Abstract
Urban blue–green spaces (UBGSs) are critical for biodiversity, yet our understanding of how plant diversity in these spaces responds to urbanization across different UBGS types remains limited. Here, we investigated plant communities in four sampled UBGSs along an urban gradient in the Minjiang River Basin, China. Because each blue–green space type is represented by a single site, all comparisons are site-specific and descriptive. Using species richness (S), the Shannon–Wiener index (H′), Pielou’s evenness index (J′), frequency (F), and importance value (I.V.), we described whether taxonomic composition and diversity patterns differed among the four sampled sites. Across all sites, based on 615 species records, angiosperms dominated (92.6%), and cosmopolitan families (e.g., Asteraceae, Poaceae, Fabaceae) were ubiquitous, consistent with a strong regional species pool influence. However, our cross-site comparison showed a contrasting pattern: the near-natural Baihetan National Wetland Park showed the highest S (12.87) but low evenness (0.49), whereas the intensively managed Huoshui Park, with the lowest richness (9.67), achieved the highest evenness (0.68). Further, a discordance between F and I.V. rankings was observed in the managed site, where ornamental species showed high spatial coverage but low I.V. ranks. We report this rank discordance only as an observational finding, not as a proven metric or indicator. This discordance suggests that richness alone may not fully capture community structure, though its diagnostic value remains unverified. We interpret these contrasting richness–evenness patterns from the four sampled sites. Because site identity is confounded with management intensity, habitat area, and landscape context, no causal attribution to ecological or anthropogenic factors can be made. Evenness and species richness showed different site-to-site patterns, but this association remains observational. We hypothesize that a resilient urban biodiversity network could combine complementary nodes, with large natural wetlands as species-rich sites and constructed parks as structurally balanced units, rather than pursuing a single diversity target. This study provides preliminary insights that may inform planning considerations under Chengdu’s Park City initiative, but these suggestions should be viewed as hypotheses requiring further validation rather than as established management guidelines. The proposed network approach, while conceptually appealing, requires empirical testing through replicated studies before it can be considered for broader application.