FloodSafe

Floods are among the most destructive climate disasters in Europe and Spain, causing massive human and economic losses that demand urgent anticipatory tools. To address this reality, a team of young engineering students from the Universitat Politècnica de Catalunya (UPC) has developed FloodSafe, a technological solution designed to mitigate the impact of these events in highly vulnerable areas such as Catalonia and the Spanish Mediterranean.

This early warning mobile application bases its predictive capacity on crossing weather forecasts with open satellite data from the European Copernicus program (Sentinel-1 and Sentinel-2). The algorithm monitors soil moisture levels in real-time—even on cloudy days—and, upon detecting saturated ground alongside imminent torrential rains, it can predict the soil's inability to drain water, forecasting the flood before it occurs.

All this complex data processing is translated into an intuitive interface that issues geolocated alerts for the user, differentiating risk levels for key locations such as their home or vehicle. Beyond just warning of danger, FloodSafe suggests immediate preventive actions to safeguard property. Utilizing global tools from the European Space Agency (ESA), the project is built with a fully scalable design, with the vision of expanding at a national and European level to protect citizens against the growing climate threat.

We are a highly motivated team of four engineering students from UPC ESEIAAT (Universitat Politecnica de Catalunya) and we all four share a strong engineering foundation and a deep passion for leveraging new platforms, spatial tools, and innovative technologies to solve real-world climate changes.

Josep Marsal Perarnau - CEO: Josep drives the overall vision and strategic direction of FloodSafe. He leads our alignment with the Cassini hackathon goals, manages strategic partnerships (such as our integration with OpenCosmos), and ensures the product directly solves the real-world needs of our business to government clients.

Mario Carreras Santos: - CFO: Mario oversees the financial health and scalability of the project. He is responsible for defending our B2G Saas pricing model, managing the operational budget (including cloud computingg, satellite data costs and also all the salaries and new talents)and preparing the company for seed investment. 

Arnau Barberan Gomez - CMO: Arnau leads our Go-to-market strategy and external communications. He focuses on B2G relations, building the FloodSafe brand, and securing our first strategic pilot programs with local and national municipalities and Civil Protection agencies. 

Quim Limonero -  COO: Quim ensures the seamless execution of our operations and product delivery. He oversees the technical pipeline - bridging the gap between the complex earth observation data extraction and the final, user-friendly application, guaranteeing a reliable and continuous flow of information. 

This project aims to provide a solution for the challenge of the disaster risk monitoring, more exactly focused on flood prevention and mitigtion of its damages. It tries to reduce the the material impacts of this type of catastrophes. 

FloodSafe uses European space data and services as the core layer for flood-risk intelligence. The system combines Copernicus Earth observation data, especially Sentinel-1 radar imagery, with water and moisture-related indicators such as NDWI and NDMI, plus a Digital Surface Model (DSM) and in-situ data. These inputs are layered and processed to identify exposed areas, estimate water-related risk, and generate a dynamic flood-risk map. This matches the technical pipeline shown in the presentation: inputs such as Sentinel-1, NDMI, NDWI, DSM and in-situ data are processed through OpenCosmos, layering, modelling and risk calculation to produce dynamic risk maps, early warning alerts and actionable insights.

A key part of the system is the use of the DSM to model how water is likely to move across the terrain. By analysing surface elevation, slopes, depressions, barriers and flow paths, FloodSafe can estimate which areas are more likely to accumulate water or become affected during heavy rainfall or river overflow. This allows the platform not only to detect existing water or moisture conditions through satellite-derived indices, but also to anticipate the likely direction and spread of floodwater. In this sense, the DSM acts as the structural layer of the model: it helps translate rainfall, surface-water and moisture information into a spatial prediction of potential flood impact zones.

Space data is therefore not used as a decorative map layer, but as the basis for understanding where flooding is likely to occur and who or what may be affected. Sentinel-1 is especially relevant because radar data can support flood monitoring even under cloudy conditions, while NDWI helps detect water presence and surface-water changes, and NDMI contributes information related to moisture conditions. When these indicators are combined with DSM-based water-movement modelling, FloodSafe can identify low-lying or vulnerable zones and translate them into practical alerts for users, such as risks affecting a car, a ground-floor home, a production plant, or a delivery route. This is reflected in the app mockups, where the platform turns risk calculations into concrete recommendations and alternative actions.

The system also integrates space-enabled services and signals through location-based functionality. Galileo/GNSS positioning can be used to understand whether a user, vehicle, facility, or route is inside or near a high-risk flood area. This enables personalised warnings such as “car in high-risk flood location,” “moderate risk at home residence,” or “moderate risk in delivery route,” as shown in the business-case slide. In this way, FloodSafe connects satellite-derived flood intelligence, DSM-based water-movement prediction and user location to generate early warnings, safer routing and actionable recommendations before the flood impact occurs.

For the space-data component, we plan to use Copernicus/Sentinel Hub APIs, especially the Sentinel Hub Process API and Catalog API, to access and process Sentinel-1 and Sentinel-2 data. Sentinel-1 would be used for flood monitoring because radar data can support observation under cloudy conditions, while Sentinel-2-derived indices such as NDWI and NDMI would support water detection and surface-moisture analysis. These satellite layers would be combined with a Digital Surface Model (DSM) to estimate water movement, identify low-lying areas and predict zones likely to be affected by flooding.

For the geospatial processing layer, we would consider using Python-based libraries such as GeoPandas, Rasterio, NumPy and Shapely to process raster and vector data, calculate risk layers and combine satellite-derived information with DSM and in-situ data. If the project is developed further, OpenCosmos/DataCosmos could also be used as a platform to access, visualise and process satellite imagery and geospatial layers.

For the application interface, we plan to develop a mobile-first web app using standard web technologies such as HTML, CSS and JavaScript, potentially with React as the frontend framework. For map visualisation, Leaflet.js would be a suitable lightweight mapping library, allowing us to display risk zones, alerts, user locations and satellite-derived overlays. The app would initially use simple GeoJSON and JSON files to represent flood-risk areas, alerts and recommendations, before connecting to live APIs in a later version.

For location-based services, we would rely on browser geolocation on Galileo/GNSS-enabled mobile devices. This would allow the app to identify whether a user, vehicle, production plant or delivery route is located inside or near a high-risk area. In this way, the system could transform space-derived risk calculations into personalised warnings and actionable recommendations, such as safe parking, alternative delivery routes, or preventive actions for homes and industrial sites.

Overall, the proposed software architecture would combine Copernicus/Sentinel data access, DSM-based flood modelling, geospatial processing libraries, and a mobile web mapping framework. This technical stack is intended to support the outputs shown in the FloodSafe concept: dynamic risk maps, early warning alerts and actionable recommendations for citizens and businesses.