Scientists at the University of Miami have developed an AI model that predicts when coral reefs experience heat stress — caused by rising ocean temperatures — in the hope of restoring them before deadly coral bleaching occurs, which can harm coastal regions.
The Florida researchers trained an open-source machine learning model with 40 years of environmental data to flag when the dangerous heating process could start up to six weeks in advance. It not only forecasts if heat stress will occur in a particular season, but estimates the exact week it is most likely to begin. It also identifies the environmental factors specific to each coral reef site that contribute to the decay, helping scientists customize action plans ahead of time.
Coral bleaching is a huge problem globally, causing catastrophic damage to the reefs that provide food and shelter to a quarter of marine life. They also protect coastal regions, where roughly 1 billion people live globally, according to the Great Barrier Reef Foundation.