The introduction of the new weather model ICON marks a milestone in Swiss weather forecasting
The Federal Office of Meteorology and Climatology (MeteoSwiss) is operationally launching the cutting-edge weather and climate model ICON (Icosahedral Nonhydrostatic Weather and Climate Model). Replacing the COSMO weather forecast model used previously, it increases the reliability and efficiency of meteorological forecasts in Switzerland.
Accurate and reliable weather forecasts and alerts are crucial for authorities, businesses and private individuals to make timely weather-related decisions. Weather models use three-dimensional computer simulations of changes in the atmosphere in the hours and days ahead to make the calculations needed for these forecasts. Especially in the Alpine region, where local weather events such as thunderstorms and mountain and valley wind systems are a frequent occurrence, these models require a very high spatial resolution. Calculations by MeteoSwiss's weather models (ICON-CH1-EPS and ICON-CH2-EPS) come up with multiple permutations for these forecasts. This means that it is possible to assess the reliability of forecasts and to take probabilities into account when making decisions.
Following an intensive development and testing phase lasting a number of years as part of the ICON-22 project, MeteoSwiss is now making operational use of the ICON weather model, replacing its predecessor, COSMO. ICON provides a variety of benefits, including a more precise reproduction of Swiss topography (which helps enhance the accuracy and quality of weather forecasting) and more efficient computing processes, improving decision-making in critical sectors of the economy such as agriculture, energy and disaster management.
What advantages does ICON have over COSMO?
ICON uses an innovative triangular grid to ensure a higher and more uniform resolution. By mapping Switzerland's complex topography more precisely, ICON makes more accurate weather forecasts than COSMO.
ICON can be used for both regional and global weather and climate simulations. It can also model certain regions, such as the Alps, in more detail in nesting mode. The algorithms deployed by ICON capture the physical processes involved in the weather more effectively than COSMO. ICON affords greater flexibility and can be extended more easily, making it simpler to take on board new scientific findings and technological advances. This facilitates continuous improvement of the model.
New weather model running on a new supercomputer
Like the COSMO model in the past, ICON is optimised for the use of powerful supercomputers at the Swiss National Supercomputing Centre (CSCS). This is a result of the long-standing close cooperation between MeteoSwiss and the CSCS. By using graphics processing units (GPUs) and GPU-optimised software, the supercomputer makes calculations not only significantly faster but also much more energy-efficiently, multiplying the computing power without adding to the cost.
While MeteoSwiss used dedicated infrastructure at CSCS to run COSMO, ICON will now be operated on Alps, CSCS's innovative computing platform, which is spread across the Lugano and Lausanne sites. This platform increases operational resilience (due to being distributed across various sites) and provides almost unlimited computing power, which MeteoSwiss can now harness more flexibly.
Successful international cooperation
ICON is the result of close cooperation between international partners. The ICON partners include the Center for Climate Systems Modeling (C2SM) together with ETH Zurich and MeteoSwiss; the German Climate Computing Center (DKRZ); the German Weather Service (DWD); the Karlsruhe Institute of Technology (KIT); and the Max Planck Institute for Meteorology (MPI-M). ICON demonstrates the real benefits of cooperation between research and national weather services. ICON is publicly available as an open-source solution, promoting innovation and enabling continuous improvement and optimisation of the model across an extensive active community of developers. This close cooperation results in high-quality weather forecasts and climate projections, benefiting both the research community and society.