Introduction

The nonhydrostatic Euler equations came into general use in the 1990s, making it possible to deal properly with atmospheric motion on spatial scales of a few kilometres. Advances in numerical methods and faster computers mean nonhydrostatic models can now be used operationally on limited areas. In the medium to long term, then, the nonhydrostatic approach is destined to be applied to all categories of model (including global models) working with grid meshes of less than ten kilometres.

It seemed, therefore, that this book, which is essentially about the primitive equations, needed to be supplemented by two examples of nonhydrostatic models based on the Euler equations. Two models were chosen that have proved their worth for operational forecasting:

• the AROME model developed at Météo-France can be thought of as the nonhydrostatic extension of the baroclinic models described in Chapter 8 and it uses the same type of time integration algorithm (semi-Lagrangian semi-implicit);

• the WRF/ARW model has been developed by US universities and organizations to be made available to various users for research and operational forecasting alike; this model uses a split-explicit time integration algorithm quite unlike those covered elsewhere in this book.