A stochastic model for landscape patterns of biodiversity

Authored by Jayme A Prevedello, Nicholas J Gotelli, Jean Paul Metzger

Date Published: 2016

DOI: 10.1002/ecm.1223

Sponsors: Brazilian National Council for Scientific and Technological Development (CNPq) United States National Science Foundation (NSF)

Platforms: R

Model Documentation: Other Narrative

Model Code URLs: http://onlinelibrary.wiley.com/store/10.1002/ecm.1223/asset/supinfo/ecm1223-sup-0009-DataS1.zip?v=1&s=12539f36d1a5641c3d68d62295e6bda4aab1d4a6

Abstract

Many factors have been proposed to affect biodiversity patterns across landscapes, including patch area, patch isolation, edge distances, and matrix quality, but existing models emphasize only one or two of these factors at a time. Here we introduce a synthetic but simple individual-based model that generates realistic patterns of species richness and density as a function of landscape structure. In this model, we simulated the stochastic placement of home ranges in landscapes, thus combining features of existing random placement and mid-domain effect models. As such, the model allows investigation of whether and how geometric constraints on home range placement of individuals scale up to affect species abundance and richness in landscapes. The model encompassed a gradient of possible landscapes, from a strictly homogeneous landscape to an archipelago of discrete patches. The model incorporated only variation in home range size of individuals of different species, with a simple suitability index that controlled home range spread into areas of habitat and areas of inter-habitat matrix. Demographic details of birth, death, and migration, as well as effects of species interactions were not included. Nevertheless, this simple model generated realistic patterns of biodiversity, including species-area curves and increases in diversity and abundance with decreasing isolation and increasing distance from patch edges. Species-area slopes (z values) generated by the model fell within the range observed in empirical studies on both true islands and habitat patches. Isolation and edge effects were stronger when the matrix was unsuitable, and became progressively weaker as matrix suitability increased, again in accordance with many empirical studies. When applied to a real data set on the abundance of 20 small mammal species sampled in forest patches in the Atlantic Forest of Brazil, the model predicted increases in abundance and richness with increasing patch size, consistent with the general pattern observed with field data. The ability of this simple model to reproduce realistic qualitative patterns of biodiversity suggests that such patterns may be driven, at least in part, by geometric constraints acting on the placement of individual ranges, which ultimately affect biodiversity patterns at the landscape level.

Tags

Habitat fragmentation Population-density Small mammals Patch size Atlantic forest landscape Species richness patterns Geometric constraints Island biogeography Area relationships Matrix quality