New national climate projections for Australia: distinguishing knowledge and data (Penny Whetton, CSIRO)

Over the past twenty-five years, CSIRO has led the provision of national climate change projections for Australia designed to serve the needs of adaptation planning. Statements were released in 1992, 1996, 2001, and (with) in 2007. These projections have been widely cited and used in Australian adaptation work.

Early in 2015, CSIRO and the Bureau of Meteorology released new national projections for Australia based on extensive analysis of the CMIP5 climate model ensemble and some additional downscaling. The projections were aimed specifically at supporting the needs of natural resource management planning, as the work was part of a larger government initiative in this sector.  However they were also designed to serve wider national needs in line with earlier projection products. The projections cover a range of variables, such as temperature, precipitation, sea level rise, humidity, soil moisture and radiation, including some aspects of extremes, such as hot days, fire weather, extreme rainfall and extreme sea level. Communication of the projections to users is through a series of brochures and reports, and through a website (http://www.climatechangeinaustralia.gov.au) which includes range of interactive tools for exploring the projections, data download facility and guidance.

Development of these projections entailed extensive consultation with natural resource management planners and associated impacts and adaptation researchers to ensure that the projections meet their needs and expectations. To serve the needs of some users, the projections provide key messages on regional climate change and our confidence in them, whereas supporting technical projection data sets were also been prepared in various forms for those users who need information of this nature. Additionally, to assist users in selecting a small set of models representative of the range of plausible change relevant to their decision making context, a software tool (‘Climate Futures’) has been developed.

An important motivation on our part was to make the projections as scientifically robust as possible, by integrating different lines of evidence into climate projection development, including understanding the processes driving regional projected changes and assessing confidence in projections. This effectively strengthened the ‘assessment’ component of the projections compared to previous products, and increased its similarity to the IPCC process. Furthermore, choosing methods of communication, we avoided presenting any detail which we felt was unlikely be scientifically robust.  In particular, this meant that ranges of change averaged over regions was the key communication device rather than maps based on gridded data (as had been used in the previous projection product), and that qualitative guidance only was provided when our confidence in the projections was very low.

As well as the above issues, the presentation will also describe a key concept underpinning the projections and their use which involves distinguishing between two types of projection information. The first of these is scientific knowledge about future climate change and the second is projection data sets tailored for use in technical risk assessments. These two products serve different purposes, and a key challenge for risk assessment is for projection providers and users to ensure that the two aspects of knowledge and data are as harmonised as possible.

Dr Whetton is an Honorary Research Fellow with CSIRO's Oceans and Atmosphere Flagship, and formerly a Senior Principal Research Scientist with CSIRO.  Dr Whetton led CSIRO's national climate change projection work from 1992 until 2014.