Operational extreme weather forecasts and early warnings are generally limited to timescales of up to around 10 days and to predicting single events, such as flooding or a heatwave. However, experimental ‘extended-range’ weather predictions that extend up to 46 days have been developed over the last decade by the world’s leading meteorological centres. A key motivation of exploring this prediction timescale is to bridge the gap between timescales, incorporate the latest ‘multi-hazard’ approaches, and improve early warnings and anticipatory actions. Currently, however, the extended-range prediction and the multi-hazard research and operational communities are largely disconnected. To date, there has been no coordinated effort to build a network that connects these disciplines and communities towards the development of operational systems. However, it is essential that these communities come together to explore windows of opportunity and instigate a step-change in the way forecasts are designed, produced and used. To address this challenge, here we present the ANTICIPATE COST Action (CA24144) that has created the first pan-European network focused on extended-range multi-hazard predictions and warnings. Over the next 4 years, ANTICIPATE will bring together existing but largely disconnected disciplines, operational practitioners and stakeholders (including extreme weather forecasting, extended-range prediction and climate dynamics, disaster risk reduction, multi-hazards, and communications) to drive forward advancements in the science, training, communication and application that will support next generation of effective early warnings that enable preparedness and action across hazards and forecasting lead times. In this talk, we will explore upcoming events and activities, and share how ANTICIPATE will provide vital leadership in multi-hazard predictions and warnings, address gaps and challenges, and help educate the next generation of forecasters and communicators for societal benefit. Further details about the ANTICIPATE COST Action are available here: https://www.cost.eu/actions/CA24144/.

How to cite: White, C., Rivoire, P., Walpole, O., Ramos, A., Wegmann, M., Russo, A., Pechlivanidis, I., Bloomfield, H., Larsen, M., Robbins, J., Arosio, M., Šakić Trogrlić, R., de Ruiter, M., De Angeli, S., O'Loughlin, F., Domeisen, D., Caltabiano, N., Ribeiro, A., and Hronček, S.: ANTICIPATE COST Action: extended-range multi-hazard predictions and early warnings, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4949, https://doi.org/10.5194/egusphere-egu26-4949, 2026.

Sustainable water and land management strategies to address land subsidence in Dutch built-up areas, as outlined in the backcasting approach of the Living on Soft Soil (NWA-LOSS) research programme, require robust projections of future subsidence and associated risks in the built-up area, such as the structural-damage risk of shallow foundation buildings, under different intervention water and land management strategies. While InSAR data, such as the ortho vertical displacement of ground surface from the European Ground Motion Service (EGMS), provides excellent records of recent vertical ground-surface deformation with millimeter accuracy, it is not a standalone tool that can be used to forecast future subsidence under varying future conditions. To bridge this gap, we introduce EGMS+, a machine learning framework that integrates the EGMS data with a Random Forest (RF) regressor to project future subsidence under various future conditions. The Random Forest algorithm was employed to learn the complex, non-linear relationships between EGMS mean annual velocity rates and a suite of relevant spatial predictors. These predictors consist of the mean lowest groundwater level, percentage of built-up area, percentage of old buildings, ground-surface elevation, Holocene soft-soil thickness, and percentage of grass cover within each 100 m grid cell across the built-up areas of Gouda and Krimpenerwaard municipalities in the Netherlands. The model achieved high predictive accuracy (R² = 0.73, Out-of-Bag score OOB = 0.73, Mean Absolute Error MAE = 0.095 mm/year) on five years of data (2019–2023). For the structural-damage risk assessment, we use a fragility curve developed by the NWA-LOSS team at TU Delft, which defines the probability of slight structural damage as a function of 5 mm crack width. This curve was used to compute building-specific structural-damage probabilities by integrating differential settlement with the short-side dimension of each building unit in the study area. Essentially, the EGMS+ framework enables future scenario projection by simulating how changes in these predictors affect future subsidence. This capability can be demonstrated by projecting future subsidence and associated risks, such as the structural-damage risk of shallow foundation buildings under several NWA-LOSS targeted future states, such as those involving raised water tables and intervention targeting shallow-foundation building units. This EGMS+ framework provides quantitative estimates of the effectiveness of various mitigation strategies, offering a powerful, dynamic decision-support and spatial planning tool that can evaluate and prioritize sustainable pathways of addressing land subsidence in the Dutch built-up areas.

How to cite: Hammad, M. and Stouthamer, E.: EGMS+: A Machine Learning Framework for Projecting Future Land Subsidence and the Associated Structural-Damage Risk of Shallow-Foundation Buildings: A case study of Gouda and Krimpenerwaard municipalities, the Netherlands, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18341, https://doi.org/10.5194/egusphere-egu26-18341, 2026.

Climate change brings emerging complex risks, subtle and weak, starting to manifest in some regions around the world, followed by the recurrence of preventable tragedies across regions. For instance, in Macau in 2017, people drowned in flooded underground car parks as they tried to save their vehicles. Tragically, similar preventable tragedies have since recurred in South Korea (2022) and Spain (2024). Before these incidents, local disaster risk reduction strategies in Macau, South Korea, and Spain did not cover specific guidelines addressing the resilience of underground spaces to extreme weather. Although local governments eventually enhance their regulations, such action is typically a reactive measure, triggered only by catastrophe rather than proactive foresight.

The primary obstacle to foresight is the challenge of identifying ‘unknown unknowns’—rare, variable-severity emerging risks. Our study directly addresses this critical gap in the early warning chain by demonstrating a systematic methodology that leverages cross-regional knowledge of analogous events to identify ‘unknown unknowns’ for regions without prior experience, thereby transforming them into foreseeable risks and enabling proactive preparation and strengthening response capabilities.

This study utilizes Natural Language Processing to analyze 7.7 million news articles across four dimensions—public awareness, priority of human needs, level of severity, and scope of influence—identifying 639 emerging climate threats, subsequently refined by an expert intervention to pinpoint the most critical tail-end risks. The findings uncover a wide spectrum of lesser-known emerging risks across diverse sectors, such as health, food, infrastructure, finance, transportation, and wildlife-related threats. An example of the findings is a paradox, first identified in peer-reviewed research and subsequently reported by the media. This paradox reveals that mercury in fish is increasing even as oceanic mercury declines, a phenomenon driven by warmer seawater that compels fish to migrate to cooler regions, which in turn elevates their energy consumption and accelerates bioaccumulation.

Ultimately, this research provides a practical decision-support tool for a range of stakeholders. By translating ‘unknown unknowns’ into actionable insights, our methodology enables a paradigm shift from reactive post-disaster response to proactive risk management. Specifically, these identified risks can be used to inform targeted risk communication strategies and establish triggers for anticipatory action. This provides a crucial component for the UN’s ‘Early Warnings for All’ (EW4All) initiative, enabling communities and disaster managers to prepare for emerging complex risks before they manifest as localized crises.

How to cite: Zhang, Y., Yarime, M., Lau, A. K. H., Chan, J. W. M., Fung, J. C. H., Shun, C. M., and Chan, K.: A systematic approach to identify 'unknown unknowns' for impact-based early warning systems, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1454, https://doi.org/10.5194/egusphere-egu26-1454, 2026.

In June 2017, a large landslide in Umiammakku Nunaat (Karrat), West Greenland caused a huge tsunami wave of about 90 m height on the opposite fjord slope and reached the village of Nuugaatsiaq 32 km far away. The tsunami caused severe property damage and the death of four people in the village. After the tsunami, the two settlements Nuugaatsiaq and Illorsuit were evacuated due to the still high risk of another potential tsunamigenic landslide in the fjord. To this day the two settlements are still under evacuation but none of the villages have been permanently relocated so far.

This disaster highlighted the urgent need for natural hazard monitoring systems in this region. In 2021, Geoprevent and the local responsible authorities made a feasibility study and installed the first ever natural hazard monitoring system in Greenland. This monitoring system in Umiammakku Nunaat runs year-around. Two deformation cameras were installed at the counter slope with a view on three regions of interest. An additional camera was installed next to one of these unstable slopes. These deformation cameras take multiple high-resolution images per day. Cross-correlation based algorithms are then used to identify differences between these images and estimate the deformation of these areas.

A deformation analysis can be compared to a timelapse with which one can see slow processes that a human eye cannot see. The local experts use the information provided by these monitoring systems for a continuous risk assessment. The continuous monitoring helps to evaluate the sitiation constantly and supports authorities in their decision making related to the evacuation of certain settlements.

From a technical point of view, Greenland presents quite some challenges to maintain a monitoring station under these harsh conditions. Remoteness, cold temperatures, heavy winds and polar night are only a few of them. In order to have enough power during polar night, the system is running on a methanol-based fuel cell solution with the option of solar charging during the sunny months. Moreover, communication with the station and the data transmission are satellite-based, so that the station can be controlled remotely.

How to cite: Wetter, J., Carrel, M., Stitelmann, O., St. Pierre, T., Von Wartburg, J., Mätzler, E., and Petersen, J.: Year-Around Monitoring of Slope Instabilities in Umiammakku Nunaat (Karrat), West Greenland with Deformation Analysis , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6834, https://doi.org/10.5194/egusphere-egu26-6834, 2026.

How to cite: Strehlow, K., Ghabrous Larrea, C., De Beni, E., Ganci, G., Cannavò, F., and Baladima, F.: Impact-based forecasting for volcanic eruptions: science driving financial preparedness , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20318, https://doi.org/10.5194/egusphere-egu26-20318, 2026.

For early warning systems (EWS) to translate into timely, protective early actions, particularly in the contexts marked by deep social, linguistic, and structural inequalities, effective risk communication and community engagement (RCCE) are essential. In Nepal, the Resilience, Adaptation and Inclusion in Nepal (RAIN) programme demonstrates a holistic, community-led approach for strengthening RCCE by embedding behavioural and psychological insights, fostering trust, and creating inclusive communication pathways that target the most vulnerable groups. RAIN, which is designed to support transformative impact for resilience, adaptation, and inclusion through a community-led approach, places community organisations, local governments, and at-risk populations such as landless communities, persons with disabilities, ethnic minorities, women, and girls at the centre of the early warning and early action system. This abstract examines how RAIN put RCCE into practice to improve the accessibility, credibility, and behavioural effectiveness of warnings across multiple hazards. 

RAIN addresses a core challenge in Nepal’s EWS landscape, i.e., existing alerts are highly technical, often inaccessible to non-native Nepali speakers, and do not convey clear, actionable behaviour. Realizing that people act on warnings only when they trust the source, understand the message, and see its relevance, the programme has redesigned communication flows to be community-centred, multi-lingual, and multi-modal. Behavioural insights ranging from simplifying messages and making actions concrete to tailoring messages to literacy levels and reinforcing social norms through trusted local actors to shape how communities receive and interpret alerts. Community-based organisations (CBOs) and committees become active co-designers and disseminators of warnings, leveraging their embedded trust to increase credibility, reduce uncertainty, and motivate action.

To overcome structural and psychological barriers such as low-risk perception, fatalism, gender norms restricting mobility, and limited trust in government systems, RAIN strengthens risk communication channels. These include Interactive Voice Response (IVR) systems, door-to-door dissemination, mobilisation of community volunteers, sign language videos, and accessible formats for people with disabilities to support different communication needs. The programme also integrates locally relevant languages and culturally grounded communication approaches, acknowledging that linguistic relevance and cultural resonance are crucial for behavioural uptake. By incorporating Organisations of Persons with Disabilities (OPDs) and diverse CBO networks, RAIN enhances inclusive communication, adaptive behaviour, and equitable access to life-saving information. 

At the system level, RCCE is institutionalised through collaboration with the Department of Hydrology and Meteorology (DHM), the National Disaster Risk Reduction and Management Authority (NDRRMA), provincial governments, and local governments, ensuring standardised message templates, impact-based forecasting, and a strengthened communication flow that connects scientific information to community understanding from information producers to at-risk communities. The programme’s localisation model will build trust over time by enabling communities not only to receive warnings but also to shape how warnings are generated, translated, disseminated, and acted upon. 

Overall, RAIN offers a scalable model for RCCE that demonstrates how deeply rooted behavioural insights, trusted community actors, inclusive communication technologies, and systemic coordination can together ensure that early warnings effectively reach and are acted upon by the people who need them most.

How to cite: Khadka, P., Neupane, S., Pradhanang, A., and Manandhar, V.: From Warnings to Early Action: Community-Led Risk Communication and Engagement in Multi-Hazard Early Warning Systems, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1971, https://doi.org/10.5194/egusphere-egu26-1971, 2026.

The flood disaster in western Germany in July 2021 revealed substantial shortcomings in the communication and understanding of official warnings and risk information, which contributed to severe impacts (DKKV, 2024; Thieken et al., 2023). In response, the German federal and state governments initiated the development of a Natural Hazards Portal to provide a central, authoritative access point for harmonized information on natural hazards. The Deutscher Wetterdienst (DWD), Germany’s National Meteorological Service, was commissioned to design and implement the portal.

The Natural Hazards Portal integrates official warnings with preparedness, impact-related information, and behavioural guidance across multiple natural hazards. Its objectives are to improve the visibility and comprehension of warnings, strengthen individual and societal preparedness, and contribute to long-term resilience. To this end, the portal combines event-driven warning information with contextualized data on local hazard exposure, impact-oriented indicators, and recommended actions before, during, and after hazardous events. All content is designed to be clear, accessible, and comprehensible for diverse user groups.

This presentation presents the conceptual framework, current implementation status, and future development of the portal. We discuss key challenges related to the integration and standardization of heterogeneous data sources from multiple authorities, as well as the role of the portal within an ecosystem of specialized, hazard-specific platforms. Particular emphasis is placed on the transition from hazard-centered to impact-based warning and risk communication and on the portal’s potential to support anticipatory action and informed decision-making.

The Natural Hazards Portal represents a joint, holistic response by German authorities to increasing natural hazard risks under climate change. By providing localized, impact-relevant information and official warnings through a single, central access point, the portal aims to strengthen preparedness and resilience without replacing existing specialized warning services.

References

DKKV (2024). Governance und Kommunikation im Krisenfall des Hochwasserereignisses im Juli 2021, DKKV-Schriftenreihe Nr. 63, Januar 2024, Bonn. https://dkkv.org/wp-content/uploads/2024/01/HoWas2021_DKKV_Schriftenreihe_63.pdf.

Thieken, A. H., Bubeck, P., Heidenreich, J., von Keyserlingk, L., Dillenardt, J., & Otto, A. (2023). Performance of the flood warning system in Germany in July 2021 – insights from affected residents. Natural Hazards and Earth System Sciences, 23, 973–995 https://doi.org/10.5194/nhess-23-973-2023.

How to cite: Erhardt, B., Brendel, C., Hafer, M., Haller, M., Hüser, I., Koziar, C., Lengfeld, K., Rode, D. K., Rauthe-Schöch, A., Reetz, B., Riede, H., and Walawender, E.: Developing the Natural Hazards Portal for Germany – a central access point for warnings, preparedness and resilience, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11232, https://doi.org/10.5194/egusphere-egu26-11232, 2026.

Early Warning Systems (EWS) are widely recognized as a cornerstone of disaster risk reduction; yet, their effectiveness depends not only on scientific accuracy but also on how warnings are translated into a collective understanding and timely action at the community level. In many hazard-prone contexts, early warnings fail not because data is unavailable, but because communication infrastructures do not align with local languages, temporalities, and practices of attention. This paper presents a situated design research project in development in Maroantsetra, Madagascar, which reframes EWS as a socio-spatial and relational infrastructure.

The project is being developed through a collaboration between EPFL–ALICE Lab, the NGO Medair, local communities in Ambinanitelo, Ankofa, and Anjanazana, and national disaster management institutions (CPGU). It explores the co-design of a Sensitive Risk Warning Infrastructure (SRWI) that integrates institutional early-warning protocols with vernacular communication systems and environmental indicators, including town criers, drums, conch shells, and animal behaviour. Rather than replacing scientific alerts, the approach focuses on translation, rehearsal, and trust-building as key conditions for effective anticipatory action.

Methodologically, the ongoing research combines walk-along interviews, participatory mapping, role-play rehearsals, and low-tech prototyping to identify breakdowns in the warning chain and to co-design hybrid warning practices. Preliminary findings indicate that warning communication is inherently spatial and material, unfolding through proximity, sound, visibility, and shared places such as schools and community halls. By foregrounding these dimensions, the SRWI aims to advance a community-centred and impact-oriented approach to EWS, enhancing comprehension, ownership, and timely response.

The paper contributes to ongoing discussions on community engagement, last-mile communication, and anticipatory action by presenting design as an interface between scientific warning systems and situated action. Developed in parallel with the Architectures of Emergency research and an Atlas of Inhabiting Emergency, it connects multiple geographies of risk and positions design as a form of spatial inquiry that supports infrastructures of care, enabling communities to sense, interpret, and rehearse risk collectively.

How to cite: Mompean Botias, E.: Co-Designing a Sensitive Early Warning Infrastructure in Maroantsetra, Madagascar. , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9637, https://doi.org/10.5194/egusphere-egu26-9637, 2026.

Effective Early Warning Systems (EWS) does not only rely on natural hazard forecasting but also on how actors are prepared, coordinated, and respond in the different stages of the warning chain. We here present a scenario‑based methodology designed to assess emergency response at different responsibility levels before, during, and after multi-hazard events.

The method involves systematic information gathering during facilitated scenario exercises, followed by synthesis of findings to improve action cards and emergency plans. Study area is rather remote regions of Norway, and the approach effectively consolidated existing knowledge, initiated cross‑level dialogue, and revealed clear gaps in preparedness, coordination, and resource allocation.

The workshop objectives were to evaluate current response protocols, identify training needs among responders and communities, and propose interactive, scenario‑based training approaches. A three‑phase scenario—covering the warning, action, and recovery stages—was developed using local knowledge and recent events to ensure realism and relevance.

Key findings include: (i) inconsistencies in responsibility distribution and inter‑agency coordination, (ii) missing competencies and resource limitations, and (iii) the need for clearer communication pathways throughout the evolving event. Even though there is a strong and knowledgeable commitment from participants, improvement areas were identified.

Conducted in the aftermath of Extreme Weather Hans (2023) and Amy (2025), the work demonstrates the value of scenario‑based evaluation as an integral component of EWS development. This contribution forms part of the Norwegian research project Beredt! – Scalable Services & Risk-Based Governance for Climate-Driven Natural Hazards in Norway and highlights the importance of continuous training and assessment to enhance disaster preparedness and resilience.

How to cite: Hegdahl, T. J., Gilbert, G., Devoli, G., Müller, K., Syndes, A. K., and Sydnes, M.: A Scenario‑Based Framework for Evaluating Emergency Response and Communication Throughout Multi-Hazard Events, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16980, https://doi.org/10.5194/egusphere-egu26-16980, 2026.