AbstractThe fundamental question of this thesis addresses the suitability of spatial point processes for the modelling of plant communities with the aim of understanding the mechanisms that sustain biodiversity and allow species coexistence through an analysis of a biodiverse plant community in Western Australia.
Terrestrial plants are non-motile organisms and hence interact mainly with their nearest neighbours. Therefore, plant community dynamics need to be modelled in a spatial context. Spatial point process models describe the arrangement of objects in space and derive information on inter- and intra-species interaction through an analysis of the objects’ relative position. However, there has been a lack in appropriate methodology since previous applications of spatial point processes have analysed data sets of much smaller complexity and research has been mainly theory driven.
This thesis develops new methodology suitable for this context. It provides parsimonious tools for the exploratory data analysis and derives multivariate methods for spatial point pattern data for the first time. In addition, three models of increasing complexity are fitted to the data and their suitability is assessed.
This thesis has provided a number of statistical tools that may also be successfully applied in other situations, notably where highly multivariate data sets of spatial patterns occur. The methods are suitable in the context of plant community dynamics and their application to the study data set has made contributions to the development and validation of existing ecological theories on biodiversity. Furthermore, they informed specifically on intra- and inter-species interaction structures in the study data set and are ultimately contributing to conservation.
|Date of Award||Feb 2006|
|Supervisor||John W. Crawford (Supervisor)|