Publication Date

Summer 2023

Degree Type


Degree Name

Master of Science (MS)


Meteorology and Climate Science


Minghui Diao; Qian Tan; Jie Gong


Cloud thermodynamic phase distributions play a crucial role in accurately representing cloud radiative effects and feedback in a changing climate. The partitioning of cloud thermodynamic phases (ice, liquid, and mixed phase) significantly influences Earth's surface temperature and its ability to mitigate the impact of global warming. Satellite-based cloud phase data are frequently used for the evaluation of global climate models, yet validation of them against in-situ observations is still lacking. This study examines global cloud phase distributions and their determinant factors by validating three satellite-based cloud phase products against an extensive in-situ airborne dataset. CALIPSO exhibits the most similar ice phase profiles, CloudSat tends to overestimate mixed phase frequency, and DARDAR overestimates ice phase frequency. The comparison results reveal variations in ice phase frequency across latitudes and seasons. Spatiotemporal mismatches have minimal impacts on the main findings, emphasizing statistical robustness. Machine learning techniques are employed to explore key determinant factors, such as temperature, relative humidity, vertical velocity, and aerosol effects. Temperature is the most influential factor in cloud phase distribution, while relative humidity determines in-cloud and clear-sky conditions. This study provides guidelines for globally validating satellite-based cloud products and enhances our understanding of factors influencing cloud phase distributions.

Included in

Meteorology Commons