A Cluster Analysis Approach To Classifying Irish Native Woodlands

Abstract

The National Survey of Native Woodlands in Ireland, which was initiated in 2003, will provide two key resources identified in the National Biodiversity Plan: an inventory of native woodlands and a woodland classification system. The project has taken advantage of recent advances in GIS and remote sensing to develop a GIS-based inventory of potentially native woodland sites. Field survey and characterisation of a large subset of these sites is currently ongoing. In this paper, a proposed methodology is presented for developing a national woodland classification based on data from this field survey using current best practice in statistical techniques. The applicability of this method is demonstrated using the data gathered thus far. The method uses hierarchical agglomerative cluster analysis and indicator species analysis with validation using multi-response permutation procedure (MRPP) and non-metric multidimensional scaling ordination (NMS). A two-tier classification system has been produced, with five broad groups of woodland being each divided into a number (between five and eight) of more specific woodland types. The five main vegetation groups were named, using the results of indicator species analysis, as: Quercus petraea Á Luzula sylvatica , Quercus robur Á Hedera helix , Corylus avellana Á Thamnobryum alopecurum, Alnus glutinosa Á Filipendula ulmaria and Betula pubescens Á Molinia caerulea. This classification defines a greater number of woodland types than previous studies based on the more subjective central European phytosociological approach. However, for applied purposes any evaluation based on the statistical significance of these groups may need to be tempered by ecological significance. It is apparent that the identification of native woodland communities may be hampered by the modified nature of Irish woodlands and the presence of non-native species in particular. With the dataset still expanding, the results must be regarded as preliminary, but it is concluded that this is an effective method for the objective classification of Irish native woodlands.

A hierarchical analysis of vegetation on a Mojave Desert landscape, USA

“…Using a vegetation data set from western Montana, USA, McCune and Grace (2002) reported that a mean P-value criterion for indicator species optimized at four groups, whereas the number of significant indicators was high between three and eight groups and was zero after 12 groups. These studies, combined with others (e.g., Perrin et al, 2006), suggest that there can be considerable variability in the results of this technique among communities, which have seldom included desert vegetation, but communities at intermediate hierarchical levels are typically best discriminated through ISA. Similarly, hierarchical ISA can help identify species distribution patterns among communities at different levels of a hierarchy (Dufrêne and Legendre, 1997).…”

Re‐defining the natural range of Scots Pine (Pinus sylvestrisL.): a newly discovered microrefugium in western Ireland

“…The concept of what constitutes a native species, and its resultant conservation value, has become increasingly relevant due to the recent challenges of biodiversity loss from invasive nonnative species (Simberloff et al, 2013) and the call for biodiversity inventories to inform conservation policies (Perrin et al, 2006). Pinus sylvestris L. is one of the most widely distributed tree species in the world.…”

Realising the ecosystem-service value of native woodland in Ireland

“…There is an opportunity to realise significant economic value where concentrated strategic planting can replace the current fragmented wooded landscape with diversity and connectivity. Perrin et al (2006) ii http://ec.europa.eu/environment/nature/biodiversity/ comm2006/2020.htm iii The figure was provided by Coillte (http://www. coillte.ie/aboutcoillte/recreation) for visits to its own landholding, although this represents the bulk of accessible forest in Ireland.…”