Urbanization represents a global trend that is often defined by dramatic human-induced environmental changes. These changes directly influence a number of factors including the regional and local processes collectively responsible for structuring local plant communities. Studies often attribute community-level shifts in composition to local effects such as competition or environmental sorting, and thereby do not consider the importance of dispersal as a regional driver of community assembly. Yet if different dispersal strategies are favored among local habitat patches, the trajectory of assembly will reflect this variation well before local processes influence community structure. We used database seed dispersal trait measures to explore whether community composition and diversity of dispersal strategies shifted between two vacant lot types in Baltimore, Maryland. PERMANOVA tests were performed to identify divergence in community-level composition based on seed dispersal mechanisms. Comparisons of mean community weighted trait values for each strategy were assessed using mixed effects modeling with lot number as a random effect.
We found clear differentiation in community-level trait composition between urban sites. These shifts were driven by significant differences in the representation of several dispersal strategies. Models incorporating the random effect of vacant lot number explained the most variance, thereby suggesting that heterogeneity within the urban environment may differentially influence the recruitment of plant species into habitat patches. These results highlight that although local environmental characteristics between these sites have been identified as potentially important filters in assembling these communities, these local processes may not be the sole driver of differences in plant community structure. Therefore, in order to make predictions about the way plant communities are structured, researchers must consider local variability of regional processes more closely.