The software COSMO members use, falls into two broad categories: official packages and other shared software. The goal of this page is to promote the sharing of both major code and all sort of useful utilities among members.

Not all software can be downloaded from this page (the model itself being the most notable case). Some of it is registered and you have to contact its owners first.

Currently, the following software descriptions are available here:

Int2lm & Cosmo model (COSMO software)

 

Last updated: as in DWD ftp site.

The use of the "Cosmo" model was free of charge for research applications and conditionally for national meteorological Services.

The COSMO-Model reached the end of its life-time and got replaced by ICON-LAM (the limited area variant of global model ICON). So, instead of the COSMO-Model we are now offering licenses for ICON. For ICON, see the licensing terms

In case you are still interested in COSMO, information about eligibility and access, please contact: cosmo-licence@cosmo-model.org.

EXTPAR (COSMO software)

 

Last updated: as on GitHub site (see releases).

EXTPAR (EXTernal PARameters for Numerical Weather Prediction and Climate Application) is used to prepare the external parameter data files that are used as input for the COSMO or the ICON model.

For a general overview see the EGU 2019 poster, user and implementation manual is available here, latest status reports (03.2021, 03.2022).

The source code administartor is Jonas Jucker, working at ETHZ for C2SM.

The software is managed in the GitHub private repo C2SM-RCM/extpar; to get read access to this repo ask the code owner.

To get informed about new releases subscribe to the the extpar mailing list.

Interactive generation of external parameters for the COSMO model is possible through the WebPEP interface (maintained by the CLM community, requires authentication, only old EXTPAR release).

Interactive generation of external parameters for the ICON model is possible through this web service (maintained by the DWD, requires authentication, not updated for all new EXTPAR releases).

Only NetCDF output is provided (fieldextra v14.0.0 or better can be used to translate NetCDF to GRIB).

Fieldextra (COSMO software)

 

Last updated: Feb 2024

Fieldextra is a generic tool to manipulate NWP model data and gridded observations; simple data processing and more complex data operations are supported. Fieldextra is designed as a toolbox; a set of primitive operations which can be arbitrarily combined are provided.

A strong focus of this program is the production environment, with a lot of effort being put in the robustness of the code, in an extensive reporting of exceptions, and in memory and CPU optimization. This program is used, in particular, for production at MeteoSwiss, at DWD, at IMS and for COSMO-LEPS at ECMWF.

Full installation of the latest releases of fieldextra are available at ECMWF, on cca, in /ec/res4/hpcperm/chcosmo/projects/fieldextra (UNIX group cfxtra).

A self-contained package of the latest major production release is available through the link in the "Downloads" section below. This package contains the source files and all external libraries required to build the code. All resources files used at run time are also provided, as well as comprehensive documentation. A set of command line tools, based on fieldextra, is also part of the package.

In order to participate to the development of the code, or to access to the latest development release, you need to have access to the GitHub platform. For that, you first have to create a GitHub account, which is free of charges, and then ask the code owner for access to the fieldextra repository.

To get informed about new releases subscribe to the the fieldextra mailing list.

Information about usage (in particular for ECMWF platform), software primer documentation, documentation of releases, and  more are available at the dedicated fieldextra GitHub wiki.

 

Fieldextra requires a license for users outside of the COSMO Consortium; this license is free of charge for R&D users, but is provided without support.

For R&D users the license can be set between any COSMO partner and the R&D institution (but the COSMO StC and the SCA have to be informed).

The license template is available here.

 

SNOWE (full featured snow analysis software)

 

Last updated: Dec 2016

SNOWE is a standalone software package to prepare the initial snow water equivalent and snow density fields required by the COSMO model. It combines satellite derived snow map, SYNOP standard observations, and COSMO forecast fields within an intermittent assimilation cycle.

Because snow density observations are almost never available, one major issue faced by any snow analysis scheme is the transformation from observed snow height into snow water equivalent. One specific feature of this package is to use observations driven 1-dimensional snow model at each SYNOP location to derive the snow pack density, in order to transform observed snow height into snow water equivalent. This approach has proven to be successful to significantly reduce the T2m bias at the boundary of the snow pack over the Russian teritory.

Roshydromet is further developing and improving this software package. Both the original software resulting from the PT SNOWE (version 1.0) and the most recent software used at RHM (currently version 2.0) are available for download.

 

VERSUS and VAST (verification software)

 

Last updated: Dec 2016

VERSUS, consisting in an Italian Air Force Meteorological Service issued system, is the official verification dedicated tool within COSMO consortium. Versus is a flexible and configurable modular system, able to analyze mathematical models behaviour by studying statistical indexes and time series. This software is used for the production of the common verification products (Common Plots) for the Cosmo Consortium.

Full installation package of the latest releases of Versus is available on Italian Air Force Met Service ftp server (ftp.meteoam.it). The installation package is in the directory VERSUS_INSTALLATION/VERSUS_5.1.8 and it includes the patch_versus-5.1.8.tar.gz file for the patch installation and the versus5.1.8.tar.gz file for the installation from scratch. In order to get the latest available release, you need to have access to the ftp.meteoam.it server. For that, you first have to ask the code owner for access credentials to the server.

 

 

In connection with VERSUS, another verification software (VAST, Versus Additional Statistical Techniques) is available. Its architecture has been produced after a deep study of the IDL code of the Ebert software for fuzzy verification. VAST can be used to verify the precipitation output of the COSMO model according to a multi intensity and multi scale approach.

Categorical scores can be calculated through different upscaling approaches (normal upscale, yes/no, minimum coverage, fuzzy logic, etc.). Non categorical scores are also included (FSS, BSS and ETSr). Output of VAST are png or pdf plots produced through R routines.

 

DACE (data assimilation software)

 

Last updated: Nov 2018

DACE (Data Assimilation Coding Environment) contains the source code for the data assimilation used at DWD. It comprises the operational scheme used for the global model (ICON, EnVar) and the regional model (COSMO, LETKF).

The source code can be accessed via git repository (dace_code) at MPI Met Hamburg. For information on how to install DACE at arbitrary sites see the README.install file in the repository. Significant changes in the code, relevant for the regional data assimilation, shall be collected in the file SIGNIFICANT_CHANGES_FOR_COSMO.

In order to access the latest development release, you need to have access to the git repository. For that, please ask the code owner for access.

 

CALMO meta-model (model calibration)

 

Last updated: Feb 2017 (more recent software is available contact SCA)

The calibration method used by the CALMO project optimizes a global model performance score by adjusting the values of a set of unconfined model parameters. A central element of the calibration process is the so called meta-model, which represents with a simple mathematical function the dependency of some representative model fields on the selected model parameters.

The mathematical function at the core of the meta-model is calibrated by a set of full model simulations over a time period long enough to represent the variability of the atmospheric conditions. Once fully specified, the meta-model supports a fast sampling of the parameter space to find an optimal combination of the model parameters.

In order to participate to the development of the code, or to access to the latest development release, you need to have access to the GitHub platform. For that, you first have to create an account, which is free of charges, and ask the code owner for access to the fieldextra repository.

 

Bufr2netcdf

 

Last updated: Aug 2019

Bufr2netcdf is an open source software package which can convert various types of meteorological observations coded in BUFR format to the netcdf format required by the COSMO model data assimilation system. If used in association with DB-All.e package it can also handle observations coded in BUFR with old ECMWF templates from MARS archive. A brief documentation on how to convert BUFR observation with DB-All.e and bufr2netcdf and use them in the COSMO model can be found in the DB-All.e wiki https://github.com/ARPA-SIMC/dballe/wiki/CosmoNetcdfConversion

Since bufr2netcdf requires a lot of dependencies for a stand-alone build, it is suggested to refer to the SMND helper package (https://github.com/ARPA-SIMC/smnd) for learning different strategies for building and/or deploying the software suite including bufr2netcdf.

 

TERRA standalone (COSMO externalized soil & surface module)

 

Last updated: Mar 2022

TSA is an externalized version of the soil-vegetation-atmosphere transfer scheme of the COSMO model, originally developed by Felix Ament.
It consists of the soil module TERRA combined with a simplified transfer scheme, parameterizations of the radiation interaction at the surface, and the annual cycles of vegetation parameters.
Except for the transfer scheme (the Louis scheme), all components are copied from the operational model.

TSA can be computed at a single point or on a geographical domain.
It requires the associated external parameters, the atmospheric forcing, and some initial conditions.
It uses the GRIB1 format for input/output files (or some ad hoc format for atmospheric observations).

See PT Terra SAnta final report for more information (including validation and spin-up evaluatioan, pdf 0.8MiB).