18.07.2023

The challenges of man-made climate change are global in nature - even if different regions face different impacts. But not everyone is in a position to assess future impacts today and initiate the right adaptations if necessary. Targeted action first requires a sufficient data base - as well as the knowledge and infrastructure to use it accordingly. Knowledge and information about regional consequences of climate change are particularly necessary for adaptation to climate change. How will future heat waves affect agriculture in Ghana, Costa Rica or northern Germany, for example?

Much data is available today (e.g., the European Copernicus program [1]), but often not easily accessible and not with the same accuracy for all parts of the world. With an information system that provides global climate projections at a spatial resolution of up to 1 km, possible climate developments could be better assessed. This includes the development of extreme events such as heat waves, heavy rainfall and storms for any region and different scenarios of a warming world. In addition, societal data, such as population density, industry and agricultural production, will be integrated into the system to provide indications of useful measures to reduce greenhouse gases and adapt to climate change.

Today, for the first time, it is now possible with Earth system models to reproduce the entire global climate system with its interactions at resolutions of up to one kilometer - but so far only for relatively short simulation periods [2]. If climate research had supercomputers of today's highest performance class - so-called exascale systems - at its exclusive disposal, scenario calculations could be carried out over several decades with such models and thus urgently needed data could be calculated.

In addition, the ever-increasing sources of observational data can be better assimilated and processed. If these "data sources" are now combined with the rapid progress of machine learning, a virtualization of the Earth based on huge data streams is possible. The first approaches to this are already being shown by tech giants such as Nvidia [3].

At the Berlin Summit, there was therefore now also a discussion in the scientific environment about how this mass of data and possibilities can be made accessible to a broad range of stakeholders. On the basis of a concept paper [4], in addition to technical solution options, it was discussed which stakeholders could benefit directly or indirectly from EVE.

Cloud services and AI methods are to be used so that the knowledge gained can also be distributed and used beyond the boundaries of the scientific community (cf. ChatGPT). In this way, interactive and visual access to data would become feasible, as Nvidia vividly illustrates using high-resolution ICON climate simulations [5]. Such technical and scientific advances can and should be leveraged by EVE to create a novel global climate information system.

Of course, the question arises as to what technical and human resources would be required to realize EVE. A worldwide network of several multi-national centers is envisaged, each equipped with an exascale computer system and several hundred employees to develop and operate Earth Virtualization Engines. The model is thus similar to the large-scale research facility CERN, where a globally unique multinational technical infrastructure for particle research is developed and operated. Not least for this reason, Charlotte Warakaulle, CERN Director of International Relations, was invited to the Berlin Summit to present the governance and operating model of this international research infrastructure.

The Berlin Summit for EVE was supported by the German Federal Ministry of Education and Research through funding from the German WarmWorld project. The ETH Zurich-led project EXCLAIM, the Hamburg Cluster of Excellence CLICCS, the Horizon2020 project nextGEMS, the Institute of Atmospheric Physics of the Chinese Academy of Sciences, and the Max Planck Institutes for Chemistry and for Meteorology provided further support.

Further information:

[1] https://www.copernicus.eu
[2] https://www.dkrz.de/de/kommunikation/aktuelles/50-jahre-blue-marble
[3] https://www.nvidia.com/en-us/high-performance-computing/earth-2/
[4] https://owncloud.gwdg.de/index.php/s/rNWYNJSdJ19iwbJ
[5] https://www.youtube.com/watch?v=8cQoYcbUG_M