Evaluating secondary ice production by raindrop fragmentation upon freezing in mixed-phase clouds using modeling and observations

Field observations of mixed-phase clouds frequently reveal ice particle number concentrations that exceed what can be explained by primary ice nucleation alone. This discrepancy is commonly attributed to secondary ice production (SIP) processes. Despite their recognized importance, the efficiency and representation of SIP mechanisms in numerical weather prediction models remain highly uncertain. One such mechanism is raindrop fragmentation upon freezing: when a supercooled drop freezes, excess internal pressure can cause it to shatter, releasing ice splinters that may subsequently grow to secondary ice particles. A comprehensive understanding and representation of both primary and secondary ice formation processes in mixed-phase clouds is crucial, as these processes strongly influence cloud properties such as precipitation formation and cloud lifetime, and thus affect numerical weather and climate predictions.

In this study, we examine the impact of secondary ice production by raindrop fragmentation upon freezing on cloud glaciation and precipitation within the ICON model framework. We focus on a warm front observed during the Evaluating Microphysical Pathways Of Midlatitude Snow Formation (EMPOS) field campaign in Hyytiälä, Finland, in February 2024, where raindrop fragmentation was directly observed by the Video In Situ Snowfall Sensor (VISSS). This case is ideal for studying raindrop fragmentation upon freezing: frozen hydrometeors fall through a warm layer, melt partially or completely, and refreeze as they enter the sub-zero layer below, creating conditions favorable to raindrop fragmentation upon freezing. We specifically assess potential limitations of the model representation of raindrop fragmentation upon freezing, including the production of sufficient raindrops of relevant size ranges, their refreezing under suitable thermodynamic conditions, and the resulting efficiency of ice splinter generation. We compare the performance of parameterizations for raindrop fragmentation upon freezing in the two-moment microphysics scheme by Seifert and Beheng (2006) – specifically Sullivan et al. (2018) and a newly implemented parametrization based on Phillips et al. (2018) - against observations at the study site. In-situ measurements from the VISSS allow us to constrain and evaluate both model parameterizations of SIP. To complement the bulk microphysics simulations and to gain deeper physical insight into the underlying SIP process, we additionally present first results from the Monte-Carlo super-particle cloud microphysics scheme by Seifert (2018), which better represents mixed-phase states of hydrometeors that bulk-microphysical schemes cannot capture.