Physical processes controlling ice concentrations in cold cirrus near the tropical tropopause

Previous ice nucleation calculations have suggested that in the presence of wave‐driven temperature perturbations typical of the tropical tropopause layer (TTL), homogeneous freezing should produce ice concentrations well in excess of measured values. A statistical ice cloud parameterization that includes effects of sedimentation was recently used to show that if the wave amplitudes are not too large, a quasi steady state may be established wherein loss of ice crystals by sedimentation is balanced by nucleation of new ice crystals, and the resulting cloud ice concentrations agree well with observations. Here, we use numerical models to further evaluate the evolution of ice concentrations in TTL cirrus, including a range of cloud physical processes (homogeneous and heterogeneous ice nucleation, sedimentation, and radiatively driven dynamics). We use a one‐dimensional microphysical model with bin microphysics to show that as a result of gravitational size sorting, the mean ice concentrations over the life cycle of the clouds are considerably smaller than the peak ice concentrations produced by ice nucleation events. However, the mean ice concentrations predicted here are considerably higher than either those reported based on the statistical model or those indicated by the observations. With the baseline wave amplitudes, ice crystals nucleated heterogeneously do not quench rising supersaturation in cooling air parcels and prevent homogeneous nucleation that produces high ice concentrations. We also use a three‐dimensional cloud resolving model to show that radiatively driven internal circulations and entrainment do slowly shift the ice concentrations toward lower values, but the time required to dilute ice concentrations produced by homogeneous freezing to values comparable to measured ice concentrations is of the order of 12–24 h, which may be longer than typical TTL cirrus lifetimes. Key Points Cloud processes such as sedimentation and entrainment reduce ice concentrations Detailed cloud model simulation ice concentrations exceed parameterization Simulated ice concentrations exceed observed values