While effects of built-up structure and woody vegetation cover on avian diversity are well studied in urbanized landscapes, how other environmental features (urban land-use types, e.g., commercial or residential area, building heights, and noise levels) affect avian diversity remains unclear. Relatively little attention has also been given to medium-sized cities in subtropical/tropical regions of Asia where open green space with crop or herbaceous vegetation is often found. We examined relationships between avian species (richness and occupancy) and those environmental characteristics within a subtropical urbanized landscape in southern China. We established 57 sample points across the city of Nanning, Guangxi, and performed bird surveys during the winter and the summer of 2016. The number of people observed and noise level were recorded as measures of human disturbance. We surveyed local scale environmental features including heights and urban land-use types of buildings within a 200m-circular area. The amount of woody vegetation, non-woody vegetation (crops at small farming areas and herbaceous vegetation at other open areas), and open water were also determined at the local scale and the landscape sale (a 500m-circular area). We used a Bayesian multi-species occupancy model with five principal components (PC1~5) summarizing variations in these environmental features as covariates.
During the winter, species richness and occupancy of all 26 species decreased with increasing built-up structure (PC1). Increasing non-woody vegetation (PC2) or woody vegetation (PC3) with deceasing building height and human disturbance influenced species richness positively. Occupancy by 11 species showing weak/strong responses to PC2 was high at areas where non- woody vegetation and 1-2 story buildings were dominant. Most species associated with PC3 showed positive responses. Urban land-use type (PC4) and amount of open water (PC5) did not significantly influence on avian species richness and occupancy. In the summer, observed species richness was low (only 11 species used for analysis). Species richness and occupancy by 7 species showed strong positive responses to PC2; however, there were no significant responses to other 4 variables. Our results partly support the finding of previous studies, i.e., the positive effect of woody vegetation on avian species in urbanized landscapes; however, our study further suggests that building height, which is related to human disturbance, may be one of the urban environmental features affecting avian species. Our results also indicate that open green space including small patches of crops often observed in small/medium-sized cities in Asia can play a positive role in conserving avian diversity.