Optimizing a Novel <scp>eDNA</scp> ‐Based Framework for Reef Fish Biodiversity Monitoring Using an Autonomous Filtration System and in situ Nanopore Sequencing

ABSTRACT Environmental DNA (eDNA) metabarcoding emerged as a powerful method for biodiversity monitoring, offering non‐invasive, reproducible, and scalable assessments. However, its implementation in marine environments and remote regions still presents significant technical and logistical challenges. Here, we tested several methodologies to propose a comprehensive field‐adapted eDNA metabarcoding workflow, from sampling to taxonomic analysis, designed for the specific challenge of fish biodiversity monitoring in coral reefs, but transferable to various aquatic ecosystems. We use large‐volume water sampling using a custom‐built autonomous underwater filtration system, which reduces contamination risks, simplifies logistics, and enhances sampling consistency. In this frame, we assess the efficiency of our custom underwater filtration system, different filter porosities, and water volumes to optimize DNA yield, showing that 20 L filtered on a 1.2 μm pore‐size filter was optimal for collecting fish eDNA. We chose the mitochondrial 12S rRNA for DNA amplification then Oxford Nanopore Technologies sequencing for its portability and compatibility with tropical fieldwork. We developed a high‐resolution 12S reference database encompassing around 400 fish sequences, representing over 50% of the known fish biodiversity in French Polynesia. Finally, we benchmark eDNA‐based biodiversity data against traditional visual census observations across multiple reef habitats and temporal replicates. Our approach enables high‐resolution monitoring both at spatial and temporal scales while providing a more comprehensive assessment of fish biodiversity richness compared to visual census methods. Diel sampling revealed pronounced temporal structure, with nocturnal taxa enriched at night. This study validates a fully integrated eDNA metabarcoding workflow, from field collection to taxonomic analysis, in diverse and remote marine ecosystems, providing a scalable and cost‐efficient framework solution for long‐term biodiversity monitoring in remote regions.