In brief
- StormScope EU is the Europe-focused version of StormScope, a model developed for short-term forecasting.
- Using EUMETSAT satellite observations, the model forecasts cloud conditions across 16 Meteosat channels at 10-minute intervals.
- The interactive demo shown at GTC Berlin updates forecasts as new satellite observations arrive and presents results in map tiles.
- NVIDIA Earth-2 models are also used in applications beyond weather forecasting, including air quality and solar energy forecasting.
In a blog post published on 22 October 2026, NVIDIA introduced StormScope EU as the latest addition to its Earth-2 open model family. Trained on European weather and climate data, the model predicts changes in cloud conditions for short-term forecasts. It uses EUMETSAT satellite observations to generate forecasts at 10-minute intervals across 16 Meteosat channels. NVIDIA is showcasing the model in an interactive weather demo developed with Earthmover at GTC Berlin.
How does StormScope EU work?
StormScope EU is a regional version of NVIDIA’s StormScope model. The demo uses Earth2Studio to run forecasts and Earthmover’s data platform to manage and serve real-time results. Results are updated every 10 minutes as new satellite observations arrive. Users can explore forecasts for Europe through map tiles without downloading the entire dataset.
Earth-2 models can be combined to meet regional requirements. In NVIDIA’s example, StormScope is positioned for short-term forecasting, while CorrDiff adds local detail to coarser weather data. NVIDIA’s open-source PhysicsNeMo framework also provides tools for teams to train and customize models using their own data and expertise. According to the company, this approach can be adapted for environmental applications such as air quality modeling.
Use cases in Europe
The examples shared by NVIDIA show that Earth-2 is not limited to a single forecasting scenario. AXA uses FourCastNet 3 across more than 1,000 simulated seasons to examine regional cyclone risk beyond historical records. Vaisala Xweather combines its own observations, lightning data and forecasting expertise with Earth-2; by adapting StormScope with lightning data, it aims to improve short-term forecasting alerts.
Researchers at the University of Manchester adapted CorrDiff for air quality modeling in the United Kingdom. The model was trained on the NVIDIA-supported Isambard-AI supercomputer and run locally on NVIDIA DGX Spark systems. TotalEnergies is evaluating StormScope-based short-term solar irradiance forecasts at two photovoltaic sites. Space42 used CorrDiff and PhysicsNeMo to develop SARDiff4Height, which predicts surface elevation from a single radar image and also produces uncertainty maps.
What it means for architecture and visualization workflows
StormScope EU is not being presented as architecture or visualization software. However, up-to-date local weather forecasts may be relevant to projects that use climate and environmental data to inform decisions, particularly for teams assessing solar energy and site conditions. NVIDIA’s examples show model outputs being evaluated in fields such as energy, transportation and insurance; no direct integration for architecture firms has been announced.
Teams considering using this kind of data in a visualization or design pipeline should first assess data access, model adaptation requirements and the computing infrastructure needed to run it. NVIDIA says Earth-2 technologies run on its own computing infrastructure; the blog post does not specify pricing, detailed licensing terms or compatibility with particular architecture software. The model should therefore not be assumed to integrate directly into project workflows.
Sources
1 sourceSource texts are not republished; short quotes are marked, everything else is our own summary and commentary.
StormScope EU’s main value for architecture firms lies less in being a feature that can be added directly to design software and more in showing how local weather and environmental data can be customized. The approach is worth following for teams analyzing solar energy or site conditions, but the announcement describes neither integration with architecture tools nor a ready-to-use project workflow.
For firms in Turkey, data access, the NVIDIA computing infrastructure required and licensing terms should be clarified before implementation. Since the announcement does not include pricing or detailed licensing information, its practical benefits should be assessed cautiously until these points are addressed.
Frequently asked questions
What is StormScope EU?
It is a regional StormScope model added to NVIDIA’s Earth-2 family and trained on European weather and climate data. It uses EUMETSAT satellite observations to forecast short-term cloud conditions.
How often does StormScope EU generate forecasts?
The model forecasts cloud conditions across 16 Meteosat channels at 10-minute intervals. Results in the GTC Berlin demo are also updated every 10 minutes as new satellite observations arrive.
Can StormScope EU be used with architecture software?
NVIDIA’s announcement does not specify integration with any particular architecture software. The model is described in the context of weather forecasting, environmental modeling and solar energy forecasting.
