Understanding the drivers of variability in plant diversity from local to landscape spatial scales is a challenge in ecological systems. Environmental gradients exist at several spatial scales and can be nested hierarchically, influencing patterns of plant diversity in complex ways. As plant community dynamics influence ecosystem function, understanding the drivers of plant community variability across space is paramount for predicting potential shifts in ecosystem function from global change. Determining the scales at which stress gradients influence vegetation composition is crucial to inform management and restoration of tidal marshes for specific functions. Here, we analyzed vegetation community composition in 51 tidal marshes from the San Francisco Bay Estuary, California, USA. We used model-based compositional analysis and rank abundance curves to quantify environmental (elevation/tidal frame position, distance to channel, and channel salinity) and species trait (species form, wetland indicator status, and native status) influences on plant community variability at the marsh site and estuary scales. While environmental impacts on plant diversity varied by species and their relationships to each other, overall impacts increased in strength from marsh to estuary scales. Relative species abundance was important in structuring these tidal marsh communities even with the limited species pools dominated by a few species. Rank abundance curves revealed different community structures by region with higher species evenness at plots higher in the tidal frame and adjacent to freshwater channels. By identifying interactions (species–species, species–environment, and environment–trait) at multiple scales (local, landscape), we begin to understand how variability measurements could be interpreted for conservation and land management decisions.