The presence of a seasonally variable biogenic habitat (eelgrass, Zostera marina) increased the spring-summer variability of associated nekton relative to unvegetated bare tidal flat. This spatio-temporal pattern emerged because most eelgrassassociated taxa tracked the decline in eelgrass biomass from summer to spring, but in one case reached greater density in spring when predation intensity was low. Among 21 taxa (26,884 individuals) captured, a strong correlation arose between structure association and summer dominance, and certain functional traits, in particular morphology and on-vs. off-bottom position, were strong predictors of eelgrass association. Structure-associated taxa were slender-bodied and pelagic schooling fishes, while habitat generalists or bare-associated taxa were more consistent seasonally, primarily benthic, and cryptically colored with sand. Estuarine use (transient, reproducing, or estuarine resident) was not a strong predictor of structure association or seasonality. Because an identical sampling design was used in five regions of Washington State, USA, coarse-scale (> 100 km) differentiation in nekton assemblages was identified, representing less of the total variation than across seasons but more than across different habitats. While regional nekton differences were attributable in part to geographic distance and eelgrass morphotypes, the most seasonally variable nekton were at sites with morphotypes adding the least vertical structure but highest density. These results support two mechanisms that increase seasonal variability of taxa using structured habitats, including both bottom-up provision of habitat and resources, and subsequent possibilities for negative interspecific interactions and top-down control.