Living in a variable climate

Integrative commentary
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

Historical variability in rainfall

As described above, in eastern Australia, the interaction of the La Niña phase of ENSO and the 'cool' phase of the Pacific Decadal Oscillation or Inter-decadal Pacific Oscillation (Power et al.1999, Mantua and Hare 2002) has resulted in sequences of wet years (early 1890s, 1916–18, early 1920s, mid-1950s, early 1970s and late 1990s). These wet periods were likely to have led to biased expectations of livestock carrying capacity, agricultural production and water availability, leading to government and community support for inappropriate land use (such as cropping of marginal areas and small grazing property sizes; Heathcote 1965, Russell 1988, Condon 2002). Dry conditions in various rural regions followed these wet sequences (drought periods 1896–1902, 1919–20, 1926–31, mid-1960s, early 1980s, 2001–present). The rapidity of decline in rainfall, and the large contrast between the drier and wetter rainfall periods, demonstrated the high variability of the climate. In several of these drought episodes, there have been major re-examinations of appropriate land use as well as the public call for major infrastructure development to 'drought-proof' regions.

From the viewpoint of understanding longer term climate change or 'regime shifts' in the natural climate system, the instrumental record for most locations in Australia is relatively short (about 100 years). Importantly, where historical records commenced after the Federation drought (ending after 1902), they do not include the extreme variations of the wet, early 1890s and the later severe drought (Gibbs and Maher 1967). For some locations, such as north-east Queensland, proxy climatic information can be reconstructed from coral records (Lough 2003). The reconstructed rainfall record suggests that the driest and wettest years in the past 230 years (1754 to 1985) have occurred in the twentieth century (1902 and 1974, respectively). However, the driest ten- and 30-year periods occurred at the end of the eighteenth century (1766–75 and 1770–99 respectively) prior to the period of northern settlement and instrumental record (post-1870). The wettest ten- and 30-year periods occurred in the last 50 years (1972–81 and 1950–79, respectively).  Thus climatic extremes of rainfall deficit that may be outside the community's experience would appear to have occurred in the natural climate system. These extremes will undoubtedly challenge the capacity for managing climate risk should they occur again.

The direction and magnitude of rainfall changes or trends for the last one hundred years over the continent vary regionally, with increases in north-western Australia and decreases in south-western Australia and coastal Queensland. Given the influence of decadal and multi-decadal variability in the climate system on Australian rainfall , the start and finish year of analysis also affect the reported magnitude and direction of the trends (1900–2005 compared with 1970 to 2005 in eastern Australia; see Figure 1 and Figure 2). The choice of period for analysis has been determined by either scientific need (accuracy of records, behaviour of climate system) or practical application (recent resource management issues).

Figure 1: Annual rainfall for Australia, 1900–2005

 Annual rainfall for Australia, 1900–2005

Source: Bureau of Meteorology, January 2006.

Figure 2: Trends in Annual Rainfall 1900-2005 and 1970-2005.

 Trends in Annual Rainfall 1900-2005 and 1970-2005.

Source:  Bureau of Meteorology, June 2006.

The community, through its funding of research, expects (or hopes) that climate science will become more confident in the future projections of climate, integrating: (1) multi-decadal variability; (2) human-induced climate 'forcings'; and (3) associated predictions of future development (such as energy use) and mitigation responses. The current emphasis on the use of historical rainfall variability for planning could then evolve to include the comparison of current climate trends (last 30 years) with the expected trajectory for the future (typically 30 to 70 years). Management and infrastructure changes can then be made in terms of considering not only necessary responses to current trends and conditions (for example, multi-year drought) but also future expectations of climate change (for example, lower or higher average rainfall).