The mass-flux approach to the parametrization of deep convection
March 1997

By D. Gregory

1 . Introduction

Although the principles of the mass-flux approach to convection were formulated in the 1970s and formed the basis for the well known "Arakawa-Schubert" scheme (Arakawa and Schubert 1974) it has only been recently that such schemes have been used in many GCMs due to their expense compared to simpler schemes (moist convective adjustment and Kuo-type schemes). The approach has a stronger physical basis than these earlier methods and provides an understanding of how convection affects the large-scale atmosphere. It also allows more physically based microphysical treatments to be employed and vertical transports of tracers by convection to be estimated, of importance when considering the impact that aerosols have upon climate.

This paper will outline the mass-flux approach to the parametrization of convection, concentrating "bulk" type schemes such as those used in the ECMWF forecast model (Tiedtke 1989) and the UK Met. Office Unified Model (Gregory and Rowntree 1990). Important issues relating to this approach are discussed by reference to the Tiedtke scheme at ECMWF, although reference is made to other schemes as appropriate. Only brief comments are made concerning the more complex "spectral" cloud model approach, of which the Arakawa-Schubert scheme is the best known example.

11.06.2002