Compositional dissimilarity data for the National Climate Risk Assessment

Compositional dissimilarity, or the potential degree of ecological change, is the proportion of species at a given location expected to be absent from the same location, and potentially replaced by other species, under a changed climate scenario. Compositional dissimilarity ranges from 0 (no change in species composition) to 1 (100 % change in species composition, Ferrier et al., 2007).
The data presented here are pixel-based dissimilarities in species composition between the baseline scenario, Australian climate in 1990, and two future climate change scenarios, ACCESS 1.0, and GDFL-ESM2M, at a 2050-centred climate under RCP (Repressentative Concentration Pathway) 8.5 (Harwood et al., 2014). Compositional dissimilarity is estimated in the absence of the effect of land use change by assuming the land is in highest condition, or “reference” condition. As such the expected change in species composition is based on climate change alone.
A model of compositional dissimilarity for vascular plants was fitted using generalized dissimilarity modelling (GDM) (Ferrier et al., 2007) based on 1990-centred species records, climate, and other environmental data (Mokany et al. 2018). Compositional dissimilarity was estimated between 1990 and each of the two climate scenarios, and summarised as the arithmetic mean of the estimated compositional dissimilarity for both climate scenarios (ACCESS 1.0 and GDFL-ESM2M).
For the National Climate Risk Assessment (NCRA) project, the mean model outputs were summarised (arithmetic mean) by various regions (Aggregate Ecosystem Groups (AEGs), and NCRA regions), available in the links below. The AEG and NCRA region data were reprojected to the projection of the dissimilarity raster, and arithmetic mean dissimilarity was calculated per region of the interaction between AEG and NCRA region.