Very-high resolution coral reef mapping in Mayotte using Satellite Imagery
Résumé
Tropical coral reefs are the most biodiverse ecosystems on the planet, but increasing human pressures are rapidly degrading those sensitive ecosystems. Marine habitat maps, often used to manage coastal environments, tend to be available at spatial resolutions too coarse for the needs of management, impacting management strategies and monitoring efforts. Here, we developed an innovative approach for conducting very high-resolution (sub-meter) mapping of benthic habitats semi-automatically and applied it across the 195 km of Mayotte's reef flat. The mapping was performed using 50-cm resolution Pleiades satellite imagery and LiDAR-derived bathymetry. Satellite images were supplemented by underwater images acquired along 193 transects using a GoPro camera fixed to a diver propulsion vehicle. To achieve a very high spatial accuracy, underwater images were positioned with a few centimeters accuracy using a Global Navigation Satellite System (GNSS) system and post-processed kinematic (PPK) corrections. Geo-referenced image frames were extracted from the videos and integrated for annotations into CoralNet, a free online benthic image analysis software using semi-automatic deep learning analysis. About 10% of the video frames (3323 images) were annotated, allowing the analysis of over 31,500 images and calculate habitats' covers on the reef. Satellite images were segmented using an Object-Based Image Analysis (OBIA) approach and segments properties were used with other variables in a Random Forest pixel-based classification. To ensure replicability, data processing and analyses were conducted using free and open source software Orfeo ToolBox, QGIS, and R Stats. Thirteen benthic habitats, extending from coastal mangroves to the outer reef slope, were mapped at the pixel level. CoralNet classifiers achieved 65% accuracy, while the Random Forest models reached an accuracy of 76%. The composition of benthic assemblages varied significantly around the island, with changes in live corals accompanied by a rise in muddy habitats, turf, and dead corals as anthropogenic pressures intensify. The detailed habitat maps produced through this project will serve as an essential management tool to inform the management of Mayotte's coastal waters and the approach developed could help produce more regular and systematic habitat maps in support of coastal zone management.