Description
💎 Idea
Combined sewer systems across Europe regularly overflow, releasing untreated sewage into rivers, lakes, and fjords. This happens not because infrastructure fails, but because rainfall overwhelms the system before operators can react.
SewerCast is a satellite-powered early warning system that gives water utilities 6 to 48 hours of notice before an overflow.
The concept is simple. When soil is already saturated, new rainfall cannot be absorbed and instead flows into sewer systems. SewerCast tracks soil moisture across entire catchments using Sentinel-1 radar data, combines it with ground sensors and a 48-hour rainfall forecast, and calculates a real-time overflow risk score for each outfall.
Operators receive early alerts and can take action such as diverting flows or warning downstream facilities.
With stricter EU wastewater regulations and ongoing compliance pressure, municipalities need predictive tools. SewerCast turns overflow risk into something measurable and manageable.
🛰️ EU Space Technologies
Copernicus Sentinel-1 SAR, the foundationSentinel-1 uses C-band radar to measure soil moisture across entire catchments at 10-meter resolution, regardless of weather or time of day. Data is accessed through the Copernicus openEO API. This matters because the conditions that cause sewer overflows often block optical satellites. Sentinel-1 fills that gap and replaces the need for hundreds of additional ground sensors, making city-scale monitoring feasible.
KINEIS satellite IoT network, the ground truth layer. We install buried soil moisture sensors at key points in the catchment. These send real-time saturation data via the KINEIS satellite network, ensuring coverage even where mobile networks are unavailable. This ground data helps calibrate and validate the satellite measurements, improving accuracy at the local level.
Open-Meteo / ECMWF, the predictive layer- A 48-hour precipitation forecast, based on ECMWF data at 1 km resolution, turns monitoring into prediction. Without this, you only know the ground is saturated. With it, you know whether overflow is likely. Combining forecast rainfall with current soil conditions creates a meaningful early warning window.
All three data streams are fused into a single overflow risk score for each outfall. This is delivered through the SewerCast API and a real-time dashboard built with React and FastAPI. A working prototype was developed during the hackathon.
🌊 EU Space for Water
SewerCast supports disaster risk monitoring and pollution prevention.
Sewer overflows release untreated wastewater containing harmful pollutants into waterways. By predicting these events in advance, SewerCast enables prevention instead of cleanup.
This reduces environmental damage, protects water resources, and helps cities meet new EU regulatory requirements. The solution is scalable across thousands of municipalities with aging sewer infrastructure.
🤼 Team
Tarik Sørensen - Environmental engineer focused on hydrological modeling.
Maks Dunajski - Data scientist specializing in AI/ML and sensor networks.
Ingeborg Briskodden - Leads municipal relationships and operational coordination across projects.
Mats Åmot - Oversees day-to-day execution, delivery, and cross-team coordination.
https://github.com/maksdunajski/Cassini