Prof Simon Parsons, University of Lincoln; Prof Ana Ivanovic, University of Aberdeen.

Trawling is known to release substantial amounts of CO2 by disturbing carbon-rich seabed sediments – but we currently lack effective ways to monitor these admissions at scale. 

This project aims to explore an innovative solution: using satellite imagery and AI-based computer vision to detect turbidity plumes generated by trawlers and assess whether these plumes can serve as reliable indicators of C02 release.

Recent research shows that large-scale CO2 emissions, such as those from power plants, can be detected using hyperspectral satellite data. Similarly, seabed-disturbing activities like dredging create visible turbidity plumes detectable in conventional RGB satellite imagery. 

However, it remains unknown whether the smaller CO2 signals produced by trawlers can be directly or indirectly measured from space. Your challenge is to help answer that question.

You will develop and test machine-learning models across three interconnected tasks.

First, using public RGB imagery, you’ll train and evaluate models capable of identifying sediment plumes associated with trawling. This offers a scalable, cost-effective method to locate potential emission events and link them to individual vessels, especially when combined with vessel-tracking data.

Second, using public datasets (eg Carbon Mapper) and hyperspectral imagery, you’ll explore whether trawler-associated CO2 emissions can be detected directly. Although potentially more accurate, this approach faces challenges including lower spatial resolution and higher data costs.

Finally, by combining insights from RGB and hyperspectral analysis, you will design approaches to quantify CO2 emissions – either indirectly via plume characteristics or through hybrid detection pipelines.

This project sits at the intersection of AI, environmental monitoring, and sustainable food systems. Its outputs will support Scope 3 emissions reporting in the seafood industry, and contribute to global emission-mapping initiatives. 

If you’re passionate about applying AI to real-world sustainability challenges, this project offers a unique opportunity to innovate.