Thesis of Robin Miri

Defense of thesis Robin Miri - laboratory LOA

Abstract :

This thesis focuses on the analysis of atmospheric fluorescence measurements by lidar for characterizing aerosols and studying their interactions with water vapor and clouds. The LILAS (LIlle Lidar Atmospheric Study) lidar, located on the ATOLL (ATmospheric Observations at LiLLe) platform at the LOA (Laboratoire d’Optique Atmosphérique) in Lille, is a Mie-Raman-fluorescence lidar and a member of the European EARLINET-ACTRIS network (European Aerosol Research Lidar NETwork - Aerosol-Clouds and Trace gases Research InfraStructure). It is capable of profiling aerosol-induced fluorescence and has a high level of automation. These features were used for the development of FLARE-GMM (Fluorescence Lidar Aerosol REcognition based on Gaussian Mixture Model), an automatic aerosol classification algorithm based on a machine learning model. This algorithm identifies the type of aerosols detected by the lidar using their depolarization signature, fluorescence, and atmospheric humidity. Following validation, a statistical study of aerosols present in Lille from 2021 to 2023 was conducted using FLARE-GMM results.
The atmospheric fluorescence profile database was then used to improve the estimation of the hygroscopic properties of aerosols. Independent of the presence of water vapor, fluorescence allows for tracking aerosol concentration changes in a given layer, assuming the aerosols nature in that layer is homogeneous. This ability allows for compensating potential variations in aerosol concentration within the layer, enabling more accurate estimation of hygroscopic properties. This method, evaluated in different scenarios, demonstrated the advantages of using fluorescence. The hygroscopic properties of aerosols in Lille were subsequently estimated through an automatic detection method for hygroscopic growth events, revealing significant uncertainties primarily related to relative humidity estimation. The dependency of hygroscopic property estimation on relative humidity was modeled to better account for this phenomenon and understand how to minimize it.
Finally, fluorescence signals in clouds were studied. This part remains more exploratory, as many unknowns persist regarding the impact of clouds on fluorescence signals. Nevertheless, several hypotheses explaining how clouds might influence fluorescence signals were proposed and discussed. Additionally, a method for inverting the elastic backscatter of aerosols within the cloud layer was proposed. Although this approach currently has large uncertainties and relies on strong assumptions, it offers an original and promising perspective for studying aerosols in cloudy conditions through remote sensing.

Keywords : Remote sensing, Aerosols, Clouds, Lidar