Predicting the impacts of global change on highly dynamic ecosystems requires a better understanding of how communities respond to disturbance duration, frequency and timing. Intermittent rivers and ephemeral streams are dynamic ecosystems that are recognized as the most common fluvial ecosystem globally. The complexity of the drying process can give rise to different annual and antecedent hydrological conditions, but their effect on aquatic communities remains unclear. Here, using aquatic invertebrates from 33 streams across a flow‐intermittence gradient, we assessed how annual (drying duration and frequency) and recent drying characteristics (duration of the last dry period and flowing duration since the last rewetting) affect the density and diversity metrics of communities and trophic groups while controlling for other key abiotic factors (dissolved oxygen and altitude). We characterized invertebrate communities using taxonomy and functional traits to capture biological features that increase vulnerability to drying. In addition, using structural equation modelling (SEM), we evaluated pathways by which drying characteristics directly impact invertebrate density and whether diversity indirectly mediates such relationships. We show that drying frequency drove reductions in diversity at the community level and within trophic groups, whereas both the drying duration and frequency had a negative influence on density metrics. Reductions in taxonomic richness were linked to increased annual drying duration, whereas functional diversity declined in response to annual drying frequency. Filterer, predator and shredder trophic groups exhibited the strongest negative responses to drying. Recent drying characteristics had a minor effect on density and diversity metrics. Our SEM results demonstrated that diversity mediates the negative impacts of annual drying duration and frequency on invertebrate density through reductions in their taxonomic richness and functional diversity. Our results underscore the importance of considering multiple drying characteristics together with the interdependence of density and diversity to better anticipate drying responses in freshwater ecosystems.