Landscape connectivity dynamics based on network analysis in the Xishuangbanna Nature Reserve, China

• Forest dynamics of Xishuangbanna Nature Reserve during 1970–2000 were presented.
• Connectivity changes and impact of land use changes on connectivity were measured.
• The potential connectivity network were displayed by effective paths.
• Functional connectivity indices are significantly correlated with structural indices.

Lack of landscape connectivity and habitat loss is major threats to biodiversity and ecosystem integrity in nature reserves aimed at conservation. In this study, we used structural pattern and functional connectivity metrics to analyze the spatial patterns and landscape connectivity of habitat patches for the Shangyong sub-reserve of the Xishuangbanna Nature Reserve from 1970, 1990, and 2000. On the basis of vegetation and land cover data, we applied the equivalent connected area ECA(PC) indicator to analyze the changes in forest connectivity. Four distance thresholds (2, 4, 8, 12 km) were considered to compare the patch importance of connectivity by dECA values. The results showed the declining trends of landscape connectivity measured by ECA(PC) index from 1970 to 2000. The importance of connectivity in each forest patch varied with the increment of dispersal distances at the patch level, and some important habitat patches, which exhibit a potential to enhance landscape connectivity, should be given more attention. The least-cost pathways based on network structure were displayed under four dispersal distances in three periods. The results showed that the number of paths among the fragments of forest patches exhibited radical increases for larger dispersal distances. Further correlation analyses of AWF, ECA (IIC), and ECA (PC) showed the weakest and least-frequent correlations with the structural pattern indices, while H presented more significant correlations with the PD fragmentation metric. Furthermore, Kendall's rank correlations between the forest patch area and functional connectivity indicators showed that dECA (PC) and dAWF indicators should provided the area-based prioritization of habitat patches. Moreover, the low-rank correlations showed that dF and dLCP can be considered as effective and appropriate indicators for the evaluation of habitat features and network patterns.