Distance-based parametric bootstrap
tests for clustering of species ranges
Christian Hennig and Bernhard Hausdorf
November 2002
Abstract
This paper deals with species range data, i.e., n species (taxa) are
characterized by their presence or absence on c units into which a
map is
subdivided. Such data occur often in biogeography. We propose some tests for
the existence of clusters of species according to their ranges.
We define some distance-based test statistics for the presence of
clustering, we propose a null model for the generation of a species and an
alternative model for clustering.
The models include a parameter governing the spatial autocorrelation of its
occurrence in the cells and they account for the species richness of
the individual cells. The distribution of the test statistics can be
estimated by a parametric bootstrap simulation (Monte Carlo with estimated
parameters) from the null model. The
validity of the p-values and the power of the tests are considered by
exemplary simulations.
We discuss also, but do not focus on, the determination of the clusters.
Download:
Compressed Postscript (144 Kb)
PDF (992 Kb).
Go back to the
Research Reports
from
Seminar für Statistik.