Institute of Computational Mathematics and Mathematical Geophysics

Тезисы докладов

Математическое моделирование процессов в атмосфере и гидросфере

For more than two decades atmospheric boundary layer scientists have been using 3-D dynamic models derived from the Navier-Stokes equations. Every modeller using mesoscale modelling tools knows well that the way model results approach local observations is often disappointing : computed and observed temperature and wind fields show major differences even after a series of tune-up procedures. These discrepancies seem to be scale dependent, becoming more obvious when looking at smaller atmospheric scale events. This problem is very complex and hardly understood but important for practical applications of atmospheric models. This paper tries to have a look to some basic causes.

These discrepancies appear for many reasons. Some are intrinsic to the modelling procedure : approximations in model formulation, inaccuracies in numerical schemes, non-linear amplifications of errors. However, in some cases, these effects can be minimised, producing accurate answers when proper initial and boundary conditions (I&BC) are used (as in industrial flows). In atmospheric flows, we have to cope with a wide span of spatial scales, non-stationarity, and a semi- infinitely open domain.

I&BC can be subdivided in external forcing by the large scale meteorological fields (which also suffer of inaccuracy at that scale) and internal forcing from surface characteristics (e.g. topography, land use, water availability) influencing model behaviour through the sub-grid parameterisation.

In addition, the turbulent nature of the atmospheric flow is treated through parameterisations of fluxes and appropriated exchange coefficients. These were developed from homogeneous microscale observations and over-simplified cases such as cloud free condition, stationarity and horizontal homogeneity which do not hold in mesoscale application of the models. Large turbulent eddies (at BL scale) are generated by local features over which we still have few insights on their relation with topography and surface features.

The present paper considers the possible influences of I&BC and of turbulence at various spatial scales on model behaviour. The list is certainly not exhaustive but we try to look to different possible effects. In the actual state of the art, possible partial improvements are suggested.

Ваши комментарии Обратная связь

[Головная страница] [Конференции]

© 1996-2000, Институт вычислительных технологий СО РАН, Новосибирск
© 1996-2000, Сибирское отделение Российской академии наук, Новосибирск
    Дата последней модификации: 06-Jul-2012 (11:52:46)