Predicting Water Pollution Using AI and Remote Sensing in a UNESCO Biosphere

A new collaboration between CGI and Ordnance Survey is demonstrating how artificial intelligence and satellite-based analysis can be used to proactively address water pollution in sensitive natural environments. The initiative focuses on the North Devon UNESCO Biosphere Reserve, an ecologically significant area in southwest England.

The project was developed as part of CGI’s Sustainability Exploration Environmental Data Science (SEEDS) programme, a research initiative launched in cooperation with the United Nations. SEEDS brings together industry, academia and public-sector partners to explore how data science can support environmental sustainability across multiple regions and sectors.

At the core of the project is an AI-driven predictive model designed to identify the conditions most commonly associated with sewage overflow events. By analyzing historical pollution data alongside environmental indicators, the system can forecast when and where pollution incidents are most likely to occur—before visible damage is done.

The initial phase of development has already delivered promising results. According to CGI, the model has demonstrated prediction accuracy exceeding 90%, enabling a shift from reactive cleanup to preventive environmental management. The second phase now aims to validate these results in real-world conditions within the North Devon Biosphere.

Covering more than 55 square miles, the reserve includes Braunton Burrows, England’s largest sand dune system. Protecting water quality in this area is critical not only for wildlife habitats but also for local communities whose livelihoods depend on healthy rivers and coastlines.

The solution integrates multiple data sources, including:

• Satellite imagery from open-source platforms
• Environmental sensor data
• Historical pollution records from the UK Environment Agency
• Geospatial datasets from CGI, Ordnance Survey and local biosphere partners

These inputs are processed using CGI’s GeoData360 platform, which maps pollution patterns over time and highlights areas of elevated risk.

Stakeholders such as farmers, water utilities, regulators and local authorities stand to benefit from the system’s insights. Early identification of potential pollution events allows for targeted interventions, reducing environmental damage and supporting long-term waterway protection.

Ordnance Survey views the initiative as a practical test of how advanced geospatial intelligence can support conservation goals. By combining authoritative mapping data with AI analytics, the partners aim to deliver actionable, scalable and regularly updated insights for monitoring waterways across the UK.

Looking ahead, CGI and its SEEDS partners plan to adapt and deploy the tool in other regions. If successful at scale, the approach could contribute to reducing the global impact of untreated sewage discharge and improving environmental resilience in vulnerable ecosystems worldwide.