Abstract
Understanding the drivers of seasonal community assembly in dynamic coastal ecosystems remains a central challenge in marine ecology. A null-model approach was used to investigate whether seasonal changes in the functional diversity of a demersal fish bycatch assemblage in Guanabara Bay, Brazil, are driven by deterministic trait-based filtering or stochastic processes. Four-dimensional functional spaces were constructed from morphological, reproductive, trophic, and habitat traits. Seasonal functional richness (FRic), evenness (FEve), and divergence (FDiv) were compared to null distributions generated under richness- and abundance-constrained random assembly. Observed seasonal assemblages were found not to deviate from null expectations, and pairwise seasonal changes were indistinguishable from stochastic variation. However, null expectations for FRic were observed to shift significantly between seasons, indicating a two-tiered filter-then-shuffle dynamic: broad-scale environmental filters deterministically reshape the species pool, while seasonal assemblage formation proceeds stochastically. These findings underscore the scale-dependency of assembly processes and highlight the utility of null models in disentangling deterministic from stochastic signals in highly dynamic coastal ecosystems.