Living in a variable climate
Dr Greg McKeon, CRC for Greenhouse Accounting, Queensland Department of Natural Resources and Water
prepared for the 2006 Australian State of the Environment Committee, 2006
Simulations by global climate models – of future temperature changes due to increasing greenhouse gas concentrations—suggest a rise of 0.4 to 2.0°C by 2030 across much of Australia (CSIRO 2001, Lindesay 2003). By 2070, temperatures are expected to be between 1.0 and 6.0°C higher than in 1990 (Figure 4). The largest increases are projected to occur in summer. Projected changes in regional rainfall are spatially and seasonally variable. Under enhanced greenhouse conditions, most of the climate models that have been used so far have simulated decreased rainfall in southern and eastern Australia in winter and spring (CSIRO 2001, Whetton and Suppiah 2003). Increased carbon dioxide concentrations in the atmosphere, and any further declines in potential evaporation, would tend to reduce the impact of any reduction in rainfall on plant growth (Howden2002).
Source: CSIRO (2001)
Even though some of the projected changes in rainfall ranges appear small (plus or minus ten per cent), it is important to recognise that these changes are of the same magnitude as those that have occurred over historical 30-year periods. For example, in the Murray Darling Basin averaged 30-year rainfall was eight per cent lower for 1917 to 1946, but was ten per cent higher for the subsequent period 1947 to 1976. Given the difficulties that agricultural land use experienced in this region during the drier 30-year period and the beneficial impact of the subsequent wetter period (Condon 2002, McKeon et al. 2004), small projected changes of ten per cent in average rainfall in some regions should not be dismissed as unimportant for land use or biodiversity.
A major uncertainty in projections of future climate change for Australia has been the direction of future rainfall trends. This uncertainty relates in part to the uncertainty in the response of ENSO to global warming. Whetton and Suppiah (2003, p. 134) commented on likely future changes in ENSO as a result of the enhanced greenhouse effect, based on the review of the Intergovernmental Panel on Climate Change (IPCC 2001). Whetton and Suppiah (2003, p. 134) noted a 'tendency for most models to show a pattern of change in the average sea surface temperature in the Pacific that was 'El Niño-like' (meaning a greater warming of the central and eastern Pacific Ocean). They also found that simulated pressure anomalies near Australia did not follow those that are usually associated with El Niño events (higher pressures over eastern Australia). They concluded, 'this result strongly suggests that caution should be applied when translating apparent model simulated changes in ENSO to rainfall changes in the Australian region' (p. 136). Thus simplistic statements that 'more El Niños under climate change will mean more droughts' may correctly alert the public to future risks associated with climate change, but may not necessarily be scientifically sound.
More recent examination of the simulation of ENSO by 20 atmosphere-ocean global climate models has indicated 'that those models that have the largest ENSO-like climate change also have the poorest simulation of ENSO variability' (Collins 2005, p. 89). They concluded that 'the most likely scenario … is for no trend towards either mean El Niño-like or La Niña-like conditions'. They estimated a small chance (16 per cent) of a change to El Niño-like conditions under a climate change regime that resulted from a one per cent increase per year in carbon dioxide. The apparent variation between global climate models in forecasting ENSO behaviour under climate change, and the rapidity of improving understanding that is occurring in climate science, highlights the real difficulty that science and the community are having in understanding the likely mechanisms of future climate change impacts.
Another major uncertainty in future rainfall projections is the role of human-induced forcings other than increasing greenhouse gas concentrations (such as, stratospheric ozone depletion and aerosols). The Intergovernmental Panel on Climate Change in 1990 (Houghton et al. 1990, p. 7) indicated that scientific understanding of the climate system was still at an early stage and that 'the complexity of the [climate] system means that we cannot rule out surprises'. One such 'surprise' has been the emerging hypothesis that Antarctic stratospheric ozone depletion has been affecting atmospheric circulation in the Southern Hemisphere and rainfall over Australia (Pittock 2003, pp. 43, 51, Watkins 2005, Cai 2006). A current challenge for climate science in constructing climate change scenarios (such as for 2030 and 2070) is, therefore, how to include the interactive effects of stratospheric ozone depletion, anthropogenic aerosols and increasing greenhouse gas concentrations on atmospheric circulation, especially in the Southern Hemisphere, and on Australia's climate. As indicated in Beer (2006), the decline in stratospheric ozone has stabilised, although the time required for recovery is uncertain.
Scientific understanding and model representations of important physical, chemical and biological processes in global climate models have greatly advanced over the last 30 years. For example, global climate models have been extensively used for climate change projections for more than 30 years (for example, Manabe and Wetherald 1975) and have proven accurate in their general predictions of the late twentieth-century temperature rise. Uncertainty remains at regional scales, and particularly for variables such as rainfall. Despite this achievement of climate science, currently some regional climate changes would appear to be occurring faster than climate science can understand and predict them. The public can therefore be forgiven for being confused about the cause of current regional rainfall deficits, which have been variously ascribed by scientists in the media to a range of natural and human-induced effects. Nevertheless, global climate models continue to be perceived as the best hope to address this uncertainty. The projections of future rainfall and temperature are derived from simulations conducted with global climate models that are also currently being used in seasonal climate forecasting (for example, Alves et al. 2003, Syktus et al. 2003). The continuing assessment of the accuracy of global climate models, and their operational use in seasonal and annual decision-making, will help the community develop confidence in climate change predictions that are derived from these global climate models.