Use of trends in the SOI

Compiled by Col Paull Senior Extension Officer, Risk Management and Drought, QDPI.

Recent trends in the SOI are used in the Australian Rainman computer package to calculate the probabilities of receiving particular amounts of rainfall at particular locations; the maps provided in 'The Long Paddock' display this information in a spatial format, with maps showing the chance of receiving greater than median rainfall across Australia and just for Queensland.

Better estimate of rainfall probabilities

The AUSTRALIAN RAINMAN computer package is not a forecasting program in the strict sense of the word; it does not look at synoptic weather patterns to give short-term forecasts.

AUSTRALIAN RAINMAN can, however, look at the patterns of rainfall over the past 50 to 120 years, and work out mathematically the probability of getting a certain amount of rain over a defined period.

A probability of 70% means that the event has occurred in seven years out of ten - in the past. A probability of 50% means five years out of ten (or 50 out of 100) - not every second year.

A probability of 50% means even chances; 10% means very little chance; 95% means a very good chance - but still not a certainty.

Trends in the SOI

Recent trends in the SOI can be used to calculate more accurately the probabilities of receiving particular amounts of rainfall at a particular location; over the next few months. The main trends used in AUSTRALIAN RAINMAN are:

Average SOI over the last 90 days; and
The SOI Phase as determined by the change in average monthly SOI over the two pervious months.

SOI Phases

The phases of the SOI were defined by Dr Roger Stone, QDPI, who used cluster analysis to group all sequential two-month pairs of the SOI (from 1882 to 1991) into five clusters (see Figure 1). The phases are:

Phase 1 consistently negative
Phase 2 consistently positive
Phase 3 falling
Phase 4 rising
Phase 5 consistently near zero

Image 1

References

Stone, R.C., Hammer, G.L and Marcussen, T. (1996) Prediction of global rainfall probabilities using phases of the Southern Oscillation Index. Nature, 384, 252-255.
Stone, R.C., Hammer, G.L., and Woodruff, D.R (1993) 'Assessment of risk associated with climate prediction in management of wheat in north-eastern Australia' Proceedings 7th Australian Agronomy Conference, Adelaide, September, 1993. Australian Society of Agronomy, Parkville, Victoria, 174-177.
Stone, R.C., Nicholls, N., and Hammer, G.L. Frost in NE Australia: trends and influences of phases of the Southern Oscillation. Journel of Climate (in press).

The boundaries between phases (clusters) were defined by Dr Roger Stone then further developed by Dr Jeff Clewett, QDPI, by plotting the distribution of the clusters and mathematically defining the boundary curves to minimise errors (see Figure 2). The curves are also used in AUSTRALIAN RAINMAN; where the SOI phase is calculated over 3- or 4-month periods, average SOI values for the first and last months of the period are used.

Figure 2

SOI Average Vs SOI Phase

Figure 3

Figure 4

Generally use of SOI phases to calculate future seasonal rainfall probabilities gives a more accurate result than using SOI averages.
For example, if at the end of November you wanted to determine the probability of receiving at least 310mm at Kingaroy during the December to February period:
using the SOI average during October-November, the probabilities can range from about 45% to 52% (see Figure 3); and
using the SOI phase during the October to November period, the probabilities can range from about 35% to 60%.(see Figure 4).
Thus, the increased accuracy of rainfall probabilities derived from the SOI phase may significantly improve weather-related decisions.
Reference: AUSTRALIAN RAINMAN Version 2.1 Users Guide.

Data sourced from Queensland Department of Primary Industries and Fisheries.

Last updated: 08 April 2010