Mapped Vegetation class extents from the New South Wales State Vegetation Type Map (SVTM) were used to classify BIAP’s baseline GDM into vegetation classes by applying the most similar class to each pixel in NSW.

The BIAP baseline GDM’s transformed environmental predictors were first reprojected to the GDA94 Australian Albers coordinate system (ESPG:3577) and bilinearly resampled to 90 m from their original 0.0025DD (~250 m) pixel resolution. An equally spaced sampling grid was intersected with SVTM mapped class extents assigning vegetation classes to 101,224 samples (~1:740 sampling rate) across the state. The standard GDM link function (Ferrier et al. 2007) was then applied using the GDM transformed environmental predictors to estimate every pixel’s original floristic compositional similarity to each class sample.

The derived layer shows the BIAP’s baseline GDM classified into vegetation classes by applying the class with the most compositionally similar sample from its current extent to each pixel. Vegetation classes are coloured to reflect their compositional similarity so that compositionally similar classes (those with more shared species) are given similar colours.

Candidate Native Grasslands and Riverine Plain Grasslands were not present in the class similarity matrix so were not assigned a colour (Grey). Alpine fjaeldmarks were not sampled therefore were not allocated to any pixels. The next most compositionally similar classes fill areas where this class occurs.

The RGB colour values allocated to each class are stored in the raster attribute table.

Ferrier S, Manion G, Elith J and Richardson K (2007) ‘Using generalized dissimilarity modelling to analyse and predict patterns of beta diversity in regional biodiversity assessment’, Diversity and Distributions, 13(3):252–264.