Modelling the spatial pattern of biodiversity utilizing the high-resolution tree cover data at large scale: Case study in Yunnan province, Southwest China

Abstract

Explicit biodiversity level at large spatial scale is often indicated by species richness. The spatial variability of species richness has been postulated to be related with influencing factors such as climate, habitat and human activities using many geostastical methods. Understanding the relationship between species richness and its influencing factors is of critical importance for its conservation. As tree cover data with high resolution can reflect the terrestrial habitat with environment, its efficiency for modelling spatial pattern of biodiversity was investigated in this study. The spatial distribution of tree cover and its influencing factors in Yunnan province, and the relationship between tree cover and species richness of mammals and birds were analyzed by linear regression method and Geographically Weighted Regression (GWR) model. The results showed that tree cover in this study region exhibited a substantially spatial heterogeneity, which was affected by human activities, topography and climate. Linear regression analysis showed that tree cover was significantly correlated with the species richness of mammals and birds. The explanatory power (local R2) of GWR model between tree cover and species richness of mammals and birds showed great spatial variability. The predicted spatial distribution of species richness of mammals and birds was consistent with the actual observed values. Our study revealed that the GWR model provides local details of spatial variation in species richness of mammals and birds. These results will be helpful to understand the impact of tree cover on variation of species richness of mammals and birds under global change and will be of theoretical significance for better conservation and management of biodiversity.