WarmWorld: Development of a new extremely high-resolution global climate model that can realistically simulate the weather in a warming climate for the first time.

Global climate models have steadily improved over the last three decades. Their results depict the development of the global climate with high accuracy and are the basis for all national and international climate protection measures. However, their representations have so far been insufficient to make reliable statements about future regional and local climate development. This applies in particular to extreme events or circulation anomalies, which have occurred frequently in recent years. Thus, there is a lack of reliable statements on the development of weather patterns in a warming climate, which is, however, urgently needed for adaptation to climate change.

The performance of climate models has always been limited by the power of computers. Future developments in High Performance Computing (HPC), both in hardware (exascale class) and software (including Artificial Intelligence, AI) now allow the next step in performance: the development of a climate model that performs simulations of the climate on a one-kilometre grid. This increase in performance makes it possible to describe the decisive small-scale processes, especially in clouds, but also in the ocean, according to physical laws or to parameterise them with new, much smaller-scale approaches than before. With such high-resolution climate simulations, measures for climate adaptation and climate protection can be designed much more effectively at the regional level.

WarmWorld draws on the results of the successful BMBF funding measure HD(CP)² for modelling cloud and precipitation processes (2012-2019) and builds on the convection-resolving atmospheric model ICON, which was further developed in this funding measure. The results from the WarmWorld funding measure will be new, high-resolution, local information products that are of enormous importance for climate impact and adaptation research. German climate research will thus set new standards.

The initiative is designed for two phases: a set-up phase (2022-2026) and an application phase (2026-2029).

The immediate goals of the funding measure are:

• Provision of a model system that is geared towards the new supercomputer generation (Exascale class), using the latest HPC software technologies as well as artificial intelligence (AI), and will thus strengthen Germany as a technology location.
• Significant reduction of uncertainties in the global climate projections, especially for the climate variable precipitation.
• Regional climate projections: Reliable statements on possible developments (depending on the emission pathway) on a regional scale.
• Provision of new, high-resolution, local information products that are of great importance for impact and adaptation research.
• Bringing together the scientific measurement and modelling competencies in Germany in order to optimally utilise the opportunities associated with the comparable spatial resolutions of WarmWorld simulation data with those of global measurement systems (for example from satellites) and to be able to develop new potential uses (also beyond climate modelling).

For the successful development of the climate model, the funding measure builds on an implementation concept that was developed by national and international experts, especially from the field of supercomputing and AI. This concept was also reviewed internationally. It provides for a so-called consortium part with:

• the coordination as well as the development and fine-tuning of critical model components (Module 1 "Better"),
• the transformation of the code base into an open, scalable, modularised and flexible code (Module 2 "Faster"), and
• the development of new methods to make the information (simulation results) visible, accessible and interoperable (Module 3 "Easier").

and an associated research programme (see announcement from 18.01.2023) with:

In this announcement (funding guidelines) that has now been published, associated research projects with a focus on algorithm development, mathematical space-time discretization methods and "in-situ" data analysis, also using artificial intelligence (AI), are to be selected for the "WarmWorld" funding initiative and from 2024 be funded for three years.

The overall technical coordination of WarmWorld lies with the Max Planck Institute for Meteorology (MPI-M) in Hamburg, the German Climate Computing Centre (DKRZ) in Hamburg and the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI) in Bremerhaven.

The BMBF is supporting the funding measure in the start-up phase with around 15 million euros.