1. Summary and Recommendations from Working Group 1: model uncertainty representations in convection-permitting / shorter lead-time / limited-area ensembles
- Author
-
Plant, Bob, Hally, A, Lock, S.-J., Ahlgrimm, M, Arpargus, M, Bauer, W, Cafaro, C, Chen, I, Fannon, J, Feddersen, H, Fleury, A, Frogner, I-L, Gebhardt, C, Hamalainen, K, Hieronymus, M, Kauhanen, J, Kouhen, S, Kuntze, P, Lean, H, Mamgain, A, McCabe, A, Milan, M, Petch, J, Puh, M, Schraff, C, Stirling, A, Tempest, K, Tsiringakis, A, Weidle, F, Wimmer, M, Xu, Z, Met Eireann, ECMWF, DWD, MeteoSwiss, Met Office, CWB, Taiwan, DMI, Meteo France, Norweigan Meteorological Institute, FMI, NCMRWF, KNMI, ZAMG, and Chinese Academy of Meteorological Science
- Abstract
WG3 discussed both the pros and cons of existing schemes as Working group 1 considered the treatment of model uncertainty (MU) in high-resolution ensembles, at grid spacings of order 1-5 km. These systems are often run for regional weather forecasting, perhaps over a single country, and for lead times of up to 5 days. Looking ahead, ECMWF’s strategy seeks to deliver global medium-range ensemble forecasts with 3-4 km grid spacings by 2030. It is questionable for what grid spacing we should dispense with a deep convection parameterization, but it will be either switched off or damped in these systems, such that deep convection can be assumed to be dominated by explicit motions. One of the problems with limited-area ensemble systems at this scale is that spread depends not only on the modelling system itself but also on the variability inherited from the large-scale boundary conditions. There is often thought to be a lack of spread in our high-resolution EPS (ensemble prediction systems), but this could reflect a lack of diversity on larger scales. The relative importance of lateral-boundary diversity and the model uncertainty mechanisms is regime dependent. The lateral boundaries will generally be more important in midlatitude winter but less so for summertime convection in relatively weak synoptic flow.
- Published
- 2022