Cloud formation, convection, and stratospheric dehydration

Abstract Using the Modern‐Era Retrospective Analysis for Research and Applications (MERRA) reanalysis winds, temperatures, and anvil cloud ice, we use our domain‐filling, forward trajectory model combined with a new cloud module to show that convective transport of saturated air and ice to altitudes below the tropopause has a significant impact on stratospheric water vapor and upper tropospheric clouds. We find that including cloud microphysical processes (rather than assuming that parcel water vapor never exceeds saturation) increases the lower stratospheric average H2O by 10–20%. Our model‐computed cloud fraction shows reasonably good agreement with tropical upper troposphere (TUT) cloud frequency observed by the Cloud‐Aerosol Lidar with Orthogonal Polarization (CALIOP) instrument in boreal winter with poorer agreement in summer. Our results suggest that over 40% of TUT cirrus is due to convection, and it is the saturated air from convection rather than injected cloud ice that primarily contributes to this increase. Convection can add up to 13% more water to the stratosphere. With just convective hydration (convection adds vapor up to saturation), the global lower stratospheric modeled water vapor is close to Microwave Limb Sounder observations. Adding convectively injected ice increases the modeled water vapor to ~8% over observations. Improving the representation of MERRA tropopause temperatures fields reduces stratospheric water vapor by ~4%.