Premium
Extension of the gambin model to multimodal species abundance distributions
Author(s) -
Matthews Thomas J.,
Borregaard Michael K.,
Gillespie Colin S.,
Rigal François,
Ugland Karl I.,
Krüger Rodrigo Ferreira,
Marques Roberta,
Sadler Jon P.,
Borges Paulo A. V.,
Kubota Yasuhiro,
Whittaker Robert J.
Publication year - 2019
Publication title -
methods in ecology and evolution
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.425
H-Index - 105
ISSN - 2041-210X
DOI - 10.1111/2041-210x.13122
Subject(s) - range (aeronautics) , computer science , sample size determination , multimodality , software , extension (predicate logic) , r package , statistical model , machine learning , artificial intelligence , data mining , statistics , mathematics , materials science , computational science , world wide web , composite material , programming language
Species abundance distributions ( SAD s) are one of the most widely used tools in macroecology, and it has become increasingly apparent that many empirical SAD s can best be described as multimodal. However, only a few SAD models have been extended to incorporate multiple modes and no software packages are available to fit multimodal SAD models. In this study, we present an extension of the gambin SAD model to multimodal SAD s. We derive the maximum likelihood equations for fitting the bimodal gambin distribution and generalize this approach to fit gambin models with any number of modes. We present these new functions, along with additional functions to aid in the analysis of multimodal SAD s, within an updated r package (“ gambin ”; version 2.4.0) that enables the fitting, plotting and evaluating of gambin models with any number of modes. We use a mixture of simulations and empirical datasets to test our new models, including tests of the sensitivity of the model parameters to the number of individuals and the number of species in a sample. We show that the new multimodal gambin models perform well under a variety of circumstances, and that the application of these new models to empirical SAD and other macroecological (e.g., species range size distributions) datasets can provide interesting insights. The updated software package is simple to use and provides straightforward yet flexible statistical analyses of multimodality in SAD ‐type datasets.