The Nereus System

THE IDEA:

Water pollution events happen faster than traditional monitoring can respond. By the time authorities collect samples and identify contamination, the window for effective intervention has already closed.

The Nereus System solves this by continuously monitoring water bodies using Copernicus Sentinel-2 satellite imagery. We compute spectral indices — NDWI for water extent, NDCI for algal blooms and organic contamination, and a turbidity ratio for sediment and chemical discharge — and automatically trigger alerts when anomalies exceed defined thresholds.

The system targets the Bega river in Timișoara as its primary demonstration area, a waterway with documented recurring pollution events caused by sewage overflows during heavy rainfall. Beyond satellite detection, citizens can submit geotagged pollution reports using Galileo GNSS positioning, providing ground truth that validates satellite data and enables hyper-local response.

Authorities receive structured alerts with: location, pollution type, estimated affected area, severity level, and a direct link to the satellite-derived index map. Early warning turns a reactive cleanup operation into a proactive intervention.



EU SPACE TECHNOLOGIES:

Sentinel-2 MSI (L2A) — Bands B03, B04, B05, B08:

- NDWI = (B03 - B08) / (B03 + B08) → water body extent and change detection

- NDCI = (B05 - B04) / (B05 + B04) → algal blooms and organic contamination proxy

- Turbidity ratio = B04 / B03 → sediment load and chemical discharge detection

Data accessed via Copernicus Data Space Ecosystem using the openEO Python client. Scenes are processed automatically on demand — no manual download required.

Spectral indices are further analyzed using an Isolation Forest machine learning model trained on clean water baseline values derived from Sentinel-2 reference data. The model assigns an anomaly confidence score (0-100%) to each detected event, classifying pollution type automatically without requiring labeled training data.

Galileo GNSS — used for authenticated geolocation of citizen pollution reports submitted through the web interface. Browser-based high-accuracy positioning (labeled explicitly as Galileo/GNSS in the UI) allows citizens to pin exact pollution locations for ground truth validation of satellite detections.

EU SPACE FOR WATER:

The Nereus System addresses Challenge #2 directly by providing automated near-real-time monitoring of water pollution on the Bega river, Timișoara, using Copernicus Sentinel-2 satellite imagery and Galileo-based citizen reporting.

The problem is real and recurring: on July 5, 2021 and July 11, 2025, heavy rainfall overwhelmed the AQUATIM wastewater treatment plant, causing untreated sewage to discharge directly into the Bega canal. Fish mortality was widespread. By the time Garda de Mediu arrived to take samples, 20+ hours had passed and the pollution was undetectable. No intervention was possible.

Our system would have flagged the anomaly using NDCI and turbidity indices computed from Sentinel-2 — within the next satellite overpass over the affected area. Automated alerts would have been sent to ANAR, Garda de Mediu, and local municipalities within minutes of detection, not hours or days.

Citizens can submit geotagged ground reports using Galileo/GNSS positioning via our mobile-friendly web interface, providing real-time ground truth that complements satellite detection and enables faster, targeted response.

The Bega canal extends 114km from Timișoara to the Serbian border at Otelec — pollution detected early protects not just one city but an entire transboundary waterway.


TEAM:

Adrian-Ciprian Tăian — Backend Developer. Second-year Mechatronics & Robotics student at Universitatea Politehnica Timișoara. Built the full backend architecture, Copernicus satellite data pipeline, and system integration from the ground up.

Șerban-Cristian Raica — Frontend Developer. Second-year Mechatronics & Robotics student at Universitatea Politehnica Timișoara. Designed and built the entire user interface, interactive map experience, and real-time alert dashboard.

Dan-Florinel Vegea — Data & Research Lead. First-year Mechatronics & Robotics student at Universitatea Politehnica Timișoara. Led the satellite data analysis strategy, identified the optimal Sentinel-2 spectral band combinations for pollution detection, and validated the scientific methodology behind our spectral indices.