William J. Wright, National Climate Centre, Bureau of Meteorology, GPO Box 1289K, Melbourne 3001. e-mail: w.wright@bom.gov.au Phone: (03) 9669 4781 Fax: (03) 9669 4678

This talk briefly reviews how Australia’s climate has varied over the 20^th Century, with emphasis on the longer-term variations in rainfall and temperature. It is intended as a scene-setter for the deliberations to follow. Some of the impacts - on agriculture, land management, etc - are also described. The point is made that the significant magnitude of some of these past variations, along with the prospect of future climate changes, have important implications for the management of climate-sensitive enterprises and resources in the future.

Australia is a country with a generally low rainfall, which in many areas is subject to high variability. For this reason climate sensitive industries, most obviously agriculture, are highly susceptible to climate fluctuations, and these fluctuations have, over the past century, exerted a big influence on our activities.

The fluctuations can be identified on at least three different time-scales. The strong interannual variations in rainfall associated with the ENSO phenomenon are well known. These are felt most strongly and regularly over the eastern half of Australia, and contribute greatly to Australia’s reputation as a land of "droughts and flooding rains". Secondly, it has been increasingly recognised in recent years that climate varies on decadal and multi-decadal time-scales. For instance, the mid to late 1920s and the 1930s were a period of generally low rainfall over most of Australia. In such a period a particular location will not experience dry conditions every year - but rainfall will be frequently below the long-term average, and there are often runs of years with recurrent drought. These decadal and multi-decadal variations have amounted to as much as a 40% difference in seasonal rainfall amounts between successive 30 year periods - large enough to significantly affect the practise, even the viability, of agriculture, and other industries, particularly in climatically-marginal areas.

Some of these fluctuations will be Anatural@ cycles; others may be symptoms of long-term climate trends, such as global warming. Such long-term trends form the third category. They constitute what most people think of as "climate change", and form a backdrop against which shorter-period fluctuations occur. The question is: are such trends the consequence of man’s inadvertent influence on the atmosphere, and therefore more or less here to stay, or are they yet another, longer-term, cyclical variation? And if our climate is indeed being anthropogenically-forced – what interaction is there with the shorter-period variations?

Despite the often highly-localised character of rainfall, major fluctuations in rainfall have occurred on a continent-wide scale. Fig. 1 shows rainfall averaged over the entire country during the 20^th Century. The main fluctuations are now described.

Fig. 1: Continental averaged rainfall over the 20^th Century, showing century average (black horizontal line), year to year fluctuations (light line), 5 year running mean (binomially weighted – solid line), and trend (dashed line).

The earliest years of the 20^th Century saw the culmination of the so-called "Federation Drought", which began over eastern Australia in 1895, and reached its nadir in the disastrous year of 1902, possibly Australia’s overall driest year of the 20^th Century. Rivers in western Queensland dried up; the Darling River virtually ran dry at Bourke. Further south, towns near the Murray River such as Mildura, Balranald and Deniliquin - at that time dependent on the river for transport - suffered badly. The Australian wheat crop was all but lost. This drought decimated stock numbers in Australia - sheep numbers fell from 91 million in the years before the drought to 54 million, and cattle from 11.8 million to 7 million.

The next 15 years or so featured an extended El Nino period from 1911-15, associated with generally dry conditions, culminating in the severe drought of 1914. In that year, wheat crops failed in both eastern and western Australia, a rather rare occurrence. The years 1916 and 1917 were strong La Nina years, with heavy rainfall and flooding, and outbreaks of a mysterious and often fatal disease, later to become known as "Murray Valley Encephalitis".

The period 1918-1920 was also dry, and followed soon after by a very long, dry period over most of the country, extending from the mid 1920s to the mid-1940s. Many serious droughts took place in this time, notably those of the mid-to-late 1920s over eastern Australia, and between 1937-45. The latter period was noteworthy for its frequent dust-storms, which on occasion coated the major capital cities near the coast with dust thick enough to require house- and street-lighting to be turned on. Such episodes represented devastating losses of topsoil from the country=s agricultural areas, and promoted a serious re-think about land management practices in Australia.

The period was also noteworthy for devastating bushfires in southeastern Australia, notable in January 1939 and the 1943/44 season. The fires between January 8^th and 13^th in Victoria, which followed an intense drought, killed 71 people. The ferocity of the fires was enhanced by the century’s most extreme temperatures in southeastern Australia, which claimed the lives of over 400 Victorians in one week from heat-related causes.

A very wet period set in over the eastern half of the country in the 1950s. This period was noteworthy for its frequent and widespread flooding, especially in 1950, 1952, and 1954-56. The wet 1950s in fact marked the start of an apparent decadal periodicity in rainfall over the eastern States, with the following 1960s being generally dry (with major and widespread drought between 1965-68).

The 1970s (especially the early 1970s) were exceptionally wet across the whole country - easily the wettest period of the Century, with widespread flooding, and outbreaks of diseases to crops and humans. January 1974 stood out, with cyclone "Wanda" flooding Brisbane, and huge areas of the interior becoming inland seas, as some 400-600mm of rain inundated the southern Northern Territory and southwest Queensland – about twice the average ANNUAL rainfall in one month. Lake Eyre filled in the autumn of 1974, and retained water well into 1977. At the end of 1974, cyclone "Tracy" essentially destroyed Darwin.

Rainfall over eastern Australia declined again after the late 1970s, though not to the levels seen in the earlier part of the Century. This effect was felt most strongly in Queensland, and may be related to the increased frequency of El Nino events since that time (although the years 1998 and 1999 seem to have reversed that trend on both counts). The second half of the 20^th century was somewhat wetter than the first half over the eastern States, though there is evidence that this is not unique, and may in fact represent a return to conditions that prevailed in the second half of the 19^th century.

These continent-scale rainfall variations may represent nothing more than natural variability in the climate system, and have been linked to the so-called Interdecadal Pacific Oscillation. Multi-decadal variations have also been documented over New Zealand, associated with apparent changes in broadscale circulation patterns.

Superimposed on these continental-scale patterns are some notable regional variations:

• Northwestern Australia experienced relatively low rainfall from the 1930s to the 1960s, then an extended wet period from the 1970s onward, apart from a brief decline in the later 1980s. The 1990s in general have been exceptionally wet. Rainfall over the arid inland areas of Western Australia has also been consistently above the long-term normal during the 1990s, at least in part reflecting an unusually high frequency of tropical cyclones tracking inland over the area.

• Central Australia had a severe 10 year drought between the 1950s - late 1960s.

• Rainfall over southwestern Australia was, in the early part of the century, well above what it is now, and flooding was quite frequent. Rainfall began to decline about 50 years ago, a trend that intensified after the late 1960s. This change was accompanied by clear and well-documented shifts in atmospheric circulation patterns. In particular, there has been a general increase in sea level pressure over western Australia and the southern Indian Ocean, and a poleward and westward contraction of the belt of winter cold frontal and cyclone activity responsible for most of southwestern Australia=s rainfall. The decline since the late 1960s has amounted to some 15% of the average annual total over the wetter areas of the southwest, and poses a planning dilemma for water authorities in the area.

In contrast to the rainfall variations described above, trends in temperature appear to be longer-lasting. This has led to the widespread suspicion that the temperature changes represent human-induced climate change. Following a brief cooling trend in the 1940s-early 1950s, there has been a sustained increase in continent-wide temperatures since 1950. The rises have been greatest in overnight minimum temperatures, amounting to an increase of 0.85C between 1950 and 1996. Smaller increases have occurred in the daytime (0.39C). The temperature rise has steepened since 1980, with the 1990s being the warmest decade since records commenced in 1910, with the 1980s the second warmest. The number of extremes have changed, too, with very warm days more frequent, and cold nights less frequent.

The rise in minimum temperatures has been greatest over the northeastern interior, and has been reflected in a marked decrease in frost frequency in this area since the 1970s. The average date of the last frost has also become earlier, so that a longer period of the spring growing season is frost-free. There is, as yet, no evidence of a significant decline in snowfall over the higher areas of the Australian Alps, but there is some sense that lower level resorts – which would be more sensitive to temperature changes – are doing it tougher.

5. Discussion and concluding remarks.

We enter the 21^st Century knowing that climate has changed considerably over time-periods roughly corresponding to generations. We guess that some of these changes are part of the natural variability of the climate cycle and will eventually reverse - there may or may not be comfort in that. But some of the changes may be linked to global warming, which is ever-more widely believed to be due to increases in so-called Greenhouse gases. Science does not yet allow us to distinguish well the different time-scales. For instance, southern Victoria has recently endured a four year drought, characterised by repeated shortfalls in the critical winter-spring rainfall, and associated with obvious changes in the behaviour of subtropical high pressure systems in our area. We are unable at this stage to say whether this is a short-term fluctuation, or part of a multi-decadal decline similar to what occurred in southwestern Australia, or something more permanent. I believe that an important part of future climate research will be to address such questions. The challenge for enterprise managers meanwhile will be to work together with the climate scientists to find ways of managing within this envelope of uncertainty.