Situation analysis, action planning and decision support: making extreme data usable and actionable

This project has received funding from the European Union’s Horizon Europe programme under grant agreement number 101092749. www.crexdata.eu Follow us on social media: Machine Learning models, Complex Event Forecasting Visualization (Visual Analytics) Explainable AI (XAI) Video data Temperature sensors Forest Fire Index Wind conditions … Data from the hazardous situation Data Science and AI Situational Awareness Objectives The project CREXDATA, funded by the European Commission, offers the opportunity to experiment with novel artificial intelligence technologies. One of the starting points is the ARGOS situational awareness geo-information system from the EU project ANYWHERE, which is already being used operationally by civil protection authorities, municipalities and companies in Spain and Ireland for weather data and forecasts and impact analyses. The technologies used range from data evaluation, e.g. from rescue robots, to an “explanatory” layer (Explainable AI/XAI) – which can be displayed on the situation map in the command center or made visible in the field via Augmented Reality. Motivation Extreme weather – extreme situation – extreme data: Although extreme weather conditions can be anticipated based on constantly improving weather forecasts, the actual effects are difficult to estimate. Even with foresightful and anticipatory operational planning, decision-makers face the extreme challenge of forming an up-to-date picture of the situation from global weather data and local situation assessment – and then prioritizing the deployment of limited resources. The effect of measures in prevention and preparedness phases must go hand in hand with response measures to better prepare communities for such situations and strengthen their resilience to climate change. Development of a toolbox for the collection and integration of extreme data as well as realistic simulation models and tools. Real-time prediction, including federated learning, prediction of complex events under uncertainty, and techniques for “Prediction-as-a-Service”. Reducing the perceived complexity through graphical design of workflows and visual analytics combined with understandable AI approaches and augmented reality. Conceptual Approach CREXDATA develops new approaches for the efficient use of local and global data sources to better understand and respond to complex events such as natural disasters. In particular, the handling of extreme weather events is the focus of the Fire Department of Dortmund, the German Rescue Robotics Center (DRZ), the Disaster Competence Network Austria (DCNA), the Ministry of the Interior in Finland and Paderborn University. In a network of 15 partners, including the Finnish Meteorological Institute and HYDS with ARGOS, the system will combine technologies such as Complex Event Forecasting into a “Prediction-asa-Service” system. The system will be tested in three case studies: in addition to decision support in extreme weather situations, the focus is on collision avoidance at sea and the containment of infection in pandemics. ►Does information generated by artificial intelligence algorithms need to be visualized differently than information from traditional situation assessment and evaluation? ►Does it make a difference whether the uncertainty of information is based on algorithms or witnesses on the ground? ►How does the algorithm used to generate a recommendation influence the planning of measures and decision-making? Results and outlook In October 2025, the first parts of the final evaluation were carried out based on a wildfire in Innsbruck and a flood in Dortmund. User feedback points to the high practical value and potential of the CREXDATA platform. A trial in Kuopio in November 2025 will complete the evaluation. Situation analysis, action planning and decision support: making extreme data usable and actionable! Research Questions ARGOS „Event Forecasting“ – e.g. changed premises in situation assessment Beispielfoto: Internet DRZ UAV UGV situation picture in the command and control support system satellite image „now“ archive (satellite, …) sensors „now“ nowcasted „now“ weather forecast impact forecast decision support Augmented Reality in the Field 𝑡𝑡𝑡0 𝑡𝑡0 𝑡𝑡−𝑖𝑖 𝑡𝑡−𝑛𝑛 𝑡𝑡+1(𝑖𝑖) 𝑡𝑡+2(𝑖𝑖) 𝑡𝑡+𝑚𝑚(𝑛𝑛) Trials Dr.-Ing. Jens Pottebaum, Marcel Ebel, Deniz Özcan, Niklas Döhner Prof. Dr.-Ing. Iris Gräßler