Interactions between Heterogeneous Ice Nucleation and the Hallett Mossop Process in Dust-Laden Tropical Convective Clouds

The largest atmospheric dust loading in the world occurs over the Tropical Atlantic where mineral dust from the Sahara is lofted into the atmosphere and transported towards the Americas. Desert dust is an efficient ice nucleating particle (INP) owing to its large particle size and mineral composition. Despite its importance for cloud glaciation, precipitation and lifetime, the role that heterogeneous ice nucleation by dust plays in tropical Atlantic convective clouds, which have impacts on both climate and tropical storm formation, is not well understood. Observations of tropical maritime clouds with relatively high cloud top temperatures (>-12°C) frequently show concentrations of ice crystals far exceeding those of INP, indicating the importance of a secondary ice production mechanism, such as the Hallett Mossop process. The Ice in Clouds Experiment - Dust (ICE-D) field campaign from Cape Verde in August 2015 provides observations of tropical convective clouds in high dust loading environments. In this work we show the relative importance of primary and secondary ice formation for cloud glaciation and cloud field scale properties based on high resolution simulations. The importance of sedimentation and initial ice crystal number for the triggering of secondary processes in deep convective clouds observed during ICE-D is explored. The cloud glaciation and anvil properties are sensitive to both the chosen representation of heterogeneous ice formation and the presence of a secondary ice production mechanism. Changes to primary and secondary freezing rates at low cloud levels induce changes in condensed water mass distribution at all levels. Results indicate that reduced rates of ice nucleation and secondary freezing at low cloud levels can increase the flux of cloud droplets to low temperature regimes, increasing freezing in the homogeneous nucleation zone. This results in an increase in ice crystal mass at upper levels with implications for convective anvil properties. This work serves to improve our understanding of tropical cloud glaciation and the interactions of desert dust with deep convective clouds.