First version of ICON-A model released

The first version of the atmospheric component of the ICON-ESM (ICON-A) has been finalized and documented for use in climate studies by the community. The climate simulated by ICON-A is comparable to its well and long-established predecessor, ECHAM6.3, but comes with the advantage of the more flexible and computationally efficient/scalable ICON framework. This, the first documented release of ICON-A is a major milestone in the institutes development of its new Earth system model (ICON-ESM).

The construction, tuning and performance of ICON-A is published in a two part paper led by scientists from the department „The Atmosphere of the Earth System“ at the Max Planck Institute for Meteorology (MPI-M). Dr Marco Giorgetta and colleagues described the model and the applied tuning strategy (Part I) and Dr Traute Crueger and colleagues (Part II) evaluated it.

ICON-A has a completely new dynamical core, solving the compressible equations of motion in physical (rather than spectral) space and using a vertical discretization in geometric (rather than in pressure) levels. In contrast to ECHAM6.3, by virtue of the new dynamical core, ICON-A is flexible enough to be run at grid spacing ranging from a few tens of meters to many hundreds of kilometers, better suited to the representation of orographic features, and computationally constructed to better leverage high-performance computing. This flexibility can make parametrizations of dynamical processes obsolete in high-resolution experiments. Poor representations of clouds and convection in atmospheric models are at the origin of some of the most outstanding and systematic biases in climate studies. The possibility to resolve these processes with the ICON-A model offers opportunities for new insights into the behavior of the climate system.

In the paper by Crueger et al., ICON-A is configured as a conventional climate model, with parameterized deep convection, and evaluated at two different – still rather low - resolutions (160 km and 40 km grid spacing). As a reference, two ECHAM6.3 versions with similar grid spacing (100 km and 200 km), as well as observations are utilized. Both ICON-A and ECHAM6.3 share similar parameterizations. Especially the 160 km ICON-A experiment provides a compelling representation of the mean climate (see Figure). Overall, the simulated climate is similar to that simulated by the ECHAM6.3 model. Whereas improvements are achieved especially in the northern extra-tropics, in the Tropics, ICON-A has somewhat larger biases.

The results of this first comprehensive evaluation, namely that ICON-A is able to maintain ECHAM6.3’s representation of the climate using a better and more flexible dynamical core, is a major achievement.  Fine tuning and identification of bugs through the course of its further use and development will, if experience is any guide, contribute to continued improvement in the climate of the atmospheric component of the institute’s new Earth system model – ICON-ESM.

Fig.: A performance index is used to assess the simulated mean state of the models. An index of unity denotes the performance as good as the low-resolution ECHAM6.3 version. Values larger than unity represent larger departures from the observations and indicate a worse performance relative to ECHAM6.3 (200 km), and smaller values imply a better performance. The index is based on a comparison of the annual simulated climatology of a bunch of single atmospheric variables with satellite or reanalysis based estimates of the true climatology, weighted with the observed inter-annual variance of each field. The ICON-A (160 km) model performs similarly, or slightly better, than the ECHAM6.3 (200 km) model. It   is especially good in the northern extra-tropics, where it even outperforms ECHAM6.3 (100 km). The performance of ICON-A (160 km) is less satisfactory in the tropics, but not much worse than ECHAM6.3 (200 km). Comparing ECHAM6.3 (200 km) with ECHAM6.3 (100 km) and ICON-A (160 km) with ICON-A (40 km) reveals a strong discrepancy: while the higher resolution ICON-A experiment is clearly worse than its low-resolution counterpart, the opposite is the case for the corresponding ECHAM6.3 configurations.

More information:


Original publications:

1.) Giorgetta, M. A., R. Brokopf, T. Crueger, M. Esch, S. Fiedler, J. Helmert, C. Hohenegger, L. Kornblueh, M. Köhler, E. Manzini, T. Mauritsen, C. Nam, S. Rast, C. Reick, D. Reinert, M. Sakradzija, H. Schmidt, R. Schnur, L. Silvers, H. Wan, G. Zängl, and B. Stevens (2018), ICON-A, the atmosphere component of the ICON Earth System Model. Part I: Model description, J. Adv. Model. Earth Syst., doi: 10.1029/2017MS001242.

2.) Crueger, T., M. A. Giorgetta, R. Brokopf, M. Esch, S. Fiedler, C. Hohenegger, L. Kornblueh, T. Mauritsen, C. Nam, A. K. Naumann, K. Peters, S. Rast, E. Roeckner, M. Sakradzija, H. Schmidt, J. Vial, R. Vogel, B. Stevens (2018), ICON-A, the atmosphere component of the ICON Earth System Model. Part II: Model evaluation, J. Adv. Model. Earth Syst., doi: 10.1029/2017MS001233.


Dr Marco Giorgetta
Max Planck Institute for Meteorology
Phone: +49 (0) 40 41173 358
Email: marco.giorgetta@we dont want

Dr Traute Crueger
Max Planck Institute for Meteorology
Phone: +49 (0) 40 41173 105
Email: traute.crueger@we dont want