UMaine researchers use AI to analyze marine snow

University of Maine researchers Meg Estapa and Chaofan Chen are developing AI tools to determine the chemical contents of marine snow, the mix of organic matter, minerals and other material that sinks from the ocean surface into deeper waters. The National Science Foundation awarded nearly $700,000 for the project, which is scheduled to begin in January 2027 and run through December 2029.

Underwater cameras can capture large numbers of particles and reveal traits such as size, shape and transparency, but images alone usually cannot show what the particles are made of. Estapa, Chen and University of Rhode Island collaborator Melissa Omand will use data from six major oceanographic field campaigns, including images from waters off West Africa, the North Atlantic and tropical regions, to test whether visible particle features can predict chemical composition.

The team will build a database pairing marine snow images with information on particle contents, microplastic levels and sample locations, with some samples also analyzed in the laboratory. Chen will develop neural networks designed to explain their conclusions, helping researchers spot when AI interpretations conflict with scientific understanding. The approach could reduce manual image classification and improve estimates of how carbon, nutrients and pollutants move through the ocean.