Concentrations and size distributions of aerosol ... - Wiley Online Library

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 115, D17202, doi:10.1029/2009JD013481, 2010

Concentrations and size distributions of aerosol particles at Maitri during the passage of cyclonic storms revolving around the continent of Antarctica Vimlesh Pant,1,2 Devendraa Siingh,1 and A. K. Kamra1 Received 31 October 2009; revised 12 February 2010; accepted 23 April 2010; published 3 September 2010.

[1] The number size distributions of aerosol particles in the size‐range of 0.003–20 mm

diameter have been measured at Maitri (70° 45′ 52″S, 11° 44′ 03″E, 117m above mean sea level), Antarctica, when two circumpolar cyclonic storms passed close to the station during February, 2005. As a storm approaches toward Maitri, concentration of coarse particles increases by about an order of magnitude and the number size‐distribution frequently shows a coarse mode at ∼ 2 mm, a broad Aitken mode from 0.04 to 0.1 mm and, occasionally, a nucleation mode at 0.018 mm diameter. When the storm is going away from Maitri, in addition to the coarse mode at ∼ 2 mm and a peak at 0.08 mm diameter, a nucleation mode frequently appears at < 0.01 mm diameter and the mode existing at 0.018 mm diameter shifts to 0.02−0.04 mm diameter. Particles in the range of 0.008–0.03 mm diameter grow at the rate of 0.2−0.6 nm h−1 in the case of Storm I, but no appreciable growth is observed in the case of Storm II. The peak at 0.02–0.04 mm is often so dominant that it envelops the peak at 0.08 mm diameter. Results are interpreted in terms of the mixing of continental and oceanic air masses with the subsidence associated with the storm. The nucleation mode at 0.01 mm diameter has been associated with the new particles formed in the outflow at the top of clouds and the coarse mode at 2 mm diameters with the re‐suspension of particles from the surface. Citation: Pant, V., D. Siingh, and A. K. Kamra (2010), Concentrations and size distributions of aerosol particles at Maitri during the passage of cyclonic storms revolving around the continent of Antarctica, J. Geophys. Res., 115, D17202, doi:10.1029/2009JD013481.

1. Introduction [2] The continent of Antarctica is covered with clouds of Aitken particles which are mostly composed of the photochemically produced sulphuric acid formed by oxidation of the dimethyl sulphide (DMS) emitted by the surrounding oceans [Shaw, 1988; Ito, 1993]. The cyclonic storms revolving around the continent bring the marine air and aerosol particles into the interiors of the continent. These particles are mostly comprised of the coarse sea‐salt particles produced over sea surface and modify the physical and chemical properties of the Antarctic aerosols. Sudden enhancements in aerosol concentrations during such storms have been observed not only on coastal stations but also on the South Pole [Hogan, 1975; Hogan and Barnard, 1978; Ito and Iwai, 1981; Ito, 1993; Deshpande and Kamra, 2004]. For example, Hogan and Barnard [1978] noticed in their South Pole data that the frontal passage aloft during the summer months, and/or strong subsidence of dry air are often accompanied with a large increase of fine par1 Instruments and Observational Techniques, Indian Institute of Tropical Meteorology, Pune, India. 2 Now at Modeling and Ocean Observations Group, Indian National Centre for Ocean Information Services, Hyderabad, India.

Copyright 2010 by the American Geophysical Union. 0148‐0227/10/2009JD013481

ticles. On the other hand, frontal passage at lower levels and advection of warm moist air from the Weddell Sea are accompanied with an increase in coarse particles. From their measurements at a coastal station, Syowa, Ito and Iwai [1981] observed that while acid particles descend to the ground along a cold frontal surface in one storm, maritime aerosols are transported into the inland of the Antarctic continent in other storm. [3] Considering the predominance of coarse particles at coastal sites and of fine particles at continental sites, the chemical composition of Antarctic aerosols has been extensively investigated in recent years in order to study the contribution of different substances to the aerosol particles in various size ranges at both coastal and continental stations [Hillamo et al., 1998; Minikin et al., 1998; Wolff and Cachier, 1998; Kerminen et al., 2000; Teinilä et al., 2000; Koponen et al., 2003; Hara et al., 2004, 2005; Virkkula et al., 2006a; 2006b; Tomasi et al., 2007]. Seasonal variation of sea‐salt particles showed lower concentration in Antarctic coasts and that majority of aerosol particles consist of sulphate particles in summer. However, these studies generally lead to the conclusion that the coastal sites are strongly influenced by the surrounding oceans and thus aerosol size distributions at such sites often show a strong component of coarse sea‐salt particles during the austral summer. On the contrary, continental

D17202

1 of 15

D17202

PANT ET AL.: AEROSOLS DURING CYCLONE AT MAITRI

D17202

Figure 1. (a) Map of Antarctica, showing the location of Maitri and some other stations close to Maitri on the Antarctic continent. (b) Location of instruments (Kamet observatory) at Maitri and the wind‐rose showing the magnitudes and directions of wind speed during the period of January to February 2005.

stations are mostly influenced by the subsidence from free troposphere and generally exhibit a prevalence of non‐sea‐ salt (nss) sulphate and methane sulfonic acid (MSA) particles. Any significant concentrations of sea‐salt particles on the South Pole are episodic and are observed only when the marine air associated with large storms can penetrate up to the station [Shaw, 1988; Bodhaine, 1992]. [4] In order to understand the role of the cyclonic storms in changing the characteristics of coastal aerosols, high resolution measurements of atmospheric aerosols in the size‐ range of 0.003 to 20 mm diameter were made during two periods when cyclonic storms circulating around the continent of Antarctica, passed close to Maitri (70° 45′ 52″S, 11°

44′ 03″E, 117 m above mean sea level) station during the 24th Indian Scientific Expedition to Antarctica from January 1 to February 28, 2005. The results are presented and analyzed here in an attempt to identify the sources and sinks of the aerosols accompanied with such storms.

2. Instrumentation and Measurement Site [5] Measurements of the number size distribution of atmospheric aerosol particles were made in the size range of 0.003 – 0.7 mm diameter with a Scanning Mobility Particle Sizer (SMPS, TSI Model 3936) and in the size range of 0.5 – 20 mm diameter with an Aerodynamic Particle Sizer

2 of 15

D17202


Figure 2. The average number size distribution of aerosol particles for the month of February, 2005 (excluding the cyclonic storm period) when the wind directions are within the cone of ±20° from the direction of Novolazarevskaya (curve with squares) station or wind is from other directions (curve with circles).

(APS, TSI Model 3321). A Long Differential Mobility Analyzer (LDMA) or a Nano‐Differential Mobility Analyzer (NDMA) was used in the SMPS system depending upon the particle size‐range to be measured. Long DMA was used for measurement of particles in the size‐range of 0.016 to 0.7 mm for 574 h in January and in the size range of 0.01 to 0.4 mm for 350 h in February, 2005. Nano‐DMA was used only for 210 h during the whole campaign period for measurement of particles in the size range of 0.003 to 0.16 mm. A butanol‐based Ultrafine Condensation Particle Counter (UCPC, TSI Model 3025A) was used along with the SMPS. The APS and SMPS systems have resolutions of up to 64 and 32 channels per decade, respectively, and a measurement cycle of 10 min for the whole size range. Both instruments were kept inside the Kamet observatory, a (2 m × 1.5 m × 2 m) hut, located approximately 300 m upwind of the living modules and generator complex of the Maitri station. Air samples for the APS and SMPS systems were drawn at the rates of 5 lpm and 3.3/6.6 lpm, respectively, through two conductive silicon tubes, each of 0.5 cm internal diameter. Since the instruments were placed inside but close to the wall of Kamet hut, lengths of only 1.0 m and 0.5 m were used for the APS and SMPS systems, respectively, to bring the air samples from outside to inside of the hut. Inlets of the tubes were cleaned daily with butanol and allowed to dry. Loss of particles by diffusion, as calculated from Fuchs [1964] formula, is less than 15% for 0.003 mm particles and less than 2% for 20 mm particles. Loss of coarse particles is, however, mainly by gravitational settling and impaction in the tube. Assuming the density of wet sea‐salt particles to be 1200 kg m−3, the loss of coarse particles in our experimental setup is calculated, following the approach of Fuchs [1964], to be 7% for particles of 10 mm diameter but as large as 30% for particles of 20 mm diameter. It may be noted, however that particles > 10 mm were generally not observed at Maitri. To keep the impac-

D17202

tion losses to the minimum in our experimental setup, the tubes were kept almost straight with no sharp turns. However, loss of some large particles is not ruled out. [6] Measurements of the air temperature, pressure and wind speed and direction were made at 5 m height above the ground level and the cloud amount was observed visually every hour. The NOAA HYSPLIT 4 model has been used to get the 5‐day backward trajectories of air mass [Draxler and Hess, 1998] (also R. R. Draxler and G. D. Rolph, HYSPLIT (Hybrid Single‐Particle Lagrangian Integrated Trajectory) Model, 2003, http://www.arl.noaa.gov/ready/hysplit4.html 2) (hereinafter Draxler and Rolph online model, 2003). The National Center for Environmental Predictions (NCEP) reanalysis data has been used to get the surface wind patterns and vertical wind analysis over the region. [7] All measurements were made at Maitri which is located in Schirmacher oasis of Dronning Maud Land in East Antarctica and is 90 km away from coastline in summer. Positions of Maitri and some other stations close to Maitri on the Antarctic continent are shown in Figure 1. The nearest station to Maitri is ∼ 6 km away in the southeast direction of Maitri. Also shown in Figure 1, is the location of our observatory in Maitri station. The ground surface under the instruments is covered by sandy and loamy sand type of soil so that dust particles are generally not airborne in winds < 15 m s−1. There are steep cliffs, frozen lakes and an ice‐shelf extending to about 90 km in the summer on the northern side of Maitri station. On the other hand, polar‐ice and ice‐rock interface on it’s southern side. However, because the cliffs are far away and not very high, oceanic air is not obstructed from reaching the station. Detailed description of Maitri station is given by Deshpande and Kamra [2004] and Siingh et al. [2007]. [8] Figure 1 also shows a wind‐rose for the entire period of our stay at Maitri. Any chances of pollutants released from the Novo air strip to the observatory are remote with the predominant southeasterly winds at Maitri. Since Novolazarevskaya station is located in the southeastern sector at 100° southeast of Maitri, the possible contamination of our data with the anthropogenic emissions from this station need be examined in view of the reports of Warren and Clarke [1990] and Hagler et al. [2008]. To examine such a possibility, we have computed two average size distribution curves from the entire data obtained in the month of February 2005 (excluding the periods affected by the cyclonic storms), one for the periods when the wind directions are within a cone of ±20° from the direction of Novolazarevskaya station, and other is for the periods when wind is from the other directions. The two size distributions plotted in Figure 2, do not show any significant difference, indicating that the strength of any pollutants released from the Novolazarevskaya station, are diluted by dispersion in the atmosphere to the negligible concentration levels during their transport to Maitri.

3. Observations 3.1. Cyclonic Storm I 3.1.1. Meteorological Aspects [9] During the months of January and February, the atmospheric temperature and pressure at Maitri typically vary from −8°C to +8°C and from 942 hPa to 982 hPa,

3 of 15

D17202


D17202

Figure 3. Variations in meteorological parameters and total concentration of particles from 0.5 to 20 mm diameter and 0.01–0.4/0.003–0.16 mm diameter, recorded at Maitri from Day 31 to 60, 2005 during the passage of cyclonic Storm I and Storm II. (TC is the total particle concentrations.) respectively. Occasionally, atmospheric pressure drops down to even lower values under the influence of cyclonic storms circulating around the continent of Antarctica. Cloudiness quickly changes and clear skies quite often change to overcast with low clouds in a short time. Surface winds are dominantly southeasterly that bring continental air to Maitri. [10] Figure 3 shows variations of meteorological parameters recorded at Maitri during February, 2005. An eastward propagating cyclonic storm passed close to the coast of Maitri from Day 31 to 36, 2005 and caused a pressure drop of ∼ 20 hPa on Day 33 and strong southeasterly winds of 6–16 ms−1 at Maitri. Clear skies changed to almost overcast from the morning of Day 33 onwards. Figure 4 shows the surface wind‐flow patterns and positions of cyclonic storm at 0000 UT of each day from the NCEP reanalysis data. Center of the storm was closest to Maitri on Day 33. Maitri receives back‐flow of the air from cyclonic storms once they pass from West to East of Maitri. The air mass over Maitri was strongly influenced by the airflow coming from lower

latitudes before Day 33 and by the outflow of the storm after Day 33. Figure 5 shows the 5‐day backward trajectories for the air mass arriving at Maitri at 500 m level at 0000 UT and 1200 UT obtained from the NOAA HYSPLIT transport and dispersion model and READY website (http://www.arl. noaa.govt.ready.html) using the model vertical velocity method. In all cases, the air mass over Maitri approaches from the southeast direction and travels along the coastline for 1–3 days just prior its arrival at Maitri. Moreover, in all cases, the air mass arriving over Maitri undergoes a vertical displacement of 1–3 km. 3.1.2. Total Particle Concentrations and the Number Size Distributions of Aerosol Particles [11] Figure 6 shows contour plots of the size distribution of particles measured by the APS and SMPS systems from Day 31 to 36 i.e., from the day when atmospheric surface pressure at Maitri began to decrease, attained its minimum value on Day 33, and then restored back to almost its original value. Coarse particles in the size range of 0.5 to

4 of 15

D17202


D17202

Figure 4. The surface wind‐flow patterns around Maitri from NCEP reanalysis data at 0000 UT each day from Day 31 to 36, 2005 during the passage of Cyclonic Storm I. Position of the Maitri station is marked with a dark square. 20 mm diameter were measured throughout this period. However, fine/ultrafine particles were measured in the size ranges of 0.013 to 0.7 mm diameter from 00 UT to 1000 UT on Day 31, 0.01 to 0.4 mm from 0100 UT to 2400 UT on Day 36, and of 0.003 to 0.16 mm from 1600 UT on Day 32 to 1900 UT on Day 35. The data ‐ gaps in time series of the SMPS measurements are mostly because of no data being collected during those periods either due to (i) cleaning and drying of inlet tubes, or (ii) changing the DMA in the SMPS system for changing the size‐range of particles to be measured. In spite of large data‐gaps, a continuous time series of 0.003 to 0.16 mm diameter particles was obtained for a more than 3‐day period simultaneous to a continuous series of coarse particles of 0.5 to 20 mm diameter for the entire period of 6 days. [12] The APS measured total particle concentration of 0.5 to 20 mm diameter particles varies from 0.1 to 0.8 cm−3 during this period and its average value generally increases as the storm moved from the West to East of Maitri. During this period, the tendency of atmospheric surface pressure

at Maitri changed from decreasing to increasing. Periods of enhanced particle concentration lasted from a few hours to a day when the cyclonic storm was either approaching toward or departing away from the Maitri station. Particle concentrations did not show any correlation with wind speed at Maitri during this period. The particle size distributions in this size range normally show a peak between 0.7 and 1 mm diameter. Occurrence of this peak, however, could not be confirmed as the APS has the weakness of underestimating the concentration of particles of 0.5 to 0.7 mm diameter. Our simultaneous measurements with the SMPS in January, 2005 did not show any peak in this overlapping size range. Occasionally, another peak develops at ∼ 2 mm diameter when wind speed exceeds ∼ 10 m s−1 and atmospheric temperature is higher than 0°C. This peak in a size distribution is mostly lower than the peak at 0.7–1 mm diameter and it exists only for a period of a few hours. However, during the period of its appearance, the coarse particle concentrations of diameter > 2 mm increase and show large variability with time.

5 of 15

D17202


D17202

Figure 5. The 5‐day backward‐trajectories of the air mass arriving at Maitri at 500 m altitude at 0000 UT and 1200 UT each day from Day 31 to 36, 2005. Back trajectories are drawn using NOAA HYSPLIT Model (Draxler and Rolph online model, 2003). Thin continuous lines show the geographical boundary of the continent of Antarctica, dotted lines and circles show the latitudes and longitudes, and various symbols show the trajectories on different days. [13] The 3‐day high‐resolution time series of size distribution of particles in the size range of 0.003 to 0.16 mm diameter with a 10‐min time‐resolution is a unique one to study the evolution of new particles during a storm. In the following, size distribution and its temporal changes are discussed in modes. Figure 7 shows the hourly ‐ averaged size distributions of dN/dlogD of 0.003–0.16 diameter particles for Day 33, as an example, for better illustration of the results. Particles have been categorized in nucleation mode (d ≤ 0.02 mm), Aitken mode (0.02 < d < 0.1 mm), accumulation mode (0.1 < d < 1.0 mm) and coarse mode (d ≥ 1 mm). The size distributions plotted in Figure 7, clearly demonstrate that nucleation mode is dynamic with regard of both its number concentration and mean diameter. The same holds for Aitken mode. Particles in both modes show a continuous growth to larger sizes. Nucleation mode at 0.018 mm (NM18) appears prominently for most of the period of three days. However, nucleation mode at ∼ 0.01 mm (NM10) starts appearing only between 0800 – 1000 UT and lasts till afternoon hours. However, whenever NM10 appears, NM18 slightly shifts to a larger size (0.02–0.04 mm). A depression with much lower particle concentrations develops between the peaks for NM10 and NM18, and it continues to appear as long as NM10 appears. This periodicity in the nucleation mode appearance

strongly suggests the effect of solar radiation in accelerating the new particle formation and growth processes. This aspect together with the observation that the particles < 0.005 mm diameter were rarely observed in our data, will be further discussed in Section 4. The nucleation mode particles in NM10 show large variability even on a time‐scale of an hour or less. On the other hand, the peak at 0.08 mm diameter always appears. Although, mean value of the mode diameter for this peak lies in the Aitken mode size range, the upper limits of individual size distributions in this size range frequently extend to the accumulation mode size range. So, this peak is likely to correspond to the accumulation mode. Somewhat lower value of this mode diameter may be because of the lower concentration and growth rate of smaller particles resulting from the lower values of the concentration of condensing gases and cloud coverage available on the Antarctic continent. [14] The depression developed between NM10 the NM18 starts filling‐up from 1200 UT onwards and particles in this size‐range constantly grow with time until 1800 UT when the depression is completely filled‐up and the two peaks merge together into one peak at 0.018 mm (NM18). When the two peaks merge together, the magnitude of NM18 is initially larger than those of two peaks before their merger,

6 of 15

D17202


D17202

Figure 6. Time variations of the size distributions of (a) coarse and (b) ultrafine/fine particles during the passage of Storm I from West to East of Maitri. Data gaps in time series for ultrafine/fine particles are mainly due to making arrangements for changing the size‐range of particles to be measured. Also shown are four events of the formation and growth of nucleation mode particles. but then it keeps decreasing with time at a slow rate. Rate of growth of the 0.008 – 0.03 mm diameter particles, calculated with the procedure of Kulmala et al. [2004], using our high resolution measurements, lies between 0.2 and 0.6 nm h−1. These growth rates are roughly of the same order as observed in earlier investigations at the coastal stations in Antarctica [Gras, 1993; Koponen et al., 2003; Virkkula et al., 2007] but lower than those observed at continental stations in Antarctica [Kulmala et al., 2004]. Similar trends in the appearance and growth of the particles, as shown in Figure 7, are also observed on Day 34 and 35 with some difference in timings. [15] Events of the new particle formation with a nucleation mode at 0.01 mm diameter frequently appear during the periods when the Sun is high. The mode at 0.02– 0.04 mm diameter becomes very prominent in the later period of our observations, particularly on Day 35, when the nucleation mode particle concentrations are comparatively higher. However, the peak at ∼ 0.08 mm which continues to appear before 1000 UT on Day 34 as well, either disappears or is, most likely, enveloped by a comparatively much higher peak at 0.02–0.04 mm diameter. During this period, the cyclonic storm was departing away from Maitri

and the region is under the subsidence associated with the storm. Subsidence at Maitri is also confirmed from the NCEP vertical velocity analysis at 0000 UT with a latitude – longitude resolution of 2.5 × 2.5 degrees (Figure 8). While the downward motion extends to 600 hPa level on Day 34, it extends only up to 850 hPa level on Day 32 (Figure 8) when the storm is approaching toward Maitri. [16] Our observations at Maitri have many similarities with those made at the Finish station, Aboa, 130 km away from the coastline [Koponen et al., 2003]. For example, the higher particle concentrations observed at Aboa, have also been attributed to the presence of nucleation mode particles and the measured size distributions show an accumulation mode at 0.07–0.15 mm diameter and an Aitken mode at 0.03– 0.05 mm. Furthermore, observations at Aboa also exhibit a nucleation mode in more than half of the measured size distributions and, occasionally, these size distributions also exhibit two nucleation modes. Earlier investigations in coastal Antarctica, did not extend to the particles of < 0.01 mm diameter. However, the size distributions in these investigations demonstrate the presence of a broad mode at 0.1–0.2 mm diameter and a nucleation mode below 0.02 mm diameter, during the summer [Jaenicke et al., 1992; Gras, 1993; Ito,

7 of 15

D17202


D17202

Figure 7. The hourly‐averaged size distribution (dN/dlogD) of 0.003–0.16 mm diameter particles observed at Maitri for Day 33. Note the dynamic nature of the nucleation mode particles of < 0.01 mm diameter with regard to both time and particle size. Also to be noted is the growth of particles in each subsequent hour and the development and filling‐up of depression from 0800–1500 UT. 1993]. These measurements, though broadly consistent with our measurements, did not have sufficient temporal and size resolutions to identify all modes in modal structure of spectra because of comparatively less resolutions in both time and particle size. 3.2. Cyclonic Storm II [17] Cyclonic Storm II caused a pressure drop of 30 hPa at Maitri on Day 56. Figures 9, 10 and 11 show the surface wind patterns around Maitri, the 5‐day backward trajectories terminating at 500 m altitude at 00 and 1200 UT at Maitri and the contour plots of the size distributions of coarse (0.5 to 20 mm diameter) and ultrafine/fine (0.01 to 0.4 mm diameter) particles, respectively, for Days 53 – 58, 2005. The data‐ gaps in Figure 11 are due to the adverse weather conditions when measurements could not be continued. Unlike in case of Storm I, the air masses mostly originated over sea or near to

the coastline and most of these descended from an altitude of 1000 – 3000 m before reaching Maitri (Figure 10). Unfortunately, there are only 5 h of data of coarse particles and no data of ultrafine/fine particles, available on Day 56, the day of the minimum surface pressure at Maitri. Available data, however, indicate an increase in total number concentration of coarse particles as the storm approaches Maitri. Although, the coarse particle concentration increases and shows large variability, as in the case of Storm I, the coarse mode at 2 mm diameter appeared only when Storm II is departing away from Maitri. It is significant, however, that although the atmospheric temperatures remained less than 0°C for most of the time during the month of February, these increased to above 0°C on Days 56 and 57 when the coarse mode at 2 mm diameter appeared in case of Storm II. Similar to the case of Storm I, the peaks in the coarse particle concentration appear both when the surface pressure has either increasing

8 of 15

D17202


Figure 8. Vertical velocities around Maitri drawn from the NCEP reanalysis data for Days 32 and 34 at different vertical levels. Continuous contours show the upward velocities and dotted line contours the downward velocities. Position of Maitri is marked with a dark square.

9 of 15

D17202

D17202


D17202

Figure 9. The surface wind‐flow patterns around Maitri from NCEP reanalysis data at 0000 UT each day from Day 53 to 58, 2005, during the passage of cyclonic Storm II. Position of Maitri is marked with a dark square. or decreasing tendency. However, peaks in nucleation mode at < 0.015 mm diameter appear and strengthen the peak at 0.02–0.04 mm diameter only when the surface pressure has an increasing tendency. Moreover, the nucleation mode becomes relatively stronger and the peak at 0.02–0.04 mm almost envelops the peak at 0.08 mm when the storm is departing away from Maitri. However, the observations made during Storm II do not show any persistent trend of particle growth for more than one or two hour periods. Furthermore, atmospheric temperature increases from sub‐ freezing to above 0°C. The coarse particle concentrations are almost twice as large in this case as in the case of Storm I. Also, the coarse particle sizes often extend to ∼ 10 mm diameter, especially during the period when the surface pressure at Maitri has an increasing tendency. The vertical velocity analysis of NCEP also shows that subsidence motion extends to 500 hPa level or even higher during the later period when the storm is going away from Maitri (Figure 12).

3.3. Mean Number Size Distribution During the Storm I and Storm II Periods [18] Figure 13 shows the mean number size distributions of combined data of the SMPS and APS for the periods of Storm I and Storm II. For comparison, it also shows the mean number size distribution of the combined data averaged for the period of whole month of February, 2005 except for the periods of Storm I and Storm II. The aerodynamic sizes of particles measured with APS have been transformed to the mobility size. [19] Figure 13 shows that the concentrations of coarse particles of all sizes are higher during the both storm periods than that of the February concentrations. Concentrations of the nucleation and Aitken mode particles of < 0.04 mm diameter also are higher than the February concentrations in case of Storm I. In case of Storm II, however, concentrations of Aitken mode particles of 0.02–0.04 mm diameter are a

10 of 15

D17202


D17202

Figure 10. The 5‐day backward‐trajectories of the air mass arriving at Maitri at 500 m altitude at (left) 0000 UT and (right) 1200 UT each day from Day 53 to 58, 2005. Backward trajectories are drawn using NOAA HYSPLIT Model (Draxler and Rolph online model, 2003). Thin continuous lines show the geographical boundary of the continent of Antarctica, dotted lines and circles show the latitudes and longitudes, and various symbols show the trajectories on different days. little lower but of nucleation mode particles of < 0.02 mm diameter are approximately equal to the February concentrations. Significantly, concentrations of the Aitken and accumulation mode particles in the size range of 0.04–0.4 mm diameter do not exhibit much change during the storm periods and approximately equal to the February concentrations.

4. Discussion [20] The aerosol content in the air mass over Maitri is mostly determined by two types of airflow. First the sloped topography of the Antarctic continent and the widespread sloped surface inversion together produce an exceedingly persistent surface airflow which transports the aerosol particles along with katabatic winds blowing from the icy continent of Antarctica to the coastal zones [Parish, 1988; King and Turner, 1997]. The secondary sulphate particles formed by the gas‐to‐particle conversion processes and grown for several days through the condensation of the low vapor pressure gases over the continent are transported by these southerly and southeasterly winds from the Antarctic continent. Second, the oceanic airflow transports the primary aerosol particles being generated by the bubble breaking processes during the high wind speed periods over ocean

[Blanchard and Woodcock, 1957; Blanchard, 1963] and the particles generated and transported from midlatitudes by the long‐range transport processes. In addition, the aerosol precursors associated with oceanic bioactivity during the austral summer, [Cavalli et al., 2004; O’Dowd et al., 2004] can also significantly contribute to the new particle formation process at Maitri. Relative concentrations of particles from the Antarctic continent and ocean are determined by the extent of mixing of the two air masses. [21] Our aerosol measurements at Maitri in January– February, 2005, typically exhibited the number size distributions with one peak at 0.78 mm diameter and another at 0.08 mm diameter. As discussed in Section 3.1.2, while the peak at 0.78 mm diameter is doubted as an artifact of the APS, the peak at 0.08 mm diameter is likely to be for the accumulation mode. Frequently, however, an Aitken mode at 0.02 – 0.04 mm diameter or a nucleation mode at ∼ 0.018 mm diameter also appeared. Low concentrations of coarse mode particles at Maitri show that the contribution of mechanically ‐ generated particles or their transport from others places, is small. Aitken mode is mostly contributed by the secondary aerosols generated photochemically on the continent of Antarctica during the polar day [Harvey et al., 1991]. On the other hand, the accumulation mode mainly consists of sul-

11 of 15

D17202


D17202

Figure 11. Time variations of the size distributions of (a) coarse and (b) ultrafine/fine particles during the passage of Storm II from West to East of Maitri. Data gaps in time series are due to adverse weather conditions when the measurements could not be continued. phate particles with some contribution of sea‐salt minerals during intrusions of marine air into the continent. Comparatively large size of the particles in nucleation mode suggests that such particles are not generated close to the station. [22] The aerosol characteristics in our observations are significantly modified during the passage of the cyclonic storms circulating around the continent. Mixing of the continental air mass with two different types of air masses before and after the passage of storm is proposed to result in two different types of aerosol size distribution during these periods. Before the arrival of storm close to Maitri, the continental air mass mixes with the warm moist air advected from the lower latitudes (Figures 4 and 9). During this period, the coarse particle concentration slowly increases with time and exhibits large variability. Aerosol size distributions frequently show a coarse mode at ∼ 2 mm diameter when wind speeds are > 10 ms−1 and atmospheric temperatures are above 0°C. While the enhanced and variable concentration of coarse particles can be associated with the salt particles being advected with the marine air, the coarse mode at 2 mm diameter may appear due to re‐suspension of particles or snow fragments from surface. High wind speeds can strip‐off some water droplets or snow fragments from the water‐ice surface and resuspend the aerosol particles earlier deposited on the surface. How-

ever, the Aitken mode is broadened and extends from 0.04 to 0.1 mm during this period. [23] Just after the passage of storm past Maitri, the air descending from the upper atmosphere to the boundary layer with the subsidence associated with the storm, also mixes with the air mass at Maitri. During this period, the coarse mode at 2 mm diameter rarely appears in case of Storm I but continues to frequently appear in case of Storm II. This may be associated with the higher wind speeds and the above‐ freezing temperatures that prevail during this period in case of Storm‐II. The mode at 0.08 mm diameter, although regularly appears but is often enveloped by a dominant Aitken mode at 0.02–0.04 mm whenever a nucleation mode appears at ≥ 0.01 mm diameter. Higher concentration of the Aitken mode particles observed in the departing stage of the storm when the subsidence associated with the storm prevails in the region, favors the downward transport of the nucleation and/or Aitken mode particles during this period. The periodicity in periods of the enhanced formation of the nucleation mode particles, as discussed in Section 3.1.2., may be associated with enhancement in solar radiation and UV radiation in presence of the storm ‐ associated subsidence just after the passage of storm past Maitri. The probability of new particle formation and their subsequent growth depend on the characteristics of the air mass transported to

12 of 15

D17202


D17202

Figure 12. Vertical velocities around Maitri drawn from the NCEP reanalysis data for days 54 and 56 at different vertical levels. Continuous contours show the upward velocities and dotted line contours the downward velocities. Position of Maitri is marked with a dark square. the observation site. It is possible, therefore, that the vertical mixing introduces active precursor gases from the layers of air at different heights of the boundary layer. Alternatively, the particles already formed aloft during the morning hours,

are transported to the ground level due to vertical mixing of the atmospheric boundary layer. This hypothesis is also supported by the fact that the particles smaller than 0.005 mm in diameter were rarely observed in our measurements,

13 of 15

D17202


Figure 13. The mean number size distributions of combined data of the SMPS and the APS for the periods of Storm I, Storm II, and the month of February, 2005, except for the periods of Storm I and II. The different particle size ranges measured during different period are discussed in text. suggesting that such particles are transported from some other place. With the observed rates of particle growth in our observations, the new particles formed in the outflow of the cloud ‐ tops can be transported down to the ground and grow to a size of 0.005 mm diameter in a period of a few hours if an average downward velocity of ∼ 0.1 ms−1 is assumed. Higher rate of growth of the particles brought down in the subsidence of the storm may be due to some excess amount of the condensing trace gases also brought down from the upper levels along with the newly formed particles. Furthermore, our observations of the bimodality observed between Aitken and accumulation modes suggest that growth of particles by cloud processes in non‐precipitating clouds [Hoppel et al., 1994] play a prominent role in shaping the size distribution of aerosols in this region. To explain their observations made during the passage of cyclonic storms close to Syowa, Ito and Iwai [1981] also suggest such transportation of fine particles from the free troposphere. It must be added, however, that mixing of the free tropospheric air into mixed layer of the atmospheric boundary layer is favorable but not sufficient condition for the new particle formation at Maitri. Conditions of higher concentration of aerosol precursors (e.g., H2SO4 and DMS) and low concentration of pre‐existing particles are essentially required for the new particle formation. On the other hand, the observation that coarse particles are observed in both approaching and departing stages of the storm, suggests that these particles are generated during the high wind speed periods and then these are transported to Maitri. [24] Differences in the mean number size distributions observed during the periods of Storms I and II, and plotted in Figure 13, can be understood in terms of the formation and growth of new particles. During the Storm I period, the nucleation mode particles form and subsequently grow to

D17202

the size of Aitken mode particles (Figure 6). On the other hand, during the Storm II period, the nucleation mode particles may form but do not grow to the size of Aitken mode particles (Figure 11). Lack of particle growth during the Storm II period may be due to the absence of sufficient concentrations of condensing gases. Higher concentrations of coarse particles observed during the both Storm I and II periods as compared to the February concentrations, can be associated with the stronger winds observed during the storm periods. Absence of any significant change in the mean concentrations of 0.04–0.4 diameter particles during different periods in Figure 11 may be related to the almost constant concentration of background aerosol particles often reported to exist in the Antarctic environment. Our high‐ resolution measurements, thus, reinforce the conclusion of robustness of the Antarctic aerosol system. The cyclonic storms circulating around the continent can add the short‐ lived nucleation and/or coarse mode particles to the Antarctic environment. Some of these nucleation mode particles, of course, can grow to the size of Aitken particles and contribute to the population of Aitken and accumulation modes of particles. [25] The frequently observed high concentration of nucleation mode particles, observed by Bates et al. [2000], over the southern Indian Ocean south of Australia during the Aerosol Characterization Experiment (ACE I), are also associated with the passage of cold fronts and high convective activity in the region. In their observations as well, the Aitken mode becomes prominent with the peak concentrations reaching in the range of 103 to 104 cm−3, and the particles of less than 0.005 mm diameter are not observed. Recent measurements of Koponen et al. [2003] at Aboa, Antarctica also show evidence of the formation of new particles and their growth up to a diameter of 0.04 mm in marine/coastal air masses. Our observations support their conclusion that such particles can significantly contribute to the overall particle budget of the Antarctic boundary layer. Furthermore, these new particles can be a source of Aitken and accumulation mode particles which can affect the Antarctic climate.

5. Conclusions [26] Our measurements show that concentration and number size distribution of aerosol particles at Maitri are significantly modified during the passage of circumpolar cyclonic storms revolving around the continent of Antarctica. Total concentration of coarse particles increases by about an order of magnitude as the storm approaches toward Maitri. Under conditions of strong winds > 10 ms−1 and atmospheric temperatures > 0°C, the number size distributions of coarse particles show a coarse mode at 2 mm diameter, which can be associated with the re‐suspension of particles. Formation of new particles occurs in the region of subsidence when the storm is going away from the station and subsequently they grow at the rate of 0.2 to 0.6 nm h−1. This nucleation mode which sometimes appears at < 0.01 mm diameter, is dynamic in both its number concentration and mean diameter, and it shifts the mode at 0.018 mm to 0.02–0.04 mm. The mode at 0.02–0.04 mm is sometimes so dominant that it envelops the accumulation mode at 0.08 mm. Our results suggest that the nucleation mode particles formed in the upper regions of

14 of 15

D17202


cyclones can grow to the Aitken mode particles in the Antarctic boundary layer and can affect the climate of Antarctica. [ 27 ] Acknowledgments. Authors gratefully acknowledge the National Centre for Antarctic and Ocean Research (NCAOR) Goa, India, for participation in the 24th Indian Scientific Expedition to Antarctica (ISEA) and the India Meteorological Department for providing the Metrological data. The authors gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model and READY Web site (http://www.arl.noaa.gov.ready.html) and NCEP reanalysis data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their Web site at (http://www.cdc.noaa.gov) used in this publication. One of us (A.K.K.) acknowledges the support under the INSA Senior Scientist Programme.

References Bates, T. S., P. K. Quinn, D. S. Covert, D. J. Coffman, J. E. Johnson, and A. Wiedensohler (2000), Aerosol physical properties and process in the lower marine boundary layer: A comparison of shipboard submicron data from ACE1 and ACE2, Tellus, Ser. B, 52, 258–272, doi:10.1034/j.16000889.2000.00021.x. Blanchard, D. C. (1963), Electrification of the atmosphere by particles from bubbles in the sea, in Oceanography, vol. 1, edited by M. Sears, pp. 73–197, Pergamon, New York. Blanchard, D. C., and A. H. Woodcock (1957), Bubble formation and modification in the sea and its micrometeorological significance, Tellus, 9, 145–158, doi:10.1111/j.2153-3490.1957.tb01867.x. Bodhaine, B. A. (1992), The U. S. aerosol monitoring program in Antarctica, in SIF Conference Proceedings, vol. 35, edited by M. Colacino, G. Giovanelli, and L. Stefanutti pp. 15–25, Ital. Phys. Soc., Bologna, Italy. Cavalli, F., et al. (2004), Advances in characterization of size‐resolved organic matter in marine aerosol over the North Atlantic, J. Geophys. Res., 109, D24215, doi:10.1029/2004JD005137. Deshpande, C. G., and A. K. Kamra (2004), Physical properties of aerosols at Maitri, Antarctica, Proc. Ind. Acad. Sci., 113(1), 1–25. Draxler, R. R., and G. D. Hess (1998), An overview of the HYSPLIT4 modeling system for trajectories, dispersion and deposition, Aust. Meteorol. Mag., 47, 295–308. Fuchs, N. A. (1964), The Mechanics of Aerosols, 408 pp., Pergamon, New York. Gras, J. L. (1993), Condensation nucleus size distribution at Mawson, Antarctica: Microphysics and chemistry, Atmos. Environ., 27, 1417–1425. Hagler, G. S. W., M. H. Bergin, E. A. Smith, M. Town, and J. E. Dibb (2008), Local anthropogenic impact on particulate elemental carbon concentrations at Summit, Greenland, Atmos. Chem. Phys., 8, 2485–2491, doi:10.5194/acp-8-2485-2008. Hara, K., K. Osada, M. Kido, M. Hayashi, K. Matsunaga, Y. Iwasaka, T. Yamanouchi, G. Hashida, and T. Fukatsu (2004), Chemistry of sea‐salt particles and inorganic halogen species in Antarctic regions: Compositional differences between coastal and inland stations, J. Geophys. Res., 109, D20208, doi:10.1029/2004JD004713. Hara, K., K. Osada, M. Kido, K. Matsunaga, Y. Iwasaka, G. Hashida, and T. Yamanouchi (2005), Variations of constituents of individual sea‐salt particles at Syowa station, Antarctica, Tellus, Ser. B, 57, 230–246, doi:10.1111/j.1600-0889.2005.00142.x. Harvey, M. J., G. W. Fisher, I. S. Lechner, P. Issac, N. E. Flower, and A. L. Dick (1991), Summertime aerosol measurement in the Ross sea region of Antarctica, Atmos. Environ., 25, 569–580. Hillamo, R., I. Allegrini, R. Sparapani, and V.‐M. Kerminen (1998), Mass size distributions and precursors gas concentrations of major inorganic ions in Antarctic aerosol, Int. J. Environ. Anal. Chem., 71, 353–372, doi:10.1080/03067319808032638. Hogan, A. W. (1975), Antarctic aerosols, J. Appl. Meteorol., 14, 550–559, doi:10.1175/1520-0450(1975)0142.0.CO;2. Hogan, A. W., and S. C. Barnard (1978), Seasonal and frontal variation in Antarctic aerosol concentration, J. Appl. Meteorol., 17, 1458‐1465, doi:10.1175/1520-0450(1978)0172.0.CO;2.

D17202

Hoppel, W. A., G. M. Frick, J. W. Fitzgerald, and R. E. Larson (1994), Marine boundary layer measurements of new particle formation and the effect which non‐precipitating clouds have on the aerosol size distribution, J. Geophys. Res., 99, 14,443–14,459, doi:10.1029/94JD00797. Ito, T. (1993), Size distribution of Antarctic submicron aerosols, Tellus, Ser. B, 45, 145–159. Ito, T., and K. Iwai (1981), On the sudden increase in the concentration of Aitken particles in the Antarctic atmosphere, J. Meteorol. Soc. Jpn., 59, 262–271. Jaenicke, R., V. Dreiling, E. Lehmann, P. K. Koutsenoguii and J. Stingl (1992), Condensation nuclei at the German Antarctic station ‘Georg von Neumayer,’Tellus, Ser. B, 44, 311–317. Kerminen, V.‐M., K. Teinilä, and R. Hillamo (2000), Chemistry of sea‐salt particles in the summer Antarctic atmosphere, Atmos. Environ., 34, 2817–2825, doi:10.1016/S1352-2310(00)00089-3. King, J. C., and J. Turner (1997), Antarctic Meteorology and Climatology, J, 409 pp., Cambridge Univ. Press, Cambridge, U. K. Koponen, I. K., A. Virkkula, R. Hillamo, V.‐M. Kerminen, and M. Kulmala (2003), Number size distributions and concentrations of the continental summer aerosols in Queen Maud Land, Antarctica, J. Geophys. Res., 108(D18), 4587, doi:10.1029/2003JD003614. Kulmala, M., H. Vehkamaki, T. Petaja, M. Dal Maso, A. Lauri, V.‐M. Kerminen, W. Birmili, and P. H. McMurry (2004), Formation and growth rates of ultrafine atmospheric particles: A review of observations, J. Aerosol Sci., 35(2), 143–176, doi:10.1016/j.jaerosci.2003.10.003. Minikin, A., M. Legrand, J. Hall, D. Wagenbach, C. Kleefeld, E. Wolff, E. C. Pasteur, and F. Ducroz (1998), Sulfur‐containing species (sulfate and MSA) in coastal Antarctic aerosol and precipitation, J. Geophys. Res., 103, 10,975–10,990, doi:10.1029/98JD00249. O’Dowd, C. D., M. C. Facchini, F. Cavalli, D. Cebrunis, M. Mircea, S. Decesari, S. Fuzzi, Y. J. Yoon, and J.‐P. Putard (2004), Biogenically driven organic contribution to marine aerosol, Nature, 431, 676–680, doi:10.1038/ nature02959. Parish, T. R. (1988), Surface wind over the Antarctic continent: A review, Rev. Geophys., 26, 169–180, doi:10.1029/RG026i001p00169. Shaw, G. E. (1988), Antarctic aerosol: A review, Rev. Geophys., 26, 89–112, doi:10.1029/RG026i001p00089. Siingh, D., V. Pant, and A. K. Kamra (2007), Measurements of positive ions and air‐Earth current density at Maitri, Antarctica, J. Geophys. Res., 112, D13212, doi:10.1029/2006JD008101. Teinilä, K., V.‐M. Kerminen, and R. Hillamo (2000), A study of size‐ segregated aerosol chemistry in the Antarctic atmosphere, J. Geophys. Res., 105, 3893–3904, doi:10.1029/1999JD901033. Tomasi, C., et al. (2007), Aerosols in polar regions: A historical overview based on optical depth and in situ observations, J. Geophys. Res., 112, D16205, doi:10.1029/2007JD008432. Virkkula, A., K. Teinilä, R. Hillamo, V.‐M. Kerminen, S. Saarikoski, M. Aurela, I. K. Koponen, and M. Kulmala (2006a), Chemical size distributions of boundary layer aerosol over the Atlantic Ocean and at an Antarctic site, J. Geophys. Res., 111, D05306, doi:10.1029/2004JD004958. Virkkula, A., I. K. Koponen, K. Teinilä, R. Hillamo, V.‐M. Kerminen, and M. Kulmala (2006b), Effective real refractive index of dry aerosols in the Antarctic boundary layer, Geophys. Res. Lett., 33, L06805, doi:10.1029/ 2005GL024602. Virkkula, A., A. Horsikko, M. Vana, P. P. Alto, R. Hillamo, and M. Kulmala (2007), Charged particle size distributions and analysis of particle formation events at Finish Antarctic research station, Aboa, Borel, Environ. Res., 12, 397–408. Warren, S. G., and A. D. Clarke (1990), Soot in the atmosphere and snow surface of Antarctica, J. Geophys. Res., 95, 1811–1816, doi:10.1029/ JD095iD02p01811. Wolff, E. W., and H. Cachier (1998), Concentrations and seasonal cycle of black carbon in aerosol at a costal Antarctic station, J. Geophys. Res., 103, 11,033–11,041, doi:10.1029/97JD01363. A. K. Kamra and D. Siingh, Instruments and Observational Techniques, Indian Institute of Tropical Meteorology, Dr. Homi Bhabha Rd., Pashan, Pune 001‐411 008, India. ([email protected]) V. Pant, Modeling and Ocean Observations Group, Indian National Centre for Ocean Information Services, Hyderabad 500055, India.

15 of 15