Stochastic Parameterizations of Equatorial Gravity Waves

Deterministic parametrizations in global climate models rest on the assumption (or hope) that a statistical-dynamical equilibrium exists between the resolved and subgrid-scale unresolved flows. A particular choice of parameter values then takes into account only the mean of the unresolved quantities while intrinsically neglecting all higher order moments of the distribution. In the past some studies have suggested, that representing higher order moments in parameterizations of unresolved flows may improve GCMs. Convection is thought to be the most important source of gravity waves in the tropics. Unfortunately convectively triggered gravity waves are episodic in nature and highly variable in strength. Furthermore circulation of the tropical stratosphere is strongly affected by wave-mean flow interactions. It therefore seems reasonable to study the effects on the tropical stratospheric climate of a stochastic gravity wave scheme. A version of the Hines gravity wave scheme with a stochastic source strength is developed. The scheme is implemented in a 55 level version the Unified Model. The effects of different stochastic source strengths are examined. In some cases the source is assumed to represent the intensity distribution of equatorial convection. The effects of including higher order moments of the distribution are also examined.