The spatial variation of soil bacterial community assembly processes affects the accuracy of source tracking in ten major Chinese cities.

Urban soils harbor billions of bacterial cells and millions of species. However, the distribution patterns and assembly processes of bacterial communities remain largely uncharacterized in urban soils. It is also unknown if we can use the bacteria to track soil sources to certain cities and districts. Here, Illumina MiSeq sequencing was used to survey soil bacterial communities from 529 random plots spanning 61 districts and 10 major cities in China. Over a 3,000 km range, community similarity declined with increasing geographic distance (Mantel r=0.62), and community composition was clustered by city (R2=0.50). Within cities (<100 km), the aforementioned biogeographic patterns were weakened. Process analysis showed that homogenizing dispersal and dispersal limitation dominated soil bacterial assembly at small and large spatial scales, respectively. Accordingly, the probabilities of accurately tracking random soil sources to certain cities and districts were 90.0% and 66.7%, respectively. When the tested samples originated from cities that were more than 1,265 km apart, the soil sources could be identified with nearly 100% accuracy. Overall, this study demonstrates the strong distance-decay relationship and the clear geographic zoning of urban soil bacterial communities among cities. The varied importance of different community assembly processes at multiple spatial scales strongly affects the accuracy of microbial source tracking.