Satellite Fusion Sharpens Fish Biodiversity Mapping

A framework combines hyperspectral imagery, the Google Satellite Embedding dataset (GSED), environmental DNA (eDNA), and machine learning to predict fish taxonomic richness. By linking biological observations with habitat information, the approach improves accuracy over conventional multispectral methods and offers a scalable, noninvasive pathway for biodiversity mapping, conservation planning, and freshwater surveys.