Categories: Sciences

Global warming: a digital Earth to simulate it

Scientists are starting to build a digital twin of our planet to better understand global warming developments

A digital twin of our planet was designed to simulate the Earth’s climate system and how it will evolve in the future. It will therefore be a tool to support politicians in taking the most appropriate measures to better prepare for extreme events.

Global warming: 2050 target

To become neutral from the point of global warming by 2050, theEuropean Union launched two ambitious programs: the “Green Deal” and the “DigitalStrategy”. As a key component of their successful implementation, climate scientists and computer scientists have launched the initiative “Destination Earth“, Which will start in mid-2021 and should last for another ten years. During this time, a highly accurate digital model of the Earth. A true digital twin of our planet, to map the development of climate and extreme events as accurately as possible in space and time.

A digital climate system

Earth’s digital twin wants to be a computer system which develops and tests various development scenarios to give more information to policy makers. The observational data over the years will be continuously incorporated into the model in order to make the “Digital Earth”More accurate to monitor the evolution and predict possible future trajectories. But in addition to the observation data used for the climate simulations, the researchers also want to integrate new data on human activities into the model. The new “Earth system” will represent virtually all processes on the Earth’s surface in the most realistic way possible. The influence of humans on the management ofwater, of food and of‘energy and the processes in the physical system of the Earth. In a recent publication in Nature Computational Science, the team of researchers stresses the need for the Earth Science marry the computer science looking for the best challenges and solutions to solve them.

Meteorological and climatic models as a basis

In their article, the researchers look at the steady development of weather models since the 1940s, a success story that took place in silence. THE meteorology they have experimented, as it were, with simulations of physical processes on the largest computers in the world. As a computer physicist, Schulthess, an author of the article, is therefore convinced that today’s weather and climate models can get better and better. To identify new methods for integrating the various technologies, the authors will need to use supercomputers and the most innovative technologies. While the models on the global warming current typically neglect small-scale processes or focus on a few variables, the digital twin will bring together various areas and enable high-resolution simulations of the entire Earth system. To achieve this, i codes simulation programs need to be adapted to new technologies that promise much greater computing power.

Boost efficiency for global warming

Con i computer and algorithms available today, the most complex simulations come to see in detail events within a maximum radius of one kilometer. Research on the global warming it has, in fact, benefited from ever higher performance through generations of processors without having to radically change their program. This performance increase it stopped about 10 years ago. As a result, today’s global warming programs often use only 5% of the performance of current CPUs. To improve this process it will be necessary to be able to develop more hardware and algorithms simultaneously. Scientists also propose to separate the codes to solve the scientific problem from those that perform the computation on the respective system architecture. This programming more flexible it would allow a faster and more efficient transition to future architectures. It will also be necessary share codes and programs to facilitate the work of various research groups on global warming.

Take advantage of artificial intelligence

In the near future, i supercomputer processor-based graphics appear to be the most promising option. The researchers estimate that a large-scale digital twin would require a system with approx 20.000 GPU, which consume approx 20 MW of power. For both economic and ecological reasons, such a computer should be used with zero-emission generated electricity CO2. The authors therefore see great potential in artificial intelligence. L’artificial intelligence it can be used, for example, for the assimilation and processing of data using also graphic and visual methods. Artificial intelligence therefore allows you to speed up simulations and filter the most important information the study of global warming. Furthermore, the researchers assume that the use of machine learning not only makes calculations more efficient, but can also help describe the processes more accurately.