2018, Papagiannopoulos, Nikolaos, Mona, Lucia, Amodeo, Aldo, D'Amico, Giuseppe, Gumà Claramunt, Pilar, Pappalardo, Gelsomina, Alados-Arboledas, Lucas, Guerrero-Rascado, Juan Luís, Amiridis, Vassilis, Kokkalis, Panagiotis, Apituley, Arnoud, Baars, Holger, Schwarz, Anja, Wandinger, Ulla, Binietoglou, Ioannis, Nicolae, Doina, Bortoli, Daniele, Comerón, Adolfo, Rodríguez-Gómez, Alejandro, Sicard, Michaël, Papayannis, Alex, Wiegner, Matthias

We present an automatic aerosol classification method based solely on the European Aerosol Research Lidar Network (EARLINET) intensive optical parameters with the aim of building a network-wide classification tool that could provide near-real-time aerosol typing information. The presented method depends on a supervised learning technique and makes use of the Mahalanobis distance function that relates each unclassified measurement to a predefined aerosol type. As a first step (training phase), a reference dataset is set up consisting of already classified EARLINET data. Using this dataset, we defined 8 aerosol classes: clean continental, polluted continental, dust, mixed dust, polluted dust, mixed marine, smoke, and volcanic ash. The effect of the number of aerosol classes has been explored, as well as the optimal set of intensive parameters to separate different aerosol types. Furthermore, the algorithm is trained with literature particle linear depolarization ratio values. As a second step (testing phase), we apply the method to an already classified EARLINET dataset and analyze the results of the comparison to this classified dataset. The predictive accuracy of the automatic classification varies between 59% (minimum) and 90% (maximum) from 8 to 4 aerosol classes, respectively, when evaluated against pre-classified EARLINET lidar. This indicates the potential use of the automatic classification to all network lidar data. Furthermore, the training of the algorithm with particle linear depolarization values found in the literature further improves the accuracy with values for all the aerosol classes around 80%. Additionally, the algorithm has proven to be highly versatile as it adapts to changes in the size of the training dataset and the number of aerosol classes and classifying parameters. Finally, the low computational time and demand for resources make the algorithm extremely suitable for the implementation within the single calculus chain (SCC), the EARLINET centralized processing suite.

2018, Pandolfi, Marco, Alados-Arboledas, Lucas, Alastuey, Andrés, Andrade, Marcos, Angelov, Christo, Artiñano, Begoña, Backman, John, Baltensperger, Urs, Bonasoni, Paolo, Bukowiecki, Nicolas, Collaud Coen, Martine, Conil, Sébastien, Coz, Esther, Crenn, Vincent, Dudoitis, Vadimas, Ealo, Marina, Eleftheriadis, Kostas, Favez, Olivier, Fetfatzis, Prodromos, Fiebig, Markus, Flentje, Harald, Ginot, Patrick, Gysel, Martin, Henzing, Bas, Hoffer, Andras, Holubova Smejkalova, Adela, Kalapov, Ivo, Kalivitis, Nikos, Kouvarakis, Giorgos, Kristensson, Adam, Kulmala, Markku, Lihavainen, Heikki, Lunder, Chris, Luoma, Krista, Lyamani, Hassan, Marinoni, Angela, Mihalopoulos, Nikos, Moerman, Marcel, Nicolas, José, O'Dowd, Colin, Petäjä, Tuukka, Petit, Jean-Eudes, Pichon, Jean Marc, Prokopciuk, Nina, Putaud, Jean-Philippe, Rodríguez, Sergio, Sciare, Jean, Sellegri, Karine, Swietlicki, Erik, Titos, Gloria, Tuch, Thomas, Tunved, Peter, Ulevicius, Vidmantas, Vaishya, Aditya, Vana, Milan, Virkkula, Aki, Vratolis, Stergios, Weingartner, Ernest, Wiedensohler, Alfred, Laj, Paolo

This paper presents the light-scattering properties of atmospheric aerosol particles measured over the past decade at 28 ACTRIS observatories, which are located mainly in Europe. The data include particle light scattering (σsp) and hemispheric backscattering (σbsp) coefficients, scattering Ångström exponent (SAE), backscatter fraction (BF) and asymmetry parameter (g). An increasing gradient of σsp is observed when moving from remote environments (arctic/mountain) to regional and to urban environments. At a regional level in Europe, σsp also increases when moving from Nordic and Baltic countries and from western Europe to central/eastern Europe, whereas no clear spatial gradient is observed for other station environments. The SAE does not show a clear gradient as a function of the placement of the station. However, a west-to-east-increasing gradient is observed for both regional and mountain placements, suggesting a lower fraction of fine-mode particle in western/south-western Europe compared to central and eastern Europe, where the fine-mode particles dominate the scattering. The g does not show any clear gradient by station placement or geographical location reflecting the complex relationship of this parameter with the physical properties of the aerosol particles. Both the station placement and the geographical location are important factors affecting the intraannual variability. At mountain sites, higher σsp and SAE values are measured in the summer due to the enhanced boundary layer influence and/or new particle-formation episodes. Conversely, the lower horizontal and vertical dispersion during winter leads to higher σsp values at all low-altitude sites in central and eastern Europe compared to summer. These sites also show SAE maxima in the summer (with corresponding g minima). At all sites, both SAE and g show a strong variation with aerosol particle loading. The lowest values of g are always observed together with low σsp values, indicating a larger contribution from particles in the smaller accumulation mode. During periods of high σsp values, the variation of g is less pronounced, whereas the SAE increases or decreases, suggesting changes mostly in the coarse aerosol particle mode rather than in the fine mode. Statistically significant decreasing trends of σsp are observed at 5 out of the 13 stations included in the trend analyses. The total reductions of σsp are consistent with those reported for PM2.5 and PM10 mass concentrations over similar periods across Europe.