Abstract: Rapid urbanization is shaping urban forests and greenery in China, playing a fundamental role in driving vegetation characteristics and landscape patterns. In this study, we classified the level of urbanization intensity based on the proportion of impervious surface area and investigated 184 plots in the builtup area of Nanchang City. Vegetation and landscape indices were measured and analyzed within each plot. The results showed that landscape patterns and vegetation characteristics greatly differed among different urbanized areas. With increasing urbanization intensity, the landscape indices, including mean perimeter-area ratio, mean Euclidean nearestneighbor distance, and landscape division index, increased by 130.8%, 63.2%, and 92.4%, respectively, and the percentage of landscape type, largest patch index, and aggregation index decreased by 82.8%, 85.5%, and 22.3%, respectively. Moreover, the lowest tree height, diameter at breast height, and herbaceous richness were observed in heavily urbanized areas, whereas woody species density exhibited the opposite trend. The results of variation partition analysis revealed that the impervious surface area and landscape indices explained 0.8% and 17.1% of the variation in vegetation characteristics, respectively, whereas their interactions explained up to 82.1%. The results of Pearson correlation and redundancy analyses showed that changes in forest vegetation characteristics were strongly associated with mean perimeter-area ratio, percentage of landscape type, patch density, and landscape division index. Increasing the mean perimeter-area ratio could increase the richness and cover of herbaceous species in moderately and lowly-urbanized areas. From the perspective of landscape pattern, we proposed suggestions for protecting urban forests and enhancing plant diversity, aiming to provide references for maximizing the ecological services of urban forests.

Key words: urbanization intensity, urban forest, impervious surface area, species diversity, landscape pattern