Interactive comment on “ 30-year lidar observations of the stratospheric aerosol layer state over Tomsk ( Western Siberia , Russia ) ”

This paper uses an extensive lidar dataset to characterize multiple eruptions over a northern mid-latitude site. The Russian site is quite far from other lidar sites so is unique is that respect. The analysis is straight forward and described well. I recommend that the paper be published with minor corrections. The content of the paper is well written although the English grammar should be edited.


Introduction
Long-term studies show that the presence of various types of aerosol in the stratosphere is mainly caused by powerful volcanic eruptions (Robock, 2000;Robock and Oppenheimer, 2003). Volcanic eruptions are ranked in the volcanic explosivity index (VEI) category from 0 to 8 (Newhall and Self, 1982;Siebert et al., 2010). During Plinian or, more rarely, Vulcanian explosive eruptions with VEI ≥ 3, volcanic ejecta and gases can directly reach the stratospheric altitudes, where 5 the volcanogenic aerosol stays for a long time. Then this aerosol spreads throughout the global stratosphere in the form of clouds. The volcanogenic aerosol perturbs the radiation-heat balance of the atmosphere, and thus, significantly affects the atmospheric dynamics and climate (Timmreck, 2012;Driscoll et al., 2012;Kremser et al., 2016). The injection of volcanogenic aerosol particles into the stratosphere leads to a considerable increase of their specific surface area and, therefore, to activation of heterogeneous chemical reactions on the surface of these particles. The reactions can result in, e.g., 10 stratospheric ozone depletion (Hofmann and Solomon, 1989;Prather, 1992;Randel et al., 1995). Moreover, the long-term presence of volcanogenic aerosol clouds in the stratosphere also leads to cooling of the underlying surface and near-surface atmosphere due to the aerosol scattering and extinction of the direct solar radiation (Stenchikov et al., 2002). The latter effect is the basis for several geoengineering projects on artificial climate control (Crutzen, 2006;Robock et al., 2009;Kravitz and Tomsk (56.48° N, 85.05° E, Western Siberia, Russia) is located in the central part of the Eurasian continent. The information on the atmosphere over the vast area of Siberia is poorly presented in various databases. Therefore, the lidar measurements time series accumulated in Tomsk are definitely unique and can be useful, e.g., in studying climate change (Mills et al., 2016). A new lidar station was designed and implemented at the IAO in 1985 for continuous monitoring of the SAL volcanogenic perturbations and other stratospheric parameters over Tomsk. The monitoring started at the end of 1985 5 and is ongoing at the present time (i.e. more than 30 years). In 2004, the lidar station in Tomsk was integrated into the Lidar Network for atmospheric monitoring in the Commonwealth of Independent States (CIS-LiNet;Chaykovskii et al., 2005;Zuev et al., 2009). The CIS-LiNet has been established by six lidar teams from Belarus, Russia, and the Kyrgyz Republic.
Note that the CIS-LiNet station located in Minsk, Belarus, is also integrated into the European Aerosol Research Lidar Network (EARLINET; Wandinger et al., 2016). 10 The detection of high aerosol concentration in the stratosphere over Tomsk after the Nevado del Ruiz volcano eruption (Colombia, 13 November 1985; VEI = 3) marked the beginning of routine lidar observations in 1986 (El'nikov et al., 1988).
Definitely, the detection and subsequent monitoring of strong SAL perturbations by volcanogenic aerosol after the Pinatubo eruption were the major events during the first decade of lidar observations in Tomsk. The data of lidar measurements made in Tomsk over the 1986-2000 period were summarized and analyzed by Zuev et al. (1998) and Zuev et al. (2001). retrievals in the troposphere and stratosphere. The receiving telescopes with the main mirror diameters of 2.2, 1, 0.5, and 0.3 m and lasers operating in the wavelength range 271-1064 nm are used at the SLS for these purposes.
The SLS aerosol channel we consider uses a Nd:YAG laser as the channel transmitter and a Newtonian telescope with a mirror diameter of 0.3 m and a focal length of 1 m as the channel receiver. The laser (LS-2132T-LBO model, LOTIS TII   Co., the Republic of Belarus) can operate at wavelengths of 1064, 532, and 355 nm with 200, 100, and 40 mJ pulse energies,  5 respectively, at a pulse repetition rate of 20 Hz. The backscattered signals from altitudes up to the stratopause (∼50 km) are registered with a vertical resolution of 100 m by R7206-01 and R7207-01 PMTs (Hamamatsu Photonics, Japan) at used wavelengths of 532 and 355 nm, respectively. The PMTs operate in the photon counting mode. Two shutdown periods of the SLS aerosol channel from July 1997 to May 1999 and from February to September 2014 were due to the maintenance of the channel laser, and the rearrangement and improvement of the SLS. A more detailed technical description of the SLS aerosol 10 channel and its data acquisition electronics can be found, e.g., in (Burlakov et al., 2010).
We use the scattering ratio R(H) to describe the stratospheric aerosol vertical distribution, i.e.
( ) a H π β are the molecular (Rayleigh) and aerosol (Mie) backscatter coefficients, respectively; π denotes an angle of π radian, i.e. the angle of the backscatter lidar signal propagation. The SLS aerosol channel makes it possible to 15 receive almost undisturbed backscattered signals from altitudes of ~40-45 km. At higher altitudes, the signal-to-noise ratio is too low. Therefore, altitudes of ~30-35 km, where the stratosphere is considered to be aerosol-free, were used as the calibration altitudes H 0 . Thus, the detected lidar signals were calibrated by normalizing them to the molecular backscatter signal from altitudes H 0 ≥ 30 km. The calibration method of lidar signals against the molecular backscatter coefficient ( ) m H π β is described in detail by, e.g., Measures (1984). 20 We use the integrated aerosol backscatter coefficient a B π to describe the temporal dynamics (time series) of stratospheric aerosol loading over Tomsk. The coefficient is calculated for a certain range of stratospheric altitudes (H 1 ; H 2 ) 2 1 ( ) .
Here H 1 is the local tropopause altitude or slightly above, where upper-tropospheric aerosol does not contribute to the value of a B π , and H 2 corresponds to the calibration altitude H 0 = 30 km. Tomsk is located near the southern boundary of subarctic 25 latitudes, where the tropopause altitude can significantly vary, e.g., due to migration of the Arctic stratospheric jet stream within the Tomsk region. Sometimes one can observe a double (or even multiple) tropopause. For this reason, we consciously removed the interval of the tropopause altitude variations to observe the stratospheric perturbations only. As the tropopause altitude over Tomsk varies from ∼11 to 13 km, depending on season, we set H 1 = 15 km.
Various data on volcanic eruptions were taken from the Smithsonian Institution Global Volcanism Program (GVP; http://volcano.si.edu/; Section: Reports; Subsections: Smithsonian/USGS Weekly Volcanic Activity Report and Bulletin of the Global Volcanism Network). To study the SAL volcanogenic perturbations, we also analyze air-mass backward trajectories started from aerosol layers observed over Tomsk. All the trajectories were calculated by using the NOAA's http://ready.arl.noaa.gov/HYSPLIT.php) and the HYSPLIT-compatible NOAA meteorological data from the Global Data Assimilation System (GDAS) one-degree archive.

Time series of the integrated stratospheric backscatter coefficient (1986-2015)
The 30-year time series of the integrated stratospheric backscatter coefficient a B π , obtained from the SAL lidar observations 10 performed at λ = 532 nm in Tomsk from 1986 to 2015, is presented in Fig. 1. The backscatter coefficients are integrated over the 15-30 km stratospheric layer described above. We divided the time series into the following four intervals. The 1986-1991 period (I) reflects the final SAL relaxation after the explosive eruption of El Chichon volcano (Mexico, 29 March 1982, VEI = 5) together with small SAL perturbations after several less powerful volcanic eruptions during the period (see Table 1). The next 1991-1999 is mainly determined by the strong perturbation and subsequent long-term relaxation of the SAL after the Pinatubo eruption. As noted above, the results of aerosol lidar observations at the SLS during the periods I and II were described by Zuev et 5 al. (1998) and Zuev et al. (2001). Next, we consider the temporal dynamics of stratospheric aerosol loading over Tomsk during the periods III and IV. consider the state of the SAL over Tomsk as background during the period III, when the annual average a B π values were less than those in the pre-Pinatubo period (1989)(1990)(1991). Note that taking into account the spectral dependence of a B π , its 20 minimum annual average value observed in Tomsk at λ = 532 nm in 2004 was close to that determined for Garmisch-Partenkirchen at λ = 694 nm in 1979 and considered as the background by Trickl et al. (2013).
Both inter-and intra-annual variations of a B π values in the stratosphere over Tomsk during the periods III and IV are presented in Figs. 2 and 3. Figure 2 shows the inter-annual a B π variations separately averaged over the warm (April to September) and cold (October to March) half-years. The a B π values are mostly higher in the cold half-year than those in the 5 warm one. Furthermore, these "cold" and "warm" average a B π values are modulated by the quasi-biennial oscillations (QBO; http://www.geo.fu-berlin.de/en/met/ag/strat/produkte/qbo/). The behavior of both a B π curves is seen in Fig. 2 to clearly demonstrate the influence of the Brewer-Dobson circulation on the aerosol state of the mid-latitude stratosphere.
Stratospheric aerosol loading is minimal in the warm half-year, when the zonal air mass transport dominates. On the other hand, the meridional air mass transport from tropical into extratropical (middle) latitudes intensifies in the cold half-year 10 and, therefore, it provides the mid-latitude stratosphere with additional aerosol mass from the stratospheric tropical aerosol reservoir (Hitchman et al., 1994). Note that the minimum "warm" average  The influence of the Brewer-Dobson circulation on background aerosol loading in the stratosphere over Tomsk can also be discovered by analyzing the intra-annual variations of the monthly average a B π values. For example, Fig. 3  Peak eruption), and also April and August-October 2011 (after the Merapi, Grimsvötn, and Nabro eruptions) were not taken into account. The exclusion of these perturbed data allowed us to extend the analyzed period of the background aerosol loading variations up to 16 years and, therefore, to improve the statistical reliability of the a B π data series. As seen in Fig (Table 1). Soon after, in 2006, two relatively strong eruptions of the Soufriere Hills and Rabaul tropical volcanoes (Table 1) Table 1). In the next sections we consider contributions of plumes from the volcanoes erupted in the period IV to the SAL volcanogenic perturbations over Tomsk, and also discuss other possible sources of the SAL perturbations.  (Table 1). Solid green circles denote the 10-day average a B π values. The red curve denotes the a B π values smoothed by five-point averaging.

Detection of plumes from northern volcanoes in the stratosphere over Tomsk in 2008-2010 5
Detection of volcanic plumes over Tomsk is based on: 1) the use of the scattering ratio R(H) profiles retrieved from the lidar measurements between 12.5 and 30 km and 2) the assignment of observed aerosol layers to volcanic eruptions via the HYSPLIT model trajectory analysis, when possible.

Okmok and Kasatochi
In summer 2008, two Aleutian volcanoes Okmok and Kasatochi started to erupt at 19:43 UTC on 12 July and between 23:00 10 UTC on 7 August and 05:35 UTC on 8 August, respectively (both VEI = 4). The plumes from these volcanoes considerably   The HYSPLIT trajectory analysis also showed that both Okmok (Fig. 6a) and Kasatochi (Fig. 6b) plumes passed close to the Minsk lidar station. This explains the similarity of the R(H) profiles presented in Fig. 5. Owing to the westerly transport of air masses, the volcanic plumes passed over Minsk three days earlier than over Tomsk. Figure 6c shows the backward trajectories which allowed us to find the connection between two aerosol layers (thick red lines in

20
The volcanoes started to erupt in the Aleutian Islands on 12 July and 7 August 2008, respectively. It should be noted that due to the westerly zonal transport of air masses in the Northern Hemisphere lower stratosphere during summer seasons and vast geographical distance between Tomsk and the Aleutian Islands, both backward trajectories (Figs. 6a and 6b) could hardly be expected to be equal to or shorter than two weeks. Therefore, these trajectories are slightly longer than those usually used in the HYSPLIT model and, thus, can be considered only as probable ones. Nevertheless, we made the trajectory analysis to assign the observed aerosol layers to the corresponding volcanic eruptions. 5

Redoubt and Sarychev Peak
The SAL perturbations over Tomsk in 2009 were caused by the eruptions of two northern volcanoes Redoubt (Alaska, 15 March to 4 April; VEI = 3) and Sarychev Peak (the Kuril Islands, 11-16 June; VEI = 4). The Redoubt plumes caused insignificant SAL perturbations over Tomsk during the first two weeks of May 2009 (Fig. 7). Stronger and longer-lasting SAL perturbations were related to the Sarychev Peak volcano eruption. According to the GVP data, the MPA was within the 10 range of 8-16 km or even reached 21 km (GVP, 2009). The Sarychev Peak plumes were reliably detected in the stratosphere over Tomsk during July and August (Fig. 8), and weakly observed up to November 2009. For a trajectory analysis, we considered an aerosol layer observed over Tomsk at an altitude of ∼13.1 km on 7 July at 02:30 LT (6 July, 19:30 UTC). This layer is seen in Fig. 9 to be: 1) associated with the backward trajectory passed over Sarychev Peak volcano at an altitude of

Eyjafjallajökull
During April-May 2010, there was a series of explosive eruptions of the Icelandic volcano Eyjafjallajökull. These eruptions are noted for the subsequent extensive air travel disruption across large parts of Western Europe. According to the GVP data, the MPA occasionally reached 9 km (GVP, 2010), but did not exceed the local tropopause (GVP, 2010). However, lidar observations, performed in Tomsk on 20 and 26 April 2010, detected the presence of aerosol layers in the troposphere and 5 lower stratosphere at altitudes up to 15 km (Fig. 10). As a comparison, aerosol lidar measurements at Garmisch-Partenkirchen revealed that the upper boundary of the observed aerosol layers from the Eyjafjallajökull volcanic plumes was ∼14.3 km on 20 April, whereas the average altitude of the local tropopause was of ∼10.2 km (Trickl et al., 2013).

Detection of volcanic plumes in the stratosphere over Tomsk in 2011
High values of a B π were detected during the SAL lidar observations in Tomsk from February to April and from August to 5 December 2011. The "first" wave of the SAL perturbations in the winter-spring period was caused by the Merapi volcano eruption (Indonesia, 4-5 November 2010; VEI = 4), whereas the "second" wave was due to the eruptions of the northern volcano Grimsvötn (Iceland, 21 May 2011; VEI = 4) and the tropical volcano Nabro (Eritrea, 13 June 2011; VEI = 4).

Merapi
High values of a B π were detected in the stratosphere over Tomsk from February to April 2011, i.e. 3-5 months after the 10 Merapi volcano eruption. Figure 12 presents the observed after-effect of the Merapi eruption, i.e. several perturbed scattering ratio profiles retrieved from the SLS aerosol lidar measurements between 28 February and 18 April 2011. The Merapi plume (Table 1)

Grimsvötn and Nabro
In 2011, two volcanoes with VEI = 4 Grimsvötn and Nabro started to erupt on 21 May at 19:25 UTC and 13 June after 22:00 UTC, respectively. According to the GVP data, Grimsvötn volcano erupted ash clouds and gases directly into the 5 stratosphere at an altitude of 20 km, whereas the Nabro volcanic plume did not exceed the local tropopause altitude.  (2014) showed that the initial Nabro plume was directly injected 10 into the lower stratosphere at altitudes up to 18 km (Fromm eat al., 2014). The SAL perturbations by volcanogenic aerosol after the eruptions of both volcanoes were observed in the lower stratosphere over Tomsk from August to November 2011 ( Fig. 13). All the scattering ratio profiles shown in Fig. 13, with equal probability, represent superpositions of plumes from both Grimsvötn and Nabro volcanoes.

Polar stratospheric clouds and the after-effect of the 2006 Rabaul eruption
Occasional perturbations of the mid-latitude SAL can also be related to the occurrence of polar stratospheric clouds (PSCs) 5 in winter periods. PSCs are known to form at extremely low temperatures (lower than -78 °C) mainly on sulfuric acid (H 2 SO 4 ) aerosols, acting as condensation nuclei and formed from sulfur dioxide (SO 2 ; Finlayson-Pitts and Pitts, 2000). Therefore, injections of volcanogenic H 2 SO 4 aerosols or/and SO 2 into the stratosphere can lead to PSC formation, if the air temperature < -78 °C. The direct positive correlation between PSC formation and volcanogenic nitric and sulfur acid aerosols loading was shown, e.g., by Rose et al. (2006). However, it should be noted that, in contrast to Rose et al. (2006), 10 Fromm et al. (2003 showed little (or even negative) correlation between PSC events and ambient aerosol loading.
The Northern Hemisphere stratosphere is usually cooled to the required low temperatures inside the Arctic stratospheric polar vortex in cold seasons (Newman, 2010). The Arctic polar vortex sometimes deforms and stretches to mid-latitudes including Siberian regions. Hence, the stratospheric temperature over Tomsk can occasionally be cooled lower than -78 °C, when Tomsk is inside the polar vortex. Thus, the detection of aerosol layers in the stratosphere at extremely low 15 temperatures can be indicative of the presence of PSCs. The first lidar PSC observations over Tomsk were made at λ = 1064 nm in January 1995 (Zuev and Smirnov, 1997).
More precisely, some dense aerosol layers were detected at altitudes in the range of 15 to 19 km on 24 and 26 January. The maximum scattering ratio R(H) was more than 14 at an altitude of 18.1 km. The stratospheric temperature was lower than -80 °C. The cold pool presence and PSC events near the Tomsk longitude during the northern winter of 1994/95 were also 10 reported by Fromm et al. (1999). The formation of these dense PSCs was caused by high concentrations of residual post-Pinatubo aerosols.
Another event of PSCs over Tomsk was observed at λ = 532 nm on 27 January 2007 (Fig. 14). As seen in Fig. 14 Fig. 14). Thus, PSCs were detected at least twice (in 1995 and 2007) during 30 years of stratospheric aerosol lidar measurements in Tomsk. 20

The latest SAL perturbations over Tomsk (2012-2015)
In summer 2011, the annual average a B π value started to decrease and the SAL state over Tomsk started to relax to its background one (Fig. 1). However, a marked increase in a B π value was observed in the winter of 2015. Figure 15 shows several perturbed scattering ratio profiles retrieved from the SLS aerosol lidar measurements between 29 January and 30 March, 2015. During that period of time, the Kelut volcano eruption could probably be a source of the SAL perturbations 5 over Tomsk.
An explosive eruption of the tropical volcano Kelut occurred in East Java, Indonesia, on 13 February 2014 (Table 1)  Thirty years  of lidar monitoring of the SAL state over Tomsk definitely showed that explosive eruptions with VEI ≥ 3 of both tropical and extratropical (northern) volcanoes represent the main cause of the northern mid-latitude SAL perturbations. Moreover, the tropical volcanoes, rather than the northern ones, have a dominant role in volcanogenic aerosol loading of the mid-latitude stratosphere. Indeed, major explosive eruptions of tropical volcanoes are able to enrich the 5 stratospheric tropical reservoir with volcanogenic aerosol. Additional aerosol loading of the tropical reservoir can usually lead to an increase in the annual average a B π value in the Northern Hemisphere mid-latitude stratosphere via the meridional transport in the cold seasons (October to March; Hitchman et al., 1994). For example, plumes from both Merapi and Kelut volcanoes additionally supplied the stratospheric tropical reservoir with volcanic aerosol and gases ( Table 1). As a result, the increased annual average a B π values (i.e. the SAL perturbations) were detected over Tomsk in 2011 and 2015, respectively 10 (see Sects. 3.3.1 and 3.5). On the other hand, by contrast to tropical volcanoes, the narrow volcanic gas, aerosol, and ash plumes from northern volcanoes can either pass over a lidar station or pass it by. Owing to this, a certain part of northern volcanoes eruptions into the stratosphere did not perturb the SAL over Tomsk and, therefore, was not detected there. It is clear that an extensive network of lidar stations in the territory of the Russian Federation is required to obtain objective data on the mid-latitude stratospheric aerosol loading. 15 In cases of the Eyjafjallajökull and probably Okmok and Kasatochi eruptions, the HYSPLIT air-mass backward trajectories, started from the altitudes of aerosol layers detected over Tomsk with the SLS aerosol lidar, passed over these volcanoes at altitudes back. ∼26 km that is 9 km higher than MPA H (GVP, 2014;Sect. 3.5). Based on these facts, we can offer the following explanation of the inconsistencies between the altitudes back. traj.
H and MPA H . During Plinian explosive eruptions, solid and liquid ejecta, ash, and gas-vapor emissions intermix with each other, heat, and ascend inside the "convective thrust region" of an eruption column. Then the heated air together with erupted materials is known to expand, cool, and form the "umbrella region" of the eruption column (Woods, 1988;Scase, 2009). The most heated fraction of gas-vapor emissions from the "convective thrust 25 region" has the highest speed and, therefore, can penetrate through the higher-density "umbrella region" of the eruption column and reach altitudes higher than H MPA due to the cumulative (jet) effect (Raible et al., 2016). The secondary atmospheric H 2 SO 4 aerosols are formed via oxidation of SO 2 contained in volcanic gas-vapor emissions. The currently available visual and radar methods for determining volcanic plume altitudes can detect only the large-sized volcanic ash particles. At the same time, these methods are not sensitive to the small-sized atmospheric H 2 SO 4 aerosols. Nevertheless, the 30 submicron H 2 SO 4 aerosol particles can be easily detected by lidars.
In addition to volcanoes, PSCs also represent a cause of significant SAL perturbations. However, the temperature condition required for PSC formation (air temperature should be < -78 °C) rarely holds in the mid-latitude stratosphere.
Only two PSC events in January 1995 and January 2007 were observed over Tomsk during the 30-year period of lidar observations in Tomsk.
Extensive forest (bush) fires could be another cause of occasional increases of the a B π value. Combustion products (gases 5 and aerosol particles) can reach the stratospheric altitudes via convective ascent within pyro-cumulonimbus (pyroCb) clouds (see, e.g., Fromm et al., 2006). For example, the smoke plumes from the strong bush fire, occurred near the Australian city of Melbourne on 7 February 2009, were observed in the local stratosphere at an altitude of ~18 km (Siddaway and Petelina, 2011). Due to the climate warming, the number and intensity of massive forest fires have considerably increased in the last few years (Wotton et al., 2010). For example, about 137 strong forest fires were registered in the Northwest Territories of 10 Canada in July 2014 (CBC News, 2014). The smoke-filled air masses frequently enter the stratosphere over the South of Western Siberia from North America, where extensive forest fires occur. Their smoke plumes are most likely to be detected as the SAL perturbations over Tomsk. However, more detailed information about the pyroCb events is required for their correct identification. It is quite possible that some after-effects of strong forest fires occurred, e.g., in North America could be detected over Tomsk, but not identified during lidar observations in Tomsk (1986Tomsk ( -2015. from Fuego to Eyjafjallajökull, and beyond, Atmos. Chem. Phys., 13, 5205-5225, doi:10.5194/acp-13-5205-2013Phys., 13, 5205-5225, doi:10.5194/acp-13-5205- , 2013.