AI Helps Scientists Track Ocean Carbon and Microplastics

Scientists have long watched a mysterious snowfall happening beneath the waves, but now they're finally learning what it's made of.

Tiny particles constantly drift from the ocean's surface to the deep sea, carrying carbon, nutrients, and unfortunately, pollution along the way. University of Maine researchers Meg Estapa and Chaofan Chen just won nearly $700,000 from the National Science Foundation to develop AI tools that can identify what these particles contain just by looking at underwater images.

The stakes are higher than you might think. This sinking "marine snow" shapes everything about the deep ocean, from its acidity to oxygen levels to which creatures can survive there. Understanding what sinks where helps scientists track how carbon moves through our planet's largest ecosystem.

For decades, studying these particles meant the slow, painstaking process of collecting water samples at sea, hauling them back to labs, and spending months analyzing them by hand. One of Estapa's graduate students recently spent months just classifying particles in underwater images. The AI approach could accelerate that timeline dramatically, freeing scientists to focus on discoveries instead of data entry.

The three-year project starting in January 2027 will train AI models using data from six major ocean expeditions, including waters off West Africa, the North Atlantic, and tropical regions. The team will pair thousands of particle images with lab analysis showing exactly what each speck contains.

Why This Inspires

What makes this project special isn't just the speed. Chen is designing neural networks that explain their reasoning to humans, not just spit out answers. Scientists will see which particle characteristics led the AI to each conclusion, helping them spot when the model gets confused and needs refinement.

This transparency matters because every new scientific tool needs skepticism alongside excitement. By understanding where the AI might mislead them, researchers can use it wisely and safely.

The technology could help track microplastics as they journey from surface waters into the deep ocean and through marine food webs. It could improve estimates of how carbon cycles through the seas. And the approach might extend beyond oceanography to any field struggling with complex, time-consuming image analysis, from ecology to medicine.

The project will also train graduate students in both artificial intelligence and oceanography, preparing the next generation of scientists to bridge these increasingly connected fields.

For Chen, it represents his first major project using AI for scientific discovery. "It'll be an exciting new thing for me," he said. That sense of possibility, of opening new windows into understanding our world, is exactly what good science should feel like.

Estapa's hope is simple but profound: "We hope this project will help us understand our ocean better, and our impacts on the ocean directly and indirectly." In an era when our oceans face mounting pressures, better understanding is the first step toward better protection.