Node taxonomic resolution affects the robustness of marine macrophyte–epifauna networks

• DNA metabarcoding increases node resolution in marine macrophyte–epifauna networks. • Higher node resolution reveals hidden macrophyte–epifauna association patterns. • Metabarcoding-based networks show lower robustness to macrophyte loss. • Aggregating nodes in morphology-based networks overestimate network robustness. Ecological networks provide a powerful framework for investigating species interactions in natural communities and addressing a wide range of ecological questions. But assembling realistic ecological networks depends on our ability to identify the species in question accurately. This often leads to the reconstruction of simplified networks, where nodes are aggregated at a high taxonomic level or by functional groups. However, such aggregations not only reduce the size of the network (in terms of number of nodes and links) but might also cover species-specific patterns of interaction, altering network structure and properties, and hence affecting our understanding of the study system. Here we explore the issue in a small set of real-world ecological networks (marine macrophyte–epifauna networks from the Baltic Sea) which were first assembled using a morphological species identification, leading to simplified networks where some species were pooled into higher taxonomic units. We then reassembled these networks using a molecular approach (DNA metabarcoding), which allowed resolving most nodes to the species level. We show that increasing the taxonomic resolution of network nodes with DNA metabarcoding can provide a substantially different picture of these systems, revealing hidden patterns of association which, in turn, affect the system’s robustness to macrophyte loss.