PCAPS at the Workshop on Diagnostics For Global Weather Prediction, ECMWF, 9 - 12 September 2024
The European Centre for Medium-Range Weather Forecasts (ECMWF) organised a workshop on diagnostics for global weather prediction in Reading, UK from 9 - 12 September 2024. Diagnostics of a forecasting systems help to identify problems, provide a deeper understanding of forecasting errors, and aid data assimilation and model development. The workshop’s presentations covered a wide range of diagnostic topics, such as predictability, model errors and uncertainty, tropical waves, AI modelling, and the impact of resolution and community datasets, to name a few.
From the PCAPS perspective, Steffen Tietsche (ECMWF) presented the ICECAP toolbox, a verification tool for sea-ice forecasts from days to seasons. The aim of the tool is to provide meaningful information about forecast quality alongside sea-ice products. This tool will be used in PCAPS to evaluate sea-ice forecasts, as part of the Verification Task Team’s work.
Luise Schulte (ECMWF/Deutscher Wetterdienst/University of Cologne) presented a poster about the errors in Arctic clouds in ECMWF forecasts, using measurements of liquid water paths from the MOSAiC campaign. Sensitivity experiments for different cloud parameterisations using a single-column model was presented. Barbara Casati from Environment and Climate Change Canada and a PCAPS SG member, presented a poster evaluating relationships between different variables (long- and short-wave radiation, 2-metre temperature, near-surface humidity). The relationships were compared for forecasts from the Canadian Meteorological Centre with YOPP super site observations in the Arctic.
These two posters are examples of how diagnostics can be used to get a better understanding of processes that needs to be improved, in order to provide better forecasts in the Arctic. This type of work is closely connected to the Verification and Processes Task Teams in PCAPS.
More information about the workshop, including presentations and recordings, can be found here.