Oral Presentation AMOS Annual Meeting and International Conference on Tropical Meteorology and Oceanography

The Maritime Continent prediction system (#139)

Gary B Brassington 1 , Gary Dietachmayer 2 , Frank Colberg 2 , Stefan Zeiger 2
  1. Bureau of Meteorology, Sydney, NSW, Australia
  2. Bureau of Meteorology, Melbourne, VIC, Australia

The maritime continent is a region that significantly influences Australia’s climate, economy and security and requires a wide range of services including routine high resolution real-time atmospheric, marine and ocean forecasts.

A one-year pilot project was undertaken to explore the feasibility of forecasting the atmospheric, oceanic and wave condition in this region. A common region defined by 114.3E to 142.5E and 17.75S to 7.85N was resolved by all component models at a uniform 1/50° resolution. The atmospheric model was based on the Unified Model v10.6 with a full Euler (non-hydrostatic) formulation, 80 model levels and an explicit treatment of convection. The ocean model was based on the Regional Ocean Modeling System (ROMS) with 30 vertical levels and SRTM30+ bathymetry. The wave model used WaveWatch III with an unstructured grid and boundary conditions from AUSWAVE.

The three models were assessed for a parallel forecast period of one month, March 2018. This period included two tropical cyclones, TC Kelvin (4-9 March) and TC Marcus (14-24 March). The atmospheric model demonstrated improved representation of tropical convection but a reduction in general precipitation. The wave model demonstrated improvements associated with resolved narrow passages and islands. A reduction in errors for extreme wave heights was obtained relative to altimetry observations. The ocean model was compared with 16 tide gauges showing correlations better than 0.94 for all gauges except those along the boundary in Torres Strait and one in the Gulf of Carpentaria. Mean absolute differences of 0.67 degrees and 0.141 PSU for temperature and salinity respectively were obtained from ~8000 profiles.

The results obtained were based on downscaled models without data assimilation. The performance was determined to be sufficient to advance to a full development stage. An overview of this project will be presented and opportunities for collaboration and stakeholder engagement will be outlined.