Part of ECMWF's 50th anniversary celebrations
Background
Data assimilation is traditionally about optimally blending observations and model information to provide the best estimate of the current state of the Earth system for both monitoring and prediction purposes. However, the recent advent of machine learning in numerical weather prediction (NWP) and climate studies has shown that data assimilation and observations can be effectively and more directly used to improve NWP and climate prediction than just for providing initial conditions for physics-based and data-driven machine learning forecast models.
Workshop description
The focus of this two-day workshop was to discuss current cutting-edge development directions in both traditional data assimilation and observations for NWP and climate and the emerging areas of hybridising aspects of the data assimilation workflow or, possibly, fully replacing data assimilation with machine learning technologies.
Format
The workshop was structured around two main thematic areas.
During the first day of the workshop, we took stock of the current status of operational Earth System Data Assimilation systems in major NWP Centres, discussed the sources and limits of weather predictability from increasingly accurate initial conditions, and evaluated the prospects of improvements in this area from, e.g., advances in resolution, model complexity, coupling, new observations.
The second day of the workshop was devoted to discussing the emerging area of hybridising traditional Data Assimilation methodologies with Machine Learning techniques. The aim of this part of the workshop was to give an up-to-date snapshot of the main development directions in this fast evolving field and provide an informed perspective on its expected medium-term landing points.
Panel discussions were scheduled at the end of both days of the workshop involving all speakers and the on-site and on-line audiences.
Poster presentations took place on both thematic areas.
