Biomass burning events measured by lidars in EARLINET - Part 2: Optical properties investigation.

. Biomass burning episodes measured at 14 stations of the European Aerosol Research Lidar Network (EARLINET) over 2008-2017 were analysed using the methodology described in "Biomass burning events measured by lidars in EARLINET - Part 1: Data analysis methodology" (Adam et al., 2020, this issue). The smoke layers were identified in lidar optical properties profiles. A number of 795 layers for which we measured at least one intensive parameter was analysed. These layers were 35 geographically distributed as follows: 399 layers observed in South-East Europe, 119 layers observed in South-West Europe, 243 layers observed in North-East Europe, and 34 layers observed in Central Europe. The mean layer intensive parameters are discussed following two research directions: (I) the long-range transport of smoke particles from North America, and (II) the 2 smoke properties (fresh versus aged), separating the smoke events into four continental source regions (European, North American, African, Asian or a mixture of two), based on back trajectory analysis. The smoke detected in Central Europe (Cabauw, Leipzig, and Hohenpeißenberg) was mostly transported from North America (87% of fires). In North-East Europe (Belsk, Minsk, Warsaw) smoke advected mostly from Eastern Europe (Ukraine and Russia), but there was a significant contribution (31%) from North America. In South-West Europe (Barcelona, Evora, Granada) smoke originated mainly from 5 the Iberian Peninsula and North Africa (while 9% were originating in North America). In the South-East Europe (Athens, Bucharest, Potenza, Sofia, Thessaloniki) the origin of the smoke was mostly local (only 3% represented North America smoke). The following features, correlated with the increased smoke travel time (corresponding to aging) were found: the colour ratio of the lidar ratio (i.e., the ratio of the lidar ratio at 532 nm to the lidar ratio at 355 nm ) and the colour ratio of the backscatter Ångström exponent (i.e., the ratio of the backscatter-related Angstrom exponent for the pair 532 nm – 1064 nm to the one for 10 the pair 355 nm – 532 nm) increase, while the extinction Ångström exponent and the colour ratio of the particle depolarization ratio (i.e., the ratio of the particle linear depolarization ratio at 532 nm to the particle depolarization ratio at 355 nm) decrease. The smoke originating from all continental regions can be characterized on average as aged smoke, with a very few exceptions. In general, the long range transported smoke shows higher lidar ratio and lower depolarization ratio compared to the local smoke. that the climate change induces an increase in the number of fires. Flannigan et al. (2000) modelled the climate change impact, demonstrating an increase of forest wildfire activity. Carvalho et al. (2011) modelled the impact of forest fires in a changing climate on air quality (a case study on Portugal) showing a strong impact on ozone and PM10 (particulate matter with size diameter below 10  m). One of the current challenges is in evaluating accurately the role of BB in climate change. Besides the BB impacts on 25 climate change, Keywood et al. (2013) describe the impacts of climate change on BB (e.g., fire severity, increase of fuel consumption). The authors state that, based on the BB impact on air pollution, climate, poverty, security, supply and biodiversity, a more effective control of the fires is needed, along with continuous and improved and Thessaloniki stations. For the South-West region, we identified a number of 197 fires in North America (8.7%) and 2066 elsewhere, most of the latter being located in the Iberian Peninsula and North Africa. Most of the fires occurred in the region [0° 10°W] x [35°N 43°N] that corresponds mostly to the Iberian Peninsula. Other fires were located over [0° 20°E] x [30°N 40°N], corresponding to North Africa (mostly North Algeria) and Sicily in South Italy. Most of the measurements were taken at Granada station. 25 For the Central Europe region, we have found 1420 fires originating in North America (86.9 %) and 214 elsewhere, most of the latter located in East Europe. Most of the fires occurred over [80°W 75°W] x [51°N 53°N] region, which corresponds to North America (East Canada). Most of the measurements in Cabauw station were performed over the LRT of smoke from North America, contributing to the histogram peak indicating North American locations. The stations of Hohenpeißenberg and Leipzig contained a ~24 % and ~79 % LRT of smoke from North America, but their number of measurements is much smaller 30 than that of the Cabauw. increase of the EAE (Fig. 6b), while the CR LR decreases (Fig. 6c). Based on panel b), we observe that except the case with low EAE (<0.5) and CR PDR <1 which indicates aged particles and larger depolarization at 355 nm, the depolarization at 532nm can be higher for either fresh or aged smoke. The dataset is not statistically significant, but increased number of samples in future studies is expected to reveal the statistical significance of this correlation. A slight 25 decrease of the CR PDR with smoke travel time was observed (see lower values for EUNA, NA, AS and EUAS), while the CR BAE maintained similar values for all the source regions. An increase of EAE versus decreasing CR LR (Fig. 6d), evident especially for the North-East region (Fig. 7d), was reported also by Samaras et al. (2015) and Janicka et al. (2019). No clear relationship between CR BAE and CR LR (Fig. 6e), CR BAE and EAE (Fig. 6f) and CR BAE and CR PDR (Fig. 6a) was found. Veselovskii et al. (2015) showed that the relationship between EAE and BAE is not straightforward, pointing out that while 30 EAE depends mainly on the particle size, BAE depends on both particle size and complex refractive index. The relationship between BAE and EAE was analysed from the relative humidity (RH) perspective by Su et al. (2008) and Wang et al. (2019). layers were labelled as smoke based on a combined analysis of Hysplit backtrajectories and the FIRMS fire locations. The smoke was further labelled as ‘mixed’ if multiple fire sources contributed to the smoke measurement. For the smoke originating in North America, the smoke was labelled as ‘pure North America’ or ‘mixed’ (with contribution from fires in Europe). We demonstrated that in most of the 25 cases the smoke was mixed and the quantification (based on number of fires and detections) of the contributing fires to the mixture explains the wide range of values obtained for the intensive parameters. The statistics over all LRT events from North America revealed no significant difference between the measurements where the smoke was originating solely from North America and the measurements with mixed smoke (having origin in both North America and local). This suggests that the contribution of the local smoke is not significant. Based on the LR values, a moderate 30 absorption at 355 nm (46 sr) and a high absorption at 532 nm (66 sr) were observed. The mean CR LR and EAE suggest aged measurement region, fresh smoke from the EUNA source region and a mixture of fresh and aged smoke originating from the EUAS was measured. In the North-East, region fresh smoke originating from EUAF was measured. For the South-West region with European or African source regions we obtained a CR LR of 0.8 and an EAE of 1. We assumed that the smoke measured was aged based on the high RH (in agreement with Veselovskii et al., 2020). The lowest absorption 25 was determined for the Central region (LRs < 36 sr). The South-West region displayed a highly absorbing smoke (61 sr < LR@355 < 79 sr and 64 < LR@532 < 91 sr). The South-East region displayed smoke with a medium/high absorption at 532 nm (50–72 sr) and a low/medium absorption at 355 nm (31–48 sr). The smoke measured in the North-East region has a medium to very high absorption at 532 nm (57–91 sr) and a medium to high absorption at 355 nm (46–78 sr). The quite diverse absorption was determined for the different measurement’s regions, even for smoke from the same 30 continental source region, which may be related, among others, with different RH conditions (e.g., Veselovskii et al, 2020). In line with previous studies, we showed that BAE and further CR BAE do not show specific values based on sources and no trends, and thus, they cannot be used to identify the smoke type. In order to easily quantify the smoke type, LR (CR LR ) and EAE are essential. The aerosol typing algorithm developed by Papagiannopoulos et al. (2018) based on 3 backscatter and 2


Introduction
The biomass burning (BB) context was given in Adam et al., 2020 (Part 1, this ACP issue). Therein, the information on BB was reviewed, its importance and role on radiative transfer, air quality and human health, were highlighted, and an overview of the fire monitoring perspective was discussed.
There is a direct link between climate change and forest wildfires. The European Union reports of fires occurrence over Europe 20 (http://effis.jrc.ec.europa.eu/reports-and-publications/annual-fire-reports, last access 13 July 2021) indicate that the climate change induces an increase in the number of fires. Flannigan et al. (2000) modelled the climate change impact, demonstrating an increase of forest wildfire activity. Carvalho et al. (2011) modelled the impact of forest fires in a changing climate on air quality (a case study on Portugal) showing a strong impact on ozone and PM10 (particulate matter with size diameter below 10 m). One of the current challenges is in evaluating accurately the role of BB in climate change. Besides the BB impacts on 25 climate change, Keywood et al. (2013) describe the impacts of climate change on BB (e.g., fire severity, increase of fuel consumption). The authors state that, based on the BB impact on air pollution, climate, poverty, security, food supply and biodiversity, a more effective control of the fires is needed, along with continuous and improved monitoring.
EARLINET (European Aerosol Research Lidar Network; https://www.earlinet.org/ last access: 10 July 2021; e.g., Pappalardo et al., 2014) provides high temporal and spatial resolution ground-based aerosol measurements, and represents a valuable tool 30 for smoke monitoring. EARLINET is part of the Aerosol Cloud and Trace Gases Research Infrastructure (ACTRIS) (https://actris.eu, last access: 13 July 2021). There are numerous studies describing various BB events over Europe, most of https://doi.org/10.5194/acp-2021-759 Preprint. Discussion started: 18 October 2021 c Author(s) 2021. CC BY 4.0 License. them focusing on the optical properties of either fresh/local aerosol (e.g. Balis et al., 2003;Alados-Arboledas et al., 2011;Sicard et al., 2012;Nicolae et al., 2013;Stachlewska et al. 2017a,b;Osborne et al., 2019) or aged/long range transported aerosol (Wandinger et al., 2002;Mattis et al., 2003;Müller et al., 2005;Ortiz-Amezcua et al., 2017;Stachlewska et al., 2018;Vaughan et al., 2018;Hu et al., 2019;Sicard et al., 2019, Baars et al., 2019. The aim of this study is to find specific features of the smoke originating from North America and investigate different 5 continental origin of the smoke for each of the four considered geographical regions. The smoke origin is assessed by backtrajectory analyses and the FIRMS product. The analysis is made using intensive parameters (referred to as IPs), which are independent of the aerosol load and are solely aerosol type dependent. This paper presents the Part 2 of investigation of biomass burning episodes as measured by EARLINET, and it focuses on results interpretation. Part 1  described in detail the methodology used to analyse lidar data. Nonetheless, a short overview of the methodology is given in 10 Section 2. In Section 3, we analyse the results for the smoke originating in North America. In Section 4, we focus on results from four European geographical regions, with different continental smoke origin. In Section 5, we provide the summary and conclusions. A list of acronyms used in the current work is given in Appendix A. The location of the EARLINET stations along with the chosen geographical regions are given in Appendix B.

Methodology 15
The methodology steps are shown in Fig. S1 (Fig. 2 in Adam et al., 2020). The input for the analysis is the EARLINET/ACTRIS so-called backscatter (b) and extinction (e) files providing the vertical profiles of particle backscatter coefficient, particle extinction coefficient, and particle linear depolarization ratio (when available). In general, for most of the stations the range resolution of profiles is 3.75 m for backscatter coefficients and 60 m for extinction coefficients and the profiles are averages of 1 h (i.e., various resolutions were used by the stations; Adam et al., 2020). The files are allocated by 20 the stations to the Forest Fire category in the EARLINET/ACTRIS database when an investigation at the station level highlighted the potential presence of smoke layer. The aerosol layer assignment is made manually by the EARLINET stations and it is typically made by means of investigation of intensive parameters (Ångström exponent, lidar ratios, linear particle depolarization ratio, etc), model outputs, backward trajectory analyses, and ancillary instruments data if available. Data are quality assured following the EARLINET Quality Check (QC) procedures. Most of the data used for this paper are the 25 EARLINET data reported in Forest Fire category labelled as Level 2 data, where 2341 files out of 3589 files (input data) were compliant with all the QC v2.0, at the date of 23 April 2019 .
Additional data check procedures were applied for the specific purposes and analysis, as described in detail Part 1, here recalled in short. i) For the analysis, a distinct peak in signal amplitude well above the SNR was considered as essential for a layer identification. ii) For identifying the layer(s) affected by smoke a ten days backtrajectory was computed per each layer using 30 the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT; Stein et al., 2015;Rolph et al., 2017). The meteorological model applied was the Global Data Assimilation System (GDAS), with 0.5 resolution. The identification of https://doi.org/10.5194/acp-2021-759 Preprint. Discussion started: 18 October 2021 c Author(s) 2021. CC BY 4.0 License. the smoke layers was assessed based on the hypothesis of an existing fire within 100 km and  1 h from the location and time of the air mass, respectively. The location of the fires was provided by the Fire Information for Resource Management System (FIRMS) (https://firms.modaps.eosdis.nasa.gov/, last access: 13 July 2021) that uses satellite observations of the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard the Aqua and Terra satellites (Davies et al., 2009). In the current study, a fire was defined by a specific location (given by latitude and longitude) and a specific time. According to MODIS, 5 latitude and longitude are the middle points of 1 km grid (centre of 1 km fire pixel) but not necessarily the actual location of the fire as one or more fires can be detected within the 1 km pixel. The uncertainties related with the Hysplit backtrajectories or the FIRMS database were not considered. iii) The layer's mean optical properties were calculated only for sufficient signal to noise ratio (SNR  2) and the number of available data in the layer being  90% . The number of layers available after each criterion was presented therein (Table 2). iv) As a last step, outliers were removed from the data, i.e., the 10 mean intensive parameters in the layers were discarded when outside the following boundaries: 20 sr ≤ LR@355 ≤ 150 sr, 20 sr ≤ LR@532 ≤ 150 sr, -1 ≤ EAE ≤ 3, -1 ≤ BAE@355/532 ≤ 3, -1 ≤ BAE@532/1064 ≤ 3, 0 ≤ PDR@355 ≤ 0.3, and 0 ≤ PDR@355 ≤ 0.3. BAE represents the backscatter Ångström exponent and PDR represents the linear particle depolarization ratio (see Appendix A). As mentioned in Part 1, there was a low number of IPs removed (outliers) based on the above predefined ranges (3.7%). In general, the number of the optical properties analysed is lower than the number of layers due to 15 the following reasons: a) profiles of some optical properties are not available, b) some profiles do not cover the entire altitude range, c) the mean values are calculated only if 90% of the data are available while the SNR  2. Thus, we analysed a number of 795 layers for which we identified at least one intensive parameter.
The mean, median, minimum and maximum values of the intensive parameters for all of the stations providing at least one parameter (except Sofia station) are shown in Table 1. The number of available values for each variable is shown as well (# 20 lines).
In the current study, the smoke is considered fresh if LR@355 > LR@532 and EAE > 1.4 (Nicolae et al., 2013). Conversely, the smoke is considered aged when LR@532 > LR@355 and EAE < 1.4. LR denotes the lidar ratio and EAE the extinction Ångström exponent. These findings by Nicolae et al. (2013), based on lidar measurements were confirmed by measurements with an aerosol mass spectrometer which allowed to estimate the degree of oxidation in BB aerosol. The colour ratio CR 25 (spectral ratio) is the ratio of an optical parameter or an intensive parameter at two wavelengths (Appendix A). Here we refer to the colour ratio for the following intensive parameters: LR, BAE and PDR. We investigate the values of the CRs and we expect that they can be associated with fresh or aged smoke (short versus long distance smoke transport) and further be a fingerprint of the smoke as compared with other types of aerosols.
An event represents a series of BB measurements over a specific period of time. Thus, an event and a period are 30 interchangeable. A measurement represents the data acquisition for a specific time, where the time is the average over 1 h. A measurement can contain one or more layers over the vertical profiles.

Biomass burning events originating in North America
In total, 24 events (periods of measurements) of smoke originating in North America were identified, for which at least one intensive parameter was retrieved. The events occurred in 2009 and 2011 and during 2012-2017. Eight events represented measurements of smoke coming solely from North America ('pure North America'), while the others represented 'mixed' smoke (mixture of North American and local smoke, i.e., fires were found along backtrajectory both in North America and 5 locally). "Local smoke" refers to smoke originating in European locations, in general. In a few cases, the smoke came from North Africa or Middle East. The number of fires as well as the number of their detections (a fire can be detected more than once) are quantified.   Table 2. Compared to the values found in the rather limited existing literature for smoke originating in North America and measured over Europe (only tropospheric measurements were considered), we noted several IP values (especially for BAE@355/532) that fall outside of the range reported. The large value for the mixed smoke EAE may be due 20 to the contribution of the local, fresh smoke. At a closer look, the large 'pure N America' EAE value, recorded on 4 July 2013 in Thessaloniki in a layer at ~ 3.6 km altitude, corresponds to air masses reaching ~ 9 -11 km over the fires in North America.
It is possible that the fires did not reach that altitude and thus the measurements for that layer may come from other sources.
On the other hand, biomass burning particles can be found even in the lower stratosphere (e.g., Hu et al., 2019). The smallest EAE value (negative) may be due to dust contamination for a measurement performed in Granada on 19 August 2013 at 20:45 25 UTC, when fires in Portugal and North America were found along the backtrajectory. We also observe that mean PDR values are in general smaller if compared to the mean over the values reported in literature. However, still within the extreme values for smoke originating in North America (see Fig. 1 and Table 2). The minimum value reported for PDR@355 was 0.010.001 and for PDR@532 0.0230.003 ). An EAE extreme value of -0.3 was reported by    shown that the layers accounted for ~ 40%, 30% and 70% of the total AOD for the three stations, respectively. Colour ratio of lidar ratios was around 2 while EAE was < 1.  (mostly in Ukraine). 5 In summary, the main fire sources are located in: East Europe (especially Ukraine and West Russia), South Europe (Iberian Peninsula, Italy, Balkan region) and North America. Wildfires in the West Russian regions and Ukraine occur each year from March to October. Events of small particles transport, in the boundary layer, from these regions to the North-West Europe (Belarus, Poland, Germany, Nordic countries and European Arctic) are regularly recorded (Lund Myhre et al., 2007). Such transport of biomass burning aerosol can be extremely fast and affect relative humidity within the boundary layer (Stachlewska 10 et al., 2017b). Transport of such particles to Arctic regions is contributing to arctic haze by significantly alternating the arctic aerosol properties (Stachlewska and Ritter, 2010), and thus contributing to Arctic warming.
The histogram of the backtrajectories (Figs. S3) revealed some preferential air circulation patterns for three of the regions (Central, South-West and North-East), with one common pattern being the circulation over the Atlantic. For the South-West region, we identified a vortex type circulation over North Africa as the main air pathway. For the North-East region we 15 observed other patterns as well: a circulation from Iberian Peninsula, a circulation from East Europe (Caspian Sea), and a circulation over North Europe (Scandinavian Peninsula and West Russia).

Intensive parameters by geographical regions
A statistical investigation of the intensive parameters was performed, based on the continental fire source origin. As mentioned in Part 1, the following continental source origins were considered: Europe (EU), Africa (AF), Asia (AS), North America 20 (NA), and combinations of two or more (e.g., EUAF=EU+AF, etc). The statistical analysis was performed over all of the available cases, despite their low number. Thus, we label the series with less than five cases as low statistics, and thus the results are just indicative and it is not safe to draw any conclusion. We reproduce here the results for South-East for a straightforward comparison with the other three regions. To thoroughly assess the smoke type, the scatter plots of EAE and CRLR are used. 25 Assuming that the aerosol size for the smoke layers is not significantly changing (e.g., Papanikolaou et al., 2020, for 532 nm), the LR is an indication for the absorption capacity of the particles and thus, the following description of the absorption (based on the values of the LR) can be further used: low absorption for LR < 40 sr, medium absorption for 40 sr < LR < 60 sr, high absorption for 60 sr < LR < 80 sr and very high absorption for LR > 80 sr.

South-East region 30
The mean values of the IPs in South-East region are shown in Fig. 2 (Fig. S4.1).

North-East region 10
The mean values for the IPs are shown in Fig. 3. The majority of the events were recorded for the EU source region. LR@532 is slightly larger than LR@355 for EU source region, around 75sr. Based on low statistics, the two LR values for other source regions tend to be different. EAE ~1.4 is obtained for the EU source region which suggests a mixture of fresh and aged smoke.
It would be worth investigating further in the future to see if the decrease of EAE from EUAF towards EUAS and EUNA holds, based on different local contribution to the mixed smoke. BAE values are similar, except for AS (low statistics), where 15 are larger. BAE@355/532 is larger than BAE@532/1064 for all source regions, which denotes more backscatter at 355 nm.
The similarity between NA and EUNA source regions suggests a major contribution from NA to the EUNA mixture.
PDR@532 is larger than PDR@355, except for EUNA source region. As expected, the scatter plots between various IPs show a linear regression between the two PDR and between the two BAE (Figs. S4.2). Large values of LR for EU suggest more absorption, if compared to the South-East region. 20

South-West region
The mean values for the IPs for South West region are shown in Fig. 4 to EUNA is more from North America. Smaller values are observed for EUAF source region. BAE@355/532 is larger than BAE@532/1064 for all but AF source region. Based on scatter plots (Figs. S4.3), we observe a direct proportionality between the two BAEs (observed also for North-East region).

Central region
The mean values for the IPs in the Central region are shown in Fig. 5 values were performed for EU source region. For the scatter plot between the two BAEs, a mean value from NA source region was also available. 5

Statistical analysis over all regions
We perform the analysis based on the mean IP values as a function of continental source region. We consider analysing the scatter plots between the different CRs and EAE, where, for each scatter plot, the mean values correspond to the same measurements. Still, different scatter plots can refer to slightly different sets of measurements.
As a general statement, we consider that the cases where we have only one or two measurements are not statistically significant, 10 a good confidence is considered when at least five measurements are available. Therefore, the results discussed below should be carefully treated in such situations.

Observations based on scatter plots
The general observations based on the scatter plots between CR or EAE are shown in Fig. 6. Each point on the graph represents the average for one measurement region (South-East, South-West, Central and North-East Europe) and one continental source 15 (EU, AF, AS, NA, EUAF, EUAS, EUNA). All available data are averaged. We added for comparison the mean values (red circles, Fig. 6) found in literature (Table S1, Part 1). For the North-East region, the PDRs provided by the Warsaw station allowed a complete comparison. For a better visualization of the mean CR (Fig. 6) and the corresponding IPs, in Fig. 7 are shown the CR and IP values versus continental source regions (i.e., the panels a-f of Fig. 7 corresponds to the a-f scatter plots of Fig. 6). The right-hand side axis shows the number of available measurements for the scatter plots. The black line represents 20 the number of five cases.
For increasing CRPDR we found an increase of the EAE (Fig. 6b), while the CRLR decreases (Fig. 6c). Based on panel b), we observe that except the case with low EAE (<0.5) and CRPDR<1 which indicates aged particles and larger depolarization at 355 nm, the depolarization at 532nm can be higher for either fresh or aged smoke. The dataset is not statistically significant, but increased number of samples in future studies is expected to reveal the statistical significance of this correlation. A slight 25 decrease of the CRPDR with smoke travel time was observed (see lower values for EUNA, NA, AS and EUAS), while the CRBAE maintained similar values for all the source regions. An increase of EAE versus decreasing CRLR (Fig. 6d), evident especially for the North-East region (Fig. 7d), was reported also by Samaras et al. (2015) and .
No clear relationship between CRBAE and CRLR (Fig. 6e), CRBAE and EAE (Fig. 6f) and CRBAE and CRPDR (Fig. 6a) was found. Veselovskii et al. (2015) showed that the relationship between EAE and BAE is not straightforward, pointing out that while 30 EAE depends mainly on the particle size, BAE depends on both particle size and complex refractive index. The relationship between BAE and EAE was analysed from the relative humidity (RH) perspective by Su et al. (2008) and Wang et al. (2019). They showed that the relationship of BAE and EAE depends on the RH values and, thus, one can find correlated and anticorrelated behaviours. However, the RH influence is out of the scope of this study. Still, we made use of RH values for an individual case shown below.
As seen in Fig. 7, different features are observed for different measurement regions. Based on the EAE-CRBAE scatter plot ( Fig. 6f), Fig. 7f indicates for the source regions EUAF, EUAS and EUNA the following. For the North-East region, EAE 5 decreases (from EUAF towards EUAS and EUNA), while both BAE increase, but CRBAE is similar. For the South-West region, EAE and both BAE increase from EUAF to EUNA source regions, while CRBAE is similar. For the South-East region, EAE increases, while no signature is found for BAE and CRBAE. The Central region provides data for the EU and NA source regions.
Here, EAE, both BAE and CRBAE decrease from EU to NA source regions. Considering the findings of Veselovskii et al. (2015), we conclude that, for the Central region, the fine particle mode is predominant for the EU source region (as compared 10 with the NA source region), result which is expected. For the South-East region we find a larger amount of fine particles for EUNA source region as compared to EUAS and EUAF source regions. This implies a large contribution of the EU source region to the mixture. For the North-East measurement region, we also find an increase for BAE@355/532, while based on the LR and CRLR values, the absorption at 532 nm increases from EUAF towards EUAS and EUNA.

Continental source regions 15
Based on the results shown in Figs. 6 and 7, one can assess how distinct are the characteristics of the smoke coming from various continental source regions as observed in different geographical regions. The mean values are shown in Table 3 for each of the d)-f) scatter plots presented in Fig. 6. The smoke type (fresh versus aged) is assessed based on the values of the CRLR and EAE. Information on the smoke absorption and depolarization ratio (where available) is provided. There is no clear relationship between CRBAE and EAE or CRLR (see Fig. 6d-f), while a slight decrease of EAE with increasing CRLR is captured. 20 The highlighted values in Table 3 show the occurrences with low statistics and thus, more corroborated results are needed in the future to draw unambiguous conclusions.
Except for one isolated case, we obtained positive values for BAE (and CRBAE), which indicates more backscattering towards smaller wavelengths, and low depolarization ratios (all PDR < 0.1). Except for two extremes (-1.6 and 3.2), all CRBAE values range between 0.18 and 1.6. The CRPDR (available for the North-East region only) has the largest value for the EUAF source 25 region, followed by EU and EUAS. The lowest CRPDR and EAE values were found for the EUNA source region, characterized also by the highest CRLR (aged smoke; less depolarizing and more absorbing at 532 nm). The high EAE values of the smoke mixtures are likely due to the large EU contribution (EAE value for the North-East region with EUAF origin is 1.46, for the South-East region with EUAS origin is 1.5, and for the South-East region with EUNA origin is 1.9). Table 4 summarizes the key observations over BB layers according to its source and measurement region. 30 The main features are the following. In the South-East region generally aged smoke with the EU source was measured. For the other source regions (low statistics), the smoke was labelled as aged for NA and EUAF regions, a mixture of fresh and aged smoke from the EUAS source region and fresh smoke from EUNA source region. For the mixed source regions (EUAF, https://doi.org/10.5194/acp-2021-759 Preprint. Discussion started: 18 October 2021 c Author(s) 2021. CC BY 4.0 License.
provided medium absorbing particles. The other values are indicative (low statistics). The smoke from EUAS and EUNA show medium absorption (as for EU) suggesting the influence of EU smoke contribution. However, the particles size is slightly smaller, based on EAE values. The smoke from AF and NA indicates lower absorption at 355nm. High absorption is observed for EUAF source region and it does not resemble either EU or AF regions. 5 In the South-West region (low statistics except EUAF), aged and highly absorbing smoke particles from all source regions were measured. For the EU and AF source regions (which have similar EAE and LR values), we assume aged smoke, based on the high RH (where CRLR < 1 and EAE ~ 1).
In the Central region (low statistics), aged smoke from EU and NA source regions was measured, displaying a low absorption for the EU source region and higher backscatter values at 1064 nm for the NA source region. More corroborative measurements 10 are needed to draw solid conclusions.
The North-East region displayed mixed fresh and aged smoke from the EU source region (highly absorbing). For the other regions, the few values are only indicative. For EUAF source region, we measured fresh smoke (probably due to EU contributions) with less absorption at 532nm. The smoke from EUAS and EUNA was labelled as aged, showing very high absorption at 532nm. 15 Higher/lower depolarization at 355/532 nm was observed for the LRT of smoke from North America (as for the North-East region). Based on a single continental source, in all regions, aged smoke was measured, except the North-East with a mixture of fresh and aged smoke from the EU source region. Based on two continental sources (mixtures), the regions measure either aged, fresh or mixed (aged and fresh) smoke, depending on the lower or higher contribution of the local source.

Summary and conclusions 20
The present study shows results based on the biomass burning events as measured by EARLINET over the 2008-2017 period, according to a methodology described in Part 1 . The aerosol layers were labelled as smoke based on a combined analysis of Hysplit backtrajectories and the FIRMS fire locations. The smoke was further labelled as 'mixed' if multiple fire sources contributed to the smoke measurement. For the smoke originating in North America, the smoke was labelled as 'pure North America' or 'mixed' (with contribution from fires in Europe). We demonstrated that in most of the 25 cases the smoke was mixed and the quantification (based on number of fires and detections) of the contributing fires to the mixture explains the wide range of values obtained for the intensive parameters.
The statistics over all LRT events from North America revealed no significant difference between the measurements where the smoke was originating solely from North America and the measurements with mixed smoke (having origin in both North America and local). This suggests that the contribution of the local smoke is not significant. Based on the LR values, a moderate 30 absorption at 355 nm (46 sr) and a high absorption at 532 nm (66 sr) were observed. The mean CRLR and EAE suggest aged https://doi.org/10.5194/acp-2021-759 Preprint. The statistical analysis of the smoke properties (fresh versus aged and absorption capability) in four European regions (Central, North-East, South-West and South-East Europe) separating the smoke events into continental source regions (European, North American, African, Asian or a mixture of European with each of the remaining), based on trajectory analysis revealed the 5 following. The smoke detected in Central Europe (Cabauw, Leipzig, and Hohenpeißenberg) was mostly brought form North America (87% of fires). In North-East Europe (Belsk, Minsk, Warsaw), the smoke was advected mostly from Eastern Europe (Ukraine and Russia) but there was a significant contribution (31%) of smoke from North America. In South-West Europe (Barcelona, Evora, Granada) smoke originated mainly in Iberian Peninsula and North Africa, (while 9% was originating in North America. In South-East Europe (Athens, Bucharest, Potenza, Sofia, Thessaloniki) the origin of the smoke was mostly 10 local (only 3% of smoke from North America.
For each region, the IPs were analysed based on their continental source origin.
The analysis of the scatter plots revealed correlated with the increase of smoke travel time (corresponding to aging), CRLR and CRBAE increase while EAE and CRPDR decrease. These tendences, associated with the smoke characteristics, can be further used when analysing various types of aerosols and thus helping identifying the smoke among other aerosol types. The 15 variability of the mean values / standard deviation (STD) was large in general and, thus, the individual values for different source regions overlap. Based on data from Warsaw (North-East region), the depolarization at 532 nm decreases for LRT smoke from North America (while CRPDR < 1).
Smoke was found to be aged in all measurement regions (except North-East) if there is no mixture among different fires. On the contrary, when the origin of the smoke has two continental sources, either aged, fresh or a mixture of aged and fresh smoke 20 can be measured, depending on the smaller or higher contribution of the European (local) sources. Thus, in the South-East measurement region, fresh smoke from the EUNA source region and a mixture of fresh and aged smoke originating from the EUAS was measured. In the North-East, region fresh smoke originating from EUAF was measured.
For the South-West region with European or African source regions we obtained a CRLR of 0.8 and an EAE of 1. We assumed that the smoke measured was aged based on the high RH (in agreement with Veselovskii et al., 2020). The lowest absorption 25 was determined for the Central region (LRs < 36 sr). The South-West region displayed a highly absorbing smoke (61 sr < LR@355 < 79 sr and 64 < LR@532 < 91 sr). The South-East region displayed smoke with a medium/high absorption at 532 nm (50-72 sr) and a low/medium absorption at 355 nm (31-48 sr). The smoke measured in the North-East region has a medium to very high absorption at 532 nm (57-91 sr) and a medium to high absorption at 355 nm (46-78 sr).
The quite diverse absorption was determined for the different measurement's regions, even for smoke from the same 30 continental source region, which may be related, among others, with different RH conditions (e.g., Veselovskii et al, 2020).
In line with previous studies, we showed that BAE and further CRBAE do not show specific values based on sources and no trends, and thus, they cannot be used to identify the smoke type. In order to easily quantify the smoke type, LR (CRLR) and extinction input provides one category for smoke. NATALI  distinguishes between smoke, continental smoke and mixed smoke if depolarization data are available additionally. Based on the implementation of ACTRIS Research Infrastructure in the next few years, the presented methodology will be applied on a larger dataset (more automatic lidar systems expected) providing a more complete (3 backscatter + 2 extinction + 1-3 depolarization) datasets with enhanced quality control procedures. 5 The present methodology shows new approaches for smoke characterization (smoke type along with information on absorption and depolarization in the context of different continental sources) and provides valuable information for various scientific communities (modelling, satellite).. The analysis reported in the paper shows the potentialities of the used approach for identifying specific features of smoke particles in different geographical regions and for long range transported cases. The obtained results will be corroborated by the increasing number of aerosol profiling data coming into the EARLINET database 10 thanks to the implementation of ACTRIS (Aerosol Clouds Trace Gases Research Infrastructure). This process is currently reducing the time delay in data provision, improving the quality of data products and increasing also the number of multi wavelength lidar systems over Europe. This extension of the observations will allow in the near future to increase the statistics of the result obtained with the approach here presented.
Competing interests. The authors declare that they have no conflict of interest.
Special issue statement. This article is part of the special issue "EARLINET aerosol profiling: contributions to atmospheric 10 and climate research". It is not associated with any conference.

Acknowledgements:
We Data access: The FIRMS data used in the study is available upon request from https://firms.modaps.eosdis.nasa.gov/ (last 10 access: 13 July 2021). The data files used in this study as well as the output of the data analysis is available upon request (contact mail: mariana.adam@inoe.ro).           Table 3. Mean values and their STD for IPs for each region (SE, SW, CE, NE) and each continental source region (EU, AF, NA, EUAF, EUAS, EUNA). The first column block refers to the scatter plot in Fig. 6d (EAE versus CRLR), the middle column block refers to the scatter plot in Fig. 6e (CRLR versus CRBAE) and the last column block refers to the scatter plot in Fig. 6f