Abstract

Abstract: Most processes in the natural world take place in a spatial context and are often inherently local, particularly in plant ecology. As a result, spatially explicit data sets are frequently being collected in ecology. Previously this has been perceived as a difficulty since traditional statistical approaches are unsuitable in the presence of spatial autocorrelation. However, recent years have seen an increasing interest in explicitly modelling this spatial dependence.

Spatial point processes model the locations of objects such as individual plants, animals, nests in space. They take local dependence structures, the environment around and properties of the objects into account. A suitable model hence characterises the behaviour of every individual object based on a small number of interpretable parameters.

A growing number of methods and models have been developed within the statistical literature. However, a large part of this remains unknown to ecologists. This partly due to the fact that fitting these models is not straight forward and often computationally involved. In addition, many approaches have been derived from a rather mathematical perspective and are thus not necessarily appropriate in an ecological context.

This talk will present several approaches to developing point process methodology that may be suitably applied in ecology. It provides a number of examples of applications, including complex and multivariate point pattern data sets, modelling inhomogeneous data and marked point processes.

Biography: Janine Illian is an RCUK academic fellow at the Department of Mathematics and Statistics at the University of St Andrews. She studied at the Universities of Düsseldorf, Edinburgh and California at Davis. She holds an MSC in psychology and an MSc in mathematics and English from the University of Düsseldorf and a PhD in statistics form the University of Abertay, Dundee.

Janine Illian is the first author of a book recently published by Wiley that aims at providing an overview of the current state of the art of spatial point process methodology to both a statistical and a non-specialists readership.

Most of Janine’s work focuses on the interface between ecology and statistical modelling. Her main research interests include spatial statistics, in particular spatial point process modelling, functional data analysis approaches to complex data problems, multivariate methodology and assessing and measuring biodiversity.