On 14th January 2026, MFC Wednesday was held at the University of Reading (Whiteknights Campus), bringing students together for a seminar talk in the morning followed by an afternoon visit to ECMWF HQ.
Event Highlights
-
Registration (10:50–11:00)
Location: Maths Room 314 (Maths Building), Whiteknights Campus -
Seminar Talk (11:00–11:45) — Nachiketa Chakraborty
Title: Explainable AI for Environmental Science in the era of Large Models
Abstract (brief): In the era of Big Data and ever-larger models, environmental science—space weather included—is undergoing a paradigm shift. Foundation models such as Aurora and GraphCast seek to rival, traditional data assimilation (or even hybrid) techniques. As with any application of artificial intelligence, it is critical to examine rigorously the veracity of predictions of these large and complex models. Embedding an element of explainability along with precise and timeliness of forecast is very important. This demands principled approaches of building and test AI models, rooted in mathematics and statistics as well as physics in specific case. In this talk, I will highlight how fields like causality, signal processing and indeed elements of physics play a vital role in achieving this aim.
ECMWF HQ Visit (14:00–17:20)
-
Arrival / Meet & Greet / Coffee (from 13:30) — Chantal Dunikowski
-
Overview of ECMWF and physics-based NWP (14:00) — Stephen English, Acting Director of Research / Lead Scientist
-
Overview of AIFS and DOP modelling at ECMWF (14:30) — Ewan Pinnington, Scientist- Land Data Assimilation
-
ECMWF Data Centre (virtual tour) (15:00) — Computing & Software Support Team
Students then split into two parallel groups:
Group A (University of Reading)
-
Earth system modelling — Robin Hogan, Principal Scientist
- Earth system data assimilation — Philip Browne, Senior Scientist – Coupled Ocean-Atmosphere Data Assimilation
-
Ensemble forecasting — Sarah-Jane Lock, Scientist
-
Long-range/seasonal forecasting — Tim Stockdale, Principal Scientist
-
Open forum Q&A and discussion (16:30) — Stephen English, Robin Hogan, Philip Browne, Sarah-Jane Lock and Tim Stockdale
Group B (University of Cambridge)
-
Anemoi — Mario Santa Cruz, Scientist for Machine Learning
-
Earth system AIFS modelling — Nina Raoult, Scientist – Land Modelling
-
AIFL (AI-based hydrological modelling) — Kenza Tazi, Data-driven Hydrological Modeller
-
Open forum Q&A and discussion (16:15)- Ewan Pinnington, Mario Santa Cruz, Nina Raoult and Kenza Tazi
All students then reconvened for:
-
The Weather Room (16:50) — Linus Magnusson, Senior Scientist
-
Group photo / Departure (17:20) — Simon Witter & Chantal Dunikowski