Overview

OpenIFS 43r3v1 is based on the ECMWF operational IFS cycle 43r3 which was the operational model at ECMWF from July 2017 to June 2018. OpenIFS 43r3v1 is a significant and extensive upgrade from previous release OpenIFS 40r1 model releases.

Anyone wishing to use OpenIFS must have a OpenIFS license (see the list of Licensed Institutes).

Support

Please report any issues or problems with this release to either the OpenIFS User Forums or the openifs-support@ecmwf.int email.

Model documentation

OpenIFS 43r3v1 is scientifically identical to IFS 43r3. For a full description of the IFS 43r3 operational model, please see IFS manuals in the ECMWF eLibrary.

Summary of scientific changes

OpenIFS 43r3v1 includes all the changes listed below compared to OpenIFS 40r1v2. For more details of the changes introduced between IFS 43r3 and the previous operational models (since 40r1) please see: Changes in ECMWF IFS model.

The following list summaries the changes with the IFS cycle when they were introduced:

Changes introduced from operational IFS 41r1:

• New surface climate fields (land-sea mask, sub-grid orography), also affecting number of land and sea points.
• New CO2/O3/CH4 climatologies from latest MACC-II reanalysis produced at ECMWF.
• Revised semi-Lagrangian extrapolation reducing stratospheric noise.
• Revised interpolation of moist variables in the upper-troposphere/lower stratosphere (UTLS).
• Cloud scheme change of rain evaporation, auto-conversion/accretion, riming, precipitation fraction.
• Improved representation of supercooled "freezing" rain.
• Modified convective detrainment.
• Activation of the lake model (FLAKE).
• Active use of wave modified stress in coupled mode.
• Revised sea-ice minimum threshold, sea-ice roughness length and consistency between SST and sea ice concentration.
• Changes to the fields CIN and CBH to include a bitmap with missing values.
• The domain of limited-area ocean wave model is extended to the full globe.

Changes introduced from operational IFS 41r2:

• Introduced cubic truncation for the spectral dynamics and an octahedral reduced Gaussian grid.
• Increased semi-lagrangian departure point iterations from 3 to 5 to remove numerical instabilities near strong wind gradients, particularly improving East Asia (downstream of the Himalayas) and improved representation of tropical cyclones.
• Changed formulation of the horizontal spectral diffusion to a spectral viscosity with significantly reduced damping at the small scales.  
• Removed dealiasing filter on rotational part of the wind as no longer needed for cubic grid (no aliasing).
• Reduced diffusion in the sponge layer near the top of the model (above level 30 for 137 level configuration) scaled by grid resolution rather spectral resolution, due to new cubic grid.
• Improved representation of radiation-surface interactions with approximate updates every timestep on the full resolution grid leads to a reduction in 2m temperature errors near coastlines.
• Included surface-tiling for long-wave radiation interactions to reduce occasional too cold 2m temperature errors over snow.
• Improved freezing rain physics and an additional diagnostic for freezing rain accumulation during the forecast.
• Introduced resolution dependence in the parametrization of non-orographic gravity wave drag, reducing with resolution and improving upper stratospheric wind and temperature.
• Changed the parcel perturbation for deep convection to be proportional to the surface fluxes, reducing overdeepening in tropical cyclones.
• Increased cloud erosion rate when convection is active, to reduce cloud cover slightly and improve radiation, particularly over the ocean.
• Improvements of linear physics used in the data assimilation for gravity wave drag, surface exchange and vertical diffusion, improving near-surface winds over ocean in the short-range.
• Correction to solar zenith angle for the sunshine duration diagnostic. For clear sky days the sunshine duration increases by 2 hours, now in good agreement with observations. For cloudy days, sunshine duration may now be overestimated due to an existing underestimation of cloud optical thickness.
• Improved solar zenith angle calculation removes stratospheric temperature dependence on radiation timestep and reduces anomalous small amplitude fluctuations in incoming solar radiation around the equator.

Changes introduced from operational IFS 43r1:

• Changes to boundary layer cloud for marine stratocumulus and at high latitudes.
• Modifications to surface coupling for 2 metre temperature.
• New model output fields include four cloud and freezing diagnostics (for aviation), a new direct-beam solar radiation diagnostic and improvements to the sunshine duration diagnostic.
• A global fix for tendency perturbations in the stochastic model error scheme SPPT to improve global momentum, energy and moisture conservation properties.

Changes introduced from operational IFS 43r3:

• New, more efficient radiation scheme, ecRad, with reduced noise and more accurate longwave radiation transfer calculation (see more details below).
• New aerosol climatology based on ‘tuned’ CAMS aerosol re-analysis including dependence on relative humidity.
• Increased super-cooled liquid water at colder temperatures (down to -38C) from the convection scheme.
• Visibility calculation changed to use ‘tuned’ CAMS aerosol climatology.

New radiation code ecRad

A new ECMWF radiation scheme became operational in IFS 43r3, replacing the older McRad scheme which first became operational in 2007. The new ecRad scheme is more modular allowing individual components to be swapped for faster and more accurate ones. It is also much more efficient. It uses a new implementation of the McICA (Monte Carlo Independent Column Approximation) code that is less noisy in partially cloudy conditions. Improvements in longwave radiation transfer reduce biases in temperature profiles. In operational IFS cycle 43r3, ecRad brings slight improvements in forecast skill.

For further information about ecRad please see the following sources:

• ECMWF newsletter 155: A new radiation scheme for the IFS
• A flexible and efficient radiation scheme for the ECMWF model, 2018, JAMES.
• ECMWF technical report: ECRAD. A new radiation scheme for the IFS (older version)
• ecRad home page

New climate fields: climate.v015

The climate fields (the fields in the ICMCL initial file) were altered for OpenIFS 43r3v1 as was the land-sea mask. The climate fields contain surface and soil information such sea ice area fraction, SSTs, albedo, soil temperature, GRIB codes ci, stl1, al, aluvp, aluvd, alnip, alnid, lai_lv, lai_hv.

OpenIFS 43r3v1 should not be used with the older climate files used with OpenIFS 40r1. IFS 43r3 has been developed and validated against the 'climate.v015' files which must be used instead.

The script to run OpenIFS (bin/oifs_run) expects the climate.v015 files to be available.

Technical aspects

Changes to GRIB encoding

Model identifiers

The GRIB model identifiers (generating process identification number) for cycle 43r3 will be changed as follows:

Compilers

Compilation configurations are provided for the GNU, Intel and Cray compilers. The PGI  and IBM compiler configurations provided in OpenIFS 40r1 are no longer supported.

ecCodes GRIB library

The grib-api GRIB library is no longer supported. Users must install the ECMWF ecCodes GRIB library to use OpenIFS 43r3.

netCDF library

OpenIFS now depends on the netCDF library to be available in order to read and write files in both ecRad and the wave model. Compilation of OpenIFS requires the netCDF library to be available. Please see the OpenIFS User Guide for more details.

XIOS parallel I/O server: netCDF model output

The XIOS parallel input/output library allows OpenIFS to pass all output to a separately running program (XIOS), which may execute in parallel, for more efficient performance.   XIOS allows for parallel output in netCDF format directly from OpenIFS instead of GRIB output via a single process.  It is configured using XML input files and does not use the NAMFPC model namelist. In addition, XIOS may be configured to compute additional fields such as monthly averages 'on-the-fly' instead of a separate post-processing step after the model is run. 

For instructions on how to download, build and use XIOS with OpenIFS, please see the How-to use XIOS with OpenIFS guide.  By default, OpenIFS assumes that XIOS is not available. Users must have the XIOS library installed on their system and then configure the compilation to use XIOS.

A t21test directory configured to use with XIOS is provided with OpenIFS with some example files.

Please note however, that the XML files to configure XIOS must be developed by the user and are more complex to setup than the usual NAMFPC namelist in OpenIFS.  In addition, as XIOS is not used operationally with IFS we are unable to provide anything other than support for configuring and building OpenIFS with XIOS.  Users are strongly recommended to post any queries relating to XIOS to the OpenIFS Forums for help from other users.

Further reading:

Namelists

The ecRAD radiation code is now called from the same driver as the 3D model. This by default allows the expensive radiation code to be run on lower resolution grids. The result from this coarser mesh calculation are then interpolated to the full model resolution.  However, this is not desirable for the single column model, where the radiation code should be run on the same grid as the single column model. This should be set explicitly in the namelists &NAERAD and &NAMDIM.

Namelist NAEPHY has two new switches compared to SCM 40r1. They are:

LEPHYS=T,    which activates all of the physics package (default)
LEFLAKE=T,  activates the lake scheme (FLAKE), new in SCM 43r3.

The logical namelist variable LSCMEC has been deleted. The code in the single column model no longer requires this switch. The 3D OpenIFS code only uses it to prevent execution of SCM specific code. The standalone SCM code implicitly assumes it's set true.

The variable CMODID parameter is used as a string to locate the initial files. It doesn't record the version of the model. SCM forcings are not fixed to a given model cycle, these are dependent on the experiment being run.

There are two namelist files in the test-run directory:
           namelist.trref_winds_rel  and  namelist.trref_winds_rel.simpl

The former one defines non-linear physics, the latter one the simplified physics (you may be surprised how close the results from the two packages are). We expect most users to use the full non-linear physics option.

The advection of cloud has been re-activated by setting LWADVCLD=true. If this causes a problem, it can be disabled via the namelist.

Semi-Lagrangian

By default the semi-Lagrangian scheme is activated (LSLAG=.T.) with the physics being averaged along a trajectory (LSLPHY=.T.). This follows the defaults in the IFS.

If you like to use Eulerian advection set LSLAG=.F., reduce the model timestep (don't forget to change number of time steps accordingly) and disable the semi-Lagrangian physics by setting LSLPHY=.F.