Landscape matters: Deriving a generalizable understanding of population connectivity using empirical data and graph theory
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Human impacts on ecosystems have accelerated globally, driving a 10% decline in terrestrial biodiversity and a 70% decline in wildlife populations over the past five decades. These losses are closely linked to habitat modification and fragmentation, highlighting the urgent need for management strategies grounded in a clear understanding of how wildlife use landscapes and navigate human-altered areas. Connectivity between populations is critical for species persistence, and is shaped by the interplay between landscape features and species movement. Most connectivity studies pursue application-focused goals, such as designing corridors or assessing land-use effects, often targeting single species within specific landscapes. While these approaches provide depth, they limit the development of general principles that apply across species and regions. Graph theory offers a powerful framework to distill complex connectivity patterns into comparable metrics, creating opportunities to identify such generalities. In this study, we examined connectivity patterns for 11 Indian carnivore species distributed across heterogeneous landscapes. Using secondary data, we developed habitat networks from species distribution models that incorporated both habitat quality and matrix resistance. We then applied graph theory to generate networks based on connectivity between identified habitat nodes, enabling comparisons across species and landscapes. Our results show that while connectivity patterns differ markedly among species, broad trends emerge. Larger-bodied species like tigers, which in our study are often threatened species, can overcome the effects of fragmentation better than smaller bodied species, however their connectivity is dependent on the existence of high quality patches. Fragmented and heterogenous landscapes were always associated with modular, less efficient networks irrespective of the species. Importantly, landscape characteristics had greater influence on network-level connectivity properties, while species traits more strongly determined node-level structural complexity of the network. By integrating network theory with multispecies analyses across diverse landscapes, our work moves beyond single-species case studies to identify general drivers of connectivity. In terms of conservation, our approach allows us to generate broad insights into drivers of population connectivity informing strategies for lesser-known species and guide more effective, landscape-scale management in an era of rapid environmental change.