A Likelihood Framework for Modeling Pairwise Beta Diversity Patterns Based on the Tradeoff Between Colonization and Extinction

Yongbin Wu,Youhua Chen,Tsung-Jen Shen

doi:10.3389/fevo.2020.00064

Abstract

Ecological communities are assembled by various mechanisms, including a tradeoff between colonization and extinction processes. However, it is still unclear how this tradeoff influences pairwise beta diversity pattern. Herein, we propose a novel likelihood framework to disentangle the compound impacts of species’ colonization and extinction rates on community-level compositional dissimilarity which are ignored by traditional indices. The framework infers two parameters of colonization and extinction rates, allowing ecologists to study their latitudinal or broad-scale spatial variation patterns and test the relative influences of associated environmental factors. More importantly, the likelihood-based model showed that multi-site beta diversity is essentially identical to the local colonization failure or extinction rate of species in newly colonized sites, but is not related to species’ colonization rates. Profoundly, the present likelihood model allows ecologists to explicitly infer the independent species’ colonization rate parameter and test its underpinning mechanisms, as it can be proven that this parameter would be only implicitly measured if used to compute multi-site beta diversity indices. KEYWORDS: beta diversity, island biogeography theory, ecological mechanisms, community assembly, biodiversity survey

Highlights

Beta diversity, which characterizes between-site differences in species compositions, has become a key biodiversity component in understanding the maintenance of biodiversity and the community structure of species assemblages (Condit et al, 2002; Baselga, 2010; Anderson et al, 2011; Kraft et al, 2011; Legendre and Legendre, 2012; Baselga and Leprieur, 2015; Chen, 2015)
We developed a novel but simple statistical framework that can explicitly incorporate the independent influences of colonization and extinction on pairwise beta diversity patterns
There should be a decreasing trend in the estimated multi-site beta diversity indices, βJac(r, e) and βSor(r, e), which are derived from the proposed framework and the data matrix saturation (Table 2 and Figure 2)

Summary

Introduction

Beta diversity, which characterizes between-site differences in species compositions, has become a key biodiversity component in understanding the maintenance of biodiversity and the community structure of species assemblages (Condit et al, 2002; Baselga, 2010; Anderson et al, 2011; Kraft et al, 2011; Legendre and Legendre, 2012; Baselga and Leprieur, 2015; Chen, 2015). Beta diversity indices developed in most previous studies (Koleff et al, 2003; Diserud and Ødegaard, 2007; Baselga, 2010; Chen, 2015; Ensing and Pither, 2015) are non-parametric, making it difficult to deeply quantify the influences of the underpinning biological mechanisms on structuring community patterns of ecological assemblages. Among the many non-parametric beta diversity indices, Jaccard (βJac) and Sorensen (βSor) indices (Jaccard, 1912; Sorensen, 1948) might be the most widely used ones in ecology and evolution in both pairwise and multi-site calculations of beta diversity (Table 1). Baselga (2010) presented and developed many useful metrics that can be computed at the multi-site setting.

Methods

Results

Conclusion