Digital twins help prepare for climate change

Heatwaves, floods, and wildfires are among the extreme weather events which are often likely to be linked to climate change. A new initiative will use models of Earth sub-systems, so called digital twins, to assist authorities and policy makers in preparing for the future conditions. Named Destination Earth the initiative will help predict major environmental degradation and disasters.

Destination Earth will be funded through the European Commission’s Digital Europe program over the next seven years. The initiative will start this year, and implementation will be delegated to three international organizations: the European Center for Medium-range Weather Forecasts (ECMWF), European Space Agency (ESA), and European Organization for the Exploitation of Meteorological Satellites (EUMETSAT).

Destination Earth will integrate digital replicas of various aspects of the Earth’s system on topics such as extreme weather events, climate change adaptation, oceans, and biodiversity.

“If we can simulate everything that’s going on out there as best as we can with the best possible models run with the best possible specs (specifications, ed.) that are imaginable right now, supported by all the observations we have from satellites, and even new observations like from mobile phones, then you would have an instrument at your disposal which allows you to test scenarios of how to respond to climate change impacts, how to prevent and manage risk in response to extremes,” explains Dr. Peter Bauer, Deputy Director of Research at ECMWF.

Two pillars of computation

One of the European resources which will become available to Destination Earth is the GPU (Graphic Processing Unit) accelerated supercomputer LUMI operated by Finnish national research and education network CSC.

“LUMI is a very attractive platform with respect to a number of aspects that Destination Earth will rely on. One is the actual supercomputing task, so accelerated computing of, for example, Earth system models. We must rely on GPUs because they promise acceleration in very critical and costly parts of the Earth system model. It’s a computing-intensive effort, and we believe that only GPU accelerated computing can deal with that,” Dr. Peter Bauer points out.

Another aspect is machine learning, he adds: “A lot of information to be extracted for specific applications can only be done by machine learning. These two pillars of extreme-scale computing to accelerate our Earth system model simulations and the very heavy training aspects of machine learning are the crucial bits where LUMI will be invaluable for Destination Earth.”

Showing the value of supercomputing

The access to LUMI is related to another European program, European High-Performance Computing, EuroHPC for short.

“We need a dedicated allocation under EuropHPC for Destination Earth to be successful. But then it also works the other way around (…) The European community and the individual countries have made a significant commitment to HPC (…) that requires proof that the investment has been worth doing. Protecting our societies from extremes or dealing with the consequences of extremes and natural hazards and adapting to climate change is probably the number one issue right now in terms of the importance of applications apart from, of course, Covid and pandemic disease research and health aspects. It’s certainly at the top of applications and an excellent opportunity to show value,” Dr. Bauer concludes.

The above story is inspired by Anni Jakobsson’s article “Destination Earth and digital twins – a European opportunity for HPC”, first published on the LUMI website. Read the original article.