Biomass burning aerosols in the southern hemispheric midlatitudes as observed with a multiwavelength polarization Raman lidar

In this paper, we present long-term observations of the multiwavelength Raman lidar Polly conducted in the framework of the DACAPO-PESO campaign. Regardless the relatively clean atmosphere in the southern mid-latitude oceans region, we observed regularly events of long-range transported smoke, originating either from regional sources in South America or from Australia. Two case studies will be discussed. Both cases were identified as smoke events and they occurred on 5 February 2019 and 11 March 2019. For the first case considered, the lofted smoke layer was located at an altitude between 1 5 and 4.2 km, and apart from the predominance of smoke particles, particle linear depolarization values indicated the presence of dust particles in the layer. Mean lidar ratio values at 355 and 532 nm were 49 ± 12 and 24 ± 18 sr respectively, while the mean particle linear depolarization was 7.6 ± 3.6 % at 532 nm. The advection of smoke and dust particles above Punta Arenas affected significantly the available CCN and INP in the lower troposphere, and triggered effectively ice crystal formation processes. Regarding the second case, the thin smoke layers were observed at altitudes between 5.5–7, 9 and 11 km. The particle 10 linear depolarization ratio at 532 nm increased rapidly with height, starting from 2% for the lowest two layers and increasing up to 9.5% for the highest layer, indicating the possible presence of non-spherical coated soot aggregates. INP activation was effectively facilitated. The long-term analysis of the one year of observations showed that tropospheric smoke advection over Punta Arenas occurred 16 times (lasting from 1 to 17 hours), regularly distributed over the period and with high potential to influence cloud formation in the otherwise pristine environment of the region. 15

period are available at http://polly.tropos.de in near real-time.

Auxiliary data
The Sun photometer located at UMAG (53.1 • S, 70.9 • W, 25 m a.s.l., site name: Punta_Arenas_UMAG), measures the columnintegrated extinction coefficient at 8 channels from 340 to 1640 nm. Level 1.5 AOD data are used with a corresponding uncertainty of 0.01-0.02 (Holben et al., 2001). 100 For the analysis of the air mass transport the HYSPLIT model was used (Stein et al., 2015). HYSPLIT model calculates backward and forward trajectories of air masses for simulations of dispersion and deposition at a given location. In order to identify the aerosol sources and to create some statistical basis of the aerosol conditions during DACAPO-PESO, we used ensemble backward trajectories combined with a land cover classification for a temporally and vertically resolved air mass source Prediction (NCEP). Each ensemble is generated using a small spatial offset in the trajectory endpoint. Whenever a trajectory is located below the planetary boundary layer (PBL) height provided by the GDAS1 data ("reception height"), the land cover 110 is categorized using custom defined polygons according to land mass boundaries. It is therefore assumed that an air parcel is influenced by the surface type if the corresponding trajectory is below the PBL height. The residence time for each category is then the total time an air parcel fulfilled this criterion by land cover category. This calculation is repeated in steps of 3 h in time and 500 m in height in order to provide a continuous estimate on the source of the air mass and a first indication on the potential aerosol load.

115
In this section, two of the most distinctive cases with lofted aerosol layers that were identified above Punta Arenas are presented in detail. The first one presents a 3-km thick lofted aerosol layer, located in the lowest troposphere between 1 and 4.2 km.
Source attribution with TRACE indicated that the layer originated from Central and Central-South Chile, where wildfires occurred at the same time. The second case shows thin (geometrically and optically) aerosol layers, as observed by Polly XT , 120 at altitudes of 5.5-7, 9, and 11 km after a long-range transport event from Australia. The retrieved optical properties, indicate that the layers consist of smoke particles. The retrieved optical parameters for the two case studies are summarized in Table 1.

Lofted smoke layers from South America
From 4 February to 5 February 2019, several lofted aerosol layers were observed between 1 and 4.5 km height. Figure 2 shows a two-day overview of main lidar parameters, the Sun-photometer-derived AOD, atmospheric thermodynamics, and a target 125 classification measured by Polly XT . As indicated in Figure  can conclude that the layer consisted mainly of fine-mode aerosol particles (i.e. particles smaller than 1µm).
The following day, on 5 February 2019, a descending layer was observed between 07:00 and 12:00 UTC. The layer extents from 1 to 4.2 km height. In contrast to the layer from the previous day, this one is characterized by higher backscatter and 135 low to moderate depolarization ratio with the most pronounced features found between 1.6 and 2.8 km. The higher values of volume linear depolarization ratio, with respect to the previous day, indicate the presence of non-spherical particles. Similarly to the previous day, the AOD during that event was low, with a maximum and a mean value of 0.09 and 0.05 respectively (at 500 nm), which is however still three times higher than the average. Around 18:00 UTC we observe the formation of a cloud, which according to the target classification consists of ice crystals, thus indicating that ice formation is most likely supported 140 by the advection of the smoke particles acting as INP, as discussed below.
Further context of the air masses arriving at Punta Arenas is given by the HYSPLIT backward trajectory analysis shown in Figs. 3 and 4. The 10-day backward trajectories presented in Fig. 3 show that the vast majority of the air parcels arriving over Punta Arenas at 3000 m height originated from the southern Pacific Ocean. Shortly (1-2 days) before the arrival of the air parcels to Punta Arenas, they passed over the regions of Central and Central-South Chile (Figs. 3, 4a). Fig. 3c demonstrates 145 that no precipitation occurred while the trajectories passed over the land masses. Air masses crossed above lands that were categorized as forest, savanna, grass and barren (Fig. 4b) and belong to South America (Fig. 4a). According to FIRMS (Fire Information for Resource Management System), on the 2nd and 3rd of February 2019, when the air parcels were located above the aforementioned regions, active fires and wildfires occurred over the region passed by the trajectories (not shown here). During the summer months (December to February), wildfires are common in Chile, especially between the regions of 150 Valparaíso and Los Lagos, due to the extremely dry conditions and the prevailing westerlies. Relative humidity in the aerosol layer (Fig. 3b, Fig. 2e,f) was mostly below 60% and therefore we do not expect any hygroscopicity effects on the aerosol observed. As the air masses crossed above active fire regions 1-2 days before they were observed by our system, the presence of partly coated soot particles is expected. Apart from smoke particles, soil dust is also expected to be present in the observed layer due to turbulent fire-related winds that developed above the burning areas (Nisantzi et al., 2014;Wagner et al., 2018).

170
Europe, Siberia, Canada), fire type (smoldering or flaming combustion) and the age of the smoke particles are inducing the relatively large range of values reported in the literature. Within the smoke plume, linear particle depolarization ratio ( Fig.   5d) ranges from 2% to 14% (mean: 7.6 ± 3.6%) at 532 nm with a maximum being observed between 2.2 and 2.3 km. The slightly elevated particle linear depolarization ratio at this altitude can be explained by the presence of irregularly shaped soil dust particles (Nisantzi et al., 2014) along with some partly coated soot particles. The smoke-dust mixture is mostly confined 175 between 2.0 and 2.8 km and it was well identified by the target classification algorithm as a partly non-spherical mixture (Fig.   2d).
The potential impact of this event on cloud and precipitation evolution and life cycle can be investigated by means of CCN and INP concentrations. CCN and INP number concentrations have been derived for this plume (1-4.2 km), uncertainty range for this approach is of a factor of 2 and 3, respectively (Mamouri and Ansmann, 2016). The lidar-derived total CCN 180 concentration, given for 0.2% supersaturation with respect to liquid water, has a layer mean value and a standard deviation of 642 ± 192 cm −3 . The contribution of dust to the total CCN concentration is about 6% and the rest is attributed to soot particles. The obtained CCN number concentration is found to be strongly increased compared to other values reported for this convection. In the second case, several geometrically and optically thin smoke layers were observed, after long-range transport from Australia.
The differences in the obtained optical properties between the observed layers, reflect differences in the amount and the age 285 (chemical composition) of the observed smoke particles. As it can been seen from Table 1, the 532-nm particle depolarization ratio for the smoke and soil dust mixtures is slightly lower than the one for the pure smoke layer. In the first case, the smoke particles within the layer were freshly emitted (emitted 1-2 days before observation), while in the second case, the smoke particles were aged (observed 9 days after emission). Previous studies have shown that the depolarization ratio values at 532 nm for smoke freshly emitted into the atmosphere are generally low (due to the sphericity of the newly formed particles) and ). In addition, smoke particles advected from Australia at high altitudes were found to likely facilitate the formation of ice crystals.
The Southern Oceans and more specifically Punta Arenas, are usually characterized as pristine environments, where clean marine aerosol conditions prevail. However, this study demonstrates that this is not always the case. Lofted aerosol layers, transported either from short or long distances, occur regularly and influence radiation and cloud formation processes above 305 the region. Even small perturbations influence cloud properties strongly above Punta Arenas, while in the polluted northern hemisphere such perturbations would not be observable.