Optical and geometrical properties of clouds and aerosols using ground-based and satellite remote sensing techniques

• Main Authors: Voudouri, Kalliopi-Artemis, Βουδούρη, Καλλιόπη-Άρτεμις
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: Aristotle University Of Thessaloniki (AUTH) 2019
• Subjects: Νέφη παγοκρυστάλλων; Αιωρούμενα σωματίδια; Αλγόριθμοι ταξινόμησης σωματιδίων; Cirrus clouds; Aerosols; Typing algorithms; Φυσικές Επιστήμες; Φυσική; Natural Sciences; Physical Sciences; Subarctic
• Online Access: http://hdl.handle.net/10442/hedi/46654; https://doi.org/10.12681/eadd/46654

This PhD thesis aims to study the geometrical and optical properties of cirrus clouds and aerosols through ground-based measurements with LIDAR remote sensing methods (light detection and ranging). As part of this research, an automatic detection algorithm of defining cloud boundaries is developed and applied to the Laboratory of Atmosperic Physics in Thessaloniki and certified within EARLINET (European Aerosol Research Lidar Network). One part of the study focuses on the analysis of the optical and geometrical characteristics of cirrus clouds over different geographical parts of the world and the comparison with collocated cloud radar retrievals. The object is (i) firstly to provide cirrus geometrical and optical properties in different hemispheres, based on ground-based lidar data and (ii) secondly to attribute any observable differences of cirrus properties to the subarctic and subtropical counterparts. Moreover, the differences in the measured depolarization ratio values inside the cirrus layers are examined in terms of ice crystal shapes. This variability is also, examined incorrelation with temperature dependencies and optical depth. Our aim is to define a classification scheme of ice particles into different shape categories, based on lidar depolarization ratios, in order to be further applied on operation albasis on the cirrus classification algorithms on satellite missions. The other part of the study is focused on the comparison and improvement of the two algorithms, developed within EARLINET for the automated aerosol type characterization of the aerosol layers derived from Raman lidar measurements over the EARLINET station of Thessaloniki, Greece. Our aim is: (i) to check the performance of both supervised learning techniques, when applied to lidar datafrom Thessaloniki station where, typically, variable mixtures of aerosols arepresent and (ii) to investigate the reasons of typing agreement and disagreement with respect to the uncertainties and the threshold criteria applied. A third part is dedicated ...