Journal of Geophysical Research: Atmospheres

Differentiating sulfate aerosol oxidation pathways for varying source altitudes using 35S and Δ17O tracers

Authors

  • Jason Hill-Falkenthal,

    Corresponding author
    1. Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California, USA
      Corresponding author: J. Hill-Falkenthal, Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093, USA. (jhillfal@ucsd.edu)
    Search for more papers by this author
  • Antra Priyadarshi,

    1. Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California, USA
    Search for more papers by this author
  • Mark Thiemens

    1. Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California, USA
    Search for more papers by this author

Corresponding author: J. Hill-Falkenthal, Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093, USA. (jhillfal@ucsd.edu)

Abstract

[1] Due to the complex nature of the sulfur cycle, uncertainties remain in the transport rate and chemical transformation mechanisms of sulfur species, including sulfate aerosols. Here, we report oxygen isotopic anomaly and radioactive 35S measurements in sulfate aerosols collected at La Jolla, California during 2009–2010. A strong correlation results from increased levels of specific activity (up to 195 atoms of 35S/nmol non sea salt (nss)-SO4) and Δ17O (up to 1.50‰) in sulfate aerosol (fine fraction) samples during Santa Ana wind events compared to background levels. This is possibly due to an increase in mixing of free tropospheric air mass, containing higher levels of 35S specific activity and higher Δ17O, into the boundary layer. These tracers show the ability to detect changes in oxidation chemistry during high altitude air mixing events and have the potential to trace the changes in oxidation pathways of sulfur species during Stratospheric-Tropospheric exchange events. Sampling at higher latitudes where deep stratospheric intrusions are more prominent can help further parameterize how stratosphere-troposphere exchange events (STE) affect oxidation chemistry in the boundary layer.

1. Introduction

[2] Sulfate represents a major chemical constituent of atmospheric aerosols, and is of particular importance due to its potential to serve as a cloud condensation nuclei as well as its enhancement of the Earth's albedo by reflecting sunlight back to space [Kulmala et al., 2000; Alexander et al., 2012]. Our knowledge of chemical and photochemical processes that govern the chemical transformations and transport of sulfur compounds in the atmosphere is incomplete due to the complex nature of the sulfur cycle and uncertainties in understanding aerosol chemistry [Brothers et al., 2010]. Gaining a better understanding of the chemical transformations and transport of sulfur compounds in the atmosphere is essential to improve our knowledge of sulfur chemistry and predict future climate change.

[3] Sulfur-35(35S) is a unique tracer in that it can be used to follow the sulfur cycle through both gas, as SO2, and particle phases, as aerosol sulfate, and has been used to study SO2 oxidation and transport between the boundary layer and free troposphere [Tanaka and Turekian, 1995]. 35S is a naturally occurring radionuclide (half-life: 87 days) produced from cosmic ray spallation of Argon. Once produced,35S rapidly oxidizes to 35SO (lifetime∼1 ms) via O2/O3 reaction and subsequently to 35SO2 (lifetime∼1 s) in the presence of O2, O3, N2O, or OClO [Black et al., 1982; Brothers et al., 2010; Priyadarshi et al., 2012; Robertshaw and Smith, 1980]. For a global average, only about 50% of SO2 is oxidized to SO42− via homogeneous (OH) or heterogeneous (H2O2, O3) oxidation with the remainder of SO2 lost through dry and wet deposition [Chin et al., 1996]. The production rate of 35S increases exponentially with altitude, attaining a maximum in the stratosphere [Lal et al., 1960]. Priyadarshi et al. [2012] calculated steady state concentrations of 2.8*10−6 atoms cm−3 day−1, 6.5*10−5 atoms cm−3 day−1, and 1.1*10−4 atoms cm−3 day−1 in the boundary layer, free troposphere, and lower stratosphere, respectively, for La Jolla, California. 35S activities of sulfate aerosols are a result of the equilibrium between 35S production rates, radioactive decay, oxidation of 35SO2 to 35SO42−, wet and dry deposition, and air mass mixing between the troposphere and stratosphere [Priyadarshi et al., 2012]. Due to the long residence time of gases and aerosols, 35SO2 and 35SO42− produced in the stratosphere mostly decays to 35Cl before reaching the troposphere, except during STE. STE may have a significant impact on the chemical composition of both the lower stratosphere and troposphere as these events can alter water vapor content, ozone concentration, and affects the oxidizing capacity of the atmosphere, thus changing the lifetime of many gases. Since 35SO42− possesses identical chemical properties as bulk SO42−, and the concentration of 35SO42− is expected to vary significantly in the troposphere and stratosphere, 35S has the potential to be an effective tracer for understanding boundary layer chemistry and STE [Brothers et al., 2010; Cho et al., 2011; Lal and Peters, 1968; Lee and Thiemens, 2001; Osaki et al., 1999; Priyadarshi et al., 2011, 2012; Tanaka and Turekian, 1991, 1995; Turekian and Tanaka, 1992].

[4] Combined with stable oxygen isotopic measurements, 35S has the potential to study how sulfate oxidation varies in the boundary layer due to mixing with high altitude air. Lee and Thiemens [2001] were the first to show a correlation between specific activity (35S abundance normalized to the amount of sulfate in a sample) and oxygen isotopes. Most atmospheric processes obey a mass dependent isotopic fractionation, as δ17O and δ18O vary in relative proportion to their reduced masses. Any process that deviates from this relationship is termed mass independent fractionation and is quantified by Δ17O, where Δ17O ≈ δ17O−0.515 δ18O, δ = (Rsample/Rstandard−1)*1000 and R = 17O/16O or 18O/16O. In the troposphere, SO2 is presumed to be mass dependent due to equilibrium isotope exchange with water [Holt et al., 1981] as the isotopic composition of water in the troposphere is dominated by mass dependent isotope effects such as meteorological evaporation and condensation [Franz and Röckmann, 2005]. Therefore, any observed mass independent compositions found in sulfate aerosols derive from the oxidation of SO2 to SO42− [Savarino et al., 2000]. OH radicals are mass dependent even though there exist several processes that contribute nonzero MIF anomalies [Morin et al., 2007] including formation via photolysis of ozone and nitrous acid, or reaction of HO2 radicals with NO [Seinfeld and Pandis, 1998]. The fast equilibrium isotope exchange between water and OH, which is faster than all OH sinks [Michalski et al., 2003], eliminates any Δ17O anomaly originally produced in OH radicals. H2O2 and O3 have been experimentally determined to possess Δ17O values of ∼1.7‰ and 27‰, respectively [Johnston and Thiemens, 1997; Savarino and Thiemens, 1999; Thiemens and Heidenreich, 1983]. Recently, ambient measurements of Δ17O(O3) bulk have yielded a value of 22.9 ± 1.9‰ using nitrite coated filters [Vicars et al., 2012]. Although slightly lower than measurements obtained by Johnston and Thiemens [1997], the reported value is within the error of the previous studies which used cryogenic trapping to collect ozone. These laboratory measurements have been applied to field observations and include the recognition that wintertime Arctic SO2 oxidation processes are dominated by O2 via transition metal catalysis on aerosol surfaces, rather than OH and H2O2 [Alexander et al., 2009; McCabe et al., 2006]. Quantitative definition of the contributions from maritime shipping to the primary anthropogenic sulfate budget has also been reported [Dominguez et al., 2008]. Modeling efforts of Δ17O has been employed to identify specific reaction characteristics such as the role of aerosol particle surfaces above the Indian Ocean as well as show the variation in tropospheric O3and OH oxidation from pre-industrial to industrial times in ice cores [Alexander et al., 2005; Kunasek et al., 2010].

[5] In the stratosphere, both SO2 and OH are expected to contain larger Δ17O anomalies compared to their tropospheric counterparts. Stratospheric water is mostly photochemical in origin, suggesting stratospheric water may have a mass independent composition [Franz and Röckmann, 2005]. A stratospheric ozone anomaly (Δ17O) has been measured up to 40‰ [Krankowsky et al., 2000, 2007; Lämmerzahl et al., 2002; Mauersberger et al., 2001] while the isotopic composition of OH in the stratosphere has been estimated to be as large as 30‰–40‰ [Lyons, 2001; Zahn et al., 2006]. Because H2O2 is produced from HO2 radicals, which are formed through the reaction of OH and O3, stratospheric H2O2 is also expected to contain elevated Δ17O values as compared to tropospheric H2O2 [Morin et al., 2011]. These findings indicate that stratospheric sulfate should contain larger Δ17O anomalies than tropospheric sulfate. It may be anticipated that STE can affect boundary layer Δ17O chemistry in two ways; first, increased transport of stratospheric sulfate with high Δ17O values to the boundary layer, and second, increased oxidation of SO2 in the troposphere due to increased levels of ozone from a stratospheric source. Therefore, the oxygen anomaly is expected to correlate with 35S activity in stratosphere sulfate. Combining radioactive sulfur measurements and Δ17O in sulfate has the potential to enhance understanding of the extent of STE events into the troposphere and how these events affect the oxidation of sulfur oxides to sulfate aerosols.

[6] In this paper, we show the variations in 35S and Δ17O in sulfate aerosols collected during three Santa Ana wind events and three STE events and compare them to normal coastal climate samples in Southern California. STE events are not a common occurrence at low latitudes, but are sporadically seen during late winter and early spring. Santa Ana events occur in Southern California between September and March and are a result of a high pressure system building over the Great Basin in the southern Sierra Nevada Mountains simultaneously with a low pressure system offshore. The result is the creation of strong, dry winds from the east which lowers the relative humidity and increases the risk of forest fires [Guazzotti et al., 2001; Miller and Schlegel, 2006; Raphael, 2003; Westerling et al., 2004]. Even though Santa Ana events are not linked to STE, through the use of 35S measurements, Priyadarshi et al. [2012] showed that these events lead to an increase in the mixing of free tropospheric air into the boundary layer, and consequently may affect oxidation chemistry in the boundary layer.

2. Methods

[7] Samples were collected at the Scripps Pier in La Jolla, California (32.7°N, 117.2°W) using a standard high volume (Thermo-Electron), five-stage slotted impact collector with glass-fiber filter papers. This site is located on the Pacific Ocean, 10 miles north of downtown San Diego. Depending on meteorological patterns, this site collects onshore oceanic as well as offshore air masses with possible anthropogenic sulfate sources from the Los Angeles urban environment. From September to March, this site experiences occasional strong Santa Ana winds as air pressure builds over the Great Basin in Nevada leading to offshore high altitude air masses pouring into the Southern California basin.

[8] Samples were collected from February, 2009 to August, 2010 and previously measured for 35S activity in Priyadarshi et al. [2012] following the procedure of Brothers et al. [2010] with a background error of less than 5 atoms/m3. Fine particle mode samples (PM < 1.5 μm) were analyzed for oxygen isotopes to observe the potential relation between 35S and Δ17O during different climatic events in Southern California including Santa Ana and STE events. Santa Ana events were determined to occur when the relative humidity dropped below 30% and strong easterly winds prevailed. NOAA Hysplit back trajectory models (R. R. Draxler and G. D. Rolph, HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) model, Air Resources Laboratory, NOAA, Silver Spring, Maryland, 2011, available athttp://ready.arl.noaa.gov/HYSPLIT.php) confirm that air masses during these events originated from the Great Basin of Nevada before descending into Southern California. STE events were determined in Priyadarshi et al. [2012] where the 35S activity in the fine particle mode was higher than 2 sigma deviations from annual averages, at least 769 atoms/m3. Samples which did not experience a Santa Ana or STE event were also measured for background oxygen isotopic levels and are referred to as coastal climate samples. Oxygen isotope measurements were performed by separating sulfate through ion chromatography and converting to Ag2SO4. O2 gas was collected from pyrolysis of Ag2SO4 at a temperature of 1030°C and measured for δ18O and δ17O values on a MAT 251 mass-spectrometer. The value of the oxygen anomaly (Δ17O) was determined with an error of ±0.1‰ [Savarino et al., 2001].

[9] Oxygen isotopic compositions are corrected for sea salt content in each sample using magnesium as a sea salt tracer and assuming the molar ratio of SO42−:Mg2+ in seawater to be 0.5369 and sea salt sulfate to have a Δ17O of 0‰ and a δ18O of 10‰. Sodium is frequently used as a sea salt marker, but at some sampling sites, the use of sodium may be problematic due to the influence of non-marine sources of sodium, such as crustal material [Cainey et al., 1999; Martens et al., 1973; Sievering et al., 2004]. In our samples, the average ratio of Na+:Mg2+ and Cl:Mg2+ was 16.7 and 19.8 respectively, compared to the expected seawater ratios of 8.9 and 10.4, leading to overestimation of sea salt sulfate when using sodium as a sea salt proxy.

3. Results and Discussion

[10] The average sulfate concentration of bulk fine sulfate, reported in Table 1, is 21 nmol/m3, with a maximum of 63 nmol/m3, and nss sulfate concentrations averaging 20 nmol/m3, with a maximum of 62 nmol/m3. Specific activity averages 54 atoms/[nmol nss-sulfate] with a maximum of 195 atoms/[nmol nss-sulfate], whereas Δ17O anomalies in nss-sulfate range from 0.47‰ to 1.50‰ with an average value of 0.92‰.Figure 1 shows the relationship between specific activity and Δ17O. Santa Ana samples show enhanced levels of specific activity (138–195 atoms/nmol nss-sulfate) and Δ17O (1.42‰–1.50‰) relative to coastal climate samples. STE events, show specific activity and Δ17O levels which agree with coastal climate samples. These results suggest that Santa Ana events have a larger effect on boundary layer chemistry than STE events in Southern California. Δ17O shows a strong correlation with specific activity, but not with 35S activity (35S atoms/m3) likely because specific activity normalizes the radioactivity count to the sulfate concentration, thus accounting for non-radioactively produced sulfate such as those derived from anthropogenic fossil fuel production. As Δ17O is a measurement of total atmospheric sulfate, including anthropogenic and natural sources, it is expected that Δ17O would compare more favorably to specific activity than 35S activity. It is possible that specific activity better traces atmospheric chemical (oxidation) processes compared to 35S activity due to the inclusion of non-radioactive production processes.

Table 1. Summary of Sulfate Concentration, Specific Activity, and Δ17O in Fine Particles Collected at Scripps Piera
Sample IDStart DateEnd Date[SO4] (nmol/m3)NSS-[SO4] (nmol/m3)35S (atoms/m3)SA (35S atoms/nmol nss SO4)Uncorrected δ18O (‰)Uncorrected δ17O (‰)NSS-δ18O (‰)NSS-Δ17O (‰)
  • a

    Uncorrected δvalues refer to measurements before the sea salt correction is applied, whereas nss-δ values refer to measurement values after the sea salt correction.

Coastal Climate Samples
C-14/28/095/1/0962.5061.6366410.7711.326.4111.340.59
C-25/5/095/8/0945.6543.591754.018.975.158.920.59
C-35/15/095/19/0941.1640.682636.478.464.828.440.47
C-45/26/095/29/0943.6943.4464614.8710.495.9310.490.53
C-57/9/097/14/0914.2013.6239028.649.866.149.861.09
C-68/4/098/7/097.146.9532747.0617.239.7217.420.87
C-78/28/099/1/0918.1916.6554132.508.515.298.370.99
C-89/5/099/8/0914.1212.5545336.119.155.429.040.80
C-99/24/099/28/0917.7116.6036321.8611.036.6311.101.01
C-1010/8/0910/12/0913.4813.1375157.2011.176.7111.200.98
C-1110/16/0910/20/0911.2410.6133429.7014.108.1614.340.95
C-1212/1/0912/4/099.968.1953064.7112.507.2113.040.94
C-131/29/102/2/105.964.8535472.9412.517.2913.081.04
 
Santa Ana Samples
A-111/24/0912/1/093.271.8743913811.706.8812.981.50
A-212/18/0912/22/094.843.174711959.325.768.961.47
A-312/29/091/5/104.062.423361808.575.267.601.42
 
Stratospheric-Tropospheric Exchange Samples
E-13/20/093/24/0926.1225.4381932.2011.566.7111.600.78
E-25/22/095/26/0943.0242.3282019.389.895.699.890.61
E-37/6/097/9/0921.3920.8586841.6411.596.7711.630.82
Figure 1.

Specific activity plotted against Δ17O in the fine particle fraction (<1.5 μm). During Santa Ana events, enhanced specific activity shows strong correlation with increased Δ17O anomalies.

[11] The question of whether coarse mode sulfate follows the same patterns as fine remains unanswered. Due to the close proximity of the sampling site to the ocean, coarse samples consisted of sea salt sulfate compositions reaching excesses of 90%. This result magnifies the error in the sea salt correction immensely and can lead to a few per mil error in Δ17O making it difficult to determine if a relationship between specific activity and Δ17O exists in coarse mode sulfate [Patris et al., 2007].

4. Coastal Climate Samples

[12] Specific activity in coastal climate samples (referred to as samples C1–C13 in Table 1) range from 4 to 73 atoms/[nmol nss-sulfate] with an average value of 33 atoms/[nmol]. Δ17O ranges from 0.47‰ to 1.09‰ and agree relatively well with previous sulfate isotopic measurements taken at Scripps Pier [Dominguez et al., 2008; Lee and Thiemens, 2001]. Δ17O is negatively correlated with nss-sulfate, as seen inFigure 2. From analysis of Table 1, specific activity can also be seen to decrease with sulfate concentrations suggesting the influence of a non-radioactive, mass dependent sulfate component. Anthropogenic emissions contribute 70–80% of the total sulfate burden in the northern hemisphere [Rasch et al., 2000], potentially more in urban settings. NOAA Hysplit back trajectory models show the five most concentrated fine mode samples; C-1 to C-5; trace back to the Los Angeles region, which is known for high pollution levels [Su et al., 2009]. δ18O values do not correlate with Δ17O, suggesting multiple sources, not one distinct source. It's possible that the mixing line is composed of mass dependent sulfate sources including land based primary diesel sulfate (δ18O ≈ 5‰–7‰, Δ17O ≈ 0‰) [Lee et al., 2002] and primary sulfate from ship emissions (δ18O ≈ 20‰, Δ17O ≈ 0‰) [Dominguez et al., 2008]. On a global scale, primary sulfate emissions are small, however, proximal to these sources, combustion sulfate can locally contribute a significant portion of atmospheric sulfate [Holt et al., 1982]. Dominguez et al. [2008]calculated that ships can contribute up to 44% of the nss-sulfate found in fine particulate matter within Southern California. As vehicle and ship sulfate emissions are derived from fossil fuel, they possess a35S activity of zero, thus diluting specific activity along with Δ17O.

Figure 2.

The correlation between Δ17O and sulfate concentrations in the fine particle fraction (<1.5 μm). Santa Ana samples suggest mixing of a high altitude air source containing fewer aerosols into the boundary layer. The negative trend line seen in coastal climate samples suggests a mixing source of anthropogenic primary sulfate decreasing the Δ17O anomaly, yet increasing the sulfate concentration.

[13] Biomass burning (δ18O ≈ 6‰–11‰, Δ17O ≈ 0‰) [Lee et al., 2002] is another possible source of mass dependent sulfate. Southern California is highly susceptible to wildfires due to its dry climate with the wildfire season typically running from mid-May to early December. It is unknown how these biomass burning events affect radioactive sulfur activity in the atmosphere.Cho et al. [2011] has suggested a 35S turnover process in plants and soil of 66 days caused by the deposition of 35SO42− in soil and the release of H2S by plants, which subsequently oxidizes to SO4 in the atmosphere. It is plausible that biomass burning can lead to a release of 35S into the atmosphere. However, constraining the amount of 35S discharge from biomass burning during a distinct event is difficult. Dead biomass will not uptake 35S from the soil and thus is not suspected to release 35S upon burning. However, as the half-life of35S is 87 days, it is possible that a plant which has recently died might still possess trace amounts of 35S which will be released to the atmosphere upon burning. Therefore, the release of 35S is not constrained to just live plant matter. Samples (C-8 to C-10) collected during the Station Fire, which burned 160,577 acres from August 26th to October 16th, 2009 in Los Angeles County, making it the largest fire in Los Angeles County recorded history [National Climatic Data Center, 2009] (accessed 16 May 2012), do not show variation in specific activity relative to other coastal climate samples suggesting that biomass burning events have only minor effects on 35S radioactivity in the atmosphere. This assumption agrees with calculations from Priyadarshi et al. [2012], who estimated the contribution of this turnover process to account for only 2–5% of the measured 35S in the atmosphere. Therefore, the negative correlation seen in Figure 2 implies a heavy influence from fossil fuel emissions, not biomass burning.

5. Santa Ana Samples

[14] Santa Ana samples (A1–A3) possess low levels of 35S activity in the fine mode, ranging from 336 to 471 atoms/m3 with an average of 415 atoms/m3. Depositional rates are thought to increase on ocean surfaces and vegetation due to the prevailing strong gravity produced winds [Priyadarshi et al., 2012] and may account for the low observed 35S activity measurements. This is consistent with findings by Guazzotti et al. [2001], who observed a stark decrease in fine particle concentration at the onset of a Santa Ana event. It is interesting to note that although total sulfate concentrations dropped during Santa Ana events, sea salt sulfate concentrations stayed fairly constant relative to coastal climate and STE samples, resulting in an increase in the relative proportion of sea salt sulfate to the overall sulfate budget. Guazzotti et al. [2001] noticed a decrease in sea salt sulfate while sampling during Santa Ana events in Riverside, California. This variation could be due to the differences in sampling locations as data collected here was taken directly on the ocean, whereas Riverside, California is about 50 miles inland. The change from offshore winds to onshore winds typical of Santa Ana wind events probably has minimal effect on the sea salt concentrations seen here since the ocean was in such close proximity to the sampler.

[15] Specific activity measurements display large increases as compared to coastal climate samples, ranging from 139 to 195 atoms/[nmol]. Since the chemical properties of 35SO42− are thought to be essentially identical to bulk SO42−, specific activity is likely unaffected by increased deposition rates as it is assumed that 35SO42− deposits concomitantly with bulk sulfate. During Santa Ana events, Δ17O values vary from 1.42 to 1.50‰. The concurrent rise of specific activity and Δ17O suggests an increase in sulfate (high in Δ17O) which is oxidized at a higher altitude, thus implying the presence of high altitude air mixing with the boundary layer. A similar correlation between specific activity and Δ17O was observed by Lee and Thiemens [2001] at White Mountain Research Station, which is located at the remote site of Mount Barcroft (37.5°N, 118.2°S), northeast of Bishop, California and above the planetary boundary layer at an altitude of 3.8 km. Santa Ana events are thought to result in mid and upper tropospheric air mixing with the boundary layer, not stratospheric intrusions [Fosberg et al., 1966; Huang et al., 2009]. Also, Priyadarshi et al. [2012] estimated that up to 41% of total air mass sampled in the boundary layer during Santa Ana events in Southern California originated from the free troposphere. Therefore, the source of STE as a cause for the observed increases in specific activity and Δ17O is ruled out as the main source of mixing during Santa Ana events.

[16] Aqueous oxidation of sulfur dioxide via ozone is the predominant source of Δ17O anomalies in sulfate compared to oxidation by hydrogen peroxide (Δ17O O3 ≈ 27‰, Δ17O H2O2 ≈ 1.7‰) suggesting that the increase in Δ17O is a direct relation to an increase in ozone oxidation. However, at typical cloud pH levels of around 3–5 [Faloona et al., 2009], the rate of aqueous phase oxidation of S(IV) is dominated by H2O2 rather than ozone oxidation; 10−8 M/sec compared to 10−11 M/sec [Lee and Thiemens, 2001]. The ozone mixing ratio in the free troposphere would have to be much greater relative to the boundary layer to produce large changes in Δ17O sulfate anomalies. Increases in ozone oxidation of S(IV) should lead to larger δ18O values as δ18O of tropospheric ozone have been measured to be above 80‰ [Johnston and Thiemens, 1997]. A substantial increase in δ18O in Santa Ana samples relative to coastal climate samples is not observed, suggesting this is a minor process, and not the dominant mechanism driving the increase of Δ17O in free tropospheric sulfate.

[17] It is possible that the difference in observed Δ17O values arises from a dilution effect in the boundary layer. The boundary layer, in general, possesses larger concentrations of mass dependent sulfate due to anthropogenic emissions and ocean derived sulfate, which lower the Δ17O anomaly in bulk sulfate. This may account for the difference in Δ17O between free tropospheric sulfate and boundary layer sulfate rather than a change in aqueous oxidation pathways. Thermally convective transport of boundary layer air to the free troposphere is possible, however, it is expected that upward mixing is significantly decreased during Santa Ana events as large levels of high altitude air spill into the boundary layer [Priyadarshi et al., 2012]. As amplified anomalous concentrations of free tropospheric sulfate flow into the boundary layer, the resultant mixing increases the Δ17O of the bulk sample.

[18] Another mechanism to potentially account for the correlation between specific activity and Δ17O of sulfate is the transfer of free tropospheric SO2 to the boundary layer with subsequent oxidation to sulfate. To result in increased Δ17O anomalies, this mechanism requires a relative change in oxidative formation pathways of S(IV) in the boundary layer due to the prevailing Santa Ana winds. Increased wind speeds can boost mineral dust concentrations in the atmosphere, which can lead to an increase in the pH of the boundary layer. Guazzotti et al. [2001] found an increase in the concentration of mineral dust during Santa Ana events compared to normal weather conditions in Riverside, CA. Uptake of SO2 in the presence of mineral dust is thought to occur via aqueous phase oxidation of SO32− by O3 due to the increased alkalinity of mineral dust. Hydrogen peroxide oxidation is the primary oxidant at low pH values, but above a pH of ∼5.6, O3 oxidation dominates [Alexander et al., 2002]. However, nss-calcium concentrations in the fine particle mode remain consistent with coastal climate sample measurements. As nss-calcium is used as a tracer for mineral dust, this suggests that SO2 uptake via mineral dust is not a dominant mechanism in fine modal sulfate [Athanasopoulou et al., 2010; Virkkula et al., 2006]. The observations suggest the increase in Δ17O is a result of increased sulfate levels from the free troposphere and not a change in oxidative pathway of the boundary layer.

6. STE Samples

[19] Although STE samples (E1–E3) contain larger 35S activity values, ranging from 819 to 868 atoms/m3, the specific activity (average value of 31 atoms/[nmol]) of these samples are comparable to coastal climate samples. Deep STE intrusion events, where stratospheric air mixes directly into the boundary layer, typically occur at high altitudes, but are rarely detected at sea level. The STE event considered here is a shallow type of stratospheric-tropospheric mixing, where stratospheric air masses are mixed into the free troposphere and subsequently mixed into the boundary layer [Priyadarshi et al., 2012]. Although Priyadarshi et al. [2012] calculated 5–6% mixing of lower stratospheric air with the boundary layer, specific activity measurements suggest that this mixing had minimal compositional effect on the boundary layer. Δ17O measurements, ranging from 0.53‰ to 0.76‰, agree with this hypothesis as no increase is seen relative to coastal climate samples. Stratospheric mixing events are expected to lead to changes in the oxidation chemistry of the troposphere due to an increase in ozone levels, thus effecting the Δ17O measurements. Nevertheless, no spikes were observed in ozone concentrations relative to the average during the sampling periods, suggesting only weak interaction with the boundary layer. Even with admixture of anomalous stratospheric sulfate into the boundary layer, it is likely that the percentage of stratospheric sulfate to the total sulfate budget was negligibly small such that any large Δ17O signal from these particles was diluted by a predominant boundary layer sulfate component.

[20] We note that as cosmogenic nuclide production increases with latitude, there is a possibility that the large increases in 35S concentrations seen in STE samples arise due to a contribution of tropospheric air from high latitudes (>60°N) and is not necessary due to STE event. To quantify the likelihood of this scenario, the steady state model described in detail in Priyadarshi et al. [2011, 2012] was used to determine steady state concentrations for 35S at the 60°N latitude. 35SO42− concentrations were determined to be 716, 886, 1062, and 11925 atoms/m3 in the marine boundary layer, buffer layer, free troposphere, and lower stratosphere, respectively. Assuming the transport of air from higher latitudes is a result of free tropospheric air mixing into the boundary layer in La Jolla, CA., a simple mass balance can be used to estimate the percentage of free tropospheric air from 60°N necessary to replicate the measured 35SO42− values seen in the STE samples. It is found that the collected air mass must contain 53%, 53%, and 63% of free tropospheric air from 60°N for samples E1, E2, and E3, respectively, to match the measured 35SO42− values. High latitude air masses mix with air from lower latitudes as they travel south diluting the 35S concentration, thus this calculation can be considered a lower limit to the amount of high latitude air required. As stated earlier, Priyadarshi et al. [2012] estimated up to a 41% influx of upper altitude air mass from the free troposphere during Santa Ana events. This influx is aided by a large pressure gradient forcing upper altitude air into the boundary layer. Pressure gradients of this magnitude were not observed during the sampling periods that E1, E2, and E3 were taken suggesting that the increase in 35S is unlikely to occur due from free tropospheric air masses above 60°N latitude. Measurements of 7Be and 210Pb have been used to study air mass mixing between the upper altitudes and the boundary layer, and in the future, can be used in support of 35S measurements to determine the extent of stratospheric mixing [Graustein and Turekian, 1996]. To resolve the stratospheric source, sampling should also be done at locations where strong stratospheric intrusions are thought to periodically occur, such as near the poles. For example, Priyadarshi et al. [2011] measured 35S activities at Dome C, Antarctica and estimated up to a 10% mixing of stratospheric air mass with the boundary layer.

7. Conclusions

[21] Concurrent measurements of 35S and Δ17O of sulfate along the San Diego coastline are reported and reveal sulfate oxygen isotopic anomalies during Santa Ana wind events. It is observed that during Santa Ana events 35S specific activities of fine sulfate aerosols strongly correlates with the oxygen anomaly (Δ17O), indicating increased levels of mixing between the free troposphere and the boundary layer. From data presented here, it is apparent that free tropospheric sulfate may contain increased Δ17O anomalies relative to boundary layer sulfate. This results from increased levels of anthropogenic sulfate at ground level, diluting the Δ17O of boundary layer sulfate. In addition, the possibility of increased ozone oxidation of SO2in the free troposphere may produce a potential change in the oxidation chemistry of the free troposphere compared to the boundary layer. The combination of these tracers has the potential of studying variations in oxidative pathways arising from deep stratospheric intrusions, as well as tracing sulfur-oxygen chemical processes during transport.

Acknowledgments

[22] The authors thank S. Chakraborty, G. Dominguez, R. Shaheen, and T. Jackson for beneficial scientific discussions that significantly improved the manuscript.