30.09.2025
Artificial intelligence (AI) and machine learning (ML) are reshaping the field of climate prediction, offering powerful new tools to improve forecasts across a wide range of timescales. While much progress has been made in short- to medium-range prediction, research on AI/ML models for subseasonal to interdecadal forecasting is rapidly expanding. These approaches promise greater efficiency, higher resolution, and more frequent updates than traditional models—but challenges remain, including extending forecast skill beyond two weeks, ensuring physical consistency, and evaluating performance across climate extremes.
To address these opportunities and challenges, the Working Group on Subseasonal to Interdecadal Predictions (WGSIP) is launching a new webinar series in late 2025. The series will provide a collegial forum for researchers, practitioners, and early-career scientists to share advances, discuss technical approaches, and explore collaborative opportunities in data-driven climate prediction.
Objectives
- Knowledge sharing: Present the latest research, innovations, and applications of data-driven and hybrid modeling for S2S predictions.
- Community building: Connect experts worldwide and foster open discussions about successes, challenges, and future directions.
- Collaboration: Identify joint projects, experiments, and shared resources to advance the field.
Format
- One-hour live sessions on Zoom.
- Each webinar will feature two invited speakers (20 minutes each + Q&A).
- Recordings will be available exclusively to registered participants.
Provisional Calendar
-
30 September 2025, 15:00 UTC - Registrations closed
Theme: ACE2
Speakers: Oliver Watts-Meyer (AI2, Seattle), Chris Kent (UK Met Office)
Host: Stephanie Johnson (ECMWF) -
29 October 2025, 22:00 UTC - Register here
Theme: Predictability limits
Speakers: Trent Vonich (University of Washington) & Kirsten Mayer (NCAR, Boulder)
Host: Debbie Hudson (BoM) -
2 December 2025, 00:00 UTC - Register here
Theme: ENSO & MJO predictability
Speakers: Jing-Jia Luo (Nanjing University) & Daehyun Kim (Seoul National University)
Host: Kirsten Mayer (NCAR)
The series will run through the last quarter of 2025, with plans for continuation into 2026 based on community feedback.
Stay tuned for registration details and join us in discussing the future of data-driven climate prediction. Please note: we are not able to provide proof of attendance/participation certificates for these webinars.