Indicator: Weather

Analysis of rainfall data from Hall village and Canberra Airport

Rainfall statistics

The ACT was generally drier than average in the period 2000–02. Rainfall at Canberra was slightly above average (+0.5%), and that at Hall just below average (-3.4%), in 2000 (see table below). Both places recorded annual rainfall totals between 70% and 80% of normal in 2001 and 2002. Conditions at Hall were particularly dry, with rainfall 27.9% and 28.4% below average in 2001 and 2002 respectively.

Table 1: Annual rainfall statistics in the ACT, 2000–02
Rainfall Statistics ___________Hall___________ _________Canberra_________
2000 2001 2002 2000 2001 2002
Annual Total (mm) 626 500 505 694 518 514
Anomaly (mm) +3.2 -123.2 -117.8 -24.6 -200.4 -204.2
Anomaly (%) +0.5 -19.8 -18.9 -3.4 -27.9 -28.4
Mean (mm) ____________623____________ ____________718____________

ACT: Monthly rainfall 2000-03

Rainfall patterns at Hall and Canberra were similar during the reporting period (Figure 1). Most striking were the periods of below-average rainfall which lasted for several months at a time, particularly during 2002–03;. Also significant was the extremely wet month of February 2002. The prolonged period of below-average rainfall in 2002–03 is associated with the El Niño event of that time (coinciding with a negative phase of the Pacific Decadal Oscillation).

Of the 42 months in this period, 28 months (two-thirds of the record) had below-average rainfall at Canberra, and 23 months were drier than average at Hall. The driest spell was in the spring and early summer of 2002–03, when in the 11 months from March 2002 to January 2003 only two had slightly above-average rainfall; the rest were significantly drier than normal (Figure 2 and Figure 3. The driest individual month at both Canberra and Hall was May 2001, with 2.6 mm (or less than 6% of the average monthly rainfall) at Canberra and 2.2 mm (less than 4% of the average) at Hall.

Conditions were slightly wetter than normal in five months during 2000, and in winter 2001, at both places. The wettest month throughout the ACT region (and across much of south-eastern Australia) was February 2002, when near-record totals of 211 mm and 196 mm (more than 250% above average) were received at Canberra and Hall respectively.

Figure 1. Monthly rainfall totals at Canberra and Hall, January 2000 to June 2003

Graph of monthly rainfall data from Hall village and Canberra Airport [Credit: Australian National University]

Figure 2. Canberra monthly rainfall totals (blue bars) compared with the long-term monthly mean rainfall shown in yellow (all in mm), January 2000 to June 2003

Graph of rainfall data from Canberra Airport [Credit: Australian National University]

Figure 3. Hall monthly rainfall totals (blue bars) compared with the long-term monthly mean rainfall shown in yellow (all in mm), January 2000 to June 2003

Graph of rainfall data from Hall village [Credit: Australian National University]

The result of the generally dry conditions in the ACT during most of the 2000 03 period is reflected in the accumulated rainfall deficits at Canberra Airport and at Hall (shown in Figure 4). At both places accumulated monthly rainfall totals were close to normal by March 2001; from then on monthly rainfall totals lagged steadily behind the long-term average, reaching an accumulated deficit of 329 mm (more than half the long-term mean annual rainfall) at Canberra and 505 mm (70% of mean annual rainfall) at Hall by the end of June 2003. Even the extremely high rainfall in February 2002 did not significantly affect this pattern.

Figure 4. Accumulated monthly rainfall totals (blue) compared with accumulated long-term monthly mean rainfall (yellow) (all in mm), January 2000 to June 2003 at Canberra (left) and Hall (right). The vertical bar to the right of the graph represents the annual average rainfall total of 623 mm.

Graph of cumulative rainfall data from Hall village and Canberra Airport [Credit: Australian National University]

The climatological context for rainfall in the ACT

The long-term perspective: Canberra

As is typical of the region, the ACT has experienced large fluctuations in annual rainfall, including several extended periods of above- and below-average rainfall since records began (1940 at Canberra and 1904 at Hall). In general, annual rainfall was lower and less variable early in the twentieth century, followed by a period of greater variability. Rainfall variability has been lower since about 1980 (Figures 5 and 6).

Figure 5. Annual rainfall totals at Hall from 1904 to 2002 (in mm), with the long-term mean rainfall shown as the horizontal line for comparison of individual annual totals

graph of rainfall at Hall village 1904-2002 [Credit: Australian National University]

Figure 6. Annual rainfall at Hall from 1904 to 2002 shown as deviations (in mm) from the long-term mean. A five-year running mean is superimposed to highlight wetter and drier periods. Deviations were calculated as the difference between the rainfall for each year and the mean for the period 1961–1990.

graph of deviations from mean in rainfall at Hall Village 2000-03 [Credit: Australian National University]

  • Rainfall in the first half of the 1900s was generally below average (to about 1950).
  • Several wetter-than-normal years around 1950 and 1960 produced a generally wetter period then; the same is true of the mid 1970s.
  • A generally drier period around 1980 was followed by slightly above-average rainfall, with the notable exception of the dry El Niño years of 1994 and 1997. Since 2000 conditions have been drier than average.

The wettest year on record was 1950 at both Hall (1368 mm) and Canberra (1063 mm); the driest year at Canberra was the El Niño year of 1982 (262 mm), and 1944 (273 mm) at Hall.

The seasonal view

The seasonal view of rainfall variability at Hall gives a longer-term perspective than is possible from the shorter Canberra record. Rainfall at Hall has undergone some quite marked seasonal variations. The driest part of the year has shifted from summer-autumn (February-April) in the early part of the record, to winter (June-July) in the 1960s and 1970s, and back to February, although June has remained drier. The result is that June rainfall has declined by about 30 mm since the start of the record. Rainfall at Canberra follows a similar pattern for the period of overlap in the records (the Canberra rainfall record begins in 1940). These types of shifts in rainfall patterns can have important consequences for agriculture, particularly for farming systems that are reliant on the 'autumn break'.

On average the wettest month is October and the driest are February (at Hall) and June-July (at both stations), with a range of over 30 mm.

Figure 7. Seasonal variation in monthly rainfall at Canberra (left) and Hall (right)

graph of seasonal rainfall over time in the ACT 2000-03 [Credit: Australian National University]

A smoothed view of the variation of monthly rainfall (y-axis: January at bottom, December at top) through time (x-axis). The colour scale indicates the intensity of the rainfall anomaly (i.e. the difference between the individual month's rainfall and the long-term average for all months in the data set), ranging from white and brown (very dry), through yellow (slightly dry), to green (slightly wet) and dark blue (very wet). Contours are at 10 mm intervals.

View a larger version of these plots, with a fuller explanation of how they were developed.

Other important points are:

  • The wettest part of the year was June–July (winter) prior to about 1920; between 1940 and 1980 the wettest months were October (spring), and then August–October.
  • Prior to 1920 the seasonal cycle was well defined, with the dry season centred on February and the wet season on June. Between the 1920s and 1940s seasonality was less well defined, as February and the autumn months became wetter and June became drier.
  • From about 1950 a different pattern of seasonality began to develop, and by the 1960s the wettest month was in spring (October) and the driest in winter (June).
  • Since 1980 at both Hall and Canberra there has been no clearly defined peak rainfall season. At Hall, January and July November all have slightly wetter signatures than the rest of the year. February is re-emerging as the driest month of the year at Hall, as it was early in the 20th century.
  • At Canberra the pattern is different, with the wetter season starting in September and extending to March (with a maximum in November), and the drier season from April to August. There is a trend in the Canberra record for drier autumn conditions to begin earlier (the change from wetter to drier seasonal conditions began in May in 1940, and in March by 2000, a shift of two months). At the same time the drier season now ends one month earlier, in August rather than September.

Underlying trends

The underlying trend in the record of rainfall at Hall has been for a step-wise rise to the 1950s, followed by a decline through until the early 1980s. Interannual variability, as shown by the amplitudes of the medium and shorter-term components, was lower early in the record, then increased through the 1940s and remained high until about 1980. The dry El Niño year of 1982 shows a distinct signal in both medium and shorter-term components, and came at a time when the underlying trend was also relatively low.

Figure 8. Low-frequency fluctuations in monthly rainfall: Hall

graph of monthly long term rainfall 2000-03 [Credit: Australian National University]

The graph shows the underlying characteristics or trends in the monthly rainfall (y-axis, in mm) over time (x-axis). Three different smoothers have been used: a low-frequency one in black (about 40 years), that shows the lowest variability; a medium-frequency one in orange (about 20 years) that shows moderate variability; and a higher-frequency one in blue (about 10 years) showing the greatest variability.

This type of graph helps to clarify patterns in the apparently random interannual fluctuations. It should be remembered that particularly high or low values at the start or end of a record can have a disproportionate effect on the shape of the curves. Rainfall records at Canberra follow a similar trend for the duration of that record.

Rainfall and the Southern Oscillation Index

As is true throughout most of south-eastern Australia, there is a relatively weak direct relationship between the Southern Oscillation Index (SOI), a measure of the behaviour of the ENSO phenomenon, and rainfall in the ACT. This means that higher rainfall tends to occur when the SOI is strongly positive (e.g. 1955), and lower rainfall when it is strongly9 and 10), which show the degree of association between spring (Sep–Nov) rainfall totals and seasonal average values of the SOI for the same season. Although this is the season when the rainfall-SOI relationship is strongest in the region, the association is not dominant (the correlation coefficient is 0.48 at Hall) and some very wet and dry years have occurred with the SOI close to zero. The large-scale atmospheric fluctuations measured by the SOI account for between 15% and 25% of interannual rainfall variability in the ACT.

Figure 9. The relationship between rainfall at Hall and the SOI in spring (Sep-Nov). Circles indicate the years in this reporting period: red = 2000, black = 2001 and blue = 2002. A tight, generally linear grouping of points suggests a good correlation, whereas a wide, random scatter indicates a poor relationship.

scatterplot of spring rainfall and the SOI at Hall [Credit: Australian National University]

Figure 10. The time series highlights variations in the relationship between the SOI and rainfall at Hall over time. It shows the standardised deviations from mean annual rainfall as vertical bars, with the SOI plotted as a continuous line (SON = September–November).

graph of rainfall and the SOI at Hall from 1905 to 2003 [Credit: Australian National University]

Data sources and references

Temperature is measured principally at the official Bureau of Meteorology station at Canberra Airport. Data for this and other stations in the Territory may be made available through the Bureau of Meteorology. There may also be a number of privately-kept records in the Territory; these often contain invaluable information for otherwise sparsely-monitored areas.

The Canberra temperature record begins late in 1939. While this is a sufficiently long record for climatological analysis, it is too short for detecting low-frequency fluctuations in temperature. Long records from stations close to the ACT (in Tallaganda, Tumut and Yass shires, for example) may also provide insights into the temperature climatology of the Territory.

Information on Bureau of Meteorology weather stations and climate data can be obtained from:

Organisation National Climate Centre
Contact Numbers
Phone (BH) (03) 9669 4082
Fax (03) 9669 4515
Web site 

living sustainably

Click to expand sitemap