Climate Model Summary

The webpage

This webpage surveys the latest output from several reputable climate models developed at leading international climate agencies. Data from these models is analysed and displayed in graphs and dials to show how temperature patterns in the Indian and Pacific ocean are likely to change over the coming six-months. These patterns in the oceans have been shown to have a significant impact on Australian temperature and rainfall patterns (e.g. cooling of the tropical Pacific Ocean during La Niña events is typically associated with increased rainfall and cooler daytime temperatures in Australia).

When the models are in agreement, then there is higher confidence in the model outlooks. When the models are showing a large spread or large variety of outcomes, there is a larger amount of uncertainty associated with the forecast.

The models

All the models included in this page have been documented in the peer-reviewed international scientific literature and have skill in predicting El Niño-Southern Oscillation (ENSO) and Indian Ocean patterns. They are all coupled General Circulation Models, and hence simulate both atmospheric and oceanic processes using physics rather than statistics. Most have been run routinely for a number of years and are subject to regular revision and update with the aim of continuous improvement. They represent the state-of-the-art of climate prediction systems and offer useful guidance up to about nine months. The majority of models used in our survey are formally recognised by the World Meteorological Organization as Global Producing Centres for long-range forecasts.

Update schedule

The Climate Model Summary is updated on the 12th of each month (or the next working day following the 12th) when model data becomes available. Data for the Bureau's model, ACCESS–S, is updated fortnightly along with several other models that also provide more frequent updates during the month.

The ensemble technique

All models use an ensemble method, where several forecasts are run at the same time using slightly different initial conditions. While all models use ensemble techniques, the number of ensembles differs for each model.

Important note

The data presented on these pages is the 'ensemble mean' which is the average of all the model-runs performed by each model, each month. The ensemble mean therefore flattens out the potential forecast outcomes and doesn't show the full spread of the possible forecast scenarios.

The dials

The ENSO and IOD dials (from the Overview section) are calculated by simply averaging the ensemble means from all the models surveyed. Values are not representative of the potential spread of the forecast outcomes. Therefore, the dials are intended to be used as a quick reference to determine whether the models are showing a warming trend, a cooling trend, or neither. More information is provided in the graphs presented in other sections.