Predicting the climate

Climate-change scientists have forecast a gloomy outlook for rainfall and Perth metropolitan water supplies in the 21st century. A reduction in annual rainfall of up to 20 per cent by 2030 is predicted in the south-west of Australia, compared to a maximum reduction of 10 per cent in the eastern states; and Perth urban water supplies face greater insecurity (Pittock 2003). In 1998 the WA government set up the Indian Ocean Climate Initiative (IOCI), a collaboration between the CSIRO, the Bureau of Meteorology and Western Australian government agencies to assess climate-change issues facing south Western Australia, and provide interpretation of, and judgment on, the uncertainties surrounding climate-change science. This group has developed models to downscale the results of global climate-change models to the weather patterns of the south-west of Western Australia and to forecast the impact on surface-water inflows into metropolitan dams. The downscaled rainfall models have been shown to fit well to the post-1975 rainfall sequence, during which there has been an increase in the frequency of the driest winter-weather patterns and a reduction in the frequency of various wetter-winter patterns (Charles, Bates and Hughes 1999; and Steve Charles July 2007, personal communication). An 11 per cent reduction in mean rainfall is predicted for the region by the middle of the 21st century compared to the recent (post-1975) average. Using a rainfall-runoff model of Stirling Dam, researchers have predicted that the impact of an 11 per cent reduction in rainfall will be a 31 per cent reduction in streamflows in the period 2042–62 compared to the period 1982–2002. However, there is some uncertainty in the predicted outcome because of uncertainty regarding the impact of climate change on evapotranspiration. If potential evaporation increases by 10 per cent under a warmer climate regime, then the predicted streamflow reduction is 41% (Berti et al. 2004; Bari et al. 2005). These predicted changes represent expectations 50 years from now.

In its source-development planning, the Water Corporation has adopted the use of a recent historical sequence, which was initiated in a major planning document (Water Corporation 2005) when all system yields were expressed for two climate-change scenarios; one being the post-1975 sequence and the other a more pessimistic scenario represented by the post-1997 sequence (Water Corporation 2005). In response to the extreme weather conditions experienced in 2006, which saw the lowest winter inflows on record, the Water Corporation has revised its estimates of system yield for planning purposes and now uses the climate average since 2001, including 2006. Inflows in the winter of 2000 were among the largest in recent history, and if the ruler-and-pen approach included the year 2000, the mean system yield would be 28% higher than for the ‘2001–06’ scenario. In fact, the mean system yield for the 2000–06 scenario is similar to the mean system yield for 2050 predicted by the IOCI models.

In the analysis provided in this paper, four system yield scenarios are presented. These are, in order of decreasing system yield, the 1975–2006 climate sequence, the 2000–06 sequence, the IOCI worst-case scenario, and the Water Corporation’s current planning yield, based on the 2001–06 climate sequence.