• Jan Barkmeijer
    • KNMI - The Netherlands
    • Mathematics

Weather forecasting and the initial condition challenge

During the last decades the quality of weather forecasts has improved beyond any doubt. On average, forecast quality has gained one day per decade. This major achievement, recently coined ‘the quiet revolution in numerical weather prediction’, could not have been accomplished without facing fundamental scientific and technological challenges. In this talk the focus will be on one of these challenges, namely how to incorporate the vast amount of available imperfect observations into an imperfect model and still be able to improve the forecast performance. Also the implication of the chaotic behaviour of the atmosphere for this data assimilation process will be addressed.


Jan Barkmeijer is group leader of the meso-scale model group at the Royal Netherlands Meteorological Institute (KNMI). He received his doctorate in mathematics from Groningen University in 1988 with a thesis on the chaotic behaviour for a class of dynamical systems. In 1989 he started to work at KNMI in the recently formed predictability section. He became interested in the interface between data assimilation and predictability research and moved to the European Center for Medium-range Weather Forecasts (ECMWF, UK). During 1995-2002 he was closely involved in the development of the ensemble prediction system, a means to assess the intrinsic chaotic behaviour of the atmosphere. In 2002 he returned to KNMI . He has co-authored many papers and participated in various international committees, such as chair of the scientific advisory committee of ECMWF.

J.Barkmeijer Homepage