Complementary but Not Interchangeable: A Comparison of eDNA and Bulk Sample Metabarcoding for the Assessment of Groundwater Invertebrate Communities

ABSTRACT

Groundwater invertebrates exhibit a high level of biodiversity, and robust assessments are crucial for the protection of groundwater ecosystems and the preservation of drinking water quality. Molecular approaches are increasingly used to characterize subterranean biodiversity, but it remains unclear to what extent water‐derived eDNA and bulk sample metabarcoding recover comparable patterns of diversity and community composition. We therefore evaluated the agreement and complementarity of both approaches across 50 sampling locations for the assessment of groundwater invertebrate communities at multiple taxonomic levels, including OTUs, phyla, and species, with a particular focus on stygofauna and the distinction between stygobiont and non‐stygobiont species. In addition, we examined how sampling context influenced community profiles by comparing springs and hydrants, where both methods were applied to the same medium, with groundwater monitoring wells (GMW), where borehole and aquifer samples were spatially separated. Overall, the two methods recovered significantly different aquatic invertebrate community structures, showed limited taxonomic overlap, and produced distinct method‐specific patterns. Crustacean species were recovered more effectively by bulk sample metabarcoding, whereas smaller and soft‐bodied taxa were better detected by eDNA metabarcoding. Among the 54 aquatic species detected overall, 22 were classified as stygobionts, of which 100% were recovered by bulk sample metabarcoding, whereas eDNA detected only 50%. Dissimilarities between borehole and aquifer communities were largely masked by methodological differences between eDNA and bulk sample metabarcoding. These findings show that both approaches are complementary rather than interchangeable and highlight the need to account for methodological effects when assessing groundwater biodiversity and interpreting borehole‐aquifer relationships.