| Abstract | The In-Cloud Icing and Large-drop Experiment (ICICLE) flight campaign, led by the United States Federal Aviation Administration, was conducted in the geographical region over US Midwest and Western Great Lakes, between January and March 2019, with the aim to collect atmospheric data and study the aircraft icing hazard. Measurements were taken onboard the National Research Council of Canada (NRC) Convair-580 aircraft, which was equipped with more than 40 in situ probes, sensors, and remote sensing instruments in collaboration with Environment and Climate Change Canada (ECCC). In each flight, aerosol, cloud microphysics, atmospheric and aircraft state data were collected. Atmospheric environment characterization is critical both for cloud studies and for operational decision making in flight. In this study, we use the advantage of multiple input parameters collected in-flight together with machine learning and clustering techniques to characterize the flight environment. Eleven parameters were evaluated for the classification of the sampled environment along the flight path. Namely, aerosol concentration, temperature, hydrometeor concentration, hydrometeor size, liquid water content, total water content, ice accretion rate, and radar parameters in the vicinity of the aircraft. In the analysis of selected flights, we were able to identify periods of supercooled liquid clouds, glaciated clouds, two types of mixed-phase clouds, and clear air conditions. This approach offers an alternative characterization of cloud boundaries and a complementary identification of flight periods with hazardous icing conditions. |
|---|
