When projecting impacts of climate change on future water resources (e.g. rainfall, streamflow) assumptions need to be made about how and what historical statistics should change. For example, to create a precipitation sequence that is representative of a “future climate” currently accepted methodology obtains percentage changes in the mean, as compared with “current climate”, from a climate model or ensemble of climate models and applies these “change factors” to the mean of the historical data. As a result the historical standard deviation is altered by the same percentage as the mean, and the coefficient of variation is maintained – but is this the right thing to do? The standard deviation could be changed in various ways, including, by the same proportion as the mean (as described above), not change at all or change by some other amount (e.g. different changes for high and low frequency cycles). Other statistical properties such as range and skew could also change. There is some suggestion that more extreme wet and dry events may occur which could potentially have the same mean but alter the skew and range of future precipitation sequences. This paper will discuss, via the presentation of results from a recent case study in the Canberra/Murrumbidgee area, the statistical properties currently used to project future climatic or hydrologic time series, which statistics might change in the future and the challenges associated with applying changes to historical statistics to realistically simulate projected impacts of climate change on future water resources.