ACCESS-S: forecasts for weeks to seasons ahead

ACCESS-S (Australian Community Climate and Earth-System Simulator – Seasonal) is the Bureau of Meteorology's climate modelling system used for outlooks on weekly through to seasonal timescales.The system comprises a coupled atmosphere-ocean model, a data assimilation system and a strategy for generating forecast ensembles. The development of ACCESS-S supports several seasonal prediction projects (listed below).


In 2018, the first version of ACCESS-S replaced its predecessor, POAMA, as the Bureau's official seasonal forecast system. ACCESS-S1 has a number of significant enhancements over POAMA, including increased spatial resolution and improved model physics parameterisations. This upgrade brought seasonal prediction into the national ACCESS modelling framework, already in use for weather forecasts and climate change, which utilises the latest local and overseas developments. In particular, the development of ACCESS-S leverages off a key partnership with the UK Met Office.

The upgrade of the seasonal prediction system to ACCESS-S was funded in part by the 2015 Agricultural Competitiveness White Paper in a project known as the Improved Seasonal Forecast Services (ISFS) project. Since 2015, the ISFS project has enhanced the Bureau's climate outlooks service by providing higher resolution maps, more frequent updates and increased outlook accuracy. In 2019 the final phase of the project delivered multi-week forecasts to bridge the gap between 1-7 day weather forecasts and monthly to seasonal outlooks. The project aimed to assist the agricultural sector (and other climate sensitive sectors) to unlock the value of multi-week and seasonal forecasting by improving the intelligence provided to decision makers.

ACCESS-S development information


ACCESS-S1 is based on the UK Met Office's global coupled model seasonal forecast system GloSea5-GC2 (Global Seasonal forecast system version 5 using the Global Coupled model configuration 2; MacLachlan et al. 2015), but has enhancements to the ensemble generation strategy to make it appropriate for sub-seasonal forecasting, and a larger ensemble size. The GloSea5-GC2 system is made available to BoM through the Unified Model Partnership, of which BoM is a core partner.

The key components of the ACCESS-S1 forecast system are summarised in the Table below. The main differences between ACCESS-S1 and GloSea5-GC2 are the ensemble size, method of ensemble generation, the hindcast size, and the real-time configuration. Full details of ACCESS-S1 are provided in Hudson et al. 2017.

Atmospheric model

Global Atmosphere 6.0: The Unified Model (UM; Walters et al. 2017, Williams et al. 2015).

Horizontal resolution: N216 (~60 km in the mid-latitudes)

Vertical resolution: 85 levels

Land surface model

Global Land 6.0: Joint UK Land Environment Simulator (JULES; Best et al. 2011; Walters et al. 2017) with 4 soil levels

Ocean model

Global Ocean 5.0: Nucleus for European Modelling of the Ocean (NEMO ORCA25 ; Madec 2008, Megann et al. 2014).

Horizontal resolution: 0.25°

Vertical resolution: 75 levels. Level thicknesses range from 1 m near the surface to about 200 m near the bottom (6000 m depth).

Sea ice model

Global Sea Ice 6.0: Los Alamos sea ice model (CICE; Hunke and Lipscomb 2010; Rae et al. 2015).


Ocean Atmosphere Sea Ice Soil coupler (OASIS3, Valcke, 2013)

Atmosphere initial conditions

ERA-Interim (Dee et al. 2011) for the hindcasts

Operational Global NWP analysis for the real-time forecasts

Ocean and ice initial conditions

Met Office Forecast Ocean Assimilation Model (FOAM) which uses the NEMO 3-dimensional variational ocean data assimilation (NEMOVAR; Mogensen et al. 2009, 2012; Waters et al. 2015)

Land surface initial conditions

Climatological soil moisture (MacLachlan et al. 2015)

Ensemble generation

Initial condition uncertainty: BoM perturbation scheme combined with lagged initial conditions

Model uncertainty/unresolved processes: Stochastic Kinetic Energy Backscatter version 2 (SKEB2; Bowler et al. 2009).

Future Versions: ACCESS-S2/S3

Compared to ACCESS-S1, ACCESS-S2 will have a locally developed data assimilation system for the ocean and will improve on the land surface initialisation for soil moisture. It will also have a longer hindcast period (more than 30 years) and forecasts will run out to lead times of several years to facilitate an investigation of multi-year prediction.

ACCESS-S3 will improve the data assimilation (new variables assimilated e.g., altimeter data) and ensemble generation (weakly coupled Ensemble Kalman Filter) and will use an upgraded coupled model version, including utilising locally developed model physics parameterisations improvements.

Current seasonal prediction projects

Forewarned is Forearmed (FWFA): managing the impacts of extreme climate events

Australian farmers and agribusiness operate in one of the most variable climates in the world. Extreme climatic events (such as heatwaves, frost and heavy rainfall) and climate variability are amongst the largest drivers of annual agricultural production and income.

A 5-year project funded under the Australian Government Department of Agriculture and Water Resources as part of its Rural R&D for Profit program and managed by Meat & Livestock Australia (MLA), will deliver direct value to farmers by providing forecasts of extremes and equipping farmers with the information and tools to be forewarned and prepared.

The Bureau of Meteorology (BoM), working with several research partners, will develop and deliver forecasts of the likelihood of climate extremes on multi-week to seasonal timescales – beyond the 7-day weather forecast. This will provide farmers with the first ever forecasts of extreme climate events weeks to seasons ahead. The forecasts will be based on BoM's seasonal forecast system, ACCESS-S. The BoM component of the project includes research to 1) evaluate farmer and advisor needs, 2) understand large-scale drivers (e.g., El Niño, the MJO) of extreme events, 3) improve ACCESS-S to give better forecasts of extremes, and 4) develop experimental forecast products which will be trialled by users to assess value. A subset of products that have sufficient accuracy and utility will be delivered as official BoM forecasts to the benefit of agriculture.

Project partners include agricultural climate and systems analysis researchers and extension experts with expertise in the dairy, beef, sheep, grains, sugar and wine industries. They will use BoM outputs and work directly with farmers and farm consultants to interface the forecasts with agricultural decision-making systems. They will also develop risk management strategies to proactively prepare for these events, as well as extending project outputs to farmers and advisors. For more details, visit the FWFA project page.

Northern Australia Climate Program (NACP)

NACP will deliver innovative research, development and extension outcomes, through a number of partners, to improve the capacity of the red meat industry to manage drought and climate risk across northern Australia. This project includes collaboration between the Queensland Government (under their Drought and Climate Adaptation Program) and the MLA (Meat & Livestock Australia) Donor Company to fund the University of Southern Queensland (USQ), Bureau of Meteorology and the UK Met Office for a range of research, development and extension projects. For more details, visit the NACP project page.

Queensland Government DCAP2 (Drought and climate adaptation program) Projects

The Bureau is a partner in a number of DCAP2 projects. The major projects are NACP and Forewarned is Forearmed, mentioned above. We are also involved in:

  • "The Use of Bureau of Meteorology Multi-Week and Seasonal Forecasts to Facilitate Improved Management Decisions in Queensland's Vegetable Industry" (led by David Carey at the Queensland Department of Agriculture and Fisheries).
  • "The inside edge for graziers to master Queensland's drought-prone climate" (led by the Queensland Department of Science, Information Technology and Innovation)

Sea Surface Temperature Forecast Service Upgrade

The Bureau's coral bleaching risk forecast products were the first operational dynamical seasonal forecasts of their kind in the world and are integral components in the Great Barrier Reef Marine Park Authority's (GBRMPA) Coral Bleaching Response Plan. These forecast tools provide invaluable advance warning of potential bleaching conditions in the upcoming summer, allowing for proactive management and response. They are used for annual inter-agency planning, ministerial briefings, GBRMPA reporting and media releases. This project will migrate these operational products from POAMA to ACCESS-S

Ocean climate forecasts for New Zealand (OCForNZ)

OCForNZ is a collaboration between the Bureau of Meteorology and the National Institute of Water and Atmospheric Research (NIWA) in New Zealand to develop a state of the art climate forecast product for aquaculture in the region. This project will provide tailored multi-week to seasonal forecasts of relevant parameters such as sea surface temperature in order to facilitate improved decision making. The products developed will also enable advance detection of extreme conditions such as marine heatwaves that might pose a threat to the industry.


Team Leaders


  • Forecast system development and evaluation
  • Transitioning forecast systems to operations
  • Developing forecast outputs
  • Understanding the fundamental drivers and processes of climate variability in observations and the model
  • Interfacing to applications

For more information,
please contact Oscar Alves, Matthew Wheeler or Debbie Hudson