Objectives of the service
Use a bold image to start your webpage. It can be a photo or an infographic like this. Remember that you project webpage can be a powerful marketing tool.
UK water companies lack the high-resolution data needed to accurately quantify reservoir greenhouse gas (GHG) emissions and link them to specific water quality drivers. This data gap makes it difficult to prioritise catchment management investments or prove that nature-based solutions actually reduce emissions.
The SWEEP project addresses this by delivering a satellite-driven monitoring service that maps reservoir GHG emissions alongside key water quality indicators. By pairing Earth observation data with machine learning models, the service provides utilities with the transparent evidence base required to track environmental changes over time.
Over its nine-month duration, the project's primary objective is to build and validate the core data framework. We will establish the baseline monitoring capability necessary to identify which targeted interventions, such as nutrient reduction and catchment management, will deliver the maximum positive impact on both water quality and emissions. Ultimately, this activity transforms raw satellite data into actionable insights, enabling water companies to make informed, data-led decisions for long-term environmental stewardship and net-zero planning.
Users and their needs
Our primary target community comprises water quality, carbon, and catchment management teams within water companies that manage reservoir assets. The activity currently focuses on utilities in the United Kingdom, with a defined roadmap to expand services to Europe, North America, and South America.
Core User Needs
-
Accurate GHG baselines: Transparent, high-resolution data to quantify and track reservoir emissions.
-
Causal insights: Evidence linking water quality degradation directly to rising emissions.
-
Intervention guidance: Actionable data to target, prioritise, and justify catchment management investments.
Project Challenges
Meeting these needs presents distinct technical and operational challenges. The primary obstacle is translating complex EO data into user-friendly metrics that seamlessly align with existing utility decision-making frameworks. Additionally, because this is a nine-month project, establishing a reliable baseline within a short timeframe requires highly accurate ML models that can overcome historical data gaps. Finally, we must ensure the service delivers actionable insights that help users confidently plan long-term nature-based solutions and prepare for seamless deployment across diverse international regulatory and geographic environments.
Service/ system concept
The service will deliver a dashboard of high-resolution data that maps reservoir greenhouse gas emissions alongside water quality indicators like chlorophyll-a or turbidity. Once deployed, water companies can track environmental changes over time and see how water quality impacts gas releases.
In simple terms, the system acts like a health tracker for reservoirs from space. Our system automatically processes Earth Observation images from a variety of space assets, using advanced cloud-based computing to translate variations in light and colour into precise environmental measurements.
The system architecture is built on a secure, automated pipeline. First, raw data is automatically ingested from satellites. Next, cloud-hosted data processing pipelines clean and calibrate the imagery. Machine learning models then analyse the data to calculate water quality trends. The EO-derived water quality data is then used to retrain our ML emissions models, allowing the system to understand the specific drivers behind water quality variations. Finally, this information will be pushed to an intuitive web dashboard, giving users clear, visual insights to guide their catchment management decisions.
Space Added Value
The service combines data from four distinct satellite missions. Sentinel-2 provides high-resolution optical imagery to track water-quality proxies such as chlorophyll-a at a localised, within-reservoir scale. Sentinel-3 delivers coarser but highly consistent temporal data, capturing surface temperature and long-term trends across large reservoir networks. For validation, Sentinel-5P retrieves regional column concentrations of CO2 and CH4, while GOSAT-2 provides long-term greenhouse gas records.
Expected Added Value
Current emission measurement methods rely heavily on traditional localised manual sampling, which is resource-intensive and lacks spatial coverage. By blending these space assets, the service overcomes the limitations of potential competitors who rely on single-satellite or localised approaches.
Integrating high-resolution imagery with atmospheric gas monitoring delivers two critical advantages:
-
Unmatched spatial and temporal scale: It captures both localised within-reservoir dynamics and regional, long-term trends across entire utility networks.
-
Zero operational disruption: It eliminates the health and safety risks, transport costs, and staffing requirements associated with sending teams out to remote or restricted water bodies.
This combined approach provides water companies with a scalable, continuous baseline to track emissions drivers that traditional methods simply cannot replicate.
Current Status
We have successfully built the automated data pipeline to extract surface reflectance imagery. The pipeline now fully integrates cloud masking, the Normalised Difference Water Index (NDWI), and shoreline buffering to isolate clean reservoir surface pixels.
Work is currently in progress to compile and clean the historical water quality datasets that will serve as our training dataset. Immediately following this, the next activity is to train the core ML model using the extracted reflectance data to estimate water quality parameters. This step will allow us to map the key environmental drivers behind reservoir GHG emissions.
Prime Contractor(s)
