LAUSR

Abstract In the context of landscape fragmentation, a variety of ecological conservation theories have been proposed to maintain habitat connectivity. These theories focus on the connectivity of the ecological corridors in ecological space but ignore the community structure of ecological space. However, community structure is ubiquitous in ecological space. Our study predicted habitat quality of ecological patches by using an artificial neural network and species occurrence points. Then, the ecological space network (ESN) was proposed and constructed by taking large ecological patches and fragmented patches as habitat nodes. Based on island biogeography theory (TTIB) and network community detection, the ecological groups in the ESN were proposed and identified. They are considered to be collections of ecological land with close species connections. After that, the hierarchy of the ecological group was analyzed, and the existing policy was evaluated. The results show that there were 39 ecological groups in the ESN. The dendrogram of the hierarchy shows that all ecological groups can form five bigger ecological groups. The extent of the bigger groups proves that the ecological groups in the eastern part are closely connected. By analyzing the protection of ecological control lines on the linkages among ecological groups, we found that the existing policy could protect most of the ecological groups separately, but could not maintain connectivity among ecological groups, leading to habitat isolation. We believe that the concept of “ecological group” makes up for the lack of “habitat isolation” in the existing theory. The idea can understand the hierarchical structure of ecological space, which is conducive to the maintenance of ecological integrity.