Organisation of Multiple Processes of Secondary Ice Production among Basic Cloud-Types and Cloud Age

Decades ago it was first recognised that solid aerosol material typically initiates the first ice in mixed-phase clouds. Reasons for high concentrations of ice in such clouds, far beyond those of active solid aerosols have involved the notion that ‘secondary ice’ is formed somehow from pre-existing primary ice from this aerosol material. Multiple types of fragmentation of pre-existing precipitation account for this secondary ice.

In this presentation, results are shown from lab observations of an overlooked type of SIP involving fragmentation when raindrops collide with more massive ice particles. This has refined treatment of this 'Mode 2' of raindrop-freezing fragmentation in our cloud model. Another type of SIP has been represented for the first time in models, namely sublimational breakup. More recently, field observations from northern Sweden have been analysed to characterize the fragmentation of snow in collisions with graupel.

Improved treatment of four pathways of SIP are now treated in our 'aerosol-cloud' model:

• rime-splintering (the Hallett-Mossop [HM] process);
• fragmentation during freezing of rain and drizzle;
• breakup in ice-ice collisions;
• sublimational breakup.

Four contrasting cases of deep convection have been simulated. These involve basic cloud-types:

• stratiform cloud with a slightly cold base in the 'ACAPEX' field campaign;
• and convective clouds with cold, warm and very warm bases in the 'STEPS', 'MC3E' and 'GoAMAZON' campaigns respectively.

Tagged ice concentrations from the four SIP mechanisms are predicted in 3D. That for each type of SIP is plotted in the phase-space of vertical velocity and cloud-base temperature, and in vertical profiles for long-term averages for mature clouds. By contrast, tracking of young growing convective clouds in the simulations allows the classic plot of ice enhancement (IE) ratio as a function of cloud-top temperature to be predicted. This is compared with the original published plot by Hobbs and colleagues from 1980.

For long time-scales of clouds, breakup in ice-ice collisions is the most important process of ice initiation overall in all basic cloud-types simulated. Cold-based convective clouds display less ice enhancement than warm or very warm convective clouds. The other SIP processes make comparable contributions to the ice enhancement. The ice enhancement from the HM process increases with cloud-base temperature and vertical velocity. In the very warm cloud-base case, sublimational breakup dominates the average ice concentrations in the upper half of the mixed-phase region. Raindrop-freezing fragmentation makes only a weak contribution. A schematic conceptual diagram is provided of the relative roles of the four SIP processes in the phase-space of cloud-base temperature and ascent for all basic cloud-types.

For short time-scales with young convective turrets, the peak in IE ratio near -10 to -15 degC is predicted to be due to both the HM process and raindrop-freezing fragmentation. However, there is empirical uncertainty about these types of SIP.