Multivariate Approaches to the Study of Urban Biodiversity and Vegetation: An Example from a Southern Temperate Colonial City, Christchurch, New Zealand

A prerequisite for more sustainable urban design is an understanding of the current composition of urban plant communities and what 'drives' their compositional variation. Various approaches have been used in the past to describe urban plant community patterns, including phytosociological approaches in Europe and more quantitative urban-rural gradient approaches in the United States. We used multivariate statistical methods to describe compositional variation and causation in urban biotopes of Christchurch city,New Zealand. From stratified random biotopes, we collected compositional, environmental and 'social' data at a range of spatial scales. Our data analysis 'tool box' included TWo-way INdicator SPecies ANalysis (TWINSPAN), descriptive statistics, Principal Components Analysis (PCA), Principal Co-ordinates Analysis (PCoA), ordination (Detrended DCA and Canonical Correspondence Analysis CCA) and regression. In this chapter, we provide examples of our approach and how our findings can be applied to sustainable urban design and restoration.