Scientists using high-resolution satellites and machine learning have made a significant leap forward in understanding and monitoring the world’s coral reefs.
James Cook University’s Associate Professor Nicholas Murray was part of a University of Queensland-led study aiming to capture fine scale distribution of coral reefs worldwide.
”The team processed 100 trillion pixels from the Sentinel-2 and Planet Dove cubesat satellites to produce a global map of coral reefs,” said Dr Murray.
“The five-meter pixel resolution map captures fine scale detail of coral reef geomorphic zones and benthic substrate. A decade of collaborating on global mapping projects has allowed us to interpret Earth’s ecosystems from above.”
Dr Murray said the research had already triggered new efforts in coral reef conservation around the world including in Indonesia, Timor and Australia.
“This research offers a clearer and consistent picture of these ecosystems, and a powerful tool to guide and enable efforts for their continued management,” said Dr Murray.
Lead author of the paper, the University of Queensland’s Dr Mitchell Lyons said the new mapping also allowed scientists to refine the global estimate of the area of shallow coral reefs up to 20 to 30 metres deep.
“Through remote sensing and machine learning we’ve unlocked a new understanding for coral reefs. It’s more than just maps, it’s a tool for positive change for coral reefs, and coastal and marine environments at large,” said Dr Lyons.
The Australian Institute of Marine Science’s Dr Emma Kennedy said coral reefs formed complex underwater structures.
“Our classification scheme is a lens through which we capture their diversity of forms and features. In developing clear and consistent terminology for our maps, we aimed not just for accuracy but for utility, ensuring that these maps serve as valuable tools for conservationists, managers, and researchers alike.”
The high-resolution maps and associated data are now publicly accessible through the Allen Coral Atlas and Google Earth Engine.
