Advances in joint species distribution models to reveal community structure and its environmental response

Abstract

Species distribution models are commonly used in basic and applied ecological research to examine the factors driving the distribution and abundance of organisms. They are employed to quantify species’ relationships with abiotic conditions, to predict species’ response to land-use and climatic change, and to identify potential conservation areas. Biotic interactions have been rarely included in traditional species distribution models, wherein joint species distribution models (JSDMs) emerge as a feasible approach to simultaneously incorporate environmental factors and interspecific interactions, making it a powerful tool for analyzing the structure and assembly of biotic communities. Generally, the JSDMs are based on species distribution models (SDMs), with the abundance or occurrence of multiple species as response variables and environmental factors, species associations and species traits being incorporated in the modeling framework. These models commonly use generalized linear regression methods (GLM) to relate multivariate response to environmental variables, and capture species associations in the form of random effects. The limitation has been overcome by the introduction of latent variable models (LVMs). Typically, the model parameters are estimated using maximum likelihood estimation or Bayesian methods implemented by Laplace Approximation and Markov Chain Monte Carlo (MCMC) simulations, respectively. In this review, the generation and theoretical basis of JSDMs were summarized. The characteristics of different types of JSDMs and their applications in modern ecology were emphatically introduced. The problems and prospects of JSDMs were discussed. With the in-depth study of the relationship between environmental factors and multi-species interactions, JSDMs would be the focus of future studies of species distribution model.

Keywords: biogeography.; ecological niche; joint species distribution models (JSDMs); latent variable models (LVMs); species distribution models (SDMs).