Operational Multi-Model Blending of Site Specific forecasts at the Met Office

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 7 January 2015: 11:15 AM
123 (Phoenix Convention Center - West and North Buildings)
James Richard Canvin, Met Office, Exeter, United Kingdom; and C. Jones and S. Jackson

Since 2011 the public site-specific weather forecast output from the Met Office has used multi-model blending to increase accuracy, reduce jumpiness and decouple model changes from output products. To fit the needs of the operational system, an efficient method is needed to ensure that a large number of forecasts can be generated rapidly using the latest model data. Therefore an incremental update method is used each time new data are available to reduce complexity and data volume, enabling a rapid turnaround. This paper summarizes the technique used, its optimization and the resulting improvements in forecast quality.