Exploring signatures of ice nucleation events in satellite observations

Satellite observations provide instantaneous snapshots of cloud properties, such as ice water content and ice crystal number concentration, which may retain information on the prior history of cloud parcels. However, inferring cloud microphysical processes from satellite data alone remains challenging, as these processes are not directly observable.

Lagrangian models, either purely transport-based or coupled to microphysics, are commonly used to reconstruct the history of cirrus clouds and to infer their origin (in situ vs. liquid-origin) and ice formation pathways (e.g. homogeneous vs. heterogeneous nucleation). These approaches, however, strongly depend on meteorological reanalyses and microphysical assumptions, which remain particularly uncertain for cirrus clouds. While very useful, they would ideally be further constrained by observation-based information.

Here, we explore whether satellite observations can provide additional indications of ice nucleation events. For this purpose, we analyse global lidar-radar observations from the DARDAR-Nice product, focusing on spatial patterns in vertical profiles of ice cloud properties. In particular, we investigate whether strong local absolute or relative increases in ice crystal number concentration can be interpreted as signatures of nucleation events. This methodology is applied globally over one year of observations to estimate the occurrence of homogeneous and heterogeneous ice nucleation, and the statistical occurrence of different ice nucleation pathways is presented. The approach is further evaluated using high-resolution ICON simulations of ice cloud structures to assess its limitations. Finally, global results from this observation-only method are compared with nucleation diagnostics derived from back-trajectory approaches using CLaMS-Ice and FLEXPART.