Original paper
New Zealand’s plot-based classification of vegetation
Wiser, Susan K.; De Cáceres, Miquel
Phytocoenologia Band 48 Heft 2 (2018), p. 153 - 161
36 references
published: May 1, 2018
published online: Dec 15, 2017
manuscript accepted: May 11, 2017
manuscript revision received: May 11, 2017
manuscript revision requested: Apr 9, 2017
manuscript received: Dec 11, 2016
ArtNo. ESP024004802005, Price: 29.00 €
Abstract
Abstract Aims: The classification approach presented here was developed to provide a national-scale quantitative plot-based vegetation classification of New Zealand. Methods: The classification approach consists of two hierarchically nested levels: alliances and associations. Forested (forests and shrublands) and non-forested (herbaceous) vegetation were considered two distinct consistent classification sections, differentiated by the use of cover abundance values versus relative species ranks to define vegetation types respectively. Noise clustering was used to define vegetation classes. After defuzzification of fuzzy memberships, a given plot record either (1) belongs to one vegetation type only; (2) is deemed transitional; or (3) is left unassigned (i.e. belongs to the noise, outlier, class). Results and Conclusions: The main advantage of this model is that it enables extensions of the classification by recognizing that some plots in the current classification are best left unassigned until enough data are available to robustly define a vegetation type. The system now includes 29 alliances and 79 associations of forest/shrubland vegetation, and 25 alliances and 56 associations of non-forested vegetation. Although recent, the classification system has been adopted for both basic and applied research, the latter often contracted by land management and policy agencies. Abbreviation: CCS = consistent classification section; DOC = Department of Conservation; NVS = National Vegetation Survey databank. Submitted: 11 December 2016; first decision: 9 April 2017; accepted: 11 May 2017
Keywords
classification extension • classification system • compositional outliers • consistent classification section • fuzzy clustering • National Vegetation Survey databank • noise clustering • semi-supervised classification • vegetation type • transitional plots