National rainfall outlook


Drier than normal signal persists in parts of the east

Text commentary of chance of increased rainfall

Summary

  • A drier than normal season is more likely for northern Queensland, and parts of southeast Australia
  • A wetter than normal season is more likely for a small part of central Australia
  • Climate influences include a brief negative Indian Ocean Dipole, and near-average Pacific waters
  • Outlook accuracy is moderate to high over most of the eastern States, the NT, SA and patchy over WA.
  • Details are summarised in our new monthly Climate and Water Outlook video
Probability of exceeding median rainfall, large image

Details

The chances of receiving above median rainfall for August to October are less than 40% over northern Queensland, southern NSW, the far southeast of SA, and most of Victoria. In other words, this means the chances of below median rainfall is greater than 60% in these areas. For every ten August to October outlooks with similar odds to these, about three or four of them would result in above-average rainfall over these areas, while about six or seven would be below average.

Over the rest of the country, there is no significant shift in the odds towards either a wetter or drier than normal season; the main exception is an increased chance of a wetter season over parts of the red centre.

Climatologically, areas of tropical northern Australia have at least two months of the dry season left. This means that the median rainfall at this time of the year is far less than during the summer months (see the Rainfall averages tab).

Climate influences

Warming of the tropical Pacific Ocean over the past several months has primed the climate system for an El Niño in 2014. However, in the absence of the necessary atmospheric response, Pacific Ocean temperatures have either stabilised, or some cooling has occurred. Despite some further easing in the model outlooks, a majority of international climate models still indicate El Niño is likely to develop during spring 2014. While there are some differences in ENSO outlooks, the near-average to drier-than-average signal across eastern Australia is generally consistent between international models.

The Indian Ocean Dipole (IOD) index has been below −0.4°C (the negative IOD threshold) since mid-June. Model outlooks suggest the IOD is likely to return to neutral by spring. A negative IOD typically brings wetter winter and spring conditions to inland and southern Australia. It is possible that the effects of the Indian Ocean and Pacific are competing to some degree, minimising the likelihood of broader rainfall signals.

How accurate is the outlook?

Outlook accuracy for the August to October period is:

  • Moderate to high over most of the eastern States, the NT, SA, and parts of WA
  • Low to very low over the WA interior, and small parts of central Queensland

Bureau climatologists continually monitor the climate for any significant developments, with information on the likelihood of El Niño/La Niña and IOD events available fortnightly at the ENSO Wrap-Up. For a summary of Pacific and Indian Ocean outlooks, please see the Climate Model Summary.

Further information

Media
(03) 9669 4057
Enquiries

Model accuracy

Model accuracy (also known as model confidence or model skill) is a measure of how well a model has performed at that time of year in the past. One way that the Bureau measures the accuracy of its climate models is by comparing how often the outlook favoured a particular category (for example, when above median rainfall was more likely to occur than below median rainfall), and that favoured category was then subsequently observed. This measurement of accuracy is known as "Percent Consistent", and has been tested for the official seasonal outlook model over the period from 1981 to 2010.

The accuracy levels on the maps give an indication of how well the outlooks match the observed outcomes. High accuracy means that tests of the model on historical data show a strong relationship between the most likely outlook category (above or below median) and the subsequent observation (above or below median). In areas with high accuracy, historically the model has performed very well, and hence a high degree of confidence can be placed in future outlooks. On the other hand, low accuracy means the model has not performed well in these regions and therefore the outlook should be used with caution. In the places and seasons where the outlooks are most skilful, the category of the eventual outcome (above or below median) is consistent with the category favoured in the outlook about 75% of the time. In the least skilful areas, the outlooks perform no better than random chance (equivalent to the "flip of a coin").

A random forecast of above median rainfall will be correct about 50% of the time. For this reason, the green shading on the map shows areas where the model has greater than 50% accuracy only. In areas which are not coloured in green on the map, some caution should be taken when using the forecast, notably at times when there is not a strong climate influence (for example, no El Niño or La Niña is present).

The Rainfall outlook has highest accuracy during autumn and spring, while in summer and winter there is lower skill, particularly over central Australia.

As a guide, the Bureau uses the following terminology when referring to the accuracy of the outlooks:

Map keyTerminology
Legend of the model accuracy map 75% and above - very high
65 to 75% - high
55 to 65% moderate
50 to 55% - low
45 to 50% - low
Below 45% - very low

What is normal for this period?

This map shows the median (or 50th percentile) rainfall for the given three months. The median rainfall is calculated from the 1981-2010 period.

The maps will differ from other median maps on the Bureau's website. This is because the dynamical model forecasts use an averaging period of 1981-2010. The quality of the dynamical model forecasts is in-part determined by the coverage and accuracy of the observations fed into it. Therefore, to be consistent from one year to the next, the Bureau has only run the model during the modern satellite era.

Median rainfall for January to March

About the outlook

Using the outlook

  • The Bureau's rainfall seasonal climate outlooks are general statements about the likelihood of wetter or drier than average weather over a three-month period. The probabilities are generated from the Predictive Ocean Atmosphere Model for Australia (POAMA), the Bureau of Meteorology's dynamical climate model. It is important to note that they are not categorical predictions about future rainfall, and hence the success or failure of one individual outlook does not infer that the model has low skill. Skill is assessed over multiple runs of the model. Likewise, temperature outlooks give the likelihood or chance of exceeding the average maximum and minimum temperatures over the entire three-month outlook period. Information about whether individual weeks or months may be unusually hot or cold, is presently unavailable.

  • Probability outlooks should not be used as if they were categorical (yes/no) forecasts. These outlooks should be used as a tool in risk management and decision making. Greatest benefits accrue from long-term use, say over 10 years.

About the model

  • The seasonal climate outlooks are generated by the Predictive Climate Ocean Atmosphere Model for Australia (POAMA), a dynamical (physics based) climate model developed by the Bureau of Meteorology and CSIRO Marine and Atmospheric Research. This coupled atmosphere-ocean model is a state of the art seasonal forecast system. Read more about POAMA.

  • The POAMA model is undergoing continuous research and development. Advances in the science of seasonal prediction, improvements in the observations and how they are fed into the model, as well as increases in supercomputing power are just some of the ways the model's accuracy will increase over time.

El Niño and La Niña

Indian Ocean Dipole

Statistical model outlooks

The official dynamical outlooks supercede the statistical outlooks. Statistical outlook maps will continue to be available for a review period: