Harnessing high-resolution multispectral satellite imagery provided by MAXAR Technologies (R), our team developed a remote waterway monitoring system that detects pollutant signatures anywhere in the country without leaving the office. Our system re-bands multispectral rasters, masks land & water bodies, applies spectral filters for nitrate and turbidity proxies, and visualises hotspots of concern for rapid environmental monitoring. This project was awarded "Top-10 Finallist" at TakiWaehere by the MBIE.
By CESA on 12 October, 2025
TOP-10 FINALLIST PROJECT - Awarded by MBIE at TakiWaehere 2021
Small waterways in Aotearoa are vital ecosystems, but monitoring is resource-intensive and often limited to major rivers or catchments. By leveraging satellite imagery, we saw an opportunity to automate detection of water surfaces and then assess pollution signatures (e.g., elevated nitrates, turbidity) from spectral responses - enabling frequent, broad-area reconnaissance. The Taki Waehere theme of "Healthy local waterways" directly aligned with this aim.
Key drivers:
- Many minor waterways lack continuous in-situ sensors or frequent terrestrial sampling.
- Satellite multispectral data provide spectral bands (including near-infrared, short-wave IR) that allow discrimination of water versus land, and potentially polluted vs cleaner water.
- A prototype tool could empower regional councils, environmental agencies, and community science groups with near-real-time mapping of waterways health.
The team began by exploring how multispectral satellite imagery - captured in both visible and invisible wavelengths of light - could reveal characteristics of surface water. Each pixel of these images contains information about how light is absorbed and reflected, offering subtle clues about what is in the water. Clear water, sediment-laden water, and algae-rich water all have different spectral signatures. By studying these patterns, the team saw the potential to detect contamination indicators such as turbidity, sediment, and possible nutrient pollution.
CESA's prototype system automates this process. It starts with a raster geotiff image of a chosen site, captured by Maxar's satellites. The system then reconfigures the imagery into a false-colour composite that enhances features invisible to the naked eye - particularly near-infrared and short-wave infrared bands, which are highly responsive to vegetation and water. Using these enhanced layers, the system distinguishes water from land, automatically generating a map of all water bodies present in the scene.
Once the water features are isolated, the program analyses their spectral profiles to identify potential pollution. The team used known reflectance patterns associated with turbid or nutrient-enriched water - for example, increased reflectance in the red band, or specific wavelength shifts caused by suspended particles or algal growth. By combining these cues, the system assigns each water pixel a relative pollution score, highlighting areas that may require further investigation.
The resulting data is overlaid onto a natural-colour base image, allowing users to see the landscape as they would from above, but with hotspots of concern visually highlighted. The output is simple yet powerful: a colour-coded map showing where water appears clean and where it may be contaminated, ready to be viewed in any GIS or mapping tool.
The team built a proof of concept over a single weekend. The prototype demonstrated the potential of remote sensing for large-scale waterway monitoring. It showcased how advanced satellite data can be transformed into practical conservation insight through clever engineering and automation. The approach is especially valuable for regions that are difficult to access or lack permanent water quality sensors, offering an early-warning system that complements traditional field-based monitoring.
The project also reflects the interdisciplinary nature of CESA's work. The team combined expertise from environmental science, data analysis, and software design to translate raw spectral data into meaningful ecological information. This integration of engineering precision and conservation purpose is at the core of CESA's mission - to apply technology in service of Aotearoa's natural heritage.
Future iterations of the system could refine its accuracy through calibration with field samples or by integrating machine learning models to recognise more complex pollution patterns. With additional development, such a system could form part of a nationwide monitoring network, enabling agencies and community groups to track the health of waterways in near real time.
CESA's demonstrated that satellite imagery, when coupled with thoughtful engineering, can unlock new tools for environmental stewardship. The team's water quality monitoring prototype stands as an example of how innovation, collaboration, and a commitment to Aotearoa's ecosystems can produce practical solutions with national significance.
