ON THE GENESIS OF THE LOGNORMAL DISTRIBUTION OF CLOUD CHARACTERISTICS AND ITS APPLICATION TO CUMULUS PARAMETERIZATION (CONVECTION, APPARENT HEAT SOURCE, ENSEMBLE, PARAMETERIZATION SCHEME, GATE).

摘要

Radar statistics of convective elements generated in the past decade show lognormal frequency distributions of echo areas, heights and durations. An explanation for the genesis of these distributions, based on the interrelationship between cloud groups and the environment, is proposed.;A closed cumulus ensemble model that makes explicit use of the lognormal distribution function has been developed. In this model, the cloud ensemble is spectrally divided into sub-ensembles, and each of these is characterized by its cloud top pressure. The ensemble-averaged values of cloud characteristics are controlled by the large-scale environment. In return, the cloud ensemble modifies the environment through compensating subsidence, entrainment, downdrafts and lateral and final detrainments. Application of the model to the GATE (Global Atmospheric Research Program's Atlantic Tropical Experiment) data sets predicts rainfall rates and the profiles of large-scale apparent heat source and moisture sink. The results agree reasonably well with the observations under various synoptic situations.