Over 150 participants have attended this year’s numerical weather prediction (NWP) training course modules at ECMWF. We know that participants really value face-to-face learning: it is required when studying advanced materials in the short amount of time that busy schedules permit, and it gives participants the chance to talk to lecturers about course material and their own research issues and to build networks within the research community. Responding to feedback, all modules are now delivered in short one-week face-to-face courses with practical sessions to consolidate learning and supported by pre-course study, mainly in the form of newly created eLearning modules. A ‘Data assimilation collaborative course’ has been developed and delivered by the University of Reading for those interested in more introductory theory or practical sessions. The course curriculum is focused on the Centre’s objective to provide advanced training for the scientific staff of ECMWF’s Member and Co-operating States in the field of numerical weather prediction. It draws on the expertise of over 70 research staff, mainly from the Research Department but also from the Forecast and Copernicus Departments, and one or two external experts.
Feedback on this year’s courses
“The lecturers were all very enthusiastic and had great content to their presentations.”
“There’s no substitute for meeting the lecturers and other attendees in person. Discussing the lectures and ideas during coffee breaks was an important part of the course.”
“I find it easier to digest scientific content through e-learning modules rather than traditional lectures, because I can adjust the pace according to my own needs. But to meet lecturers and other course participants in person is ever so important. It is good for networking and makes it possible to have informal on- and off-topic discussions that would never have taken place otherwise.”
Focus on ensembles and the extended range
Over the last few years, we have received a large number of applications to attend the NWP course on predictability and ensemble forecasting, and this year was no exception. As we enter the era of seamless probabilistic forecasting, with national meteorological services providing probabilistic forecasts from days to months ahead, we have seen forecasters as well as researchers wanting to increase their understanding of the systems that are used to produce the forecasts and the underpinning theory of how we can make predictions beyond the medium range. Attendees this year spent time using the OpenIFS system to look at the use of ensembles within a case study of September 2012, which was challenging for forecasters when Hurricane Nadine interacted with an Atlantic cut-off low. (See the article ‘The varied uses of OpenIFS’ in this Newsletter)
Advanced numerical methods
Numerical weather prediction courses for 2020
- Advanced numerical methods
- Parametrization of the subgrid scale and diabatic processes
- Data assimilation
- Satellite data assimilation
- Predictability and coupled ocean–atmosphere ensemble forecasting
In 2014, we introduced a new course in advanced numerical methods, which has grown in popularity. This year it attracted 25 participants to learn about the latest advances to describe the dynamics of the atmosphere considering future computational and resolution requirements. Participants had time for hands-on exploration of elliptic solvers, advection schemes and spectral transform methods, as well as looking at idealised cases with OpenIFS.
Course material is downloadable and eLearning modules can be accessed from the training resource area on the ECMWF website (www.ecmwf.int/en/learning/education-material). Registration for next year’s courses will open in the summer (see www.ecmwf.int/en/learning/training for updates).