Dynamical, convective, and microphysical control on wintertime distributions of water vapor and clouds in the tropical tropopause layer

Processes that influence the humidity and cirrus cloud abundance in the Tropical Tropopause Layer (TTL) during boreal winter 2006–2007 are investigated in simulations of clouds along backward trajectories of parcels ending at the 372 K potential temperature (100 hPa) level in the tropics. Trajectories are calculated using offline calculations of seasonal mean tropical radiative heating rates along with reanalysis temperature and wind data with enhanced wave-driven variability in the TTL. The one-dimensional (vertical) time-dependent cloud microphysical model is initialized with water vapor measurements from the Microwave Limb Sounder and the evolution of clouds along each trajectory is simulated using temperature profiles extracted from reanalysis data and convective cloud top heights estimated from 3-hourly geostationary satellite imagery. Averaged over the tropics, waves dehydrate the 100 hPa level by 0.5 ppmv, while convection and cloud microphysical processes moisten by 0.3 and 0.7 ppmv, respectively. The tropical mean cloud occurrence frequencies in the middle to upper TTL agree well with those based on satellite observations (spatial correlation of 0.8). Waves and convection enhance cloud occurrence at the cold point tropopause by 4% and 2%, respectively. Temporal variability of the heating rates as indicated by the ERA-Interim 6-hourly heating rate fields dehydrates the TTL by 0.4 ppmv and decreases the cloud occurrence by 4% because parcels are more likely to encounter the coldest temperatures and dehydrate near the cold point, limiting cloud formation above. The final dehydration locations of parcels, concentrated near the dateline in the tropical Pacific, are insensitive to various model parameters.