Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


GELATO: GELATO documentation

3. Boundary conditions - climate forcings

Most forcing data comes from the CMIP6 protocol.

Greenhouse gasesUp to 2014, CMIP6 historical values of CO2, CH4, N2O, CFC11 and CFC12 from the CNRM-CM-6.1 run as described in Voldoire et al (2019). From 2015 onwards, greenhouse gas forcings follow the SSP3-7.0 scenario.
OzoneRadiation scheme sees a seasonally varying but otherwise fixed climatological ozone field. The monthly climatology of ozone concentration is the average of a previous CMIP6 run by Michou et al (2020) over the period 1995-2014.
Tropospheric aerosolsA monthly climatology of tropospheric aerosols is computed over the 1995-2014 period from a prior run using the interactive aerosol scheme TACTIC_v2 scheme. This run is described in detail in Michou et al (2020).
Volcanic aerosolsVolcanic stratospheric aerosols are the official CMIP6 forcings provided by Thomason et al (2018).
Solar forcingTime-variation of total solar insolation (TSI) is specified from CMIP6 annual mean forcings provided by Matthes et al (2017)

Detailed documentation:

Matthes, K., Funke, B., Andersson, M. E., Barnard, L., Beer, J., Charbonneau, P., et al. (2017). Solar forcing for CMIP6 (v3.2). Geoscientific Model Development, 10(6), 2247–2302.‐10‐2247‐2017

Michou, M., Nabat, P., Saint-Martin, D., Bock, J., Decharme, B., Mallet, M., Roehrig, R., Séférian, R., Sénési, S. and Voldoire, A. (2020). Present-day and historical aerosol and ozone characteristcs in CNRM CMIP6 simulatons. Journal of Advances in Modeling Earth Systems, 12, e2019MS001816,

Thomason, L. W., Ernest, N., Millán, L., Rieger, L., Bourassa, A., Vernier, J. P., et al. (2018). A global space‐based stratospheric aerosol climatology: 1979–2016. Earth System Science Data, 10(1), 469–492.‐10‐469‐2018

Voldoire, A., Saint-Martn, D., Sénési, S., Decharme, B., Alias, A., Chevallier, M., et al (2019). Evaluation of CMIP6 DECK experiments with CNRM-CM6-1. Journal of Advances in Modeling Earth Systems, 11.

4. Initialization and initial condition (IC) perturbations


4.1 Atmosphere and land

Atmosphere initialization
Atmosphere IC perturbationsNoneNone

Land Initialization

Land IC perturbationsNoneNone
Soil moisture initializationERA5ERA5T
Snow initializationERA5ERA5T
Unperturbed control forecast?NANA

Detailed documentation: see ECMWF page


4.2 Ocean and cryosphere

Ocean initialization





Ocean IC perturbationsNoneNone
Unperturbed control forecast?NANA

Detailed documentation:



5. Model Uncertainties perturbations:

Model dynamics perturbationsYes
Model physics perturbationsNo

If there is a control forecast, is it perturbed?


Detailed documentation: Batté, L. and Déqué, M., 2016: Randomly correcting model errors in the ARPEGE-Climate v6. 1 component of CNRM-CM: applications for seasonal forecasts, Geoscientific Model Development,9, 2055-2076, doi:10.5194/gmd-9-2055-2016


6. Forecast system and hindcasts

Forecast frequencymonth
Forecast ensemble size51
Hindcast years1993-2018
Hindcast ensemble size25
On-the-fly or static hindcast set?static
Calibration (bias correction) periodNA


7. Other relevant information

Ensemble start dates

The forecast uses three start dates:

  • The penultimate Thursday of the previous month (25 members in the forecasts, 12 members in the hindcasts)
  • The last Thursday of the previous month (25 members in the forecasts, 12 members in the hindcasts)
  • the 1st of the month (1 member in the forecast/hindcast)
Data assimilation method for analysis

Coupled : coupled initialization run nudged towards ERA5T (ERA5 in the hindcast) in the atmosphere and Mercator Ocean International analyses in the ocean. Sea ice initial conditions are provided by a separate NEMO-GELATO ORCA 0.25 forced run nudged towards the same ocean analyses.



Interpolation details

Météo-France System 8 is built on the CNRM-CM coupled climate model which embeds an output manager called the XIOS server (Meurdesoif et al, 2018). XIOS is an input/output parallel server software enabling to perform online field operations in the course of the model integration and writing the final result on NetCDF files. All operations such as time sampling, time averaging, horizontal regridding or vertical interpolation are specified through xml files and performed by XIOS such that no or little post-processing is needed: the only output files written on disk are the files featuring the requested C3S 1°x1° horizontal grid and pressure levels.

Horizontal Interpolation

The ARPEGE atmospheric model is a spectral model, and horizontal fields such as temperature or geopotential are originally computed as spherical harmonic coefficients, but can be converted into grid point data on the corresponding reduced Gaussian grid (e.g 0.5° reduced Gaussian grid for Tl359 spectral resolution) using an internal function of the model. The XIOS server can perform online interpolation of this grid point data on any user specified grid such as the rectilinear C3S 1°x1° grid. As for the oceanic fields from NEMO, they are internally defined as grid point data and can directly be interpolated by XIOS onto the 1°x1° grid. For both atmospheric and oceanic data, the interpolation method is a first order conservative remapping.

Vertical interpolation

ARPEGE uses sigma coordinates in the vertical, meaning that near-surface model levels follow the orography. Transformation into data on pressure levels is done directly on request by ARPEGE through a linear interpolation, or linear extrapolation for data below the surface.

Meurdesoif, Y. (2018) XIOS fortran reference guide. IPSL

8. Where to find more information

Technical implementation details can be found in