AbstractThe response of the glaciation and precipitation of a multi‐thermal cumulus cloud to changes in the aerosol concentration has been assessed in a series of sensitivity tests with the UMIST Explicit Microphysics Model (EMM). A simulation of this cloud from the Met Office cloud‐resolving model (CRM) has been utilized in these tests. This cumulus cloud was observed by aircraft during the initial stage of its growth over New Mexico on 10 August 1987. The growth of the simulated cloud is divided into two parts: a shallow phase followed by a deep phase. Maximum values of the cloud depth in these two phases were 5 and 9 km, respectively.In the EMM simulations including only the shallow phase, the precipitation efficiency was found to decrease substantially with increasing atmospheric concentrations of cloud condensation nuclei (CCN). Also, the graupel mixing ratio and total ice concentration were found to decrease as normalized CCN concentrations were increased above typical continental values. These changes are explicable in terms of: (1) the Hallett–Mossop (H–M) process at −3 to −8°C and the freezing of supercooled raindrops in collisions with ice splinters dominating the glaciation; and (2) the warm‐rain process being more significant for the overall precipitation production than the ice process in these particular simulations. The almost complete suppression of precipitation by extreme CCN concentrations corresponding to a forest‐fire plume in the EMM simulation is consistent with the analysis by Rosenfeld of satellite images of Indonesian cumuli engulfed by smoke from biomass burning.A clear tendency for ice crystals to be smaller and more numerous in the anvil was found with increasing CCN concentrations beyond typical continental values in long‐term simulations that included the deep phase. The sensitivity of the precipitation rate to the normalized CCN concentration was found to be relatively low in these deep cases. Copyright © 2002 Royal Meteorological Society.
Read full abstract