Research
We work at the join between Earth observation, computer vision and coastal ecology. The through-line across everything below is the same: automating the analysis of very large image datasets so that ecosystems can be measured at a scale fieldwork alone cannot reach.
Blue carbon stock mapping
Estimating above- and below-ground carbon across mangrove extents by fusing Sentinel-2 optical imagery, NASA GEDI spaceborne LiDAR and field allometry with deep learning. Our first site is Mwache Creek in Mombasa, where 2,151 GEDI footprints constrain canopy structure across 2,890 hectares of mangrove forest. Outputs are structured against Verra VM0033 so they can support carbon project validation.
Blue Biodiversity Integrity Index
A carbon number says nothing about whether an ecosystem is alive and functioning. Blue-BII derives a biodiversity integrity index for blue carbon ecosystems from below-water stereo-video transects, using computer vision to identify and count organisms at a rate manual annotation cannot match. The index is designed to sit alongside carbon estimates in TNFD and GRI Biodiversity Standard reporting.
Deep-BII
Extending the integrity index below 200 m, developed with the Ocean Discovery League’s 2026 Western Indian Ocean cohort. Deep-sea habitats are the least surveyed and least protected part of the ocean, and the BBNJ Treaty creates an immediate need for methods that can assess them from sparse imagery. This work builds directly on our deep-sea image analysis in the Pacific and Atlantic.
Where the methods come from
The lab’s methods have their origin in deep-sea research. Between 2022 and 2025 we built and published automated workflows for classifying seafloor imagery and detecting megabenthic fauna across the Clarion–Clipperton Zone in the Pacific and the tropical Atlantic, working with hundreds of thousands of images collected by towed camera systems. Those pipelines — AI-SCW for seafloor classification and FaunD-Fast for fauna detection — are the technical foundation for what SIEAL now applies to Kenya’s coastal ecosystems. See Publications.
Field site
