Latitudinal and Size Segregated Compositional Variability of Aerosols over the Indian and Southern Ocean during 2010 Austral Summer

To characterize the spatial variations in size segregated aerosols, samples were collected over the Indian and Southern oceans (7°N to 68°S) during the Indian expedition to Southern Ocean (January–February 2010). The chemical and mass concentrations of ions were measured in nine size classes (0.4–> 10 μm) of aerosols through ion-exchange chromatographic system and gravimetric estimates, respectively. Mass concentrations of coarse aerosols increase towards the Antarctic coast which is well correlated with the increasing humidity suggesting growth of particles. The study area was found to be predominately impacted by sea salt aerosols with a sea salt load of 91%. F, Cl and NO3 reflected anthropogenic impact within the fine mode aerosols but the marine influence dominated the coarse mode. Methanesulphonic (MSA) was predominantly of biogenic origin but a substantially low MSA/nssSO4 ratio suggested that DMS contribution to the total nssSO4 concentration was low. NH4 concentrations showed a shift from fine to coarse particles, with a trimodal distribution over the Southern Ocean. Our study reports a significant Cl depletion in the aerosols and the degree of Cl deficit was size-dependent, increasing with decreasing particle sizes. The neutralization ratios suggested differential behavior of ions with respect to the prevailing meteorological conditions. The coarse mode aerosols were neutralized by ammonia leading to the formation of NH4NO3 particles. The fine modes were predominately composed of NH4HSO4 and its formation is favored by the high humidity and foggy conditions over the study area. Factor analysis support that the Southern Indian Ocean (10°S to 59°S) had the highest loading of anthropogenic and sea salt aerosols, which lead to the formation of secondary aerosols through gas to particle transformations. The rocky Antarctic coastline, act as a source of coarse crustal aerosols over the Southern Ocean (60°S to 65°S).


INTRODUCTION
Marine aerosol is a major component of the global atmospheric aerosol, and the specific physical and chemical properties of these aerosols have a key role in atmospheric processes.They are the main route of transferring marine species onto land, and a major channel through which terrestrial species are transported to the oceans (Duce et al., 1991).Therefore, an understanding of the composition and source of marine aerosols is important to evaluate the relative impact of anthropogenic and biogenic processes on the global biogeosphere and climate.Aerosols over the remote oceans consist of a mixture of sea-salt, sulfates, and organics, with frequent contributions from continental emissions (e.g., mineral dust and biomass burning) (Andreae, 2007).Marine aerosol can be divided into two distinct aerosol types: (1) primary sea-salt aerosol produced by the mechanical disruption of the ocean surface and (2) secondary aerosol, primarily in the form of non-sea-salt sulphate (nssSO 4 2-) and methanesulphonic acid (MSA), formed by gas-toparticle conversion processes (O'Dowd et al., 1997).Sea spray is a critical component of the primary aerosol and originates from bubble bursting during whitecap formation by surface winds over open ocean waters (Monahan et al., 1986).In remote oceanic regions, sea salt aerosol is often the primary light scattering aerosol component, and it is an important source of cloud condensation nuclei (Prospero, 1996).The sulphur containing aerosols (nssSO 4 2-and MSA) could be an important source of cloud condensation nuclei in unpolluted marine atmosphere such as that of the Southern Ocean (Meskhidze and Nenes, 2006) and may interact with incoming solar radiation, affecting cloud microphysics and consequently climate (Ayers and Gras, 1991;Liss and Lovelock, 2007).Moreover, the size of aerosols is important because their scattering of radiation, ability to serve as CCN, and the lifetimes of aerosols are controlled by their size-dependent chemical composition (Hubert et al., 1996).Several studies have been carried out in different oceanic regions that report the composition and distribution of marine aerosols in polar areas.Latitudinal distribution of marine aerosols studied over the Arctic during the first Chinese National Arctic Research Expedition (Sun, 2002) reported that Na + and Cl -ions were the dominant species (60%) in marine aerosols followed by sulphate.Similar studies were also carried out from China coast to Antarctic coast (Xu et al., 2011(Xu et al., , 2007)).Studies in the southwest Indian Ocean suggest that the atmosphere is a source of trace metals and ions to the Indian Ocean which may enhance or reduce the ocean productivity (Witt et al., 2010).Latitudinal distributions of atmospheric MSA and MSA/nssSO 4 2ratios revealed that variations in temperature are insignificant in controlling their variation over the high latitude regions of the Southern and Northern Hemispheres (Chen et al., 2012).However, no studies are available for the compositional characterization of aerosols over the Indian sector of Southern ocean.The surrounding continents of South Africa, Southern Indian Peninsula and Australia at the northern side of this study area and the remote ocean at the southern side provide an ideal natural setting to examine the chemical reactions between natural particles (e.g., dust and sea salt) and gaseous pollutants (e.g., sulfuric and reactive oxidative nitrogen species).
The present study reports the chemical composition and mass concentration of coarse (diameter: 10-2.5 µm) and fine (diameter: < 2.5 µm) particles in the marine atmosphere along the transect from India to Antarctica during January-February 2010, covering three major oceanic regions (Northern Indian Ocean, Southern Indian Ocean and Southern Ocean).The study aims in understanding the distribution of primary and secondary aerosols along the transect and the major processes influencing their distribution and provenance.

MATERIALS AND METHODS
Sampling of atmospheric marine aerosols was carried out from 14 th January to 15 th February 2010 using an 8staged non viable Impactor Thermo Fischer Scientific) onboard ORV Sagar Nidhi between 7°N to 65°S.The entire sampling track can be divided into three broad oceanic regions: Northern Indian Ocean -7°N to Equator (NIO), Southern Indian Ocean -1°S to 59°S (SIO), and Southern Ocean -60°S to 65°S (SO), (Fig. 1).A total of 10 sets of 24 hourly aerosol samples (sampling stations S1-S10) were collected for the compositional and size analysis of aerosols.Each set comprised of 8 glass fiber filters (81 mm diameter) for collecting coarse mode (10 µm-2.5 µm) and fine mode (< 2.5 µm) aerosols along with the backup filter.The flow rate of the impactor was calibrated regularly and was fixed at 1 Cfm.The impactor was mounted above the bridge of the ship and the sampling was stopped when the ship turned away from the headwind or when the ship was halted for other operations.Field blanks were collected at regular intervals.All operations were done by wearing particle-free gloves and all precautions were taken to minimize contamination.Handling of the filter sets was carried out in the laminar flow cabinet and they were stored frozen in sealed bags and transported back to the onshore laboratory for analysis.
Each 81 mm glass fiber filter was cut into half using plastic scissors in a clean laminar bench.One half of the filter was extracted in three steps (10, 5 and 5 mL) with 18 MΩ deionized water following standard procedures (Arimoto et al., 2008).A Dionex ICS-2500 reagent-free ion chromatography (RF-IC) system placed in a clean room and equipped with an automated EG50 Eluent Generator Module and CD25 conductivity detector was used for the major ion analysis.The cations were separated on an IonPac CS17 (4 mm) column with methanesulfonic acid (MSA) as eluent at a flow rate of 1.0 ml/min, using the gradient method and an IonPac CG17 Guard column with a CSRS-ULTRA Suppressor.The anions were separated on an IonPac AS11-HC (4 mm) column with Potassium Hydroxide (KOH) at 1.2 ml/min as eluent, using gradient method and an IonPac AG11-HC Guard column, with an ASRS-ULTRA (4mm) Suppressor.Calibration was done using IV (Inorganic Ventures) high-purity standards.In order to confirm the quality of measurements, chromatographic standards were analyzed on a daily basis throughout the study.The precision estimated from the standard deviation of repeat measurements of standard and samples was better than 3% for Na + , K + , Mg 2+ and Ca 2+ ; 4% for NH 4 + ; 6% for Cl − , and 4% for NO 3 − , SO 4 2− and MSA.The detection limits were within 5 µg/L for Na + , NH 4 + , K + , Cl − , F -, SO 4 2-and NO 3 2-, 2 µg/L for MSA and within 8 µg/L for Ca 2+ and Mg 2+ .Under certain conditions of temperature and humidity some artifacts can occur on the filters related to the interaction between the particles collected, the interaction between gas and particles and evaporation of volatile and semi-volatile substances.These interactions can alter the composition of the collected particles.While ammonium sulfate [(NH 4 ) 2 SO 4 ] can be considered as a conservative species (i.e., not subject to adsorption or volatilization), ammonium nitrate (NH 4 NO 3 ) is a semi volatile species and exists in reversible phase equilibrium with nitric acid (HNO 3 ) in the gas phase.Hence, the concentrations of aerosol-nitrate can be affected by the evaporative loss of the semi-volatile NH 4 NO 3 (negative artifact) or adsorption of HNO 3 during or after the sampling (positive artifact); however, NO 3 -volatilization generally dominates on adsorption (Schaap et al., 2004b;Vecchi et al., 2009).The evaporative loss of aerosol nitrate has been estimated following the empirical correlations proposed in Pathak and Chan (2005) and Pathak et al. (2009).In this study, H + Total was estimated using the ionic balance of the most relevant inorganic ionic species (Lippmann et al., 2000;Pathak et al., 2009), including sulfate, nitrate, chloride, ammonium and sodium: For ammonium rich samples: For ammonium poor samples: On this basis the average fine aerosol nitrate loss for ammonium rich and ammonium poor samples was 15% and 20% respectively and the average coarse aerosol nitrate loss was 20% and 27% respectively.A total of 5 blank filters were analyzed assuming the volume of air to be 1 Cfm.The blank values for all the ions were < 0.05 µg/m 3 except for Na + and K + which showed higher values.The mass concentration of aerosol was determined by the gravimetric estimates.The filters were placed in desiccators for ~24 hrs before and after the sampling to remove the absorbed water and weighed in a controlled environment chamber using a semi-micro balance (Sartorius, Model ME 235 P).
Air mass back trajectories were calculated at 100m above sea level (Draxler and Rolph, 2003) using the HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model accessed via NOAA Air Resources Laboratory READY (Real-time Environmental Applications and Display sYstem) site.The height was chosen between the sampling height (~20 m) and the upper limits of the marine boundary layer (typically of 500 m, Argentini et al., 2005).With a view to examine the potential pathways favoring longrange transport to each of these sampling sites, the 5-day air mass trajectories were reconstructed (Fig. 1).The ITCZ (Inter Tropical Convergence Zone) is marked by North East trade winds blowing between 7°N to 5°S.The winds are predominantly north easterly till 40°S except at 11°S where a slight deflection in the wind direction from NE to NW is observed.Further south of 40°S westerly winds dominate the sampling sites, with the inland air blowing from the Antarctic continent impacting the coastal sites.These distinctly different wind fields impart extreme spatial variability in the ionic composition of the aerosol particles.Trajectories calculated every 4-5 hours from the start of a sample period is consistent throughout the sampling time.The meteorological parameters recorded by the AWS installed onboard revealed decreasing air and water temperatures with increasing humidity from north to south during the study period (Fig. 2).The foggy conditions prevailing south of 45°S during the study period are supportive for the increase in moisture content of the atmosphere.The wind speed remained quite variable ranging from 4.0-16.5 m/s with the highest at the Antarctic coastline (Fig. 2).

Aerosol Mass Concentration and Size Distribution
The average aerosol mass concentrations observed in coarse and fine modes were comparable, measuring 3.0 µg/m 3 and 2.8 µg/m 3 , respectively.This is substantially less than the value (46.5 µg/m 3 ) reported over North Atlantic Ocean (Johansen et al., 2000).This difference can be attributed to the fact that North Atlantic Ocean has a much greater contribution from continental sources in their aerosol loadings compared to the remote sector of Southern Ocean.There was an increase in the concentration of coarse particles (Fig. 3) from 7°N to 65°S and highest concentration of 7.3 µg/m 3 was observed at sampling site S9 (61°S).Such a systematic increase in concentration towards the southern latitudes can be attributed to the increasing relative humidity, sea fog and low pressure conditions towards the Southern Ocean (Fig. 2).At low relative humidity (RH), inorganic aerosols mostly exist as small crystalline particles.But as the RH increases further, water continues to condense on the droplet, causing it to grow in size in order to maintain equilibrium with the surrounding water vapor (Seinfeld and Pandis, 1998).Alternatively, it is possible that coarse particles preferentially act as condensation nuclei of sea-fog droplets compared to fine particles because of the effect of surface curvature (Sasakawa and Uematsu, 2002).This is also reflected by the average concentrations of aerosol particles which follow a sequence of: SO > SIO > NIO.
To estimate the distribution of major water soluble ions in the nine size classes of aerosols (0.4-> 10 µm), the mass size distribution was studied for all the sites (S1-S10).One representative set of size-segregated sample from each oceanic zone S2, S6 and S10 was selected for detailed mass size analyses of the entire transect (Fig. 4).Observations reveal that a mixed distribution pattern of ions is seen over the entire transect.Over NIO, Na + , SO + in the Indian and southern oceans.In the Indian Ocean the primary and secondary peak of NH 4 + is at 3.3-4.7 µm and 5.8-9.0 µm, respectively.This distribution pattern continues till 50°S and gets reversed further south towards Antarctica.Samples south of 50°S to coastal Antarctica (represented by S9 in Fig. 4(f)) indicate a trimodal distribution.The maximum concentration of NH 4 + is seen in particles of 4.7-5.8µm aerodynamic diameters followed by two secondary peaks at 9.0-10.0µm and 1.2-2.1 µm.This distribution seemed biased towards the coarse fraction.This may also indicate that NH 4 + shifts from fine to coarse particles towards the coastal Antarctica.Such variability can be attributed to the growth of nucleation  mode particles by homogenous or heterogenous processes (Kamara et al., 2003).This observation is supported by the increased humidity and sea fog conditions in SO during our study period (Fig. 2).On the contrary higher chloride concentration was observed in coarse particles as compared to fine particles, throughout the transect, indicating higher Cl -loss with decreasing particle size.This is further discussed in the section of Cl -depletion seen during our study.

Latitudinal Variations in Aerosol Composition and Processes Influencing Their Distribution
The latitudinal distribution of ions is discussed with special emphasis on three major categories of aerosol species (sea salt aerosols, sulphate aerosols and Nitrate/Ammonium aerosols) that play an important role in marine biogeochemical and climate processes (Prospero, 1996).The mean concentrations of major cations and anions are presented in Table 1.The concentration sequence of major ions over the study transect is as follows: Sea salt aerosols, produced by the bursting of bubbles at the surface of the ocean, dominate the marine aerosol and its influence over the study area was quantified using the following equation, Eq. ( 4): where 1.47 is the value of (Na + + K + + Mg 2+ + Ca 2+ + SO 4 2-+ HCO 3 -)/Na + in sea water.This assumes that all of the Na + and Cl -in the aerosol is sourced from the sea water, and all non-sea-salt components of K + , Mg 2+ , Ca 2+ , SO 4 2and HCO 3 -have been excluded (Quinn et al., 2001).The estimated sea salt content substantially increased from north to south in the study region with a total sea salt load of 91% at the sampling sites.The latitudinal variation of sea salt along the study transect (Fig. 5) showed the highest value (49.4 µg/m 3 ) from 56°S to 61°S, where the maximum wind speed (14.9 to 16.5 m/s) was recorded (Fig. 2).Minimum sea salt concentration was recorded over the NIO where the wind speed varied from 5.0-6.8m/s (Fig. 2).This study supports the fact that high wind speed assists in the production of new sea salt particles.This could be due to excessive whitecap formation during high wind speed over the ocean, which is a main source of sea salt in marine aerosols (Erickson et al., 1986;Bates et al., 1998).Na + and Cl -have similar variations till 44°S but a substantial decrease south of 44°S is observed in case of Cl -.This decrease can be attributed to the fact that sea salt aerosols can participate in heterogeneous reactions with nitric and sulfuric acids, leading to Cl -depletion through HCl volatilization (Mouri et al., 1999;Hara et al., 2004).The high SO 4 2-concentrations south of 44°S supports such finding.Mg 2+ concentration is consistent with the sea salt variation till 44°S but south of STF Mg 2+ shows a trend similar to Cl -, with its highest concentration of 3.6 nM/m 3 at 65°S (S10).Such variability indicates that although Mg 2+ is predominately contributed from sea salt some fractionation occurs at the air-sea boundary layer specifically in the Southern ocean region (Fig. 5).As seen from the air mass trajectory (Fig. 1), the sampling sites in SO (S9 and S10) are influenced from two different air masses: the inland winds, blowing from the Antarctic land mass and the south westerly marine air mass.Such a mixed air mass is most likely to cause variations in the concentrations of sea salt components.Within NIO, it was found that the concentration of Na + , K + , NO 3 -is higher whereas the concentration of Cl -, Mg 2+ , Ca 2+ is much lower as compared to the values reported in Tropical North Atlantic (Johansen et al., 2000) and Arabian Sea (Kumar et al., 2008).This variability in the ionic composition of aerosols could be due to changes in the wind direction and season during the sampling time.The data reported by Kumar et al. (2008) was for the transition period of April-May where the wind direction is predominately from south west blowing over the landmasses of North Africa and Arabian Sea thus bringing in an air mass which is mixed with continental as well as marine aerosols.However, the samples (S1 and S2) in this study were collected when the wind is from North West, which is mostly of marine origin with some anthropogenic influence from the Indian subcontinent.
In the atmosphere over remote marine regions, there are two main sources for aerosol nssSO 4 2 : reactions involving either dimethylsulfide (DMS) or primary SO 2 produced through anthropogenic emissions (Stern, 2006;Manktelow et al., 2007).To assess the anthropogenic and/or biogenic  ) was calculated by assuming Na + as a conservative indicator of marine origin.The following equation, Eq. ( 5) is used for the calculation: where 0.060 is the mole ratio SO 4 2-/Na + for sea water.Na + content has been used in preference to Cl -concentrations because the latter can be volatilized following the reaction of sea salt with acid (Legrand and Delmas, 1988).The concentrations of nssSO 4 2-stayed at relatively low levels during this study, with 70% of the values below 0.01 µM/m 3 .This is consistent with the results reported by Chen et al. (2012) in the Southern Hemisphere during the Chinese expeditions to Antarctica where, around 80% of the nssSO 4 2-values were found below 0.01 µM/m 3 .However this value is high as compared to the value reported in the Southwest Indian Ocean (Witt et al., 2010).This variation in concentration of nssSO 4 2-could be attributed to changes in the wind intensity and direction in the study area.It is supported by the fact that south-easterly wind dominated during the study of Witt et al. (2010), which could be a predominantly marine air parcel as compared to the westerly winds influencing our sampling sites in SIO (S4-S8).The highest concentrations of nssSO 4 2-reported in this study was observed at S3, near 11°S (0.061 µM/m 3 ) and S5, near 37°S (0.072 µM/m 3 , Table 1).Contribution of biogenic source to nssSO 4 2-concentrations can be understood by the concentrations of another sulphur compound MSA, the oxidation product of DMS/OH reaction pathways.The maximum value for MSA (0.25 nM/m 3 ) was seen at 56°S which is marked as the Polar front in our study transect (Table 1) confirming a high marine biological activity in this region.The MSA concentration reported in our study is less than the values reported in other oceanic regions of Southern Hemisphere (Davison et al., 1996;Chen et al., 2012).Such low values may be due to the evaporation losses of MSA during the study.These losses occur due to the high concentration of SO 4 2-and NO 3 -as compared to low NH 4 + concentrations, which tend to make the aerosols more acidic and further enhancing the evaporation (Brimblecombe and Clegg, 1988;Savoie et al., 1993).
The variations in MSA and nssSO 4 2-concentrations along the transect were not correlated and this difference in the concentration variation of MSA and nssSO 4 2-indicates that anthropogenic source is influencing the SO 4 2-concentrations.The marked peaks of nssSO 4 2-at S3 and S5 could be attributed to the contribution from Madagascar (12°25'S, 49°20'E to 25°30'S, 45°11'E) and Amsterdam islands (37°50'S, 77°30'E), respectively, as is evident from the wind direction in the trajectory analysis (Fig. 1).These two sites are impacted by the westerly winds passing over the islands hence enriching the marine air with anthropogenic aerosols.Another method to estimate the natural oceanic contribution (via DMS) to atmospheric nssSO 4 2-aerosol concentrations is the widely used approach of comparing the measured MSA to nssSO 4 2-ratio (R) in aerosol particles with an expected ratio resulting from pure atmospheric DMS oxidation (Berresheim et al., 1991;Li and Barrie, 1993;Mihalopoulos et al., 1997).The factors governing the MSA/nssSO 4 2-ratio includes: the relative magnitudes of the marine and other sulfur sources influencing the site, the location of the marine source itself, and the development of R during the air mass transportation due to deposition and oxidation processes (Kerminen et al., 2000).A high ratio may indicate that a considerable fraction of the total nssSO 4 2-burden is derived from the atmospheric oxidation of DMS, while a low ratio implies that the contribution of DMS to the total nssSO 4 2-burden is low.The ratio (0.002) of our present study is substantially lower than the ratio observed in the Indian Ocean (0.036, Saltzmann et al., 1983) as well as Cape Grim, Southern Ocean (0.08-0.5, Berreshim et al., 1990).This clearly substantiates that atmospheric oxidation of DMS has no significant contribution towards nssSO 4 2-values during the study period.A noticeable observation is that the graphical variation of F -and NO 3 -is similar to an anthropogenic tracer nssSO 4 2-as well as to a marine tracer Cl -with their highest concentrations at S3 (11°S) and S5 (35°S).Many authors have regarded F -and NO 3 -as anthropogenic species in the marine atmosphere (Harnisch, 1999;Savoie et al., 2002).It is significant to note that the Cl -concentration also spikes at these two anthropogenically influenced sampling sites (S3 and S5) indicating significant anthropogenic influence in Cl - concentrations as well.Some of the earlier studies (Erickson and Duce, 1988) have reported that the gaseous Cl -and F - (in the form of HCl and HF) could be contributed from combustion processes.But F -also exists as an important component of marine aerosol (Symonds et al., 1988).To verify the source of these ionic species the F -/Cl -ratios was calculated for the sampling sites.This ratio was found to be in the range of 8.1 × 10 5 -21.4 × 10 5 which is higher than the F -/Cl -value of sea water (6.7 × 10 5 , Wilknes and Bressan, 1971).Such changes in the ratios could be the result of enrichment of marine aerosols by the continental dust or the anthropogenic sources.NH 4 + in the marine aerosol could be anthropogenic as well as marine (Andreae and Merlet, 2001;Baker et al., 2006a).The Net Primary productivity (NPP) calculated from the MODIS data for the study period (Fig. 5) (Behrenfeld and Falkowski, 1997) revealed a positive correlation with NH 4 + concentration in the regions of SIO and SO.NH 4 + exhibits its highest concentration (0.05 µM/m 3 ) at around 42-44°S which coincides with the Sub Tropical Front (STF) in our study region.This region typically exhibits chlorophyll (Chl a) biomass enhancement (Fig. 5) which reflects a positive flux of N 2(g) which reacts to form NH 3(g) and subsequently NH 4 + in this region.However, mean NH 4 + concentration (0.015 ± 0.278 µM/m 3, Table 1) reported in our study was lower than the value reported from South Atlantic (0.412 µM/m 3 , Virrkula et al., 2006).A noticeable observation is that the sampling sites in NIO (S1 and S2) (7°N and 11°S) showed low NH 4 + concentrations despite the high productivity in the region.Such an inverse relationship could be due to the evaporation of existing NH 4 + from the aerosols particles due to high temperatures and low humidity conditions (Milford et al., 2000) prevailing over NIO atmosphere as compared to southern latitudes (Fig. 2).Considering the sum of cations and anions over the varied oceanic region covered in this study, it is observed that the total cationic load was highest over the SO with the minimum over NIO.On the contrary the SIO had the maximum anionic load.

Chemical Characteristics of Coarse and Fine Mode Aerosol
The distribution of ionic species into coarse (10-2.5 µm) and fine (< 2.5 µm) can assist in making better estimates of source of aerosol particles.The sea salt and crustal aerosols are larger in diameter, hence get deposited faster due to their size whereas the gaseous precursors of aerosol species contributed through anthropogenic or biogenic emissions reside in the finer fraction which travels to greater distances (Raes et al., 2000).The distribution of the ions in coarse and fine aerosols revealed that greater than 95% of K + , Mg 2+ and NH 4 + , 85% of Cl -, 56% of Ca 2+ and about 60% F -is found associated with the coarse fraction (Fig. 6).This is mainly because once the seawater droplets from bubble bursting are in the air, they evaporate leading to concentrated saline droplets or minute crystals of airborne sea salt with a larger radius around 4 μm and containing 4-50 pg of salt (Udisti et al., 2012).The good correlation between Na + , Mg 2+ , K + , F -and Cl -suggest the same source (sea spray) for these ions (Table 3).The sea salt ions were present in the form of NaCl and MgCl 2 crystals.Among the nitrate/ammonium aerosols, NH 4 + was dominant in the coarse mode while NO 3 -was almost equally divided in the coarse (46%) and fine (56%) modes, suggesting that the NO 3 -concentration was influenced by marine as well as anthropogenic sources.The nssSO 4 2-and MSA which are generally categorized as secondary marine aerosols or the sulphur aerosols dominate the fine fraction (78.2% and 68%, respectively, Fig. 6) but they show no correlations in both the aerosol modes (Table 3) indicating different sources in our study region.MSA being biased towards the fine particles supports that the gas to particle conversion process is responsible for the accumulation of alkyl ammonium salts in the fine mode.In a comparative analysis between these two anions it was found that greater amount of MSA relative to nssSO 4 2-was associated with particles > 2.5 μm.This could be due to a higher solubility and lower vapor pressure of MSA in larger, less acidic sea salt aerosol (Jefferson et al., 1998).Also, MSA values showed an increase in the concentration in the coarse particles from 7°N to 60°S with a value peaking at 56°S in the Polar frontal region.On the contrary, the presence of greater percentage of nssSO 4 2-in the fine fraction (78%) suggests that it is contributed by combustion processes and other anthropogenic activities of the neighboring continents (Hsu et al., 2007).The good correlation between nssSO 4 2-, Na + and K + suggest a similar source which could be related to anthropogenic sources, since the fine particles can travel greater distances carrying anthropogenic species from the neighboring continents (Table 3).

Size Dependent Cl -Depletion and Neutralization Ratios
Inter-ionic ratios for the aerosol samples provided evidence of the volatilization of Cl -.The mean CI -/Na + ratio of 0.05 (0.02-0.1) was found to be much less than that for sea water ratio (1.17, Chester, 1971).Therefore, it can be concluded that high Cl -depletion did occur in marine aerosols in the study period.The Cl -depletion percentage (% Cl dep ) was calculated using the Eq. ( 6), (Quinn et al., 2000): where Cl - ss = 1.17 × Na + meas and Cl - meas and Na + meas are the measured Cl -and Na + mass concentrations.This reveals a Cl -depletion event taking place in the aerosols over the study region.A noticeable depletion of chloride in salt particles was reported in a number of field studies conducted in both polluted and pristine marine environments (Ebert et al., 2000;Laskin et al., 2005;Hopkins et al., 2008).Moreover, the percentage of chloride lost from sea-salt particles decreased with increasing particle size in the study sites.This is confirmed by the Cl -/Na + ratios in the two modes of aerosols, where this ratio was higher in coarse mode (0.064) than in the fine mode (0.025).Further, the average chloride loss was 97% for fine particles and 70% for coarse particles.The percentage of Cl -depletion in the fine particles over NIO (95%) is similar to the percentage reported for this region by Johansen and Hoffmann (2004).This is expected, since smaller sea-salt particles have greater surface area-to-volume ratios and longer atmospheric lifetimes making it more feasible for reactions to take place Thakur and Thamban, Aerosol and Air Quality Research, 14: 220-236, 2014 229 as revealed by the mass size distribution of Cl -in the study sites (Fig. 4(e)).
To evaluate the acid displacement reactions responsible for this size dependent chloride depletion, the nssSO 4 2-/NO 3 and nssSO 4 2-/Na + ratios was calculated in the fine and coarse modes (Table 2).The nssSO 4 2-/Na + ratio in coarse particles (0.013) was 4 times lesser than in fine particles, suggesting dominance of nssSO 4 2-in the fine aerosol.This indicates a probability of replacement of Cl -by SO 4 2-in fine mode.Further, it is interesting to note that for coarse-mode particles the mean nssSO 4 2-/NO 3 -ratios was less as compared to the ratio for fine mode.Accordingly, the coarse NO 3 -concentrations always exceed coarse nssSO 4 2- . In contrast, fine nssSO 4 2-largely exceeds fine NO 3 -, revealing that coarse mode NO 3 -may form onto the sea salt.It occurs essentially through acid displacement between nitric acid and deliquesced sea salt (Quinn and Bates, 2005).Moreover, the high sea salt loading (91%) at the sampling sites especially towards the SO favors this heterogeneous reactions of gaseous species (HNO 3 ) with coarse aerosol species (NaCl), Table 2. Average values of ions with standard deviation in fine and coarse particles.All values reported in µM/m 3 except for Mg 2+ , Ca 2+ and MSA, their values are reported in nM/m 3 ).which is the alternative pathway for the formation of nitrate aerosols over the oceans.The observed nssSO 4 2-/NO 3 -in coarse and fine mode aerosol thus suggest that the nitrate present in the fine mode aerosol fraction shifts to the coarse mode, as non volatile NaNO 3 (Wall et al., 1988;Zhaung et al., 1999) while the Cl -in fine mode is replaced by SO 4 2during our study according to Eq. ( 7).NaCl (aq) + HNO 3(aq) ⇋ NaNO 3(aq-s) + HCl (g) ( 7) South of 37°S, the NO 3 -concentration increases 1.5 times in coarse particles suggesting equilibrium shift as the marine and continental air masses mix.The formation of NH 4 NO 3 by a reversible reaction of HNO 3 and NH 3 in the fine mode is ruled out in our study sites mainly because the NH 4 + /SO 4 2-ratio in our study is only 0.3 which suggests an NH 4 + poor ambient atmosphere of the SO.This indicates a more photochemically active atmosphere, where an enhanced generation of oxidants (mainly OH and O 3 ) occurs and which could prevent the formation and evaporation of NH 4 NO 3 from filters (Trebs et al., 2005).NH 4 NO 3 is formed in the presence of high concentrations of NH 3 and HNO 3 , low temperature and high humidity (Stockwell et al., 2000;Pathak et al., 2009).In samples characterized by NH 4 + /SO 4 2-> 1.5, NH 4 + stabilizes NO 3 -whereas at NH 4 + /SO 4 2-< 1.5, NO 3 -neutralization may depend on: (i) gas-phase reaction between HNO 3 and sea-salt particles or fine crustal particles (ii) heterogeneous hydrolysis of N 2 O 5 during night-time on the pre-existing (NH 4 ) 2 SO 4 or NH 4 HSO 4 particles in high relative humidity conditions (Squizzato et al., 2012).Depending upon two prevailing conditions of high sea salt loading and ammonia poor ambient atmosphere during the study period, we can suggest that NH 4 HSO 4 dominates over NH 4 NO 3 in our study sites.Moreover, the SO 4 2− and NO 3 -concentrations during the study were much higher than the NH 4 + concentration and they compete for the limited NH 3 available.In addition, the high relative humidity conditions (daily mean 35%-90%) in summer period might dissolve a significant fraction of HNO 3 and NH 3 in humid particles, therefore enhancing fine particulate NO 3 -and NH 4 + in the atmosphere (Ianniello et al., 2010;Sun et al., 2010;Ianniello et al., 2011).But due to low temperatures and low ammonia conditions prevailing in our study regions, NH 4 + is preferentially scavenged by SO 4 2-rather than NO 3 to form NH 4 HSO 4 (Watson et al., 1994).This reaction is also preferred because the precursor gas (H 2 SO 4 ) has a low vapor pressure which tends to enter into reaction with NH 4 + to form a chemically more stable secondary aerosol particle of NH 4 HSO 4 (Seinfeld and Pandis, 2006).Moreover, the acidity ratio (Engelhart et al., 2011) or neutralization ratio (NR) (Bencs et al., 2008), was used to describe the aerosol acidity, expressing the degree of neutralization of sulfate and nitrate by ammonium (expressed as equivalent).In this way NR expresses the aerosol acidity (Eq.( 8)) characteristics by considering only the possible neutralization of the two major inorganic acids (HNO 3 and H 2 SO 4 ) with ammonium.The NR ratio was found to be 0.5, confirming the formation of NH 4 HSO 4 , in the fine mode.The higher correlation coefficient (0.4) between NH 4 + and nssSO 4 2-(0.39 for total SO 4 2- ) in fine aerosol as compared to coarse aerosol (0.26) explains the existence of NH 4 HSO 4 salts in fine mode (Table 3).The NH 4 + /NO 3 -ratio in the coarse mode was found to be 1.17 indicating a complete neutralization of coarse aerosols with the formation of NH 4 NO 3 .A positive correlation coefficient (0.22) of NH 4 + and NO 3 in coarse mode as compared to negative correlation in fine aerosols supports the NH 4 NO 3 formation in coarse aerosols (Table 3).Such a reaction could be responsible for the shift of NH 4 + from fine to coarse particles with distance from the equator (Fig. 4(f)).

Source Apportionment through Principal Component Analysis (PCA)
In order to study the origin of these aerosols and the components dominating in it, factor analysis was undertaken.Aerosols over the remote oceans consist of a mixture of seasalt, sulfates, and organics, with frequent contributions from continental emissions (e.g., mineral dust and biomass burning) (Andreae, 2007).But Source apportionment is different for coarse and fine particles since different sizes of aerosol particles can act as sink for different ionic species owing to their varying properties like surface area, settling properties and atmospheric residence times (Prospero et al., 1996).The data used for PCA is converted to a zero mean, unit variance equivalent by subtracting the mean value of the variable and dividing by the standard deviation.This conversion gives each chemical parameter the same magnitude and variance and consequently the same influence in the PCA.In general, factor loadings greater than 0.5 are considered significant in source apportionment studies.The method extracted three factors associated with Eigen values > 1. Factors are shown (Table 4(a) and 4(b)) for both fine and coarse particles.The data is synthesized into three major factors, explaining about 70% of the total variance.After identifying the factors and the correlated species, the factor analysis was also carried out as a function of region to assess the effect of different sources on the sampling points.The factor scores for each sampling point was calculated and averaged according to the three major oceanic regions.The percentage contributions of each factor are graphically represented in Fig. 7.

Fine Mode Aerosols
Factor 1 accounted for 42.5% of the total variation and showed high loadings on F -, Cl -, NO 3 -and nssSO 4 2-The presence of these ionic components especially the nssSO 4 2in fine aerosols particles are indicators of anthropogenic influence over the study site (Arimoto et al., 1996;Gao et al., 1996).The trajectory analysis tends to explain such anthropogenic enrichment of aerosols.This factor was dominant over the SIO accounting for 45% of its contribution in fine particles (Fig. 7) and decreased towards the SO indicating the reduced effect of long range transportation.Factor 2 accounted for 17.1% of the total variation and was assigned as a marine source as it showed high loadings on Na + , NH 4 + and K + that are mostly of marine origin.Many studies have shown that an important fraction of sea salt is present in fine particles which form the predominant source of CNN specially in high wind condition (Putland et al., 2000) as prevailing during the study.This factor dominated the NIO and SO regions of our study with a relative contribution of 39% and 56% respectively (Fig. 7).
Factor 3 accounted for only 13% of the total variation and showed high loadings for Mg 2+ and MSA.The high loading of MSA (0.8) indicates an organic source as MSA is specifically of biogenic origin in the marine environment.Although Mg 2+ is largely of marine or continental origin, few studies suggested that small fraction of Mg 2+ in the marine environment could also be originated from biogenic sources due to their temporal correlation with DMS and Chl a (Gaston et al., 2011).This factor reflected a comparatively similar contribution over the study region ranging from 27% to 32% from SO to NIO (Fig. 7)

Coarse Mode Aerosols
Factor 1 accounted for 41% of the total variation and showed high loadings on Na + , Mg 2+ , F -, Cl -, and NO 3 -in coarse mode aerosols.This factor can be considered as the factor of marine spray owing to the high loadings of Na + and Cl -, the proxy species for sea salt.On the coarse mode F -can be present in sea salt through the bubble bursting phenomenon at the air-sea interface which deposits the terrigenous F -at the ocean surface, to the sea salt aerosol (Wilkniss and Bressan, 1971).Moreover NO 3 -of marine origin can be scavenged by sea salt particles due to their alkaline nature (Andreae et al., 2000).This factor was predominant over the SIO region (79%) of the study area (Fig. 7).Factor 2 accounted for 16.7% of the total variation and showed high loadings on NH 4 + and nssSO 4 2+ .Since these ions are a result of gas to particle reactions occurring on aerosol particles (Luo et al., 2007) this factor can be related to secondary sources which are predominantly anthropogenic in the study area.Sievering et al. (1995) found that heterogeneous oxidation of SO 2 by O 3 proceeds rapidly in freshly formed coarse sea-salt aerosol, because of the initial alkaline nature of the particles and this process could be suggested to be the major path for the formation of coarse mode nssSO 4 2-in the marine boundary layer.In our study the alkalinity required for this process is supplied by the presence of NH 4 + in the coarse particles.Moreover, the foggy condition in our study area favours the formation of coarse particle modes of ammonium salt aerosols (Yao and Zhang, 2012).This factor loading clearly supports the formation of (NH 4 ) HSO 4 particles at the study sites.Since the Factor 1 (anthropogenic aerosol) of fine mode and coarse mode PCA shows the highest percentage loading over SIO, it implies that this area is impacted by both sea spray and anthropogenic aerosols providing larger surface area for anthropogenic gases to undergo gas to particle transformations.
Factor 3 accounted for 15% of the total variation and showed high loadings on Na + , K + and Ca 2+ .Since Ca 2+ and K + are considered to be the major crustal elements in the marine aerosols, this factor can be related to crustal sources.This factor has the minimal loading since the coarse particles have high settling velocity hence cannot carry crustal elements to larger distances but can act as major sink for other atmospheric components.This factor dominated the SO region which is clearly explained by the trajectory analysis, where the winds blowing over the rocky Antarctic coast enrich the aerosols with crustal particles.

CONCLUSIONS
The aerosol mass concentration analysis along the transect revealed that the highest concentration of coarse and fine particles was found over Southern Ocean particularly in the coastal Antarctic region.These aerosols were primarily sea salt particles since the increase in mass concentration towards the south was in agreement with the increase in sea salt concentration, (> 95%) of sea salt species in the coarse particles.The size segregated ionic variation supported by back trajectory studies and statistical analysis, indicated that ions like F -, Cl -and NO 3 -were of marine origin in the coarse aerosols but were anthropogenically influenced in the fine mode.Moreover, the enhanced biological productivity of the region seems to influence the upward flux of MSA and NH 4 + .Although MSA was primarily of biogenic origin in the study transect, nssSO 4 2-was found to be more of continental origin.Along with the source apportionment of ions, the formation of secondary inorganic ions was also studied by investigating the relationship of gaseous precursors through inter-element ratios.The relatively low Cl -/Na + ratio in the study region suggests a significant volatilization of chlorine had occurred in the aerosol particles.Moreover, the degree of Cl -deficit was size-dependent, increasing with decreasing particle sizes.This confirms the fact that greater surface area -to-volume ratios and longer atmospheric lifetimes of fine particles allow maximum reactions to happen.Correlation analysis and inter-element ratios suggests that the Cl -ions were substituted by NO 3 -in the coarse particles and by SO 4 2-in the fine particles.Nevertheless, coarse and fine-mode sea salt aerosols behave in two different ways, the former reacting with HNO 3 and the latter entering into a reaction with H 2 SO 4 .The NH 4 + /SO 4 2-ratio being less than 1.5 characterized the ambient atmosphere of the study transect as an ammonium poor medium which prevents the formation of NH 4 NO 3 .But the foggy conditions favours the formation of (NH 4 )HSO 4 which is confirmed by the mass size distributions of NH 4 + .The NH 4 + /NO 3 -ratio close to unity in coarse mode suggest a complete neutralization of NH 4 + in coarse mode aerosols.Moreover the foggy conditions prevailing in the area favors the formation of NH 4 HSO 4 in the fine mode aerosols.The factor analysis reveals that F -, Cl -, NO 3 -and nssSO 4 2-were the primary anthropogenic species which were present in the fine aerosols predominantly over the SIO region.The factor loadings also reveal that the coarse mode aerosols over the SO are considerably enriched by the crustal particles from the exposed rocks of the Antarctic coastline.Thus, this study provides spatially distributed data of ionic chemistry over this poorly studied oceanic sector and also merits additional investigation to characterize physico-chemical properties of marine aerosols in this region.

Fig. 1 .
Fig. 1.Sampling Locations and Air mass back trajectories (AMBTs) calculated from the National Oceanic and Atmospheric Administration (NOAA) GDAS meteorology database, using the Hybrid Single-Particle Lagrangian Integrated Trajectories (HY-SPLIT) program.AMBTs were performed at 100 m height levels over the sampling locations.

Fig. 5 .
Fig. 5. Average concentrations of cations and anions along the study transect.The error bars show standard deviation and the vertical colour bars show the boundaries of oceans in the study transect.All units in µM/m 3 except for Mg 2+ , Ca 2+ and MSA, which are in nM/m 3 and sea salt concentration in µg/ m 3 .Note SSTF: South Sub Tropical Front, SAF: Sub-Antarctic Front, PF: Polar Front.

Fig. 6 .
Fig. 6.Ionic distributions in Fine and Coarse mode aerosols along the study transect.

Fig. 7 .
Fig. 7. Percentage contributions of different types (source and size) of aerosols over the three oceanic regions.

Table 1 .
Average values of ions at various latitudes (All values reported in µM/m 3 except for Mg, Ca and MSA, their values are reported in nM/m 3 , SD = Standard Deviation).