EGU26, observing, predicting, acting: an integrated reading of risk systems

In this direction, new potential satellite missions, including Hydroterra+ as one of the four ESA Earth Explorer 12 candidate missions, aim to fill gaps in the observation of the fastest hydrological processes, enabling continuous monitoring of variables such as soil moisture, snow water content, and atmospheric water vapour. The ability to observe these phenomena at sub-daily scales represents a decisive step toward improving the understanding and prediction of hydrometeorological extreme events, particularly in vulnerable regions such as the Mediterranean and sub-Saharan Africa (Hydroterra+: an EO mission concept to reveal sub-daily water processes).

At more local scales, understanding snow dynamics is a key element for interpreting hydrological variability in mountain environments. In regions such as the Apennines, where the snowpack is highly variable from year to year, the integration of physically based models and satellite data enables the reconstruction of consistent multidecadal time series of snow water equivalent, improving the understanding of the role of snow in regional hydrological systems (Multidecadal snow water equivalent reconstruction in Central Italy using the Multiple Snow Data Assimilation System).

This complexity also emerges through localized atmospheric phenomena such as lake-effect snowfall. A comparative analysis between the central Apennines and the Japanese region of Niigata shows how similar configurations—cold air masses flowing over relatively warm water surfaces—can generate intense and analogous events, with significant impacts on infrastructure and local systems. In both cases, these snowfall events represent both a water resource and a hazard, and are highly sensitive to climatic variability (Cold on the water: mechanisms, impacts, and future scenarios of lake-effect snowfall in Italy and Japan).

At the same time, long-term observations of carbon fluxes allow for the analysis of ecosystem functioning and its evolution over time. The comparison between abandoned grasslands and mature forests in the Alps shows how land-use changes significantly affect CO₂ exchanges between the surface and the atmosphere, highlighting the role of ecological dynamics in carbon cycling (Tracing Carbon Flux Dynamics and Ecosystem Functioning Along a Land-Use Gradient: Long-Term Eddy Covariance Observations in the western Italian Alps).

In this context, climate change is also profoundly altering the functioning of forest ecosystems. In many European regions, beech forests are shifting from energy-limited to water-limited conditions, with direct implications for both carbon and water cycles. Integrated analyses along climatic gradients highlight how these systems are increasingly characterized by trade-offs between water use and productivity, making it necessary to adopt approaches that jointly consider carbon–water cycles (From energy to water limitation: projecting carbon and water cycling in mediterranean beech forests under climate change).

At the global scale, information on drought events is often fragmented across heterogeneous databases, making consistent risk assessment challenging. The integration of these sources, combined with the use of artificial intelligence techniques for extracting data from textual reports, enables the development of more comprehensive and structured datasets, improving the representation of impacts on populations and territories and supporting the derivation of vulnerability functions (Building a Comprehensive Drought Impact Dataset by Integrating Disaster Databases and Reports with the use of Large Language Models).

This complexity is also evident in mountain environments, where drought is emerging with increasing intensity. In glacierized catchments of the Italian Alps, water availability depends on a balance between precipitation, snow, ice, and hydrological processes interacting across multiple temporal scales. Over recent decades, drought conditions have progressively propagated across these compartments, influencing the duration and intensity of events and challenging the role of mountains as stable water reservoirs (Droughts in glacierized catchments of the Italian Alps: evolution and emerging high-elevation variabilities (2000–2024)).

These effects are directly reflected in public water supply systems, where reduced availability and deteriorating water quality can compromise essential services. The analysis of European databases reveals a high variability of impacts, ranging from usage restrictions to service interruptions, highlighting the combined role of climatic, infrastructural, and socio-economic factors, and the need for harmonized approaches to severity assessment (Drought Impacts on Public Water Supply in Europe: a challenge of diversity and severity assessment).

This dimension is closely linked to water-use management, which becomes particularly critical in Alpine contexts, where different sectors compete for increasingly variable water availability. The use of high-resolution satellite data, integrated with meteorological variables, enables the estimation of irrigation water requirements and their spatio-temporal variability, supporting more efficient allocation strategies and the planning of infrastructures such as storage reservoirs (Satellite-derived irrigation water requirement as a support tool for climate-resilient water management in the Alps).

At the same time, direct measurement of evapotranspiration allows for a more realistic observation of ecosystem water consumption, capturing dynamics that are often poorly represented by models, especially under stress conditions. The development of monitoring networks distributed along altitudinal gradients helps improve irrigation demand estimates and supports more adaptive water management strategies, strengthening the link between observation and decision-making (A regional evapotranspiration network for climate-resilient water management in mountain agro-ecosystems).

In this context, the integration of physically based models and artificial intelligence is transforming hydrological modelling, enabling efficient simulation of river dynamics and the identification of threshold exceedances. The use of heterogeneous real-time data allows these models to be applied at national scale and in near-real time, strengthening operational support for civil protection (A deep learning model of river water levels).

At the same time, the evolution toward seasonal drought impact forecasting models enables the translation of climatic indicators into information that can be directly used in decision-making processes. The application of machine learning techniques, combined with explainability tools, makes it possible to identify key drivers and temporal dynamics of impacts, improving the transparency and reliability of predictions and reinforcing the link between predictive capacity and action (Towards the Operational Implementation of Seasonal Drought Impact-based Forecasting with Explainable Machine Learning).

At the national level, the development of water resource scenarios represents an innovative step toward bridging the gap between climate projections and operational resource management. In Italy, within the IT-WATER project, the development of national-scale water resource scenarios introduces a structured approach that integrates climatic and hydrological indicators into tools directly usable in decision-making processes. Through a co-design process involving institutional actors and water managers, these tools are developed based on users’ actual needs, improving the effectiveness of information supporting adaptation strategies (Decision-relevant rendering of water-supply climate-change scenarios: a co-produced portfolio).

In the Alpine context, the integration of observational data, models, and local knowledge enables the development of operational solutions for drought risk management. The Interreg Alpine Space A-DROP project demonstrates how the combination of climate data, in situ observations, and machine learning techniques can improve water availability forecasting and support more sustainable resource management practices (From data to decisions: co-developing drought risk management solutions for the European Alps).

In contexts characterized by high vulnerability, where natural risks intersect with socio-political and infrastructural factors, a multi-risk approach becomes essential to understand and manage events that do not occur in isolation, but as combinations of interacting phenomena.

In Sudan, the integration of high-resolution meteorological modelling, hydrological monitoring, and impact-based forecasting systems has enabled the development of an early warning system capable of operating even under conditions of limited data availability. Information on the expected impacts of forecasted events is translated into concrete indications of potential effects on populations, infrastructure, and territories, supporting operational decisions by authorities and humanitarian actors (Advancing Multi-Risk Early Warning in Fragile Contexts: Methodological Insights from Sudan).

These experiences are part of a broader framework in which challenges related to the implementation of early warning systems are widespread at the global scale, and where their effectiveness depends on the ability to translate forecasts into concrete actions. In the Global South, persistent challenges include data availability, institutional coordination, and the translation of information into operational decisions. Innovative approaches, based on knowledge co-production, integration of scientific and local knowledge, and transboundary collaboration, are contributing to making these systems more inclusive, effective, and action-oriented (Early Warning Systems in the Global South: challenges and innovative approaches).

In this direction, the evolution toward impact-based early warning systems represents a further step forward. Through co-development processes with local experts and institutions, the integration of hazard, exposure, and vulnerability information enables the production of quantitative impact forecasts, supporting the generation of more targeted, contextualized, and actionable warnings. This is demonstrated by applications in Eastern Africa, where drought and flood forecasts are translated into useful information for farmers, institutions, and local communities, taking into account crop seasonality and the specific conditions of the most vulnerable groups (From risk knowledge to effective early actions: a novel framework and application for impact-based early warning with a pilot study in Eastern Africa).

Beyond technical aspects, risk governance also requires participatory processes capable of actively involving communities and stakeholders, a key component of people-centred approaches to disaster preparedness and response. The MedEWSa project shows how participatory planning can improve the effectiveness of risk reduction strategies, strengthening trust between institutions and citizens and fostering the integration of scientific and local knowledge. This approach has been translated into a structured operational tool, designed to support administrations and communities in implementing participatory and sustainable planning processes over time (Participatory Emergency Planning: a practical framework from the MedEWSa project).

Finally, the functioning of science–policy interfaces (SPI) represents a crucial element to ensure that scientific knowledge is effectively integrated into decision-making processes. A comparative analysis of SPI initiatives across 16 African countries shows how social, institutional, and technical-operational factors must operate in an integrated manner to ensure their effectiveness and long-term sustainability. From stakeholder engagement to mediation capacity, from alignment with institutional frameworks to the presence of continuity mechanisms, key indicators have been identified to support integrated and sustainable science–policy interactions (Assessing Science – Policy Interfaces for Climate Adaptation and Disaster Risk Reduction in Africa: a comparative indicator-based analysis).

In the case of wildfires, fire behaviour is increasingly driven by the combination of multiple meteorological variables—such as atmospheric dryness and wind—rather than by isolated extremes considered independently. For this reason, modelling these phenomena requires multivariate probabilistic approaches capable of representing dependencies among different drivers and simulating compound events. This enables improved wildfire risk estimation at the regional scale (Event-Based Copula Modeling of Compound Fire-Weather Extremes for Wildfire Risk Assessment).

In parallel, the identification of climatic drivers helps to understand the conditions that favour wildfire ignition and spread. In northern Italy, analyses of summer wildfire variability show that the dominant factor is often accumulated water stress over preceding months, rather than high temperatures alone, highlighting the temporal and cumulative nature of wildfire risk (Defining climatic drivers for the prediction of summer wildfires in northern Italy).

However, the most extreme events are not simply an extension of these dynamics, but respond to more specific conditions. National-scale analyses show that high-intensity wildfires are associated with critical combinations of drought, temperature, and wind acting over short timescales, generating conditions that are particularly difficult to control. The use of models based on extreme value theory allows these rare but high-impact events to be isolated, improving both prediction and management capabilities (Short-Term Hydrometeorological Drivers of Wildfires in Italy: Insights from Extreme Value Modeling).

These dynamics are further influenced by vegetation structure, which plays a key role in fire propagation. The integration of high-resolution data – such as those derived from UAV-based LiDAR remote sensing – with simulation models enables a more realistic representation of the transition from surface fires to crown fires, which are characterized by higher intensity, faster spread rates, and greater suppression difficulty. This improves both preventive planning and real-time operational management (Integrating UAV–LiDAR Fuel Data into Stochastic Cellular Automata PROPAGATOR for Crown Fire Modelling).

In this context, the representation of vegetation fuel becomes a key element linking scientific knowledge with operational application. The development of dynamic fuel maps, based on the integration of environmental data and machine learning techniques, allows the complexity of vegetation flammability processes to be translated into synthetic and actionable information. By capturing seasonal variability in vegetation conditions, these approaches improve wildfire danger forecasting systems and enhance decision support (An Operational Framework for Dynamic ML-informed Fuel Mapping for Wildfire Risk Management).

In the case of climate change, narratives based exclusively on catastrophic scenarios or individual responsibility may produce counterproductive effects, particularly among younger generations, fostering anxiety and a sense of helplessness. Overcoming these dynamics requires a shift in approach. Participatory strategies, based on interactive and solution-oriented activities, have proven more effective in strengthening understanding and motivation to act, contributing to the development of trust in scientific knowledge. Communicating climate change effectively therefore means maintaining scientific rigour and accuracy while avoiding paralysing narratives and promoting a conscious and collective response to global challenges (Engaging young audiences in climate change: moving beyond fear through active science communication).

Within this perspective, educational experiences based on active participation represent a particularly effective tool. Initiatives such as the “Next-Gen COP” simulate global decision-making processes, engaging students in negotiation, analysis, and co-design of solutions. Through pathways that integrate scientific training, interaction with experts, and the development of concrete proposals through dialogue and continuous exchange, these experiences foster key competences and strengthen awareness among younger generations. They demonstrate how knowledge can be translated into active participation in addressing climate challenges, encouraging young people to become active actors in building adaptation and resilience strategies (The “Next-Gen COP” as a tool for communicating climate change and catalyze solutions from high school students).