Accept and continue

We use cookies on this site to enhance your user experience. By clicking the right button you are giving your consent for us to set cookies. No, give me more information.


Numerical Model Training Course


Deutscher Wetterdienst (DWD), the Climate Limited-area Modelling (CLM) Community and Karlsruhe Institute of Technology (KIT) kindly invite you to a scientific training course on numerical models. The aim of the course is to familiarize new users with the nonhydrostatic numerical weather prediction modelling (NWP) framework ICON and the regional climate modelling (RCM) framework COSMO-CLM. The Numerical Model Training Course takes place from Monday 30 March 20 to Friday 03 April 2020.


blue arrow Extended Agenda ICON-NWP Course

blue arrow Time Table


The Training Course will take place at the DWD Training Center in Langen, Germany.



All participants have to book their hotel rooms by themselves. The "Achat Comfort Hotel Frankfurt Airport" in Langen is located close to the Training Center. Here we organized block reservations. Please make your reservation directly with the hotel using the
booking code: DWD-Training-Course 2019 (only this grants the special price). For more information, please see:

blue arrow Hotel Website


Contents of the Training

The main focus of the ICON/NWP practical training will lie on ICON's limited area mode (ICON-LAM). ICON-LAM is of interest especially for national meteorological services and scientific users who are planning to run ICON for regional domains with LBC data from DWD or ECMWF at their home sites.
The main focus of the COSMO-CLM/RCM practical exercises is to learn how to perform regional climate studies using COSMO-CLM. The exercises cover all relevant topics starting from the preparation of the data up to visualization of the results.
Please find a preliminary schedule.


The deadline for registration is 18 January 2019. Applicants will be informed until 25 January 2019. A registration form is accessible from the web page: registration form

Please note that the ICON/NWP and COSMO-CLM/RCM practical exercises will be at the same time. It is only possible to attend one of them. Therefore you have to choose between these options in the registration form. The course is directed towards experienced scientists (PhD or post-docs) and it is limited to at most 60 participants in total (30 ICON/NWP, 30 COSMO-CLM/RCM).


Theoretical Lessons (preliminary planning)

Most of the theoretical lessons are joint lectures for ICON/NWP and COSMO-CLM/RCM.

• Getting Started
• Dynamics
• Physics Overview (ICON/NWP)
• Dynamics on long time scales (CCLM/RCM)
• Advection
• Microphysics
• Clouds & Convection
• Radiation
• SSO & Gravity Wave Drag
• Turbulent Transfer
• Turbulent Diffusion
• Lake & Sea Ice
• Soil & External Parameters
• Verification (ICON/NWP)
• Evaluation (COSMO-CLM/RCM)
• Aersols and Reactive Trace gases – ART


Practical Exercises (preliminary planning)

The practical exercises differ between the ICON/NWP and the COSMO-CLM/RCM training course.

• Running idealized test cases
• Preparing initial data
• Running real cases (global)
• Preparing lateral boundary conditions for limited-area simulations
• Running the limited-area mode (ICON-LAM)
• Programming ICON
• Visualizing ICON data

• Preparing external, initial and boundary data
• Running the COSMO-Model in Climate-Mode
• Visualization and analysis of NetCDF Files
• Troubleshooting in climate mode
• Individual exercises


Additional Courses

Please note: If insufficient numbers of students choose a particular course, then it may not be viable to run that course in 2019.

Aerosols and Reactive Trace gases extension (ICON-ART,conducted by KIT)

ICON-ART can be used to simulate the atmospheric composition and the feedback of trace gases and aerosol particles on atmospheric processes (radiation, cloud formation). The course will provide an overview on the capabilities of ICON-ART, the theoretical background and practical exercises. This additional course is conducted by KIT.

How to install ICON (including necessary libraries, conducted by the national weather service of Romania, NMA)

During this additional course, the participants will learn how to set up their environment for a successful ICON (not COSMO-CLM) installation. This includes all the necessary library packages like, e.g. OpenMPI, NetCDF, eccodes. This additional course is conducted by the national weather service of Romania, NMA.

Accelerated dynamical core of COSMO (conducted by the Swiss national weather service, MeteoSwiss)

An optional introductory course on compiling and running with the new accelerated dynamical core will be held as part of the Numerical Model Training Course 2019. The (C++) accelerated dynamical core is a new component of the COSMO model that can be used to run faster on traditional CPU system and which is required to run the model on heterogeneous GPU systems. The target group of this course are members of institutions that have already access to a heterogeneous GPU system or are planning to acquire a heterogeneous GPU system. This additional course is conducted by the Swiss national weather service (MeteoSwiss).



The lessons, presentations and exercises will be held in English language. All slides will be in English. We expect a certain familiarity with:
• Unix / Linux basic commands (cd, ls, grep, find, chmod, ssh, Shell-scripts, ...),
• FORTRAN 90/95/2003, and
• Basic knowledge in geophysical fluid dynamics.



For further questions regarding organizational matters, lodging, and registration please contact:



blue arrow The course materials of the latest course can be found below.



Please note that we expect a certain familiarity with Fortran 90/95/2003, with the basics of command line driven usage of a Unix / Linux system and with some basic tools (shell-scripting, emacs, vi, ...). If you are not yet familiar with the latter, you can find links to some introductory tutorials under "Additional Material" below.




Material from COSMO/CLM/ART Training Course 2018

The 11th COSMO/CLM/ART Training took place from 12th  - 16th March 2018 in the Meteorological Training Centre in Langen near Frankfurt/Main. The course was targeted towards new users from universities and research institutions (PhD-students, postdocs, etc.), from COSMO members and from other National Weather Services (NWS) - COSMO licensees or developing countries - who are running the model operationally. We invited students and scientists interested to work or already working with COSMO/ CLM to participate in this Training Course. Gruppenfoto_TR18


The COSMO-Model together with the preprocessing package is a state of the art, unified, nonhydrostatic model system for simulating the processes of the atmosphere, for numerical weather prediction (NWP), regional climate modelling (RCM), and dispersion modelling (ART - Aerosol and Reactive Trace gases). The COSMO-Model in RCM mode allows for coupling of the Community Land Model via the OASIS coupling software, as an alternative soil- and vegetation model to the standard TERRA_ML module.

Lectures on the different components of the model (dynamics and numerics, physical parameterizations, data assimilation) have be offered by the model developers with emphasis on standard applications as well as idealized simulations. The practical exercises offered the opportunity to compile the programs, run the model and analyse the output of NWP-, RCM- and ART runs using standard scripts.


The course was devoted to provide a basic training on the theory and usage of the model in NWP and RCM mode as well as in an idealized framework. Some extra lessons (see below) were given on Friday afternoon March 16th, 2018.

blue arrow More informations can be found here



Training materials


I. Theory Lessons (THE)


1. Model Overview by Schättler (Presentation)

2. Dynamics and Numerics by Baldauf (Presentation)

3. Physical Parameteriazations

4. Verification for NWP by Fundel (Presentation)

5. Data Assimilation for NWP by Schraff (Presentation)

6. Dynamics on long time scales by Will

7. Evaluation of CLM runs by Anders (Presentation)

8. Aerosols, Clouds and Radiation by B. Vogel et al. (PresentationI/PresentationII)


II. Practical Exercises

The schedule of the practical exercises is included in the invitation and the attached program. The COSMO_tutorial describes the compilation of the source code of the model, the structure of the required input data, and several options for running the model in the NWP-mode and in the climate-mode. The tutorial contains descriptions of exercises, tips for troubleshooting and data visualization of the model output in the GRIP and NetCDF formats. Additional exercises are offered for advanced users.


The practical Exercises have been coordinated by the following colleagues:


NWP-team: U. Schättler, D. Rieger, M. Junk

RCM-team: B. Rockel, M. Tölle, C. Steger, I. Anders, A. Will

ART-team: B. Vogel, H. Vogel, K. Deetz, S. Gruber

ITC-team: M. Baldauf, U. Blahak

0. Tutorial

1. NWP: Installation of the Model System (Presentation) U. Schättler / D. Rieger/ M. Junk

2. NWP: Preparing external, initial and boundary data - Running INT2LM (Presentation) U. Schättler / D. Rieger/ M. Junk

3. NWP: Running the COSMO-Model in NWP- / Climate-Mode (Presentation) U. Schättler / D. Rieger / M. Junk

4. NWP: Visualization and analysis of GRIB / NetCDF Files (Presentation) U. Schättler / D. Rieger / M. Junk

5. NWP: Trouble Shooting in NWP- /Climate-Mode (Presentation) U. Schättler / D. Rieger

6. Individual exercises

7. Running Idealized Cases (Presentation, Documentation, Help on runscript framework for real cases and idealized casesM. Baldauf, D. Rieger / U. Schättler / U. Blahak

8. Accelerated dynamical core (optional)

III. Configuration of the model system


The COSMO model system has more than 150 configuration parameters which control the model physics, dynamics and input/output. Several evaluated standard configurations are offered for the COSMO-CLM setup (see the above RCM presentation under 7.).


All namelist parameters are documented in the User Guides for COSMO and INT2LM. There is also the namelist tool, which offers powerful tools for namelist documentation, configuration comparisons etc.


IV. Additional Material

1. Additional namelist documentation

2. Extended Model Documentation

3. CCLM Starter Package (logged in members only)

4. Please revise your unix skills, e.g.

4.1 Unix tutorial of Univ. of Surrey:

4.2 Unix tutorial of DWD-KU: (pdf)

4.3 Unix command overview for beginners:



blue arrowPrevious training course (archive, login necessary)