2025

Destination Earth (DestinE) Data Portfolio

EUMETSAT

Observation and Context

To support the European Union’s Green Deal and Digital Strategy priorities, policymakers and scientists require highly precise computational tools to monitor, simulate, and adapt to natural and human activities. Historically, environmental data has been heavily fragmented across different systems, making it difficult to predict extreme events or build long-term sustainability frameworks. To address this challenge, the European Union funded the Destination Earth (DestinE) initiative to create a highly accurate, integrated digital replica of our planet. This framework consolidates data management using three core pillars: the Core Service Platform (DESP), the Data Lake (DEDL), and specialized high-resolution Digital Twins (DTs).

Hypothesis

The project coordinators hypothesized that by implementing an integrated open architecture framework — pairing a distributed Data Lake with cutting-edge Earth-system simulation models — they could offer seamless, near-data processing and high-performance computing capabilities. They posited that this unified structure would allow users to dynamically model multi-decadal climate adaptation projections and on-demand regional extreme weather events with unprecedented resolution and speed.

Experiment and Methodology

The experiment relies on a robust engineering framework to systematically compile, catalog, and generate environmental data:

• The Digital Twins: DestinE deploys two primary simulation models. The Weather-induced and Geophysical Extremes DT generates daily global forecasts 4 days ahead at a ~4 km resolution, alongside an on-demand regional component over Europe down to a 500–750 meter resolution. The Climate Change Adaptation DT generates global high-quality multi-decadal projections (~5 km resolution) tracking scenarios from 2020 to roughly 2050.
• Data Lake & Federation Infrastructure: The DestinE Data Lake (DEDL) establishes a shared “Fresh Data Pool” that stores frequently accessed information directly alongside active compute nodes. Simultaneously, it links external registries through data federation, drawing on platforms like Copernicus services (CAMS, CMEMS, CLMS, CEMS), EUMETSAT satellite pipelines, EuroStat, and nextGEMS simulations.
• Data Governance and Maturity Mapping: Datasets are rigorously filtered through a standardized lifecycle scale (ranging from In development and Experimental to Pre-operational and fully Operational). Each asset is cataloged following Spatiotemporal Asset Catalog (STAC) specifications and strict FAIR data principles to verify absolute findability and reusability.

Results and Data

The operational portfolio records comprehensive progress across multi-petabyte datasets:

• Generational Advancements: Generation 1 of the Climate Adaptation Twin successfully delivered approximately 6 PiB of standard and high-resolution data fields using GRIB2 and CCSDS packing formats across models like IFS-NEMO, IFS-FESOM, and ICON. Generation 2 simulations, estimated to scale up to 10–15 PiB of target variables, are scheduled for full integration by the end of Q1 2026.
• Comprehensive Parameter Structuring: Output arrays systematically parse hourly instantaneous atmospheric fields (e.g., 2-meter temperature, wind vectors, dew point, cloud cover fractions) across 19 to 21 distinct pressure levels ranging from the surface up to 1 hPa. These are combined with daily mean sea-ice boundaries and multi-level volumetric ocean salinity and potential temperature fields.
• Standardized Availability Rules: To maximize resource efficiency, standard global extreme forecast data warehouses store a rolling 15-day window of live forecast streams, while primary climate change adaptation scenario pipelines (such as the default SSP3-7.0 trajectory) enforce an absolute minimum storage retention time of 10 years.

Conclusion and Climate Impact

The document demonstrates that the structural orchestration of Destination Earth succeeds in uniting disparate Earth observation sensors and next-generation environmental simulations into a single, cohesive spatial data repository.

By transitioning multi-decadal global climate projections and sub-kilometer weather modeling out of isolated supercomputing environments into an interoperable, public-facing Data Lake, DestinE equips European and global entities with direct decision-making assets. Regional command structures can trigger custom, on-demand simulations to map localized wildfire variables, storm surges, or urban canyon heat islands within a 5-minute output step frequency. This actionable intelligence facilitates immediate civil protection planning, optimizes clean energy grid architectures, and provides concrete, evidence-based guidelines for long-term municipal climate adaptation.