beyond the hazard itself. Considering the large amount of data and the high resolution and scale of the analysis, this study advances the understanding of spatiotemporal dynamics of multi-risk events through spatiotemporal representation of impact dynamics, highlighting the added value of integrating EO and in situ data with machine learning techniques to unravel underlying vulnerability factors and enhance multi-hazard risk assessment to support adaptation and management strategies.
How to cite: Furlanetto, J., Albergo, E., Masina, M., Ferrario, D. M., Maraschini, M., Trabucco, A., and Torresan, S.: Understanding impacts of compound drought and heatwaves: A multi-risk analysis on agricultural-dominated socio-ecological systems combining Earth Observation and Machine Learning, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6549, https://doi.org/10.5194/egusphere-egu26-6549, 2026.
The increasing frequency and intensity of compound hydro-meteorological and wildfire events require advanced, operational, integrated tools capable of supporting early warning and near-real-time Disaster Risk Management (DRM). Within the framework of the EU-funded DIRECTED project, we present the development and operational implementation of a Data Fabric designed for a Real-World Lab in the Emilia-Romagna region (Italy).
The proposed Data Fabric is a cloud-native, serverless web application specifically designed to support nowcasting and short-term forecasting of pluvial and coastal flood hazards as well as wildfire propagation. The system has been co-designed in close collaboration with civil protection authorities (ARPAE) and first emergency responders (firefighters) to ensure operational relevance, usability, and direct integration into emergency workflows.
The platform integrates interoperable real-time observations provided by the ARPAE monitoring network, including weather radar, rainfall intensity, sea level, waves, tides, and wind measurements, together with meteorological and marine forecast models. These heterogeneous data streams are ingested into a scalable processing pipeline that feeds multi-hazard impact models, including high-resolution flood hazard models developed by SaferPlaces and wildfire spread models. The system produces near-real-time hazard maps at building-level resolution, enabling rapid identification of exposed and vulnerable receptors such as population, critical infrastructure, and strategic assets.
Beyond hazard mapping, the Data Fabric supports impact-based decision-making, facilitating the rapid assessment of potential consequences and the design of mitigation, such as flood barriers, and Disaster Risk Reduction (DRR) measures during evolving events. This contribution demonstrates how cloud technologies, interoperable data infrastructures, and stakeholder-driven co-design can be effectively combined to enhance preparedness, response, and resilience in complex multi-hazard contexts. Lessons learned highlight both the opportunities and challenges of deploying advanced digital solutions for operational DRM at regional scale.
How to cite: Bagli, S., Mazzoli, P., Luzzi, V., Renzi, F., Renzi, M., Redaelli, T., Cocchi, D., Cumiskey, L., Gräler, B., Dondi, C., Pancioli, V., Morollli, C., Pesaresi, A., Carlini, M., Pedron, P., Pangalli, A., Lazzari, E., and Steinhausen, M.: A Cloud-Based DataFabric for Multi-Hazard Nowcasting and Near-Real-Time Disaster Risk Management: The Emilia-Romagna Case Study within the DIRECTED Project, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6437, https://doi.org/10.5194/egusphere-egu26-6437, 2026.
Road icing during winter in South Korea is a critical disaster factor, causing numerous casualties annually. This study conducts an integrated analysis—combining traffic accident statistics with news data text mining—to understand the quantitative characteristics of icy road accidents and to deeply investigate the underlying social and structural causes and risks that are often difficult to capture through numerical data alone. First, approximately 2.15 million traffic accident records from the past decade (2014–2023) were extracted from the Traffic Accident Analysis System (TAAS). Based on this dataset, we performed a precise spatiotemporal analysis of icy road accidents, categorized by time of occurrence, road type, road geometry, and surface conditions. The results revealed a fatality rate of 2.3% for icy road accidents, which is approximately 1.35 times higher than the 1.7% observed in general accidents, confirming the extreme danger of icing. Accidents were heavily concentrated (20.8%) during the morning rush hour (08:00–10:00), and municipal roads accounted for the highest volume of accidents (33.7%) by road type. Particularly, the fatality rate was highest on national highways (7.9%), primarily due to high vehicle speeds. Regarding road geometry, fatalities were prominent in tunnels (8.3%) and on bridges (6.4%); this was attributed to the difficulty of evacuation in constrained spaces when chain-reaction collisions or fires occur following initial icing-related accidents. To identify the specific underlying causes behind these statistical phenomena, this study analyzed news articles, which provide the most rapid, accurate, and extensive contextual information regarding problems and causes in the accident process. To systematically extract meaningful information from large-scale unstructured news data, we utilized Natural Language Processing (NLP)-based text mining. This technique involves semantic analysis to identify relationships between key elements through sentence segmentation, tokenization, morphological analysis, and named entity recognition. By applying approximately 200 keywords related to accident causes—such as "delayed response," "unpreparedness," and "negligence"—to roughly 37 million news articles from the past five years (2020–2025), we identified specific "causes" behind the "phenomena" presented by statistical data. The analysis identified 16 latent risk factors in road maintenance and situational awareness, including not only drivers' difficulty in perceiving black ice but also insufficient designation of icing-vulnerable sections, inadequate snow removal measures, and lack of relevant policies and budget investments. In conclusion, this study provides multidimensional insights into icy road accidents through the complementarity of the two analytical methods. While statistical analysis scientifically pinpointed "high-risk locations" (tunnels and bridges) and "vulnerable times" (rush hour), text mining revealed that recurring accidents are rooted in administrative and human factors. This integrated approach connects policy blind spots and driver behavioral contexts that numerical statistics might overlook, providing an effective evidence base for improving regulations and establishing tailored safety information delivery systems beyond simple infrastructure improvements.
How to cite: Choi, S. and Kim, J. E.: Identifying Characteristics and Latent Risks of Icy Road Traffic Accidents through Integrated Analysis of Traffic Statistics and News Big Data-based Text Mining, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6287, https://doi.org/10.5194/egusphere-egu26-6287, 2026.
Posters on site: Tue, 5 May, 08:30–10:15 | Hall X3
Climate change has increased the intensity and frequency of hazard events, resulting in disaster risks that exceed historical experience. In highly urbanized countries such as South Korea, heavy and extreme rainfall has become a major climate-related hazard, leading to concentrated human and economic losses. While extreme rainfall can be partially anticipated through meteorological radar observations, official forecasts, and pattern-based prediction models, such hazard-focused approaches are insufficient to fully assess disaster risk in urban areas. This is because actual impacts are strongly influenced by exposure, vulnerability, and cascading effects, which may evolve into complex disasters.
In line with the IPCC and UNDRR disaster risk framework, this study emphasizes the need to anticipate secondary hazards and cascading risk events that may develop into complex disasters under extreme rainfall conditions. To address this challenge, a scenario generation method for extreme rainfall–induced complex disasters is proposed. The method integrates three key components: (1) regional exposure and vulnerability characteristics, including population distribution, industrial activities, transportation networks, and critical infrastructure; (2) secondary hazard and impact information derived from historical disaster records; and (3) interrelationships and correlations among different hazard and disaster types.
Using a weighted analytical framework, the proposed approach generates representative scenarios with high likelihood as well as extreme scenarios with lower likelihood but potentially high impacts. These scenarios support a risk-informed understanding of possible disaster pathways and provide actionable prior information for preparedness planning, emergency response, and scenario-based training. The results contribute to strengthening disaster risk reduction and enhancing urban resilience against climate-related extreme rainfall–induced complex disasters.
This work was supported by Electronics and Telecommunications Research Institute(ETRI) grant funded by the Korean government [26ZR1300, Development of Technology for the Urban Extreme Rainfall Response Platform].
How to cite: Oh, S. and Chang, Y.-S.: Predictive Methodology for Cascading Disasters/Events Induced by Extreme Rainfall in Urban Areas, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6254, https://doi.org/10.5194/egusphere-egu26-6254, 2026.
Climate-related High-Impact Low-Probability (CR-HILP) events pose a growing challenge to urban systems worldwide, as climate change amplifies the intensity, frequency, and compound nature of extreme events. These hazards, while rare, can generate disproportionate and cascading impacts on infrastructures, socio-economic processes, and environmental systems, often exceeding design thresholds and overwhelming traditional risk management approaches. This research proposes an integrated, systems-based framework to assess and enhance resilience to CR-HILP events in the Metropolitan City of Venice, a uniquely vulnerable socio-ecological system characterized by high exposure to climate hazards, strong interdependencies among critical functions, and exceptional cultural and historical value.
The study conceptualizes systemic resilience through the lens of Critical Functions (CFs), defined as the essential infrastructural, socio-economic, and environmental services that underpin societal stability and sustainability. Building on Network Science and complex systems theory, the research models these CFs as an interconnected multi-layer network, capturing both horizontal (intra-system) and vertical (inter-system) dependencies. The methodological framework is structured around three interlinked research tasks.
The first task develops a multi-layer network representation of critical functions, integrating real-world data such as GIS layers and infrastructure topology. Network Science metrics are employed to identify structurally intrinsic critical nodes and links, while different sampling techniques are used to detect minimal failure sets and structural vulnerabilities capable of triggering static systemic collapse.
The second task introduces a scenario-driven stress testing framework to assess dynamic cascading risks under CR-HILP events. High-resolution spatial hazard data, socio-economic vulnerability indicators, and stakeholder-informed narrative scenarios are combined with Percolation Theory to simulate disruption propagation across interconnected layers. This approach explicitly accounts for non-linear dynamics, interdependencies, and compound hazards, enabling the identification of tipping points, fragile configurations, and early-warning indicators for systemic failure.
The third task focuses on adaptive and resilient network design by integrating Graph Neural Networks (GNNs), Deep Reinforcement Learning (DRL), and Game Theory. A learning-based framework is developed to simulate adaptive responses and optimize resilience-enhancing interventions, such as rerouting connections, reinforcing critical nodes and edges, decentralizing dependencies, and reallocating capacities. Actor–Critic reinforcement learning methods, combined with GNN-based representations, enable agents to learn reconfiguration strategies that balance robustness, efficiency, and implementation costs. Extensions toward Multi-Agent Reinforcement Learning (MARL) allow the exploration of cooperation, competition, and negotiation among decentralized actors.
At the current stage, the approach has been tested on a single network layer, focusing on the transportation system, to validate the learning framework and intervention strategies. Future developments will extend the analysis to multiple interconnected layers, enabling the assessment of adaptive responses and cascading effects arising from interactions among different Critical Functions. By unifying network modeling, stress testing, and adaptive learning within a single framework, this research advances the understanding of systemic risk and resilience in complex interconnected urban systems. The Venice case study serves as a transferable testbed, offering methodological insights applicable to other climate-exposed metropolitan regions. The proposed approach aims to support decision-makers with actionable tools for resilience planning under deep uncertainty, contributing to more robust, adaptive, and climate-resilient urban futures.
How to cite: Casagrande, S., Ferrario, D. M., Maraschini, M., d'Antiga, F. M., Sano, M., Torresan, S., and Critto, A.: Building climate resilience to High-Impact Low-Probability events: an AI-driven modelling approach for Venice’s critical functions network, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11996, https://doi.org/10.5194/egusphere-egu26-11996, 2026.
Climate change is intensifying the frequency, magnitude, and spatial extent of climate-related hazards. Hotspot regions such as the Mediterranean basin have experienced severe impacts in recent decades, with an alarming increase in the occurrence and intensity of compound and co-occurring hazards. Despite the recognized need for multi-risk approaches, the availability of comprehensive, harmonized, and representative data remains limited, constraining the understanding of contributing risk factors. In particular, challenges in impact assessment represent a key bottleneck for quantitative multi-risk modelling and for disentangling interactions among risk drivers.
Earth Observation (EO) offers a largely underexploited opportunity in the context of multi-risk assessment, capable of providing spatially explicit, temporally consistent, and relevant, globally comparable indicators. The integration of these capabilities into coherent multi-risk assessment frameworks is an active area of research; however, significant opportunities for improvement remain in exploiting the full spatio-temporal richness of EO data through innovative methods, including artificial intelligence.
Modern representation learning techniques, such as embeddings for large spatio-temporal datasets, project system states into high-dimensional latent spaces. This enables exploitation of the full information content available from EO data and supports analysis of the entire system in which hazards occur, rather than relying on targeted regressors that may fail to capture the complexity and completeness of the underlying processes.
Here we explore and propose an EO-driven framework for the assessment of multi-risk from climate-related hazards (such as compound hot-dry extremes, wildfires, and water scarcity), with an application to the Mediterranean basin. By leveraging large amounts of remotely sensed data that describe the dynamics of the case study in both the temporal and spatial domains, this research aims to incorporate into the analysis the EO-based long-term system trajectories of the area, rather than relying solely on closely related preconditions. To this end, the study will explore the possibilities of coupling representation-learning techniques with machine learning methods to model impacts from multiple hazards in selected Mediterranean basin case studies.
The proposed approach is designed to be flexible and transferable across diverse riskscapes, including data-scarce regions, by complementing commonly used datasets and reducing reliance on incomplete impact records. By combining EO-based system representations with data-driven modelling frameworks, the research seeks to enhance the predictability of multi-risk consequences in the Mediterranean hotspot.
By considering a wider focus instead of hazard-specific situations, the ongoing work aims to contribute to the development of methodologies for climate risk analysis that may help better represent complex risk dynamics that are currently difficult to capture within traditional risk assessment approaches, with potential implications for adaptation planning, early warning, and disaster risk reduction under current and future climate conditions.
How to cite: Albergo, E., Tiggeloven, T., Furlanetto, J., Ferrario, D. M., Torresan, S., and Critto, A.: Leveraging Earth Observation and Machine Learning to Enhance Understanding of Impacts in the Mediterranean Basin, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12997, https://doi.org/10.5194/egusphere-egu26-12997, 2026.
Stress tests are a useful method for measuring a system’s exposure to multiple threats. In stress tests, scenarios are crucial, however the ones frequently utilised usually fail to consider inner, more contextual and social elements. Consequently, adaptation opportunities may be lost and hazards may be underestimated. Stress testing reveals the vulnerabilities of specific systems (projects, plans, etc.) to different risk scenarios, both climatic and non-climatic. Furthermore, traditional stress testing exercises are often limited in the stakeholder engagement. More collaborative and multi-hazard stress testing helps connect risk information with scenario planning and adaptation options by examining a wide range of scenarios. As a result, the information leveraged by projects from the humanitarian and development sectors can strengthen this approach by identifying weak points in projects and the design of activities.
This stress testing guide is a collaborative exploration to define where and how potential impacts may put excessive stress on a system. In some cases, it can also be used to test adaptation options. This guide is intended as a bottom-up exploratory approach to identifying the vulnerabilities of specific systems to various possible stressors and scenarios. It is envisioned as a flexible and generally applicable guidance document. As a flexible tool, the implementation and format of the test can vary depending on the system or unit of analysis being tested (e.g. size, type and core functions of a system), what stressors are taken into consideration (e.g. climate, urbanisation, economic, shocks, etc.), whether adaptation options should be included, and what type of information and other resources are available.
The stress testing guide provides a clear step by step process to apply this method in multi hazard risk context, including cartoons and reference to a series of hands-on processes that can be used in the testing process.
How to cite: Koelle, B., Izquierdo, K., De Boer, T., Marr, P., Kundak, S., and Atun, F.: Stress-testing systems: A guide to the assessment of compound and cascading risk, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13174, https://doi.org/10.5194/egusphere-egu26-13174, 2026.
The Dutch Climate Act mandates a 55% reduction in emissions by 2030 and net-zero emissions by 2050, in alignment with the EU Climate Law. To accelerate the just climate transition, several actions have been taken in policy and science. We identify key barriers and catalyse learning for climate-neutral and climate-proof neighbourhoods, acknowledging the interconnections among sectors and with various stakeholders. The overall aim is to accelerate the implementation of climate-neutral, climate-proof, just and healthy neighbourhoods, including the co-design of policy interventions that support a just climate transition by empowering the stakeholders. In our research, we investigate transdisciplinary learning processes and how learning can be facilitated most effectively to enable transformation at the neighbourhood level. We developed a learning cycle based on ongoing actions, de-recontextualising, developing pathways for action, and re-formulating local learning questions. This research is part of the ‘Accelerating Just Climate Transitions in Urban Regions – ACT’ project funded by KIN-NOW. In ACT, we develop local action pathways and agendas through collaboration between previously disconnected stakeholders, including residents, housing corporations, civil servants, local business professionals, and sustainability professionals. We provide a solid stepping stone for initiating and facilitating just climate transitions at the neighbourhood level and for future KIN activities.
How to cite: Atun, F., Champlin, C., Martinez, J., Flacke, J., Dijk, M., and Pfeffer, K.: Actionable knowledge generation for better implementation of innovation and technologies to achieve climate resilience , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18617, https://doi.org/10.5194/egusphere-egu26-18617, 2026.
To support data collection and sharing within the EU Horizon Europe PARATUS project, a dedicated GeoNode platform was developed (available at https://www.paratus-geonode.eu). The PARATUS GeoNode is an open-source, web-based geospatial content management platform designed to support the efficient storage, management, visualization and dissemination of spatial data required for integrated and comprehensive risk assessments. Built on the updated version of the GeoNode 4.4 framework, the platform combines established geospatial technologies with an easy-to-use interface to facilitate the collaborative management of geospatial data by both technical and non-specialist users. It provides a complete solution for geospatial data workflows by supporting many spatial and non-spatial resources, such as vector and raster datasets, maps, documents, dashboards, and interactive GeoStories.
In addition to automatic management of projections and web-based visualizations, the platform offers robust data storage capabilities of widely used geospatial formats like Shapefiles, GeoPackage, GeoJSON, KML /KMZ, and GeoTIFF. By ensuring that datasets are thoroughly documented and searchable, integrated metadata management improves data discoverability and long-term usability. Users can effectively find resources based on keywords, spatial extent, ownership, categories, and temporal attributes with the help of sophisticated search and filtering tools. Through graphical tools and support for styled Layer Descriptor (SLD) files, the GeoNode platform enables on-platform data editing and styling, enabling users to produce meaningful cartographic representations without the need for programming knowledge.
The platform is excellent at disseminating spatial information through interactive maps and dashboards with analytical widgets, and narrative driven GeoStories. These items can be embedded or shared as a link to be added to other online platforms. In this context, the Geonode enhances risk analysis by providing a spatial representation of the factors within the PARATUS Kumu Impact Chains, which are conceptual models that illustrate the interrelations among different risk factors. Through the GeoNode, Impact Chains are connected to a range of spatial datasets representing key risk components, including exposure maps (e.g. building footprints, population distribution), hazard maps (such as flood extent), and vulnerability indicators (e.g. building, land use, or socio-demographic characteristics). These spatial layers translate conceptual risk factors into spatial evidence, supporting the practical interpretation of causal relationships within the Impact Chains. The Impact Chains can be consulted through the Wiki section in the Stakeholder Hub (available at https://www.cmine.eu/topics/35391/page/impact-chains-04a0c668-7d14-4f93-a8bd-97dd9347038a).
How to cite: Hürlimann, M., Ghosh, P., Calderon, L. J. O., Cocuccioni, S., Lantada, N., Nuñez, A., Pantaleoni, N., and van Westen, C.: Linking spatial data with Impact chains through the PARATUS GeoNode, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19042, https://doi.org/10.5194/egusphere-egu26-19042, 2026.
The dynamic nature of hazards and risk drivers increases the complexity of decision-making process. In such complex situations, more expert knowledge is required to be able to set the right strategies. This requires a high level of collaboration and interaction of various stakeholders from different expertise. A serious game is one of the tools that can serve to enhance collaboration and interaction in environments dominated by uncertainties and systemic complexities.
In the PARATUS project, we developed the Systemic Risk Card Game to provide space for conversation of various stakeholders to develop a shared vision towards informed decision-making. The Systemic Risk Card Game is developed based on the principles of scenario-based simulations to engage diverse stakeholders in DRR-related decision-making. The systemic risk card game is helpful in region-specific challenges such as urban floods to give a clearer mind map to the stakeholders about infrastructure interdependencies or transboundary crises.
The core strength of the PARATUS systemic risk card game is its ability to make conversation on disaster risk tangible and accessible to a wide range of stakeholders. The structure of the game enables the players to recognise hazard(s) and simulate behavioural preparedness and reflect on real-world spatial and temporal dimensions. The game incorporates layered scenarios involving natural (e.g., earthquake, volcanic eruption), technological (nuclear disasters, airplane crashes), and social (e.g., displacement, poverty) components. The game highlights the importance of system-wide thinking in risk reduction in both urban and rural contexts, where infrastructure networks and geographical features act as conduits for cascading hazard interactions, allowing the players to visualize these interactions and simulate DRR decisions. Another valuable feature of the PARATUS serious game is its ability to delve into historical disaster events as learning cases. By reconstructing past events, players can learn how past vulnerabilities shaped the outcomes of real disasters and shape alternative strategies to prevent the impact of potential future events.
How to cite: Ghosh, P., Atun, F., van Westen, C., Koelle, B., and Kulakowska, M.: PARATUS Systemic Risk Game , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19291, https://doi.org/10.5194/egusphere-egu26-19291, 2026.
Impact-based Forecasts and Warnings (IbFWs) are used to convey the likelihood of impacts associated with different hazardous weather conditions. Generally the appraisal of impacts is made on a hazard-by-hazard basis and does not always consider the interplay between different hazards, assets and societal groups, or the cascading risks that can occur. This has the potential to lead to inaccurate assessments of the risk associated with extreme weather events. Often the predicted impacts are underestimated because the multi-risks posed by multiple hazards occurring together or in quick succession, as well as the indirect impacts, are not accounted for. This work builds on international collaboration through the UK Met Office’s Weather and Climate Science for Services Partnership (WCSSP) India programme, which supports the development of improved impact-based forecasting capability for weather-related hazards.
The development, assessment and validation of multi-risk IbFWs is constrained by the lack of suitable impact data. In India, IbFWs have been developed, however, impact data for weather-related hazards are available from different and diverse sources, whilst the data vary in their intended end uses, formats and methods of collection. The Indian Meteorological Department’s (IMD) records of impacts from historical events are limited to single events, without recording multi-hazard contexts, making it challenging to relate impacts to multi-hazard events. Using global, open-access datasets such as the EM-DAT international disaster database (Lee, 2024) and the new MYRIAD Hazard Event Sets derived using an algorithm which identifies ‘clusters’ of natural hazards (Claassen et al, 2023), we assessed the bias in impact data and showed how combining diverse sources can significantly improve data quality. Using the guidelines from Smith (2015), impact data for multi-risk, weather-related hazards in India have been collated, reformatted and verified creating a robust impact dataset for developing, assessing and validating IbFWs. This new database records, for the period 1970 to 2025, the impacts of multi-hazard events in terms of:
- Primary impacts: Deaths, missing, injured, affected people, and economic losses;
- Secondary impacts: Displaced and evacuated people, buildings damaged and buildings destroyed;
- Tertiary impacts: Damage to critical infrastructure such as hospitals, schools, public buildings, roads, as well as to agriculture, and the costs of relief.
The database also records metadata such as event timeframe, location, GLobal unique disaster IDEntifier (GLIDE) number (if appropriate), hazard type and classification. This new resource enables assessments of biases in impact recording and identification of hazard combinations that cause the most severe damage.
Resources such as this database can provide essential knowledge of the type of multi-hazard events that are responsible for adverse impacts and can be instrumental in the development of risk assessments, emergency management response plans and mitigation policies.
How to cite: Boelee, L., Coope-vickers, G., Brown, E., Lumbroso, D., Bianco, M., and Kolusu, S. R.: An assessment of the impacts of multi-risk weather-related hazards in India from 1970 to 2025, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21770, https://doi.org/10.5194/egusphere-egu26-21770, 2026.
Posters virtual: Fri, 8 May, 14:00–18:00 | vPoster spot 3
EGU26-20736 | Posters virtual | VPS14
A Four-Phase Serious Games Approach in the PARATUS ProjectFri, 08 May, 14:03–14:06 (CEST) vPoster spot 3
The increasing complexity of cascading and compounding effects, necessitates innovative tools for wide stakeholder engagement and decision-making, especially in uncertain situations. Risk communicators often struggle to successfully convey these complexities to diverse groups of actors. In the PARATUS project, we implemented a series of four serious games: High Water Pantano, Bucur Simulation, Saltum Montem, Paratus Systemic Risk Game; to address this gap through experiential process.
The structured, stakeholder-driven process used in the PARATUS project was grounded in the CompleCSUs framework and the design thinking methodologies. The development process included four phases, as follows: 1) Research and conceptualization, focused on the literature review and Miro app based mapping of stakeholder needs and PARATUs four case study areas (Caribbean, Bucharest, Istanbul, and the Alps); 2) Scenario and role design, focused on translating real-world impact chains co-developed with stakeholders into interactive storylines; 3) Prototyping and iterative testing, focused on stakeholders interacting with the prototypes and providing direct feedback to the tools; and 4) Implementation and evaluation, focused on the deployment of serious games in workshops and assessing their effectiveness.
Some benefits identified include increased transdisciplinary collaboration and the opportunity for stakeholder exploration of the results of inaction or certain decisions linked with the risk reduction, in a safe, simulated environment. However, the four-phase serious games approach in the PARATUS also resulted in certain critical lessons for the future implementation of co-design processes. These included the need for more flexibility in formats of the tools (analog vs. digital) to accommodate technical and context-based limitations; the importance of understanding the institutional hierarchies and factoring them into the process activities, and the need for multilingual support, especially in the transboundary context, for increase of the accessibility of the tools and trust levels of the participants. Following such four-step process, scientific risk assessment can be transformed into a scalable, user-centered and engaging tool for fostering long-term resilience.
How to cite: Kulakowska, M., Atum, F., Koelle, B., and Magnueszewski, P.: A Four-Phase Serious Games Approach in the PARATUS Project, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20736, https://doi.org/10.5194/egusphere-egu26-20736, 2026.
EGU26-17369 | Posters virtual | VPS14
Deep Reinforcement Learning for Operational Coastal Emergency Response With AI Agent Orchestration and Human OversightFri, 08 May, 14:06–14:09 (CEST) vPoster spot 3
Despite urgent needs for adaptive coastal risk management, operational systems still rely heavily on static triggers and fragmented information that overlook interactions between evolving hazards and response actions. Building on a completed game-like deep reinforcement learning (DRL) testbed, we present a pathway toward operational coastal decision support, progressing toward real-world case studies such as Venice in Italy and South East Queensland in Australia.
In the first phase, we developed a controllable game-like scenario that captures the essential components of coastal emergency management: a simplified representation of coastal geography and built assets, dynamic multi-hazard drivers evolving over time, and an action space reflecting plausible operational interventions under constraints. Using this environment, we demonstrated that a PPO-based DRL agent can learn adaptive policies through repeated interactions, as we gained practical lessons on state representation, constraint handling, and reward design for safety-critical objectives.
We then focus on the transition from simulation to real-world settings by outlining a set of alternative state-representation options, spanning classical dimensionality reduction and feature engineering through to learned latent-state methods. We report results for selected approaches, using autoencoders as the primary entry point to compress high-dimensional spatio-temporal hazard and exposure information into compact variables that retain decision-relevant structure while improving training efficiency and robustness. This provides a practical interface to real-world, digital-twin style environments built from geospatial and socio-economic data and forecast inputs.
Finally, we propose an orchestration layer to reduce the risk of AI-driven decision making and improve usability. A large language model (LLM) ingests DRL outputs and contextualises recommendations via retrieval-augmented generation over plans, studies, and standard operating procedures, together with API calls to dynamic data feeds. The proposed orchestration layer is intended to translate DRL outputs into human-readable and auditable decision support for a human-in-the-loop operator, grounding recommendations in retrieved local documentation and live data feeds to strengthen transparency, uncertainty communication, and operational trust.
How to cite: Sano, M., Ferrario, D., Casagrande, S., Vascon, S., Torresan, S., and Critto, A.: Deep Reinforcement Learning for Operational Coastal Emergency Response With AI Agent Orchestration and Human Oversight, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17369, https://doi.org/10.5194/egusphere-egu26-17369, 2026.
EGU26-17956 | ECS | Posters virtual | VPS14 | Highlight
Motivation and Engagement in Disaster Mapping in Europe (MEDiME): Understanding hydrogeological risks and vulnerability through serious gamingFri, 08 May, 14:09–14:12 (CEST) vPoster spot 3
This paper presents the development and pilot evaluation of Map@Me, an RPG-based serious game designed to improve understanding of hydrogeological risks and evacuation planning. Developed within the Motivation and Engagement in Disaster Mapping in Europe (MEDiME) Horizon Project, Map@Me targets a diverse audience and was tested in formal education settings, specifically middle and high schools.
The game integrates real local hazard maps, allowing players to explore their own environments and engage in realistic evacuation scenarios. With Map@Me, the player traces a realistic evacuation route that takes into account diverse mobility conditions, including disabilities, as well as advantageous and challenging factors, such as access to local knowledge and unfamiliarity with the area. Using a randomised system to determine the fictional character’s features in a real hazard map scenario, Map@Me represents a good example of how traditional disaster education can be supported by participatory methods of learning, whereby the agents can, in a controlled environment, experiment creatively with their behavioural choices and address their intrinsic biases.
During the presentation of the preliminary results from pilot sessions conducted with students, we will highlight both traditional learning outcomes—such as knowledge of evacuation sites and emergency preparedness measures—and “soft” learning outcomes, including cooperation, empathy, and collective responsibility.
The findings suggest that serious games such as Map@Me can enhance inclusive, place-based disaster preparedness, hazard map literacy and risk awareness, and overall contribute to a more socially aware approach to risk communication among younger audiences.
How to cite: Petraroli, I., Flacke, J., and Atun, F.: Motivation and Engagement in Disaster Mapping in Europe (MEDiME): Understanding hydrogeological risks and vulnerability through serious gaming, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17956, https://doi.org/10.5194/egusphere-egu26-17956, 2026.
