1. Forecast system version

System name: GloSea6-GC3.2

First operational forecast run:  2 Feb 2021

2. Configuration of the forecast model

Is it a coupled model?  Yes: Atmosphere, land, ocean and sea-ice.

Coupling frequency:  Hourly coupling between atmosphere-land and ocean-sea-ice.

The coupled model Global Coupled 3.2 (GC3.2) is described in Williams et al, 2018.

2.1 Atmosphere and land surface


Met Office Unified Model (UM) - Global Atmosphere 7.2

Joint UK Land Environment Simulator (JULES) - Global Land 8.0

Horizontal resolution and gridN216: 0.83 degrees x 0.56 degrees (approx 60km in mid-latitudes)
Atmosphere vertical resolution85 levels
Top of atmosphere85km
Soil levels (layers)


Layer 1 : 0 - 0.1 m
Layer 2 : 0.1 - 0.35
Layer 3 : 0.35 - 1.0 m
Layer 4 : 1.0 - 3.0 m

Time step15 minutes

Detailed documentation:

JULES documentation

Global Atmosphere 7.0/7.1 & Global Land 7.0: Walters et al, 2017

2.2 Ocean and cryosphere

Ocean model

NEMO v3.6 - Global Ocean 6.0

Horizontal resolutionORCA 0.25
Vertical resolutionL75
Time step20 minutes
Sea ice model

CICE v5.1.2 - Global Sea-Ice 8.1

Sea ice model resolutionORCA 0.25
Sea ice model levels5 categories + open water
Wave modelN/A
Wave model resolutionN/A

Detailed documentation: NEMO documentation, CICE documentation

Global Ocean 6.0: Storkey et al, 2017.

Global Sea Ice 8.0:  Ridley et al, 2017

3. Initialization and initial condition (IC) perturbations

3.1 Atmosphere and land

Atmosphere initialization
ERA-InterimMet Office Global Hybrid 4D-VAR
Atmosphere IC perturbationsNoneNone
Soil moisture & temperature  initializationMet Office JULES-JRA55 analysisMet Office JULES-JRA55 analysis
Snow initialization

 Met Office JULES-JRA55 analysis

Met Office JULES-JRA55 analysis
Unperturbed control forecast?NoNo

Detailed documentation:

Met Office Global Hybrid 4D-VAR: Clayton et al, 2013

JRA55 analysis:  Kobayashi et al. 2015, Harada et al. 2016

3.2 Ocean and cryosphere

Ocean initializationGloSea Ocean Sea-Ice Analysis (GS-OSIA)Forecast Ocean Assimilation Model (FOAM)
Ocean IC perturbationsNoNo
Unperturbed control forecast?NoNo

Detailed documentation:

The GS-OSIA and the FOAM system both use the Nucleus for European Modelling of the Ocean data assimilation system (NEMOVAR). This is a 3d-VAR data assimilation scheme. The GS-OSIA uses different surface forcing (ERA-interim) and observation sets as it is a historical analysis. FOAM uses surface forcing from the Met Office Global NWP model and real-time observations.

The common NEMOVAR system is described in Blockley et al, 2014. Details of the GS-OSIA can be found in MacLachlan et al, 2015.

4. Model uncertainties perturbations:

Model dynamics perturbationsNone
Model physics perturbationsAtmosphere stochastic physics scheme, SKEB2

If there is a control forecast, is it perturbed?

No control

Detailed documentation:

SKEB2: Bowler et al, 2009

5. Forecast system and hindcasts

Forecast frequencydaily
Forecast ensemble size2 per day
Hindcast years24 (1993-2016)
Hindcast ensemble size7 per start date
Hindcast start dates1, 9, 17, 25 of each month
On-the-fly or static hindcast set?on-the-fly

6. Where to find more information

GloSea5 system:

Model description references:

Initialisation references: