The often espoused view that Australia's arid zone experiences more variable rainfall over time than areas of similar climates elsewhere was tested using annual rainfall data from 407 localities spread throughout 14 arid areas, including 68 Australian sites. This represents an updated and more comprehensive analysis than previously completed. Four measures of inter-annual rainfall variability were calculated. Two of these are commonly used in the literature: the variability index (spread of the 90th and 10th percentiles divided by the median rainfall) and the coefficient of variability (standard deviation as a percentage of the mean). In addition, two simpler measures not divided by the average annual rainfall were used: the difference between the 90th and 10th percentiles, and the standard deviation of the mean. Linear relationships between rainfall variability and average annual rainfall enabled analysis of covariance (ANCOVA) to be employed, with appropriate post hoc testing of adjusted means, to determine differences between regions in terms of the various measures of rainfall variability. ANCOVA consistently revealed that four regions (Thar, Namib-Kalahari, Somali and northern Australian deserts) were significantly more variable than all others. Southern Australia (defined as south of 27°S latitude) was grouped with these other desert regions, which include the North American deserts, Sahel, northern Sahara and Karoo. Generally the two groups were divided into low latitude, summer rainfall regions and higher latitude, winter rainfall ones, although there were some clear anomalies. The findings suggest that comparative studies of biotic responses to rainfall variability between arid Australia (particular southern regions) and other deserts are warranted before paradigms and models based on the uniqueness of Australia's arid environment are accepted.

SUMMARY: The ngwayir (western ringtail possumPseudocheirus occidentalis) is an arboreal species endemic to south-western Australia. The range and population of this species have been significantly reduced through multiple anthropogenic impacts. Classified as vulnerable, the ngwayir is highly susceptible to extremes of temperature and reduced water intake. Ngwayir distribution was determined using three different species distribution models using ngwayir presence records related to a set of 19 bioclimatic variables derived from historical climate data, overlaid with 2050 climate change scenarios. MaxEnt was used to identify core habitat and demonstrate how this habitat may be impacted. A supplementary modelling exercise was also conducted to ascertain potential impacts on the tree species that are core habitat for ngwayir. All models predicted a reduction of up to 60% in the range of the ngwayir and its habitat, as a result of global warming towards the south-west of the project area, with a mean potential distribution of 10.3% of the total modelled area of 561 059 km². All three tree species modelled (jarrah, marri and peppermint) were predicted to experience similar contractions in range throughout most of the predicted ngwayir range, although their distributions differed. Populations of ngwayir persisting outside core habitat may indicate potential conservation opportunities.