The Australian Bureau of Meteorology statistical downscaling model (BoM-SDM) is a tool to deliver high-resolution daily climate grids for essential surface variables. Over the years of development, in the case of temperature, the method has proven its ability to reproduce the observed mean values with good accuracy but has shown a tendency to underestimate the observed daily and seasonal variability. A novel approach is being applied, where the BoM-SDM has been modified so that daily analogues are allowed to be searched outside the calendar season. This modification provides a larger pool of analogues to search from, and a higher likelihood of finding an analogue for the more extreme days within a season by linking them to days observed in an adjacent season. The approach is tested with the existing optimisation of the BoM-SDM and is currently providing downscaled products (temperature and rainfall) at 5 km spatial resolution over Australia from 22 global coupled climate models from phase 5 of the Coupled Model Intercomparison Project (CMIP5). It is found that this new version is able to successfully reproduce the daily variance, daily extremes and magnitude of future warming without any inflation factor. However, we are left to deal with the warmest season and the inability to produce new record temperature outside the range of past observation. This is what is being proposed here, a new form of post-processing which uses information drawn from the rest of the PDF, and how it is modified, to infer future record high temperatures.