Laboratory to cloud modeling: investigating sensitivity of ice nucleation parameterizations

Description

Ice nucleating properties of mineral dust particles are investigated in the laboratory. The laboratory data was parameterized using Classical Nucleation Theory (CNT), and the sensitivity of cloud properties to two different representations of CNT was investigated in the cloud-resolving model (CRM). The onset contact angle CNT representation is based on two sets of laboratory deposition ice nucleation measurements: Arizona test dust (ATD) particles of 100, 300, and 500 nm sizes; and 400 nm ATD and Kaolinite dust species. Relative humidity with respect to ice (RHice) required to activate 1% of dust particles as ice nuclei (IN) was used to calculate the onset single contact angles. The mean and standard deviation parameters of PDF representation were determined by fitting the CNT-predicted activated fraction to the measurements at different RHice. Results show that onset single contact angles are not much different between experiments, while the PDF parameters are sensitive to temperature and dust size. The CRM simulations show that cloud properties (ice number concentration, ice water content, and cloud initiation times) are sensitive to onset single contact angles and PDF distribution parameters, particularly to the mean value. The comparison of our experimental results with other studies shows that under similar measurement conditions the onset single contact angles are consistent within ± 2.0 degrees, while our derived PDF parameters have large discrepancies. The study implies that modeled cloud properties are sensitive to the ice nucleation parameterization representations derived from CNT. IN measurements including coating studies on aerosol particles of natural and biological sources are needed to simplify the role of aerosol substrate in the parameterizations as well as to improve the existing parameterization treatments.