AI
WeatherQuest
By
The AI Weather Quest, organised by the European Centre for Medium-Range Weather Forecasts (ECMWF), is an ambitious international competition designed to harness artificial intelligence (AI) and machine learning (ML) in advancing weather forecasting. It challenges participants to produce and submit sub-seasonal weather forecasts – covering the critical weeks between medium-range and seasonal predictions – using AI/ML models.
Join one of our two slots of the Quest Launch Webinar on March 17th to learn more about the competition's structure, resources, and rules and ask your questions live:
The competition brings together participants from around the world to tackle one of meteorology’s most complex challenges: improving the accuracy and reliability of forecasts beyond traditional limits.
The AI Weather Quest serves as a unique platform to benchmark state-of-the-art AI/ML-based sub-seasonal predictions. It focuses on operational forecasting using real-time data, leverages open data for transparency and accessibility, and fosters equity between participants.
The competition welcomes both individual participants and teams of up to 10 members.
What is sub-seasonal forecasting?
Sub-seasonal forecasting fills the gap between short and medium-range weather forecasts (up to 15 days) and long-term seasonal outlooks (2 to 6 months), focusing on predicting weather conditions in the critical weeks between. Traditional models, which use dynamical equations to predict the evolution of atmospheric processes, struggle at the sub-seasonal timescale, as their accuracy diminishes over time, and slower-changing environmental factors like sea surface temperatures provide limited insights.
Despite these challenges, sub-seasonal forecasts are essential for sectors like energy, agriculture, and disaster risk reduction. Advancing these predictions can improve resource management, enhance preparedness, and reduce the impact of extreme weather events, transforming decision-making across industries.
Why AI/ML?
AI and ML technologies are revolutionising forecasting by uncovering patterns in vast, complex datasets where traditional methods face limitations. The use of AI/ML methods is increasingly being adopted by leading forecasting centres and tech companies to drive innovation and improve predictions.
Huge, complex and chaotic Earth system
Incredible volumes of data from various datasets
Transformative potential of AI and ML in weather forecasting
The ECMWF AI Weather Quest encourages participants to harness transformative capabilities offered by AI/ML technologies, whether by building new data-driven models trained on historical data or enhancing sub-seasonal predictions by processing existing forecast data. This competition aims to push sub-seasonal forecasting to new levels of accuracy and reliability.
What is expected from participants?
The AI Weather Quest unfolds in two phases:
- The Initial Training Phase, running from March to August 2025, provides a non-competitive environment where teams can refine their AI/ML models and familiarise themselves with the submission and evaluation process.
- From August 2025 to at least September 2026, the Competition Phase challenges teams to submit weekly, real-time sub-seasonal forecasts across four 13-week periods. Participants will be challenged to provide global forecasts at a 1.5 degree resolution, focusing on quintile probabilities for weekly-mean temperature, precipitation, and sea level pressure. Forecasts cover two specific periods: Days 19–25 and Days 26–32.
The AI Weather Quest is designed with flexibility in mind, allowing teams to join for any 13-week period of the Competition Phase and participate in as many periods as they choose. Throughout the Quest, participants have ongoing opportunities to refine and improve their models.
Why is ECMWF hosting the AI Weather Quest?
ECMWF, one of the world’s leading centres for numerical weather prediction, is renowned for producing high-quality forecasts up to seasonal timescales. By inviting global participation, ECMWF seeks to discover how AI/ML techniques can complement traditional forecasting methods and address the challenges of sub-seasonal prediction. Exploiting recent breakthroughs in the use of AI/ML for weather and climate is also part of the second phase of DestinE to enhance the capabilities of the digital twins. The competition will identify, benchmark, and integrate promising innovations into operational forecasting, driving progress to meet the growing demands of industries and communities worldwide. By leveraging open data and embracing diverse perspectives, it also seeks to foster new partnerships and share insights across the global community.
Get involved!
This competition is open to participants from international corporations, forecasting institutions, AI/ML experts, meteorologists, researchers, and students from diverse fields. No prior experience in weather forecasting is required—just a passion for solving real-world challenges with AI/ML, and access to your own computational resources for developing and training models.
The AI Weather Quest is committed to equity and inclusivity, striving to create a level playing field where participants from different backgrounds and skill sets can compete and collaborate.
Why Join?
Join a competitive community of innovators who are passionate about harnessing AI for real-world applications. By participating in the AI Weather Quest, you will:
Test and advance your expertise
In AI/ML models by tackling real-world challenges in weather forecasting.
Contribute to operational advancements
In AI/ML models by tackling real-world challenges in weather forecasting.
Build valuable connections
In AI/ML models by tackling real-world challenges in weather forecasting.
Achieve global recognition
In AI/ML models by tackling real-world challenges in weather forecasting.