Trophic state modulates congruence between metabarcoding and microscopy in freshwater biomonitoring

Phytoplankton communities are key indicators of ecosystem health, yet routine monitoring, which relies on the observation of morphology and pigments under light microscopy, requires taxonomic expertise and is resource intensive and taxonomically limited. Environmental DNA (eDNA) metabarcoding offers a scalable alternative, but its performance relative to microscopy across trophic gradients remains poorly studied. Here, we combined a full-length 18S and 16S rRNA metabarcoding with light microscopy to characterize microbial communities across a broad trophic gradient in Chautauqua Lake over two sampling seasons. Both approaches were congruent in the taxa they detected, but metabarcoding consistently recovered higher diversity, particularly under hypereutrophic conditions when microscopy's diversity rapidly saturated. While microscopy and 16S rRNA diversity followed a unimodal pattern along the trophic gradient, 18S rRNA richness continued to increase even under intense cyanobacterial blooms. Metabarcoding also revealed diverse microeukaryotes and prokaryotic groups, absent or underrepresented from microscopy observations. Together, these results indicate that no single method is sufficient on its own: microscopy remains necessary for biomass and biovolume metrics, 16S rRNA improves detection of cyanobacterial and bacterial community gradients, and 18S rRNA expands microeukaryotic coverage, particularly under bloom conditions. These patterns indicate that long read sequencing is a promising complement to light microscopy for characterizing microbial communities across trophic gradients driven by decreasing costs, expanding reference databases, improved taxonomic resolution, and simplified bioinformatic workflows. However, integrating this platform into routine or statutory monitoring programs will require further calibration.