Clusters of high abundance of plants detected from local indicators of spatial association (LISA) in a semi-deciduous tropical forest.

José Ramón Martínez Batlle,Yntze Van Der Hoek,Christopher Bone

doi:10.1371/journal.pone.0208780

José Ramón Martínez Batlle, Yntze Van Der Hoek + Show 1 more

Open Access

https://doi.org/10.1371/journal.pone.0208780

Copy DOI

Abstract

Plants are rarely randomly distributed across communities, and patchiness is a common spatial pattern in most tropical forests. Clusters of high density of plant individuals are related to internal and external forces, as well as to historical events. The detection of aggregated patterns of plant individuals allows for a better understanding of the internal and external factors that guide the distribution of species. The aim of this research was to detect and characterize clusters of high abundance of plants and species richness in semi-deciduous forests in the Dominican Republic. For this, we collected vegetation data from 575 quadrats in 23 transects (2300 m2 in total) within the Ocoa river basin. Using local Moran’s I statistics, we isolated 18 quadrats of high density of individuals. We show that density of individuals can be 2.5 times larger on average than in non-aggregated quadrats, and can reach higher values for shrubs species as well as for palms and vines species. In addition, we found that shrub species are the most abundant group in aggregated quadrats, and density of tree species is significantly smaller than that of shrub species. High density quadrats are predominantly occupied by shrubs, palms and vines, following patterns of species composition and lithology. Detecting clusters of high density of individuals could help in the efficient assessment of richness in semi-deciduous tropical forests, and may support new conservation practices for this valuable but threatened ecosystem.

Highlights

Determining spatial patterns of species assemblages is key to understanding the mechanisms that control species distributions
Analyses of individual-based tree data often focus on quantifying the variation in tropical forest species related to abundance or on evaluating spatial patterns in tropical trees populations
We focus on point pattern analysis of individuals of plants, using data collected in transects from semi-deciduous tropical forest of the Dominican Republic, to detect and characterize spatially aggregated patterns of woody species using local Moran’s I, a local indicator of spatial association (LISA) proposed by Anselin [12]

Summary

Introduction

Determining spatial patterns of species assemblages is key to understanding the mechanisms that control species distributions. Spatial patterns in the distribution of communities can be shaped by internal (e.g., population dynamics) and external (e.g., environmental characteristics) deterministic forces, as well as current and historical stochastic events [1]. Analyses of individual-based tree data often focus on quantifying the variation in tropical forest species related to abundance or on evaluating spatial patterns in tropical trees populations. Distributed species are rare across a community [2, 3], and patchiness is found at all spatial scales [1]. Tree species are shown to be non-randomly aggregated in tropical.

Objectives

Methods

Results