Ziel
During the last few decades there has been a substantial increase in the demand for more precise prediction of the weather at different scales and space and in the impact of human activities on the climate. In terms of meteorology important progress was made during this period. In terms of climate, however, there is a long way to go with regard to prediction. A preliminary stage would be improved knowledge and understanding of atmospheric and/or oceanic phenomena. During this same period the following new mathematical tools have been developed: dynamic systems; advanced numerical analysis; optimal control and estimation theories; high performance computing.
The general framework of this project is the diagnosis, modelling and predictability studies of low frequency atmospheric and oceanic processes, and the development of mathematical tools enabling such studies to be carried out.
Part of the project will be devoted to the development and study of numerical schemes for the determination of numerical solutions to thermo-hydrodynamic equations. Such schemes have to be performant and well fitted to parallel computers. A crucial point in the development of numerical modelling is the impact of new computer technologies. According to the references of the participants involved in this field, this component will be an important point of the research.
In view of sensitivity studies and data assimilation, there is a tendency in recent years to consider simultaneously a numerical model and its adjoint. The participants are at the origin of this idea and will develop co-operation in this domain. Many problems in modelling require efficient optimisation algorithms which are tested with real models.
At present, there is a strong demand within the climate study communities towards applied mathematics, from the theoretical point of view as well as from the computational point of view. This project is an important opportunity for these communities to work together and must be considered as a first step towards developing a long-term programme in this field. The expected theoretical results are: how to estimate the predictability of a model of a particular situation; what information should be computed in order to achieve the predictability; how to get the best information from a set of data and from models; what is the impact of observations on the predictability of a model; and how to carry out sensitivity studies in the presence of data. The expected practical results are: intercomparison of models; intercomparison of algorithms to solve a model; intercomparison of data assimilation techniques; and intercomparison of parallelisation techniques.
Programm/Programme
Thema/Themen
Aufforderung zur Vorschlagseinreichung
Data not availableFinanzierungsplan
Data not availableKoordinator
38041 Grenoble
Frankreich