Earth system models research

Developing weather, climate, ocean and water models

Computer simulations of the real world, based on the laws of physics, are powerful tools. We use these models to:

• provide forecasts and warnings.
• understand processes that cause changes in the environment.

Our research and development work aims to improve the models and their predictions. This means better forecasts for Australia and our region.

We collaborate with international partners to do this. For example, we are part of the Momentum Partnership with the United Kingdom Meteorological Office. This partnership is developing a set of models called the Australian Community Climate and Earth System Simulator (ACCESS).

Our research and development focus

Our work focuses on finding innovative ways to:

• process observations – that is, measurements from the atmosphere, ocean, ice and land
• combine observations with model results to fill in the gaps of our observations, to give the best estimate of the current state of the atmosphere, ocean, ice or land (data assimilation)
• improve models to make them more accurate and useful
• evaluate performance of the models
• improve ensemble forecasting – methods that take into account uncertainties in how weather and climate will evolve by running more than one forecast
• investigate how artificial intelligence can be used to improve our models and predictions.

Atmospheric models and predictions

Building accurate models of our atmosphere is vital to understanding and predicting our weather and climate.

Working with partner organisations, we develop the atmospheric components for Australian weather and climate science applications.

We contribute to the ACCESS set of models, including:

• global, regional-scale, urban and tropical cyclone configurations
• models based on artificial intelligence techniques.

We also:

• evaluate how models are performing, including at different resolutions
• develop and test parameters for physical processes, like those involved in development of thunderstorms, which are smaller than model grids
• implement and evaluate ensemble techniques
• track international developments on weather models based on artificial intelligence, and how these compare with models based on physical processes.

Ocean models and predictions

Marine weather and ocean forecasts rely on ocean modelling and predictions. This work is critical for public safety, defence and a range of industries.

We develop numerical ocean and sea-ice prediction systems. This includes models for:

• the global ocean (OceanMAPS)
• waves (AUSWAVE)
• shelf and coastal seas
• storm surge
• tsunamis
• the Antarctic region.

We also:

• process observations of sea surface temperature, fluxes and waves from the Integrated Marine Observing System
• process satellite observations to generate maps of sea surface temperatures
• help develop applications for the Department of Defence, particularly the Navy, and other priority users of our services.

Atmosphere-ocean models and seasonal prediction

Long-range climate forecasts tell us about our weather in the coming weeks, months or seasons. They aid decision-making for a wide range of purposes and sectors, such as agriculture, energy and water.

Working with partner organisations, we develop our climate forecast system. It provides weekly to seasonal and longer range climate forecasts.

The current forecast system is a configuration of the ACCESS model called ACCESS-S (S stands for seasonal). It combines global ocean, atmospheric, land, and sea ice models. This combination is called a 'coupled' model. The coupled model predicts changes in the ocean and atmosphere. For more information, view our ACCESS-S page.

Our expertise helps to guide the use of ACCESS-S forecasts in applications and downstream models. For example, models that support the agriculture, fisheries and water sectors.

We also:

• implement and evaluate coupled data assimilation and ensemble generation techniques
• evaluate and verify forecast systems
• research and develop forecast calibration and post-processing techniques
• develop prototype forecast products
• explore potential for future versions of ACCESS-S to be based on artificial intelligence methods.

Land surface and water (hydrology) models

Land conditions such as soil moisture and vegetation:

• play an important role in the exchange water, energy and carbon between the land surface and the atmosphere
• have a significant impact on our weather and climate patterns
• are important in our weather and climate models.

Land surface and hydrology models are used to:

• manage water resources
• plan for potential hazards such as floods, drought and fire.

Research into land surface and water models

Our research aims to bring hydrology and vegetation changes (dynamics) into our ACCESS models, such as ACCESS-S. This would help improve forecasts for:

• streamflow and flood
• weather and climate
• agricultural applications.

We're also exploring:

• improving modelling of water movement across land – for example, in rivers and streams
• vegetation dynamics for different hydro-climatic regions across Australia
• how we can use satellites to check models and for data assimilation.

For more about land surface modelling and water research, see our Water research page.

Data assimilation

To make more accurate predictions, we combine the output from computer models with data gathered through observations. For example, observations from satellites, weather balloons, ground stations, aircraft, ships and autonomous vehicles.

This is called data assimilation. It gives us the best estimate of current conditions, from which we can start a forecast.

We develop data assimilation systems for atmospheric, seasonal, land, ocean and sea ice prediction systems. We also:

• research how to process observations and quality control for data assimilation systems
• implement and evaluate different data assimilation techniques
• quantify the impact and value of observations to numerical prediction
• develop and implement our atmospheric high-resolution regional reanalysis for Australia (BARRA).