Exploring Formation Conditions of Dusty Cirrus Using a Large-Eddy Simulator

Dusty cirrus clouds are optically thick cirrus that form under the influence of elevated concentrations of ice nucleation active mineral dust. These clouds are associated with dust plume events in which freshly emitted mineral dust particles are lifted to near-tropopause levels by baroclinic storms. At present, dusty cirrus clouds are not well represented in global climate models, primarily because aerosol-cloud interactions are inadequately parameterized and model resolutions are insufficient to resolve the convective motions within these clouds (Seifert et al., 2023, ACP).

We used large-eddy simulatior (LES) UCLALES-SALSA to investigate the formation conditions of dusty cirrus clouds, supported by a sensitivity analysis based on observations from the ML-CIRRUS (2014) campaign. The results indicate that mineral dust concentrations must be approximately 10-100 times higher than climatological values to sustain convective overturning motions. Furthermore, the sensitivity analysis shows that both sub-500 nm and super-500 nm dust particles play a significant role in producing ice crystal number concentrations consistent with in situ observations.

We further find that the choice of deposition ice nucleation parameterization has a strong influence on the simulated properties of dusty cirrus clouds and on the temperature range over which they can form. The sensitivity analysis was conducted using the Ullrich et al. (2017, JAS) scheme, which exhibits strong sensitivity to temperature and relative humidity with respect to ice within the temperature range characteristic of dusty cirrus clouds, and produces a relatively large fraction of ice concentration near the top of the cirrus layer. In contrast, simulations using the Phillips et al. (2013, JAS) scheme yield markedly different cloud structures, with more vertically uniform ice crystal number concentrations, reflecting its weaker temperature dependence compared to the Ullrich et al. (2017) parameterization.