Resources

Below you will find curated multimedia and reading materials that showcase the science, people and progress behind Weather Generator.

Each category features brief video introductions — perfect for sharing or embedding in lectures — and links to deeper technical content.

Introductory Videos

Introducing the WeatherGenerator project, its goals and why it is needed.

Inroductory Animation Video

Introductory video

Science Explainers

Explaining background knowledge and the methods being used within the WeatherGenerator project.

Even Nordhagen: Can Weather Models Predict Power Prices?

Event Videos

Bringing together the WeatherGenerator community to collaborate, interact and work on common goals of the project.

Video Highlights from Oslo: General Assembly, Dissemination & Stakeholder Day, Coding Workshop

Highlights from Our First Hackathon in Bonn

About the Themes

The overall project is organised in four Themes: Data, Model, Application and Service. Find out more about each of them, what they work on and how they interact.

Work Structure Introduction

Theme 1: Data

Theme 2: Model

Theme 3: Application

Theme 4: Service

Women Shaping Science

Female scientists from across the WeatherGenerator project share their personal experiences in the scientific world, offering honest reflections on their paths, challenges and achievements.

Jésica Piñón Rodríguez

Florentine Weber

Publications

https://arxiv.org/abs/2510.22265 This link opens in a new window • Dec 22, 2025
Error bounded compression for weather and climate applications
Error bounded compression for weather and climate applications
Developed EBCC as novel error-bound compression method for weather data, with paper published on ArXiv and submitted to PNAS.
https://arxiv.org/abs/2511.23043 This link opens in a new window • Dec 22, 2025
High-Resolution Probabilistic Data-Driven Weather Modeling with a Stretched-Grid
High-Resolution Probabilistic Data-Driven Weather Modeling with a Stretched-Grid
A probabilistic AI weather model for high-resolution, spatially coherent ensemble forecasts.

Deliverables and Reports

Every EU funded project has to keep records and other supporting documentation in order to prove the proper implementation and the costs declared as eligible. Here we present the deliverables of the project categorized by Theme: Data, Model, Application and Communication.

D1.1 • Jul 31, 2025
Data Management Plan
Opens a PDF document
D1.2
Preliminary report on dataset compression and data optimisation methods and their effects on training
Opens a PDF document
D1.3
Documentation of the first version of the training datasets of the WeatherGenerator including a description of the data
D1.4
Report on simple and efficient data loading via a uniform API including a list of datasets on the different EuroHPC machines and DestinE data bridges
D1.5
Documentation of the final version of the training datasets of the WeatherGenerator
D1.6
Report on ethical aspects of the data used and generated by the WeatherGenerator
D1.7
Code and documentation for dataset compression and optimisation methods
D2.1
Report on deployment efforts for the WeatherGenerator
D2.2
Report on online learning of the WeatherGenerator within DestinE
D2.3
Report on the coupling of the WeatherGenerator to machine learned ocean, sea-ice and wave models

Resources

Videos

Introductory Videos

Inroductory Animation Video

Introductory video

Science Explainers

Even Nordhagen: Can Weather Models Predict Power Prices?

Event Videos

Video Highlights from Oslo: General Assembly, Dissemination & Stakeholder Day, Coding Workshop

Highlights from Our First Hackathon in Bonn

About the Themes

Work Structure Introduction

Theme 1: Data

Theme 2: Model

Theme 3: Application

Theme 4: Service

Women Shaping Science

Jésica Piñón Rodríguez

Florentine Weber

No videos found!

Publications

Error bounded compression for weather and climate applications

High-Resolution Probabilistic Data-Driven Weather Modeling with a Stretched-Grid

Deliverables and Reports

Data Management Plan

Preliminary report on dataset compression and data optimisation methods and their effects on training

Documentation of the first version of the training datasets of the WeatherGenerator including a description of the data

Report on simple and efficient data loading via a uniform API including a list of datasets on the different EuroHPC machines and DestinE data bridges

Documentation of the final version of the training datasets of the WeatherGenerator

Report on ethical aspects of the data used and generated by the WeatherGenerator

Code and documentation for dataset compression and optimisation methods

Report on deployment efforts for the WeatherGenerator

Report on online learning of the WeatherGenerator within DestinE

Report on the coupling of the WeatherGenerator to machine learned ocean, sea-ice and wave models

WeatherGenerator software development roadmap

Pre-trained and documented prototype of WeGen-Atmo-S

Pre-trained and documented prototype of WeGen-Land

Report on technical implementation and performance optimisation results of the WeatherGenerator

Preliminary WeatherGenerator evaluation results

Updated development roadmap for the WeatherGenerator

Pre-trained and documented WeGen-Atmo-S and WeGen-Atmo-M model

Pre-trained and documented WeGen-Land model

Pre-trained and documented WeGen-Atmo-L model

Final release of evaluation & explainability tool suite

Reinforcement-learning training protocol for long-term predictions and experience with online learning for DestinE digital twin model output

Pre-trained coupled WeatherGenerator and documentation of decoder and tail networks across application domains

Comprehensive evaluation report comparing the capabilities of the WeatherGenerator across model sizes including the coupled model

Report detailing each application

Report evaluating the use of AIFS in applications that rely on output of a weather forecast model

Report detailing the use of WeatherGenerator output in applications

Report on the first version of machine learning solutions with tail-networks

Prototypes of full applications using the WeatherGenerator defined with data, software and documentation

Report on the revised version of machine learning solutions with tail-networks

Report comparing results from the WeatherGenerator against benchmarks and the integration of the WeatherGenerator in application workflows

Plan for dissemination and exploitation including communication activities

First Dissemination Workshop

First Gender Equality Workshop

First external Hackathon

Second external Hackathon

Second Gender Equality Workshop

Second Dissemination Workshop

Third external Hackathons

Interim report on provided services, evaluating scientific achievements of finished projects, the added value of the WeatherGenerator, and progress of ongoing projects

Report containing service descriptions with a few highlights, evaluation statistics by service applicants, lessons learned, and feedback to the core model development team

Risk and Quality Management Plan

Exploitation Plan (EP): Plan identifying types of potential pathways of market-oriented exploitation

Exploitation Plan after closure: Strategy for defining measures for exploitation "after the product" phase

Policy Brief 1: Overview of WeatherGenerator

Policy Brief 2: Final overview of WeatherGenerator