Community Research and Development Information Service - CORDIS


Among the various assumptions of approximations that are generally called for in numerical modelling, two main categories can be identified. The first one relates to approximations based on physical arguments whereas the second relates to numerical considerations. Assumptions such as the Reynolds one, the Taylor hypothesis, the Boussinesq approximation and so on, belong to the first category, whereas to the second belong those required to practically solve the conservations equations are described. One follows the classical derivation based on a series of physical and numerical approximations whereas the second is directly expressed in terms of physical arguments. The comparison of these two methods is used to understand to what extent physical assumptions imply numerical approximations and vice-versa. The new derivation proposed here, shows that assuming the validity of the Reynolds averaging rules directly leads to discretized averaged equations without having to call for any other numerical approximations

Additional information

Authors: THUNIS P, Environment Institute, Ispra (IT);GALMARINI S, Environment Institute, Ispra (IT)
Bibliographic Reference: Article: Journal of Atmospheric Sciences, 1999
Record Number: 199910704 / Last updated on: 1999-05-14
Original language: en
Available languages: en