GOOS Products and Services Bulletin

Forecasting Ocean Assimilation Model

Mike Bell, Adrian Hines, and Jon Turton
Met Office, UK

The Met Office's FOAM (Forecasting Ocean Assimilation Model) is a system that produces real-time daily analyses and forecasts of temperature, salinity and currents in the deep ocean, for up to five days ahead. The original FOAM system was introduced in 1997, and has subsequently evolved to better meet ocean forecasting requirements. Supported by an active research and development programme, FOAM is at the forefront of the development of operational forecasting of the deep ocean.

FOAM is built around nested physically based ocean and sea-ice models. It is driven by six-hourly mean surface fluxes from the Met Office's operational numerical weather prediction (NWP) system and assimilates ocean observations (in situe and remotely sensed) that are available in near real-time. Observations, from the previous 10 days are assimilated with variable weighting.

The FOAM system is fully relocatable, allowing high-resolution configurations to be set up for any deep ocean region.

FOAM includes:

• An ocean model based on the Bryan-Cox code (the same code is used by the Hadley Centre for coupled climate simulations and also for coupled seasonal forecasting);
• A sea-ice model to predict sea-ice thickness, concentration, and velocities;
• Assimilation of temperature and salinity profiles (BATHYs and TESACs) including data from the Argo profiling floats and the TAO/TRITON and PIRATA moored arrays;
• Regional configurations of FOAM also assimilate sea-surface height data from satellite altimeters such as Jason-1;
• Assimilation of ship, buoy and satellite (AVHRR) sea-surface temperature reports; and
• Assimilation of sea-ice concentration fields received from the Canadian Meteorological Centre.

FOAM configurations

A global version of FOAM with 1° horizontal resolution and 20 levels in the vertical has been run daily since October 1997. Although this resolution allows features in the upper ocean structure due to meteorological forcing and large-scale ocean processes to be analysed and forecast, it cannot represent fronts and eddies. Typically a resolution of 1/9° (11 km) or better is required to resolve mesoscale ocean eddies and a nested model system has been developed to accomplish this, with:

• basin-scale configurations (nested within the global model) for (i) the Atlantic and Arctic oceans, and (ii) the Indian Ocean, at 1/3° horizontal resolution and 20 vertical levels;
• regional configurations for (i) the North Atlantic, (ii) the Mediterranean and Black Sea, and (iii) the Arabian Sea, at 1/9° horizontal resolution and 20 vertical levels;
• high-resolution configurations for the north-east Atlantic (1/20°), Gulf of Mexico and Caribbean Sea (1/18°).

Ten-day mean current speeds (cm/s) from 1/9° North Atlantic FOAM

Assessments

The high resolution nested FOAM configurations have the potential to analyse and predict the evolution of mesoscale ocean features such as eddies and fronts.

The FOAM system aims to realise this potential through the exploitation of advances in modelling techniques combined with the complementary data provided by Argo profiling floats and satellite altimetry.

Recent assessment work has highlighted the value of FOAM temperature analyses and forecasts at 1/9° resolution. The importance of the contribution to the quality of the FOAM products from the Argo and altimeter data has also been demonstrated. 

Assessment work is continuing to examine the quality of mesoscale forecasts at higher resolutions, with particular attention to the accuracy of mixed layer predictions and forecasts of ocean currents. Prioritisation of assessment and development work is guided by the interests of the Met Office's Oceanography Customer Group which provides a forum for discussion of potential applications of the FOAM products.

 

Sea surface temperature in the Central Mediterranean (left) from FOAM 1/9° configuration for 8 January 2004,
(right) corresponding seven-day satellite SST composite.

Collaborative development

The FOAM system is an active participant in several major international, European and UK programmes aimed at developing the capacity for operational ocean monitoring and forecasting.

The Met Office is participating in the international Global Ocean Data Assimilation Experiment (GODAE), which is bringing together float data (Argo), satellite data (Jason) and high resolution modelling. Participants are producing global ocean analyses and forecasts in near real-time from a range of different ocean models for comparison.

Other projects in which the FOAM system is playing a key role include the Mersea intercomparison of real-time ocean analysis and forecast systems, and the Natural Environment Research Council (NERC) CASIX programme exploring issues relating to air-sea interactions. Insight and experience gained from these projects is invaluable in ensuring that FOAM remains at the forefront of operational ocean modelling capability.

RMS temperature errors against depth for 1/9° North Atlantic FOAM analyses from June 2002
to June 2003, with corresponding errors for climatology.

FOAM output

• Gridded fields of temperature, salinity and current profiles, mixed-layer depth, sea-ice concentration and sea-ice thickness and cover.
• These are available in real-time over the internet (e-mail or ftp) from the Met Office's Data and Products Distribution System (DPDS).
• Selected FOAM outputs are also available to research users over the internet via the NERC Environmental Systems Science Centre (ESSC) Live Access Server.
• Provision of real-time boundary data to high-resolution limited area applications such as shelf-sea models.

Future work

FOAM development work is underway to enhance the current system capabilities, and to expand the range of applications served by the FOAM system.

Recent work has enabled the coupling of the FOAM system with an ocean biology model developed in the Hadley Centre. Trials of the initial coupled system have shown it produces a realistic evolution of biological parameters in the open ocean. Future work will develop techniques for assimilation of satellite ocean-colour data to improve initialisation of the model chlorophyll fields.

Work is underway to implement an enhanced sea-ice model in FOAM, alongside the development of a state of the art sea-ice data assimilation scheme to exploit both sea-ice concentration and sea-ice velocity data.

For further information on FOAM, and how to obtain FOAM data, see 
http://www.metoffice.gov.uk/research/ncof/products.html

© Met Office, UK