The first version of the ICON Earth System Model

The article describes the coupled ICON model and explores its performance by evaluating a set of standardized climate change experiments following the protocol of the Coupled Model Intercomparison Project (CMIP). The authors used a model configuration at moderate resolution, which allows them to compare the results with MPI-M’s predecessor model MPI-ESM, and other models participating in CMIP. One CMIP experiment is the “historical” simulation running from the beginning of industrialization in the mid 19th century to present. To account for internal variability in the climate system, the authors executed an ensemble of five simulations and found that ICON-ESM performs well in reproducing the observed trajectory of climate change (Figure 1), although some discrepancies remain, such as an overestimation of warming in the Northern Hemisphere in the late 20th century (Figure 1b).

ICON stands for “ICOsahedral Nonhydrostatic” and provides a novel framework for modelling based on innovative design principles. For example, the numerical grids are created by subsequently dividing the faces of an icosahedron (hence “icosahedral”) and avoid problems known from previously used geographical grids, such as singularities at the poles. Moreover, ICON’s numerical concepts are designed to run with very high efficiency at modern computational infrastructure, such as the new “Levante” computer at the German Climate Computing Centre (DKRZ). Over the last years, scientists at MPI-M, the German Meteorological Service (DWD), and other institutions in Germany and Switzerland have developed the component models for atmosphere, ocean, land and the biogeochemical cycle. While the performance of those models had already been explored individually, coupling the sub-system together to form the first ICON Earth System Model (Figure 2) “was a challenging endeavor that required numerous sensitivity experiments to be carried out by a dedicated team of scientists and programmers from all three departments of MPI-M and the partner institutions” says Dr. Johann Jungclaus.