Core documentation of the COSMO-model

Last updated: 10 May 2022

The documentation files of the COSMO model is split in seven parts, (made available since 2003). These parts are:

No Title Version
PART I Dynamics and Numerics September 2021
PART II Physical Parameterization September 2021
PART III Data Assimilation September 2021
PART IV Implementation Specifics Not available
PART V Preprocessing (INT2LM) September 2021
PART VI Model Output and Data Formats for I/O September 2021
PART VII User's Guide September 2021
Ideal Idealized simulations January 2015

Part I-III are for the scientific documentation, which is independent of the code itself. In particular, Part VII has a supplementary part, which describes the Feedback File Definition. Below, you will find links to the available documents (in pdf format).

Documents that complement these "official" guides, i.e. reference key model components, can be found in the publications page

DOIs for Documentation of Historic Model Releases

Because some scientific journals now require that all citations in a publication are placed in a long-term storage and do at least contain DOIs, if not peer-reviewed, you can also access the historic documentation for special model releases. Links to these documents can be found below. All documents containing a DOI are marked accordingly.

The documentation containing DOIs can also be found at DWD.

Part I: Dynamics and Numerics

This part of the documentation provides information on the theoretical and numerical formulation of the model covering both the basic thermodynamics and the particular numerical schemes to solve the model equations.

Part II: Physical Parameterization

This part of the documentation reviews shorly the basic subgrid-scale physical processes and describes in detail the corresponding parameterization schemes.

Part III: Data Assimilation

The data assimilation part of the documentation describes the full analysis system of the COSMO-package. It is made-up of four components:

  1. the nudging analysis scheme
  2. the sea-surface temperature analysis
  3. the snow depth analysis
  4. the soil moisture analysis

Part IV: Implementation Specific Details of the COSMO-Model

Not Available

Part V: Preprocessing (int2lm)

For operational applications, COSMO is nested in the global models GME (until January 2015) and ICON (starting January 2015) of DWD. Initial and boundary conditions are provided by the pre-processing program INT2LM for analyses and forecasts of GME, ICON or the ECMWF global model IFS. A detailed description of how to work with GME, ICON or IFS data is provided.

Part VI: Model Output and Data Formats for I/O

This part of the documentation provides information about the output of the COSMO-Model. First, there are the output fields (in GRIB or NetCDF). Besides the prognostic variables several diagnostics are computed in the model and can be written (e.g. 2m-temperature, wind gusts, height of snow melt, etc.). Most atmospheric fields can also be interpolated to p- or z-levels. Another type of output are several ASCII output fields, that are also described. This guide also contains the description of the output from data assimilation (nudging).

In addition, the possible output formats GRIB (1/2) and NetCDF are described briefly.

Part VII: User's Guide

The user guide of COSMO provides information on code access as well as hardware and software requirements for executing the code. Short tutorials on how to install and run the model are included. The user guide contains a detailed description of all control parameters in the NAMELIST input file. These control parameters allow a flexible set-up of model runs for various applications.

The Feedback File Definition is a supplementary documentation which describes the format of the NetCDF feedobs or feedback file. This file is produced by the data assimilation part of the COSMO model and can be used for LETKF or verification purposes.

Idealized simulations

This document describes how to set up idealized simulations with COSMO. This mode is useful for researchers doing process studies, but also for model developers when testing new features. The possibilities range from simple 2D flow-over-idealized-hill simulations, 3D idealized convective cell experiments (e.g., Weisman-Klemp-type), LES-like studies with very high spatial resolution to more complex setups using realistic orography and land use.

For each 'component' of a model run (orography, other external parameters, initial (thermo)-dynamic profiles, surface fluxes, model boundary conditions, artificial convection triggers, etc.) there are different choosable options (like a kit with lots of toys), which can be more or less freely combined via extensive namelist parameters.

The documentation is in a draft stage, but you can already find a complete namelist parameter documentation, and the 'cookbook' (together with many exemplary model runscripts in subdirectory RUNSCRIPTS of the code distribution) should get you started.