Main Article Content
Eucalyptus bosistoana, habitat modelling, GIS, NZDFI, permanent sample plot network, site-species matching, stratified random sampling
Background: Understanding the relationship between sites and the plant species they support is essential for effective vegetation management. Site-species matching requires knowledge of the growth response of a given species to the full range of environmental conditions in potential planting sites. This can be achieved by repeatedly measuring species growth at a comprehensive network of sample plots that cover a range of environmental conditions, including topography, climate, and soil factors. The New Zealand Dryland Forests Initiative has established permanent sample plots (PSPs) of a plantation species, Eucalyptus bosistoana F.Muell., across New Zealand. However, these PSPs do not cover the entire range of environmental conditions available for the species and hence there is a need to expand the network of sites. The aim of this study was to determine optimal locations for new PSPs to provide more unique information to support site-species matching studies for Eucalyptus bosistoana in New Zealand.
Methods: A geographic information system (GIS) and stratified random sampling method were used to generate a model to identify optimal locations for E. bosistoana PSP establishment. The variables used in this study included topography, climate, and soil data. Redundancy between the initial set of potential explanatory variables was reduced by a multi-collinearity analysis. The potential habitat for the species was restricted to land with environmental conditions that could support E. bosistoana. All environmental variables were stratified and an initial priority index for each stratum in each variable was calculated. Then a weighted-overlay analysis was conducted to create the final priority index, which was mapped to identify high-priority areas for targeted PSP expansion.
Results: The existing PSP network for E. bosistoana generally covers the environmental conditions in low-elevation New Zealand dry lands, which are located alongside the east coast of the South Island, and the southern part of the North Island. The model identified high priority areas for PSP expansion, including several large regions in the North Island, especially in Rangitikei and Taupo Districts.
Conclusions: The model successfully allowed identification of areas for a strategic expansion of permanent sample plots for E. bosistoana. Newly identified areas expand upon the topographic, climatic, and soil conditions represented by the existing PSP network. The new area for PSP expansion has potential to provide valuable information for further site-species matching studies. The methodology in this paper has potential to be used for other plot networks of a different species, or even natural forests.