SimOcean
Simulating and forecasting southern Africa's ocean

Development

The operational initiative of SimOcean follows a 2-phased approach. The first phase (0-2 years) will use resources that are currently available in order to produce a marine forecast product and will serve as a demonstration, or 'proof of concept' project. It will downscale global ocean forecast products to higher resolution regional and limited area southern African domains and use global atmospheric forecast products as forcing. As part of an assessment of the ocean model, a hindcast simulation will be run from as far back as the chosen wind forcing product exists (for example, NCEP from 1948-, QuikSCAT: 1999-2009 and ECMWF: 1989-2010). The second phase (2-5 years) will focus on system enhancements and improvements of the forecast system, the emphasis being on research and development. Improvements to the implementation and demonstration phase will include data assimilation, ocean-atmospheric coupling, wave coupling. The potential for additional enhancements is enormous and provides huge scope for interdisciplinary and multi-institutional collaboration.

To fulfill the objectives of the two phases summarised above, the project is divided into work packages aimed toward establishing an operational ocean forecast system for southern Africa.

Phase I Implementation activities

WP-1Assessing global ocean forecast products for the southern African region
WP-2Implement regional southern Africa model
WP-3Implement pilot limited areas model of the Benguela Current region
WP-4Data dissemination and product development
WP-5Pilot coastal embayment scale and higher resolution operational forecasts
WP-6Regional and localised wave forecasting system
WP-7Biogeochemical modelling at coastal embayment and near shore scales

Phase II Operational, system development and research activities

WP-8Data assimilation

Other potential work packages:

  • Limited areas models of other southern African regions
  • Ocean-atmosphere coupling
  • Reanalyses, establishing long time series