Imagine being able to test what would happen to your city's infrastructure if global temperatures rise by 2.5 degrees Celsius. Or simulate how a new seawall would perform against storm surges predicted for 2060. Or model whether shifting agricultural zones could feed a growing population in southern Europe by 2050.
This is the promise of Destination Earth, the European Union's most ambitious climate technology project. Known as DestinE, it aims to create a complete digital replica of the Earth's climate system at a resolution fine enough to be useful for regional and even local decision making. Phase 1 was delivered in June 2024. The full system is targeted for completion by 2030.
What a Digital Twin of Earth Actually Means
A digital twin is a virtual model that mirrors a real-world system in near real time. The concept has been used in aerospace and manufacturing for years: Rolls-Royce maintains digital twins of its jet engines, updating them with live sensor data to predict maintenance needs. DestinE applies this logic to the entire planet.
The system integrates data from the Copernicus Earth observation program, weather stations, ocean buoys, atmospheric sensors, and ground-based monitoring networks. AI and machine learning models process this data alongside traditional physics-based climate simulations to create a living model that can be queried, tested, and projected forward in time.
The initial digital twins focus on two domains: extreme weather events and climate change adaptation. The weather twin simulates natural hazards like floods, droughts, and heatwaves at unprecedented resolution. The adaptation twin models longer-term climate trajectories and their effects on food production, water availability, and coastal infrastructure.
The Supercomputers Behind It
DestinE runs on two of Europe's most powerful supercomputers: LUMI in Finland and MareNostrum 5 in Barcelona. LUMI, operated by CSC Finland, delivers over 550 petaflops of computing power and was ranked among the top five supercomputers globally when it came online. MareNostrum 5 at the Barcelona Supercomputing Center adds complementary capacity for data-intensive AI workloads.
Three EU entities coordinate the project. ECMWF handles the climate and weather modeling core. The European Space Agency (ESA) manages satellite data integration. The European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) provides real-time observational data feeds. The total budget for the initiative through 2027 is approximately 315 million euros.
Resolution That Changes the Game
Standard global climate models operate at resolutions of 50 to 100 kilometers. At that scale, individual cities, river valleys, and mountain ranges are smoothed into averages. DestinE aims for kilometer-scale resolution, which means features like Alpine valleys, coastal estuaries, and urban heat islands become visible in the model.
This matters because climate impacts are deeply local. A mayor deciding where to invest in flood defenses needs to know which neighborhoods are at risk, not just that the region faces higher precipitation. A farmer needs to understand how changing rainfall patterns will affect their specific watershed, not the country average.
The Phase 1 system demonstrated the ability to run global simulations at approximately 4.4 kilometers resolution for weather and 5 kilometers for ocean dynamics. Users can interact with the system through the DestinE platform, a cloud-based interface that allows researchers and policymakers to run queries and visualize results without direct access to the supercomputers.
Where AI Fits In
Traditional climate simulation is computationally expensive. Running a full Earth system model at kilometer scale for a century of simulated time would take years even on LUMI. AI accelerates this process in several ways.
Machine learning models can emulate the behavior of expensive physics calculations, producing approximate results in a fraction of the time. Neural networks trained on high-resolution simulation data can downscale coarser models, adding realistic local detail without rerunning the full physics. AI also handles the data assimilation pipeline, processing the firehose of observational data from satellites and sensors and integrating it into the model state.
ECMWF has been at the forefront of incorporating AI into weather and climate modeling, and DestinE benefits directly from that expertise. The center is experimenting with using models similar to DeepMind's GenCast alongside its traditional Integrated Forecasting System, combining the speed of AI with the physical rigor of numerical simulation.
The Road to 2030
By 2030, the full DestinE system is expected to provide on-demand simulations across multiple domains including atmosphere, ocean, land surface, ice sheets, and human infrastructure. The EU envisions it as a tool not just for scientists but for urban planners, insurance companies, energy grid operators, and agricultural policymakers.
The project also raises important questions about data governance, access, and equity. Europe is investing heavily, but the climate crisis is global. The question of whether and how DestinE's capabilities will be shared with developing nations, many of which face the worst climate impacts, remains open. The EU has signaled interest in international cooperation but has not yet outlined specific data-sharing agreements.
What is clear is that DestinE represents a new paradigm. Rather than studying the climate as a collection of separate processes, it treats the entire Earth system as an integrated, queryable model. If it works, it will give humanity something it has never had before: the ability to look decades ahead and ask, with precision, what is coming.
Sources: European Commission, Destination Earth initiative official documentation; ECMWF DestinE technical reports (2024); ESA Destination Earth contributions; EuroHPC Joint Undertaking LUMI supercomputer specifications; Barcelona Supercomputing Center MareNostrum 5 documentation.
