Massive volcanic SO2 oxidation and sulphate aerosol deposition in ...

Vol 465 | 17 June 2010 | doi:10.1038/nature09100

LETTERS Massive volcanic SO2 oxidation and sulphate aerosol deposition in Cenozoic North America Huiming Bao1, Shaocai Yu2 & Daniel Q. Tong3

Volcanic eruptions release a large amount of sulphur dioxide (SO2) into the atmosphere1,2. SO2 is oxidized to sulphate and can subsequently form sulphate aerosol3, which can affect the Earth’s radiation balance, biologic productivity and high-altitude ozone concentrations, as is evident from recent volcanic eruptions4. SO2 oxidation can occur via several different pathways that depend on its flux and the atmospheric conditions3. An investigation into how SO2 is oxidized to sulphate—the oxidation product preserved in the rock record—can therefore shed light on past volcanic eruptions and atmospheric conditions. Here we use sulphur and triple oxygen isotope measurements of atmospheric sulphate extracted from tuffaceous deposits to investigate the specific oxidation pathways from which the sulphate was formed. We find that seven eruption-related sulphate aerosol deposition events have occurred during the mid-Cenozoic era (34 to 7 million years ago) in the northern High Plains, North America. Two extensively sampled ash beds display a similar sulphate mixing pattern that has two distinct atmospheric secondary sulphates. A three-dimensional atmospheric sulphur chemistry and transport model study reveals that the observed, isotopically discrete sulphates in sediments can be produced only in initially alkaline cloudwater that favours an ozone-dominated SO2 oxidation pathway in the troposphere. Our finding suggests that, in contrast to the weakly acidic conditions today5, cloudwater in the northern High Plains may frequently have been alkaline during the mid-Cenozoic era. We propose that atmospheric secondary sulphate preserved in continental deposits represents an unexploited geological archive for atmospheric SO2 oxidation chemistry linked to volcanism and atmospheric conditions in the past. A close temporal correlation between large igneous provinces and major mass extinction events over the last 300 million years implies a potentially causal relationship between the two6. There has been much speculation about the large amount of gases, including CO2, SO2 and halogens, that large igneous provinces could have released into the atmosphere and their secondary effects2. Despite the immense mass of CO2 released by large igneous provinces, the added CO2 is negligible compared to the large atmospheric CO2 pool2. The answer, therefore, might be in the SO2 flux and its subsequent oxidation, which will affect the thickness and residence time of the sulphate aerosol layer and also the overall oxidizing capacity of the atmosphere2,7. Although the stratospheric sulphate aerosol might have a greater impact on global climate on a longer time scale, volcanic SO2 in the troposphere can have much more acute local and regional impacts on environments and society, as recorded by the 1783 Laki basaltic eruption and the subsequent ‘‘dry fogs’’ in continental Europe8. Past eruption events offer us a window into exploring the role volcanic SO2 has played in impacting climate and environment.

Most information, such as effects on precipitation and temperature are, unfortunately, not directly recorded for events in the distant past. However, sulphate, the oxidation product, can be preserved in geological records as minerals or in ice cores. Because SO2 oxidation can take many different pathways that depend on the nature of eruption and initial atmospheric conditions, an investigation into how SO2 was oxidized to sulphate and subsequently preserved in rock records can shed light on past volcanic eruptions and the atmospheric conditions. Unfortunately, direct observational studies on chemical evolution of SO2 in volcanic plumes are rare even for modern active eruptions (see refs 9 and 10 for an exception). Unique insights into current and past atmospheric processes can often be obtained from multiple stable isotope studies of atmospheric species (for example, NO3– in ref. 11). The oxygen in sulphate is derived from water, O2 and/or O3 (and its associated compounds, such as H2O2). Triple oxygen isotope and sulphur isotope measurements of atmospheric sulphate can reveal specific oxidation pathways from which the sulphate is formed, in this case, from volcanic SO2. Highly positive 17O anomalies (D17O value up to 15.84%) were identified12,13 in sulphate extracted from an Oligocene (,28 million years (Myr) ago) ash bed (the mid-Gering ash) near the base of the Arikaree Group, western Nebraska, USA. It was proposed that a massive amount of SO2 was oxidized to sulphate in the lower troposphere, which was deposited and quickly cemented in ash beds. An extreme sulphate aerosol or dry-fog event associated with a volcanic eruption to the west (in Colorado), more severe than the 1783 Laki one, was envisaged13,14. D17O-positive sulphate has also been identified in more recent volcanic ash beds, but with much smaller 17O anomalies15. A stratospheric SO2 1 ?OH oxidation may also produce sulphate with highly positive 17O anomalies16, as observed from volcanic sulphate preserved in recent ice core records17. The large quantity of gypsum in the Gering ash bed suggests, however, that the deposited sulphate was mostly formed within the planetary boundary layer or lower troposphere, because stratospheric sulphate would be dispersed all over the globe and would thus be unlikely to contribute to the concentration level seen in the ash bed. The D17O-highly-positive mid-Gering sulphate has been a singular case in geologic records and its origin a mystery. We conducted an extensive survey of the Cenozoic continental deposits in the northern High Plains, including northwestern Nebraska, western South Dakota, and eastern Wyoming (Fig. 1). In addition to the mid-Gering ash, we have found two other ash beds that have sulphate D17O values higher than 15.0% in the Arikaree group section exposed in Scotts Bluff National Monument. A total of seven ash-rich, well-cemented tuffaceous beds, five in the Oligocene between 34 and 22 Myr ago and two in the Miocene Ogallala Group between 7 and 13 Myr ago, have been found to have a sulphate D17O value higher than 11.4% (Fig. 1 and Supplementary Information).

1 Department of Geology and Geophysics, Louisiana State University, Baton Rouge, Louisiana 70803-4101, USA. 2Atmospheric Modeling and Analysis Division (E243-03), National Exposure Research Laboratory, United States Environmental Protection Agency, Research Triangle Park, North Carolina 27711, USA. 3Air Resources Laboratory, National Oceanic and Atmospheric Administration, Silver Springs, Maryland 20910, USA.

909 ©2010 Macmillan Publishers Limited. All rights reserved

LETTERS

NATURE | Vol 465 | 17 June 2010

13–7 Myr 13–7 Myr

A potential eruption centre 28 Myr ago

27–22 Myr 27–22 Myr 28 Myr 46º N 30–28 Myr South Dakota 44º N Wyoming 34 Myr Pete Smith Hill Wildcat Ridge

42º N

Nebraska 40º N Colorado 38º N 108º W

104º W

100º W

96º W

Figure 1 | Geological context of the samples. Location of the northern High Plains, North America and a composite Eocene–Oligocene–Miocene stratigraphy with seven 17O-anomalous (D17O . 11.4%) sulphate deposition events recorded in tuffaceous beds. Solid bars indicate the relative positions of these ash beds. See the Supplementary Information for

sources of the assigned numerical ages. Fm, Formation; CPF, Chamberlain Pass Formation; PPM, Peanut Peak Member; BCCM, Big Cottonwood Creek Member; MC-H Fm, Monroe Creek-Harrison Formation. The two events in bold black, at 28 and 34 Myr ago, are the focus of this study.

Two distinct ash beds, the ‘J Ash’ bed (,34 Myr ago) in the White River Group and the mid-Gering ash bed (,28.0 Myr ago) in the Arikaree Group, have been extensively sampled from different outcrops in the field and analysed for d34S, d18O and D17O values in the laboratory. The D17O2d18O2d34S plots for the mid-Gering and the J ash beds reveal extremely heterogeneous data for sulphates in a single ash bed; yet all data seem to fall within a triangle defined by three sulphate endmembers. We treated the data with a self-consistent weighted least-squares method18, assuming a conservative mixing and a sum of one for all endmembers (Supplementary Information). We obtained for both ash beds a three-endmember mixing scenario as the best fit. Note that there are two atmospheric secondary sulphate endmembers in each ash bed (Fig. 2, Supplementary Table 4). Endmember 1 (E1) is a basket for sulphates with no 17O anomaly, including sulphate formed via tropospheric OH oxidation or metalcatalysed O2 oxidation, and/or from subsequent oxidation of directly deposited SO2 on surface. Also included is later groundwater sulphate introduced during cementation or calcification. Further apportionment of these sources is difficult within E1, but a dominance by groundwater sulphate is consistent with its higher d34S and d18O values than those of the two atmospheric secondary sulphate endmembers (see below). Endmember 2 (E2) is atmospheric secondary sulphate from the oxidation of SO2 via one set of oxidants, less dominated by the O3 pathway. H2O2 and O3 are the main oxidants known to have positive 17 O anomalies in the troposphere16,19. The H2O2 has D17O values ranging from about 11% to 12% (ref. 20), while D17O for O3 ranges from 120 to 135% (ref. 21). Thus, a lower D17O value for sulphate suggests less involvement of the O3 pathway during SO2 oxidation in the atmosphere. Interestingly, both ash beds share an E2 with almost the same D17O2d18O2d34S values (Fig. 2, Supplementary Table 4). End-member 3 (E3) is atmospheric secondary sulphate from the oxidation of SO2 via another set of oxidants, with a higher contribution from the O3 pathway than for E2. The highly positive D17O value for E3

suggests that O3 is the dominant oxidant. The E3 in the two ash beds differs mainly in the magnitude of the D17O. At no other places—not in any modern atmospheric sulphate22,23 or in other ancient volcanic ashes15,24 or in Antarctic soils or snowpack25,26—have we seen the D17 OSO4 value reach as high as 16%. Both E2 and E3 have d34S values close to that of magmatic sulphur (,0%), which is consistent with an origin via direct oxidation of volcanic SO2. E2 (less O3-dominated) has a lower d18O value (about 12%) than that of E3 (O3-dominated) (about 111%), which is consistent with a high d18O value for O3 (from about 195 to 1125%) (the d18O value for H2O2 ranging from 121.9 to 152.5%)20,21. Since the late Eocene (,40 Myr ago), volcaniclastic or tuffaceous deposits have replaced Palaeocene–Eocene palaeosol-rich deposits and become an important component of the continental deposits in the northern High Plains, North America. This shift in sedimentation coincides with the onset of many silicic volcanic centres developed in the western USA during the mid-Tertiary as part of the so-called ‘‘ignimbrite flare-ups’’27. Our data show that the late Eocene and the early Oligocene had more frequent events of sulphate aerosol of highly positive D17O values than any other time periods of the Cenozoic in the northern High Plains. It is likely that the two Ogallala ash beds have a different volcanic centre (Snake River) than that of the White River and the Arikaree ones (Southern Rockies). In the troposphere, aqueous H2O2 and O3 oxidation of SO2 can generate sulphate with D17O values as high as 11.2% and 18.0%, respectively28. But these two pathways have to compete with other oxidation paths that produce sulphate with D17O < 0. If we were able to sample atmospheric sulphate in infinitely small space-time windows, we should always be able to pick up the two endmembers in any SO2 emission event or even in the background. However, an ash bed is a medium in which atmospheric sulphate was deposited and preserved over a period of time and our sampling and laboratory procedure further mix sulphates of different origins. That our integrated sulphate samples still display two D17O-positive sulphate endmembers means that there were time windows in which one O3-dominated and one


LETTERS


b

6 E3

5

10

3

2

2

1 E2 0 0 2

1

E1 4

6 8 δ34S

5 E2 0

2

E2 4

0

6 8 10 12 E1 δ18O

e 3.0

E3

2.5

0

10 12

δ18O

Δ17O

Δ17O

3

f

E3

2.5

1.5

Δ17O

E2

1.0 0.0

E1

0

2

4

6 8 δ34S

10 12

4

6 8 δ34S

10 12

20

E3

1.5 E2

1.0

10 5

0.0 –0.5

2

E1

0.5

0.5

0

15

2.0

2.0

E1

E3

4

d 3.0

Δ17O

c 15

5

4

–0.5

E3

6

δ18O

a

E2

E1 0

4

8 δ18O

12

16

0

0

2

4

6 8 δ34S

10 12

Figure 2 | Three sulphate endmember mixings. Relationships between three sulphate stable isotope parameters—D17O and d18O (both in % VSMOW) and d34S (in % VCDT)—for sulphates extracted from the midGering ash bed in the Wildcat Ridge area, Nebraska (a–c) and the J ash bed at Pete Smith Hill, Nebraska (d–f). (In each of panels d–f, there is also one datum point from Alcova, Wyoming.) There are two sets of sulphates: small

blue diamonds for water extraction and large red squares for HCl extraction. Error bars (s.d.) are equal to or smaller than symbols. Superimposed triangles represent the endmember sulphate positions E1, E2 and E3 for the two ash beds, respectively (Supplementary Table 4). VSMOW, Vienna Standard Mean Ocean Water; VCDT, Vienna Canyon Diablo Troilite.

less-O3-dominated pathway produced large fluxes of sulphate, large and temporally separate enough to exert discrete imprints in rock records. To examine this possibility, especially the existence of a time window in which a large quantity of sulphate can be produced by a O3dominated pathway as seen in the mid-Gering case, we ran a threedimensional atmospheric sulphur chemical-transport model (US EPA Models-3/Community Multiscale Air Quality)29 that uses existing meteorological fields in western Nebraska and a pre-industrial background sulphur budget for a case of volcanic SO2 emission from northcentral Colorado. The model has the advantage of tracking sulphate chemical productions from the gas-phase ?OH oxidation and five aqueous-phase chemical reactions including aqueous O3 and H2O2 pathways that produce D17O-positive sulphate (Supplementary Information). In this model, we are looking at sulphate deposition during a 13-day period in western Nebraska as the result of a continuous SO2 emission in the troposphere. We have run model simulations with various combinations of variables, including SO2 emission rate (5,000 to 5 million tons per day), emission type (point source or dispersed over a column), meteorology (summer and winter), dissolved [Fe21] and [Mn21] content (for example, 10210, 1025 and 1022 mg m23), initial cloudwater pH (4, 6, 7, 8 and 9 set by background [Ca21]), and stratospheric addition of O3 via tropopause folding. We calculated total sulphate D17O assuming that sulphate derived from the H2O2, O3 and the rest pathway has D17O values of 11.2%, 18.0% and 0.0%, respectively (Supplementary Information). Model results show that, if other variables are fixed, higher emission rates usually result in a lower bulk sulphate D17O value. Reduced [Fe21] and [Mn21] or stratospheric input of O3 via tropopause folding helps to increase the total sulphate D17O value, but only slightly under acidic cloudwater conditions. Winter, a season with lower overall oxidizing capacity, produces far less sulphate than summer (see Supplementary Information). The model consistently demonstrates that many combinations of variables can convert a large quantity of volcanic SO2 to sulphate that has a bulk D17O value ranging from 10.5 to 12.0%. Thus, it is not difficult to produce the observed sulphate E2 in the two extensively sampled ash beds (the mid-Gering and the J ash) or the D17O-positive sulphate observed in three of the seven ash beds in Cenozoic North America (Supplementary Table 1). The problem occurs for E3 or the D17O-highly-positive

sulphate in many ash beds. We found that the only scenarios from which the model can produce a large flux of atmospheric secondary sulphate with a highly positive D17O are the ones with alkaline background cloudwater pH (Fig. 3). We note that upon the injection of massive SO2 and its subsequent oxidation to H2SO4, the cloudwater pH will inevitably decrease over time. It is the initial background cloudwater pH that exerts a critical role. This is because the rate of aqueous O3 pathway overwhelms those of other competing paths only at higher pH (ref. 3). High cloudwater pH in combination with an overall low oxidizing capacity (such as the winter meteorological condition) is shown to be the ideal scenario to produce sulphate that is (1) high in flux, (2) highly positive in D17O, and (3) discrete in time (Fig. 3). These are the three conditions required to explain the observed two atmospheric secondary sulphate endmembers in the Wildcat Ridge (,28 Myr ago) or the Pete Smith Hill (,34 Myr ago) ash beds. Alkaline conditions are rare in the modern atmosphere, except for places in pristine condition or rich in alkaline dust (see, for example, ref. 30). Modern rainwater pH is mostly below 7 in western Nebraska (Supplementary Information), which could not produce the observed sulphate deposition pattern, as shown by our modelling results. Our finding implies that the cloudwater pH in the northern High Plains of North America was alkaline at the onset of at least two major explosive volcanic eruptions in the mid-Cenozoic (,28 Myr ago and ,34 Myr ago). This further suggests that, in contrast to its modern weak acidity, this region may have had frequent alkalinecloudwater episodes in the mid-Cenozoic era. The different D17O values for E3 in the two ash beds (that is, 12.9% versus 16.0%) also suggest that the alkalinity itself varied. A frequent alkaline cloudwater condition in the past, therefore, provides a new reference for gauging anthropogenic impacts on the atmosphere. An alternative, yet more speculative scenario is that the roof rocks (for example, carbonates and calcareous shales) of the volcanoes may have helped to increase the cloudwater pH value, through physical dispersion of Ca-rich dust particles, some of which may possibly be CaO formed by thermal decomposition of CaCO3. In recent years, new satellite observational tools have just begun to look into the physical and chemical evolution of volcanic plumes in the atmosphere. The observational data and modelling results presented here demonstrate that extreme sulphate aerosol deposition can occur at a great 911

©2010 Macmillan Publishers Limited. All rights reserved

LETTERS


Winter

10,000

8 250,000

8,000 pH = 6

4,000 2,000 Sulphate flux (µg m–2 h–1)

Summer

pH = 7

4

4,000

0 0 8 300,000

0 30,000 pH = 8

6

20,000

0 0

2 100

200

300

0 8

15.

6

16.

4 2 0 8

pH = 8

17.

18. 19.

6

200,000

4 10,000

14.

2

pH = 7 6 200,000 150,000 4 100,000 2 50,000

8,000

13.

6

0 0 8 250,000

0

12.

8

pH = 6 6 200,000 150,000 4 100,000 2 50,000

6,000

12,000

11.

Δ17O total

Δ17O (%o)

Total

4

20.

100,000

0

0 0

2 100

200

300

0

21.

Time (h)

Figure 3 | Modelling results on sulphate fluxes and D17O values. Total column (0 to 2 km above the surface) sulphate flux (solid black) and D17O value (dashed red) over time in western Nebraska, obtained from a threedimensional atmospheric sulphur chemical-transport model, assuming a constant SO2 emission rate of 5 million tons a day from an eruption in northcentral Colorado, background cloudwater of pH 5 6 (upper), 7 (middle), and 8 (lower), and [Fe21] and [Mn21] 5 1025 mg m23.

22. 23.

24. 25.

Morin, S. et al. Tracing the origin and fate of NOx in the Arctic atmosphere using stable isotopes in nitrate. Science 322, 730–732 (2008). Bao, H. M. et al. Anomalous 17O compositions in massive sulphate deposits on the Earth. Nature 406, 176–178 (2000). Bao, H. M., Thiemens, M. H., Loope, D. B. & Yuan, X. L. Sulfate oxygen-17 anomaly in an Oligocene ash bed in mid-North America: was it the dry fogs? Geophys. Res. Lett. 30, doi:10.1029/2003GL016869 (2003). Loope, D. B., Mason, J. A., Bao, H. M., Kettler, R. M. & Zanner, C. W. Deformation structures and an alteration zone linked to deposition of volcanogenic sulphate in an ancient playa (Oligocene of Nebraska, USA). Sedimentology 52, 123–139 (2005). Bindeman, I. N., Eiler, J. M., Wing, B. A. & Farquhar, J. Rare sulfur and triple oxygen isotope geochemistry of volcanogenic sulfate aerosols. Geochim. Cosmochim. Acta 71, 2326–2343 (2007). Lyons, J. R. Transfer of mass-independent fractionation in ozone to other oxygencontaining radicals in the atmosphere. Geophys. Res. Lett. 28, 3231–3234 (2001). Savarino, J., Bekki, S., Cole-Dai, J. & Thiemens, M. H. Evidence from sulfate mass independent oxygen isotopic compositions of dramatic changes in atmospheric oxidation following massive volcanic eruptions. J. Geophys. Res. Atmos. 108, doi: 10.1029/2003jd003737 (2003). Noda, T. & Shimada, K. Water mixing model calculation for evaluation of deep geothermal water. Geothermics 22, 165–180 (1993). Savarino, J., Lee, C. C. W. & Thiemens, M. H. Laboratory oxygen isotopic study of sulfur (IV) oxidation: origin of the mass-independent oxygen isotopic anomaly in atmospheric sulfates and sulfate mineral deposits on Earth. J. Geophys. Res. Atmos. 105, 29079–29088 (2000). Savarino, J. & Thiemens, M. H. Analytical procedure to determine both d18O and d17O of H2O2 in natural water and first measurements. Atmos. Environ. 33, 3683–3690 (1999). Johnston, J. C. & Thiemens, M. H. The isotopic composition of tropospheric ozone in three environments. J. Geophys. Res. Atmos. 102, 25395–25404 (1997). Jenkins, K. A. & Bao, H. M. Multiple oxygen and sulfur isotope compositions of atmospheric sulfate in BatonRouge,LA, USA. Atmos. Environ. 40, 4528–4537 (2006). Bao, H. M. Sulfate in modern playa settings and in ash beds in hyperarid deserts: implication on the origin of 17O-anomalous sulfate in an Oligocene ash bed. Chem. Geol. 214, 127–134 (2005). Howell, K. J. & Bao, H. M. Caliche as a geological repository for atmospheric sulfate. Geophys. Res. Lett. 33, doi: 10.1029/2006GL026518 (2006). Alexander, B., Savarino, J., Barkov, N. I., Delmas, R. J. & Thiemens, M. H. Climate driven changes in the oxidation pathways of atmospheric sulfur. Geophys. Res. Lett. 29, 1685, doi:10.1029/2002GL014879 (2002). Bao, H. M. & Marchant, D. R. Quantifying sulfate components and their variations in soils of the McMurdo Dry Valleys, Antarctica. J. Geophys. Res. Atmos. 111, doi:10.1029/2005JD006669 (2006). Armstrong, R. L. & Ward, P. L. Evolving geographic patterns of Cenozoic magmatism in the North American Cordillera; the temporal and spatial association of magmatism and metamorphic core complexes Mid-Tertiary Cordilleran magmatism; plate convergence versus intraplate processes. J. Geophys. Res. B 96, 13201–13224 (1991). Bao, H., Thiemens, M. H. & Heine, K. Oxygen-17 excesses of the Central Namib gypcretes: spatial distribution. Earth Planet. Sci. Lett. 192, 125–135 (2001). Eder, B. & Yu, S. C. A performance evaluation of the 2004 release of Models-3 CMAQ. Atmos. Environ. 40, 4811–4824, doi:10.1016/j.atmosenv.2005.08.045 (2006). Zhang, D. D. et al. Precipitation chemistry of Lhasa and other remote towns, Tibet. Atmos. Environ. 37, 231–240 (2003).

distance from an eruption centre. Understanding the geological causes, frequency and atmospheric chemistry of these unfamiliar volcanic hazards, therefore also has environmental and societal significance. In addition to volcanic eruptions, massive release of reduced sulphur gases can also be caused by bolide impacts or ocean overturning. Multiple isotope signatures of sulphate left behind by these dramatic geological events not only record the impacts on the atmosphere and surface conditions, but also present new questions or hypotheses to be answered or tested in future observational and modelling studies. Sedimentary deposits in arid to semi-arid continental sites offer a rarely explored archive for such an effort.

26.

Received 6 May 2009; accepted 13 April 2010.

Supplementary Information is linked to the online version of the paper at www.nature.com/nature.

1.

Wallace, P. J. Volcanic SO2 emissions and the abundance and distribution of exsolved gas in magma bodies. J. Volcanol. Geotherm. Res. 108, 85–106 (2001). 2. Self, S., Widdowson, M., Thordarson, T. & Jay, A. E. Volatile fluxes during flood basalt eruptions and potential effects on the global environment: a Deccan perspective. Earth Planet. Sci. Lett. 248, 518–532 (2006). 3. Seinfeld, J. H. & Pandis, S. N. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change Ch. 6 (Wiley, 1998). 4. Robock, A.Pinatubo eruption: the climatic aftermath. Science 295, 1242–1244 (2002). 5. National Atmospheric Deposition Program. 2008 Annual Summary. 15 Æhttp:// nadp.sws.uiuc.edu/lib/data/2008as.pdfæ (NADP Program Office, Illinois State Water Survey, 2009). 6. Wignall, P. The link between large igneous province eruptions and mass extinctions. Elements 1, 293–297 (2005). 7. Ward, P. L. Sulfur dioxide initiates global climate change in four ways. Thin Solid Films 517, 3188–3203 (2009). 8. Thordarson, T. & Self, S. Atmospheric and environmental effects of the 1783–1784 Laki eruption: a review and reassessment. J. Geophys. Res. Atmos. 108, doi: 10.1029/2001JD002042 (2003). 9. Mather, T. A. et al. Oxygen and sulfur isotopic composition of volcanic sulfate aerosol at the point of emission. J. Geophys. Res. Atmos. 111, D18205, doi: 10.1029/ 2005jd006584 (2006). 10. Rose, W. I. et al. Atmospheric chemistry of a 33–34 hour old volcanic cloud from Hekla volcano (Iceland): insights from direct sampling and the application of chemical box modeling. J. Geophys. Res. Atmos. 111, doi: 10.1029/2005JD006872 (2006).

27.

28. 29. 30.

Acknowledgements H.B. thanks D. Loope, H. LaGarry and J. Swinehart for guidance in the field, M. Khachaturyan, K. Jenkins, K. Howell, I. Kohl and A. J. Kaufman for technical assistance, B. Li for statistical treatment, and the National Park Service for sampling permission at Scotts Bluff National Monument (permit numbers SCBL-2000-SCI-000, BADL-2004-SCI-0012 and SCBL-2004-SCI-0005). Financial support was provided by the NSF (EAR-0408986 to H.B.). S.Y. thanks S. T. Rao, D. Mobley, K. Schere, R. Mathur, J. Pleim, J. Godowitch and S. Roselle for comments. The United States Environmental Protection Agency through its Office of Research and Development funded and managed the part of the research that is related to sulphur chemistry modelling. It has been subjected to the Agency’s administrative review and approved for publication. D.Q.T. is grateful to S. Fine, D. Byun and R. Artz for discussion and acknowledges constructive internal Air Resources Laboratory comments. Author Contributions H.B. designed the research, and did field and laboratory studies. S.Y. and D.Q.T. did the three-dimensional sulphur oxidation and transport modelling study. H.B. wrote the manuscript. All authors contributed to manuscript revisions. Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of this article at www.nature.com/nature. Correspondence and requests for materials should be addressed to H.B. ([email protected]) or S.Y. ([email protected]).
