The Large-Scale Properties of Tropical Convection

The large-scale atmospheric conditions associated with global tropical cloud regimes derived from the International Satellite Cloud Climatology Project (ISCCP) are examined. The cloud regimes are constructed using cluster analysis, statistically independent of any atmospheric variables, on 23 years of ISCCP D1 data between 35°N and 35°S. 19-year composite-analyses with the European Centre for Medium-Range Weather Forecasts Re-analysis (ERA-Interim) data are used to characterise the thermodynamic stability, water content and dynamics of the environment in each regime. Convective regimes are generally wetter, warmer and are associated with large-scale upward motion, while cumulus and stratocumulus regimes are identified with opposite conditions. However, there are regimes that straddle between the two extremes. One regime, suggestive of congestus cumulus, possesses convective attributes but resides in a cooler and more stable atmosphere. Another regime, largely composed of thin cirrus, is generally devoid of deep convection but has properties akin to convective regimes.

The atmospheric properties of the regimes are exploited as the basis for a stochastic model of the large-scale distribution of convection over the tropical oceans. Using simple rules, a field of cloud regimes is evolved according to known atmospheric variables and its fidelity assessed through comparisons with observed regime statistics. The stochastic model can be used as a representation of tropical convection in simple models or as a stochastic constraint in the representation of convection in general circulation models.