Detection and succession properties of forested vernal pools with eDNA metabarcoding

Vernal pools are small seasonal wetlands found throughout North America, where they provide critical habitat for a variety of organisms, including many cryptic taxa of conservation concern. In forests, pools are often poorly documented and coarsely characterized, with emphasis placed on a small group of obligate, indicator fauna. Conventional monitoring methods for pool‐associated biota suffer from imperfect detection, limiting general syntheses of forest pool diversity and community dynamics. We examined the capacity of environmental DNA (eDNA) metabarcoding to detect indicator fauna and characterize biodiversity of eukaryotes in forested vernal pools. We collected water eDNA samples from 12 pools in Indiana, New York and Massachusetts from 2020 to 2021, employing different sampling and sequencing protocols to estimate local, landscape and temporal variation in eukaryotic detection, diversity and composition. Metabarcoding assays targeting microbial eukaryotes (18S), metazoans (COI), freshwater invertebrates (COI) and vertebrates (12S) detected a wide range of taxa, including (respectively) 122 genera of eukaryotes, 48 species of metazoans, 190 species of invertebrates and 40 species of vertebrates, but most sequences could not be identified to lower taxonomic classifications. Vernal pool indicator fauna were successfully detected, with eDNA abundance peaking in May for fairy shrimp ( Eubranchipus vernalis ) and in June–July for wood frogs ( Lithobates sylvaticus ) and mole salamanders ( Ambystoma spp.). Pools showed a range of eDNA community variation over space, appearing randomly distributed at a local (within pool) scale and highly differentiated between pools. Pool communities were strongly clustered over time, with turnover in eDNA every 15–30 days depending on taxonomic group. Practical implication: Our results suggest that eDNA metabarcoding can serve as a complementary tool for ‘scaling up’ vernal pool biodiversity monitoring; however, practitioners must reconcile methodological uncertainties, from sampling to sequencing, with monitoring objectives to ensure robustness and cost‐effectiveness for any given study.