Mapping the world's coral reefs using a global multiscale earth observation framework

• Published in: Remote sensing in ecology and conservation Vol. 6; no. 4; pp. 557 - 568
• Main Authors: B. Lyons, Mitchell, M. Roelfsema, Chris, V. Kennedy, Emma, M. Kovacs, Eva, Borrego‐Acevedo, Rodney, Markey, Kathryn, Roe, Meredith, M. Yuwono, Doddy, L. Harris, Daniel, R. Phinn, Stuart, Asner, Gregory P., Li, Jiwei, E. Knapp, David, S. Fabina, Nicholas, Larsen, Kirk, Traganos, Dimosthenis, J. Murray, Nicholas, Pettorelli, Nathalie, Lecours, Vincent
• Format: Journal Article
• Language: English
• Published: Oxford John Wiley & Sons, Inc 01.12.2020
• Subjects: Anthropogenic factors; Automation; Barrier reefs; Bathymetry; Classification; Classification schemes; Coral reef habitats; Coral reefs; Earth; Ecosystem assessment; Ecosystem dynamics; Ecosystems; Environmental changes; Geomorphology; machine learning; Mapping; monitoring; Remote sensing; Risk assessment; Sensors; Pacific region; Great Barrier Reef
• ISSN: 2056-3485; 2056-3485
• DOI: 10.1002/rse2.157

Coral reefs are among the most diverse and iconic ecosystems on Earth, but a range of anthropogenic pressures are threatening their persistence. Owing to their remoteness, broad spatial coverage and cross‐jurisdictional locations, there are no high‐resolution remotely sensed maps available at the global scale. Here we present a framework that is capable of mapping coral reef habitats from individual reefs (~200 km2) to entire barrier reef systems (200 000 km2) and across vast ocean extents (>6 000 000 km2). This is the first time this has been demonstrated using a consistent and transparent remote sensing mapping framework. The ten maps that we present achieved good accuracy (78% mean overall accuracy) from multiple input image datasets and training data sources, and our framework was shown to be adaptable to either benthic or geomorphic reef features and across diverse coral reef environments. These new generation high‐resolution map data will be useful for supporting ecosystem risk assessments, detecting change in ecosystem dynamics and targeting efforts to monitor local‐scale changes in coral cover and reef health. Here we present a mapping framework for coral reefs from the scale of individual reefs to the entire reef systems across millions of square kilometers of ocean. The framework can utilize a wide range of input covariate and training data sources, and outputs both geomorphic and benthic map types. The maps presented are the largest ever coral reefs maps produced from a consistent and transparent remote sensing approach.