A network theoretic study of ecological connectivity in Western Himalayas

TL;DR

This study applies network theory to model ecological connectivity and seed dispersal among habitat patches of certain tree species in the Western Himalayas, revealing scale-free and small-world properties that facilitate rapid species spread.

Contribution

It introduces a network-based model for seed dispersal in Himalayan ecosystems, highlighting the network's scale-free, small-world, and non-modular characteristics affecting ecological connectivity.

Findings

The habitat network is scale-free and exhibits small-world properties.

Seed dispersal is rapid due to short path lengths and high clustering.

Hubs in the network significantly enhance species spread.

Abstract

Network theoretic approach has been used to model and study the flow of ecological information, growth and connectivity on landscape level of anemochory plant species Abied pindrow, Betula utilis and Taxus wallichiana in the Western Himalaya region. A network is formally defined and derived for seed dispersion model of aforementioned species where vertices represent habitat patches which are connected by an edge if the distance between the patches is less than a threshold distance. We define centrality of a network and computationally identify the habitat patches that are central to the process of seed dispersion to occur across the network. We find that the network of habitat patches is a scale free network and at the same time it also displays small world property characterized by high clustering and low average shortest path length. Due to high clustering, the spread of species is locally even as seed disperse mutually among the member vertices of a cluster. Also since every vertex is only a short number of steps away from every other vertex, the species rapidly covers all the habitat patches in the component. Also due to presence of hubs in the network the spread of species is greatly boosted whenever the species establish and thrive in a hub patch and disperse to adjacent patches. However, the network is not modular due to geographical constraints, and is is negatively assortative as the high degree vertices are connected to vertices of low degree.