Combining Multispectral and Radar Imagery with Machine Learning Techniques to Map Intertidal Habitats for Migratory Shorebirds

Migratory shorebirds are notable consumers of benthic invertebrates on intertidal sediments. The distribution and abundance will strongly depend their prey landscape sediment features such as mud surface water content, topography, the presence ecosystem engineers. An understanding shorebird ecology thus requires knowledge various habitat types which may be distinguished in areas. Here, we combine Sentinel-1 Sentinel-2 imagery a digital elevation model (DEM), using machine learning techniques to map importance migratory prey. We do this third most important non-breeding area for East Atlantic Flyway, Bijagós Archipelago West Africa. Using pixel-level random forests, successfully mapped rocks, shell beds, macroalgae between areas bare occupied by fiddler crabs, an engineer that promotes significant bioturbation flats. also classified two (sandy mixed) within crab areas, according content. overall classification accuracy was 82%, Kappa Coefficient 73%. predictors were elevation, Sentinel-2-derived moisture indexes, VH band. association with DEM produced best results compared models without these variables. This provides picture composition habitats site international shorebirds. Most flats covered sandy sediments (59%), ca. 22% is crabs. likely has implications spatial arrangement invertebrate communities due engineering vastly different species assemblages. large-scale mapping product future monitoring high biodiversity area, particularly ecological research related feeding Such information key from conservation management perspective. By delivering successful comprehensive workflow, contribute filling current gap application remote sensing among challenging environments techniques.