Transport and scavenging of soluble gases in a deep convective cloud

A one-dimensional entraining/detraining plume model is used to examine the transport and scavenging of soluble gases in tropical deep convection. The model is applied to a continental system observed over Brazil during the Tr ace and Atmospheric Chemistry Near the Equator-Atlantic (TRACE-A) TRACE-A a ircraft campaign with outflows extending from 7 to 16 km altitude. Six gase s are simulated: CO (inert tracer), CH3OOH, CH2O, H2O2, HNO3, and SO2. Obse rved (simulated) convective enhancement factors (CEF) at 7-12 km altitude; representing the ratios of postconvective to preconvective mixing ratios, a re 2.4 (1.9) for CO, 11 (9.5) for CH3OOH, 2.9 (3.1) for CH2O, 1.9 (1.2) for H2O2, and 0.8 (0.4) for HNO3. Simulated scavenging efficiencies in the con vective column are 5% for CH3OOH, 23% for CH2O, 66% for H2O2, 77% for HNO3, and 28% for SO2. The large CEF for CH3OOH reflects its low solubility and its boundary layer enrichment relative to the upper troposphere. The Henry' s law constant for CH2O puts it at the threshold for efficient scavenging. Scavenging of SO2 is limited by the rate of aqueous phase reaction with H2O 2, as H2O2 is itself efficiently scavenged by Henry's law equilibrium; effi cient scavenging of SO2 requires unusually high cloud water pH (pH > 6) to enable fast aqueous phase oxidation by O-3. Both HNO3 and H2O2 are efficien tly scavenged in the lower (warm) part of the cloud, but H2O2 is released a s the cloud freezes due to low retention efficiency during riming. Signific ant scavenging of H2O2 still takes place by cocondensation with ice in the glaciated cloud but is less efficient than in the warm cloud. Inefficient s cavenging of H2O2 ill glaciated clouds may explain the observation, in TRAC E-A and elsewhere, that H2O2 is enhanced in deep convective outflows while HNO3 is depleted. Model results indicate little direct transfer of air from the boundary layer to the cloud anvil in the convective plume, because of low-level detrainment in the warm cloud and high-level entrainment in the g laciated cloud. We find instead a convective ladder effect where midlevel o utflow during the growing phase of the storm is reentrained into the convec tive plume as the storm matures.