Weather and climate are inextricably linked through fundamental multiscale processes and their nonlinear interactions. Although numerical weather prediction (NWP) and climate forecasts/projections Traditionally developed separately, mainly due to different spatial and temporal scales, this picture is changing. New capabilities in high-performance computing (HPC) allow high-resolution simulations for increasingly long integration times, blurring the distinction between weather and climate modelling. It is in this context that a global to regional modelling system has been developed Icosahedral Nonhydrostatic (ICON).

Vertical integration of model configurations A new initiative has been launched since 2020, focusing on "vertical" integration NWP, climate predictions, climate projections, and atmospheric composition modeling within the ICON framework. The goal is to uniformly handle sub-scale parameterizations. Although ICON NWP models and climate applications have shared the same dynamical core, they have differed significantly in their physical parameterizations and in the coupling of Earth system components. ICON's unique design facilitates the development of a fully integrated modeling system. All components—atmosphere, ocean, land, and sea ice—use the same type of horizontal grid (albeit with different sizes) and the same arrangement of variables. This common grid helps in budgeting and reducing interpolation errors between model component interfaces.

Key components and their integration The ICON integrated model consists of ICON NWP for the atmosphere, ICON Ocean, ICON ART for aerosols and transport of trace elements and suitable terrestrial components (such as JSBACH a TERRAAlthough the parameterizations of physical processes have historically differed between weather and climate configurations (e.g., different schemes for radiation, cloudiness, convection, or diffusion), efforts are underway to converge the number of physical parameterizations using the same atmospheric model for weather and climate.

For example, within the land models, TERRA has been used for NWP, optimized to reproduce the diurnal cycle of temperature and humidity, while JSBACH takes into account processes relevant to longer time scales of climate, such as biogeochemistry and the full carbon cycle. Integration of these modules into a single framework for modeling processes at the land surface is underway. Similarly, the ocean component ICON O (previously developed only for climate modeling) is now integrated into weather forecasts. For atmospheric composition, a ICON ART, which is used for pollen and mineral dust forecasts and will be integrated for climate time scales in the near future.

Climate predictions and future projections The primary applications of the integrated ICON are climate forecasts (from seasonal to 10-year forecasts) and coupled atmosphere-ocean forecasts on a daily basis. ICON is intended to replace the previous DWD climate forecast system and serves as the basis for Germany's contribution to CMIP phase 7 (CMIP7).

Initial results are promising. For example, in pre-industrial control experiments (DECK) there was a significant decrease in the tendency of the zonal average zonal wind in the subtropical circulation regions of both hemispheres, from 10 m s-¹ in ICON ESM to 5 m s-¹ in ICON XPP (the ICON climate configuration). Similarities in precipitation errors between one-day weather forecasts and long-term climate averages also suggest that rapid and local physical processes influence the long-term average climate. This integration allows for a better understanding of the interactions between short-term weather events and long-term climate characteristics.

Overall, ICON's integrated modeling approach opens up a new common space for basic research, which combines weather forecasting and climate research. Although a completely unified configuration for all time scales is ambitious, bringing the physical states in NWP closer to the climate and vice versa is a realistic goal that promises to improve forecasts and extend the range of weather forecasts through coupled configurations. Spring

Glossary of key terms

• ICON (Icosahedral Nonhydrostatic): A global model system developed for applications of numerical weather prediction (NWP), climate forecasts and projections, which shares a common dynamical core.
• Numerical Weather Prediction (NWP): The use of mathematical models of the atmosphere and ocean to predict weather conditions over short to medium periods (days to weeks), usually solving an initial value problem.
• Climate forecasts: Predictions of climate conditions for the near future (months to decades), which may include the influence of initial conditions.
• Climate projections: Simulations of future climate conditions on longer time scales (decades to centuries), usually solving the boundary value problem with respect to changes in forcing factors.
• Vertical integration: A concept in ICON modeling where different components of the Earth system (atmosphere, ocean, land, sea ice) use the same type of horizontal grid and common data structures, which minimizes interpolation errors and streamlines development.
• Dynamic core: The mathematical and numerical basis of a model that describes the movement of fluids (atmosphere, ocean) at different scales.
• Subgrid-scale parameterization: Representation of physical processes that are too small to be directly resolved by a model grid (e.g. clouds, turbulence), using simplified equations or statistical relationships.
• HPC (High-Performance Computing): The use of supercomputers and parallel architecture to perform complex calculations that are essential for models with high resolution and long integration times.
• German Weather Service (DWD): The German Meteorological Service, which is a partner in the ICON initiative.
• Max Planck Institute (MPI) for Meteorology: A German research institute that is a partner in the ICON initiative.
• ICON NWP: An ICON model configuration specifically designed for numerical weather prediction.
• ICON Ocean (ICON O): Ocean component of the ICON model system.
• ICON ART: ICON model configuration that includes atmospheric aerosols and reactive trace gases.
• JSBACH: A land surface model used in ICON climate applications that includes biogeochemistry, carbon cycling, and dynamic vegetation.
• TERRA: A land surface model used in NWP ICON applications, optimized to reproduce the diurnal cycle of temperature and humidity.
• Data assimilation: The process of integrating observations into a model to improve the initial conditions for predictions.
• EnVar (Ensemble–Variational): A type of data assimilation scheme that combines the advantages of ensemble and variational methods.
• LETKF (Localized Ensemble Transform Kalman Filter): Type of ensemble Kalman filter used for data assimilation.
• EnKF (Ensemble Kalman Filter): A type of ensemble filter used for data assimilation, often for ocean observations in climate predictions.
• Nudging: A data assimilation method in which the state of the model gradually approaches the observed or reanalyzed data.
• CMIP (Coupled Model Intercomparison Project): An international model intercomparison initiative that coordinates and standardizes climate model experiments.
• ICON XPP (ICON extended predictions and projections): A prototype of the ICON integrated model being developed as a basis for climate predictions and projections.
• ERA5: A global weather and climate reanalysis developed by the European Centre for Medium-Range Weather Forecasts (ECMWF), used as reference data.
• Zonal-mean zonal wind: Average wind speed in a west-east direction, averaged over length.
• Equilibrium climate sensitivity (ECS): The warming that would occur if atmospheric CO2 concentrations were doubled and the system reached a new equilibrium.
• Transient climate response (TCR): The immediate warming that would occur at the time of doubling of CO2 concentration at a linear increase of 1% per year.