Modeling the Distribution of OCS in the Lower Atmosphere of Venus

Transcription

1 Modeling the Distribution of OCS in the Lower Atmosphere of Venus Yuk L. Yung a,*, M. C. Liang b, X. Jiang c, C. Lee a, B. Bezard d and E. Marcq d a b Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, USA. Research Center for Environmental Changes, Academia Sinica, Taipei, Taiwan. c Science Division, Jet Propulsion Laboratory, California Institute of Technology, USA d LESIA, Observatoire de Paris, section de Meudon, Bât. 18, pièce 111, Meudon cedex, France * To whom all correspondence should be addressed. yly@gps.caltech.edu TO BE SUBMITTED TO JGR, JAN 31,

2 Abstract The chemical regimes in the atmosphere of Venus vary from photochemistry in the middle atmosphere to thermal equilibrium chemistry in the lower atmosphere and the surface. The primary chemical cycles are known but few details about these cycles have been fully verified by comparison between observations and modeling. Recent high quality data of OCS from ground-based and Venus Express observations provide a unique opportunity to test our understanding of chemistry and transport in the atmosphere of Venus. OCS is produced by heterogeneous reactions on the surface; the middle atmosphere is a net sink for OCS. Polysulfur appears to play a crucial role in the photosensitized dissociation of OCS. An innovative chemical scheme involving S n, where n varies from 1 to 8, is proposed. The spatial distribution of OCS in the middle atmosphere of Venus reflects a sensitive balance between chemistry and transport. Using our updated photochemical model and winds from Lee et al. s [2007] general circulation model, we explore the new chemistry in our two-dimensional chemistry-transport model. The modeling results and observations are in good agreement. 2

3 1. Introduction Sulfur chemistry is critical to the composition of the Venus atmosphere, and four sulfur species have been firmly identified: SO 2, SO, OCS, and H 2 SO 4 (vapor and in aerosols). Mills et al. [2007] has recently carried out an extensive review of chemistry in the atmosphere of Venus, and the reader is referred to this paper for a summary of previous results. As noted in this review, there are two parts to the chemistry of sulfur species in the atmosphere of Venus. On the surface of Venus (and possibly in the dense, hot lower atmosphere) the chemistry is dominated by thermodynamic equilibrium chemistry. Near and above the cloud tops, the chemistry is driven by solar UV radiation. Thus, the partitioning of sulfur among the different species represents a competition between thermodynamic equilibrium chemistry at the surface and in the lower atmosphere and photochemistry in the middle atmosphere. This underscores the importance of transport and mixing in determining the distribution of chemical species in the atmosphere of Venus. In this paper, we will focus on carbonyl sulfide (OCS), which is produced at the surface, transported to the middle atmosphere, where it is destroyed by photolysis and possibly by heterogeneous reactions. We will carry out a combined study of photochemistry and transport in a two-dimensional (2-D) chemistry-transport model (CTM) for the atmosphere of Venus. In section 2, we give an overview of the photochemistry and transport in the model. The main modeling results for OCS are discussed in Section 3, followed by concluding remarks in section Photochemistry and Transport Previous Photochemical Models Early work on sulfur chemistry on Venus by Prinn [1975, 1978, 1979] is based on a prediction by Lewis [1970] of sulfur species with mixing ratios of 60 ppmv for OCS, 6 3

4 ppmv for H 2 S, and 0.3 ppmv for SO 2. Prinn [1975] suggested a scheme of photochemical formation of sulfuric acid and polysulfur from carbonyl sulfide OCS. The primary sulfur carrier from the deep atmosphere and surface to the middle atmosphere is OCS. Near and above the cloud tops, OCS readily dissociates, releasing the S atom. OCS + hν CO + S( 1 D) (1) where S( 1 D) is the first electronically excited state of the S atom. The most likely fate of S( 1 D) is quenching: S( 1 D) + CO 2 CO 2 + S( 3 P) (2) If there is a source of oxygen atoms (i.e., photodissociation of CO 2 ) S undergoes oxidation to SO 2 (which subsequently forms SO 3 and H 2 SO 4 and condenses to form the cloud layers), as follows. The S atom gets oxidized to SO by reacting with O and O 2 S + O + M SO + M (3) S + O 2 SO + O (4) Further oxidation to SO 2 can proceed via the three-body reaction SO + O + M SO 2 + M (5) Catalytic oxidation by ClO is also possible SO + ClO SO 2 + Cl (6) 4

5 The rate for Reaction (6) has been measured in laboratory studies [Sander et al., 2006], and Reaction (6) accounts for % of the loss of SO at km altitude in recent photochemical models [Pernice et al., 2004; Mills and Allen, 2007]. Note that the net result is the oxidation of S to SO 2, and eventually to H 2 SO 4. In the absence of an oxygen source, S reacts with other S bearing species to form polysulfur, S x. S atoms can react to produce S 2 or S 2 may be produced via coupled S-Cl chemistry involving chlorosulfanes [Mills and Allen, 2007]. Production of S 3 is possible through successive addition reactions. S 3 is the chemical analog of ozone, known as thiozone. As the number of sulfur atoms increases, the polyatomic sulfur compounds tend to have lower saturation vapor pressures. It is convenient to name all sulfur species beyond S 3 polysulfur" or S x. The production of S x is part of what has been termed the slow atmospheric sulfur cycle [von Zahn et al., 1983], which is completed by decomposition reactions in the lower atmosphere. In the UV region S x absorbs strongly, and it may be the principal constituent of the unidentified UV absorber in the upper atmosphere of Venus [Toon et al., 1982]. Krasnopolsky and Pollack [1994] and Krasnopolsky [2007] presented detailed models of OCS and S x chemistry. Transport Model The general circulation model (GCM) used to calculate the stream function uses the dynamical core of the Hadley Centre General Circulation Model [Cullen, 1993] with a 5 degree horizontal grid staggered as an Arakawa B grid [Arakawa and Lamb, 1977] from pole to pole and a 31 level hybrid sigma-pressure vertical coordinate from surface to about 100 km. Forcing and dissipation are provided by linear parameterizations of the radiative forcing ('cooling to space') and boundary layer ('Rayleigh friction'). Two-dimensional module [Liang et al., 2005] of the photochemical model developed by the Caltech/JPL has been employed to simulate the meridonal distribution of OCS in the atmosphere of Venus. We extend the base model [Yung et al., 1982; Mills et al., 2007], which set the lower boundary at the cloud top, to the entire atmosphere from the surface 5

6 to 112 km. Above the cloud top, we assume the latitude-independent vertical eddy mixing coefficients equal the ones obtained previously [Yung et al., 1982]. Below the cloud top and above the planetary boundary (~2 km), the vertical eddy mixing coefficients are uniformly set at ~100 cm 2 s -1 and we assume that the transport in this region is dominated by the advection (see below). In the planetary boundary, the eddy coefficients are set at ~10 4 cm 2 s -1, a value high enough for the mixings/transport of matters between the troposphere and the boundary. The chemical scheme used in this paper is a subset of Mills et al. [2007]; 21 species (CO 2, CO, O 2, O 2 ( 1 Δ), H 2, HCl, SO 2, O, O( 1 D), O 3, H, OH, HO 2, H 2 O 2, Cl, ClO, S, SO, Cl 2, ClCO, OCS, S 2, ClCO 3 ) with about 200 chemical reactions were selected. The surface volume mixing ratios of CO 2, CO, HCl, SO 2, Cl 2, and OCS are fixed at 0.965, 10-5, , , , and , respectively. The dry deposition is allowed only for S x. Species other than the above and all species are, respectively, impermeable at the lower and upper boundaries. In the model used to derive the streamfunction shown in Fig. 1, a qualitatively realistic radiative forcing is used to drive an equatorial super-rotation and low wavenumber planetary waves in the middle atmosphere. No topography or diurnal cycle is used, and no active chemical processes are considered within the GCM. Full details of the GCM and this experiment are given in Lee et al. [2007]. Residual meridional circulation is calculated from wind and temperature from GCM [Lee et al., 2007]. The data from the GCM are in 5 5 latitude-longitude resolution. It has 31 vertical levels from 92 bar to 10-3 bar. For the quasi-geostrophic approximation, the residual velocity is defined by [Andrew et al., 1987] " 1 v = v " $!( $ v' # '/ z ) /! z (7) * 0 0 # 0 w = w +!( v' " '/" z ) /! y (8) * 0 where v and w are the meridional and vertical velocity, " o is the density, " is the potential temperature, and! 0z is the vertical derivative of the reference potential 6

7 temperature. The overbar and prime denote the zonal average and its deviation from zonal mean, respectively. A residual stream function for the transformed Eulerian-mean circulation can be determined by v * = "!# */!z, w * =!" */! y. (9) The resulting residual meridional circulation, shown in Fig. 1a, is like that for the Hadley cell in the terrestrial atmosphere. There is upwelling in the tropics, followed by downwelling in the high latitudes. Note that the latitudinal extension of the circulation is much greater than that for the terrestrial atmosphere due to the near absence of the Coriolis effect. The horizontal mixing coefficients, K yy, is calculated on the isentropic surfaces first using the Eq. (2) in Jiang et al. [2004]. Then we interpolate it from the isentropic surfaces to the pressure surfaces for use in the 2-D CTM. Fig. 1b shows a latitude-altitude plot of K yy. The values for K yy are larger in the mid-latitudes than the tropics due to the presence of enhanced wave activities in the mid-latitudes. To visualize the effect of the circulation on transport of chemical species in the atmosphere, we compute the age of air [Hall and Waugh, 1997]. The age of air is obtained by following the trend of an inert tracer, whose abundance at the surface increases linearly with time. Fig. 2 shows the age of air in the lower atmosphere of Venus derived from the circulation shown in Fig. 1. It takes about 10 (Earth) years for an air parcel released at the surface to reach the middle atmosphere at the cloud tops. Above the cloud tops, the age of air may not be correct, as the radiative forcing used to drive the GCM is not realistic. The time constants shown here are much shorter than those in previous studies [Krasnopolsky and Pollack, 1994; Krasnopolsky, 2007]. Their eddy diffusion coefficient below 30 km is K = 2x10 3 cm 2 s -1. Thus, for air to reach h = 30 km from the ground, the time it takes is ~ h 2 /K = 4.5x10 9 s, or 143 years, which is much longer than the 5-10 years given by Fig. 2. However, unlike what we derived here from a GCM, the basis for the empirical transport in these previous studies remains obscure. 7

8 3. Model Results and Implications Comparing Model to Observations Fig. 3(left) shows the vertical profiles of OCS from our model at various latitudes for a model with only loss by photolysis. Although the model correctly simulates the falloff of OCS above the cloud tops, the scale height of OCS at 30 km is close to the atmospheric scale height of ~10 km, in contradiction to the observed scale height of about 5 km. The reason is that the photolysis coefficient of OCS falls rapidly below the cloud tops. In order to produce the correct scale height of OCS at 30 km, we include an additional parameterized sink for OCS in the model. The results are shown in Fig. 3(right), which are in agreement with the observations. We will return to the physical mechanism for the additional sink later. The latitude-altitude plot for the concentration of OCS computed by our model is shown in Fig. 4. The distribution of OCS is consistent with its source at the surface, transport to middle atmosphere, where it is destroyed. The downward portion of the Hadley cell brings down air that is poorer in OCS, resulting in a meridional gradient that has been observed. Modeled OCS distribution at 33 km is presented in Fig. 5. The Venus Express VIRTIS (red) and ground-based telescope IRTF (black) are shown by special symbols. Dotted line represents the cosine function of latitude. The error bars are 1-σ standard deviation intervals, for both VIRTIS and IRTF retrievals. The variation of OCS at the probed altitude (33 km) was assumed to be caused by vertical translations of the reference profile given by Pollack et al. [1993]. A more detailed discussion of the considered set of OCS vertical profiles is available in Marcq et al. [2005]. Orbits #98, 110, 111, 134, 136, 258, 277 of the Venus Express spectra, have been used, mostly in science case #2 (off-pericenter observations), because of the need of long 8

9 integration times. This explains the N-S asymmetry in our latitudinal coverage. The spectra were provided by the high spectral resolution IR channel from the VIRTIS instrument (VIRTIS-H). The resolving power R in the used order of dispersion (order 5) is close to 1500 between 2.42 and 2.46 µm. The interpretation of these spectra is still in progress, so the retrievals in upcoming publications may differ significantly, although the main trends should persist at least qualitatively. On the other hand, the Earth-based retrievals have already been published by Marcq et al. [2005, 2006]. The used spectra were acquired on 2004/08/13 using the order 3 of the SpeX spectrometer (R ~ 2000 from 1.92 µm to 2.48 µm, which includes the whole 2.3 µm spectral window) at the NASA IRTF. Only the retrievals from slit positions #1 and #2, as defined in Fig. 1 from Marcq et al. [2006], are used. Innovative Chemical Cycle The falloff of the OCS profile around 30 km is a well-documented observation (see Marq et al and 2006 and references therein). The slope is dlog[ocs]/dlogp = 5 ± 1 (10) In order to account for the additional loss of OCS below the cloud tops (see Fig.3(right)), we propose an innovative photosensitized dissociation of OCS driven by photolysis of S n using photons at long wavelengths that are able to penetrate through the clouds 2 [ S 3 + hv S 2 + S ] (R3) S 4 + hv S 3 + S S 2 + S + M S 3 + M S 2 + S 2 + M S 4 + M (R4) (R8) (R9) 2 [ OCS + S CO + S 2 ] (R7) net 2 OCS 2 CO + S 2 (11) 9

10 The details of this innovative chemistry are summarized in Table 1 and shown in Fig.6. Note that the scheme has the following characteristics: (a) Photolysis at UV wavelengths releases S from OCS (b) S reacts with OCS to form S 2 and subsequent reactions produces S n (c) Photolysis of S n occurs at near UV and longer wavelengths, producing S (d) Cycle repeats via (b) The net result is summarized by (11). This scheme is known as photosensitized dissociation (for a discussion of this process see, e.g., section II (b) in Yung et al. 1984). The rate coefficient for the key reaction (R7) has recently been measured in the temperature range of K by Lu et al. [2006] and their expression is given in Table 1. At 500 K, k 7 = cm 3 s -1 is much smaller than the value cm 3 s -1 used by Prinn [1975] but is close to the value cm 3 s -1 estimated by Krasnopolsky [2007]. There are major gaps in our knowledge of the chemistry of S n ; new laboratory studies are needed to close the gaps. Our chemical scheme (17) for destroying OCS below the cloud tops is an alternative to that proposed by Krasnopolsky [2007]: H 2 SO 4 H 2 O + SO 3 (12) SO 3 + OCS CO + (SO) 2 (13) (SO) 2 + OCS CO + SO 2 + S 2 (14) net H 2 SO OCS H 2 O + CO 2 + CO + SO 2 + S 2 (15) A major weakness of scheme (15) is that the reactions (13) and (14) have no laboratory basis. Indeed, a reaction with so much rearrangement such as (14) is extremely unlikely (W. B. DeMore, private communication). Geochemical Implications 10

11 Our model which is consistent with the observations suggests that the global rate of destruction of OCS is ~3000 Tg-S/yr, a value that should be compared to the total volcanic source of ~10 Tg-S/yr for the Earth (see, e.g., Seinfeld and Pandis [1996]). The atmosphere is a net sink for OCS. We argue that the large implied source of OCS is unlikely to be supplied by volcanic emission. It is most likely that OCS is produced by heterogeneous reactions on the surface from CO and polysulfur (Sx), or between CO and CO 2 and surface minerals (e.g. pyrite). It remains a challenge to confirm and quantify these surface reactions (see discussion in the review of Venus chemistry by Mills et al. 2007). 4. Concluding Remarks A simple two-dimensional chemistry and transport (CTM) model is used to study the spatial distribution of OCS in the lower atmosphere of Venus. The residual circulation and horizontal eddy diffusivities are derived from winds from Lee et al s (2007) general circulation model. The GCM suggests rapid transport in the lower atmosphere of Venus. Mixing between the surface and the cloud tops occurs in as short as 10 years, which is significantly shorter than previous estimates of this time constant. The results for OCS from our 2-D CTM are compared to recent observations. There is generally good agreement between the model and data, but a new chemical scheme on photosensitized dissociation of OCS has to be postulated in order to explain the smaller scale height of OCS near 30 km. The sulfur chemistry in the lower atmosphere of Venus remains highly uncertain. New laboratory experiments and observations are urgently needed. 11

12 Acknowledgements We thank K. Baines, W. B. DeMore, P. Drossart, J. Moses and R. L. Shia for helpful discussions, D. Crisp for providing UV absorber profiles, and F. K. Li and X. Zhang for assistance in preparing the manuscript. This research was supported by NASA grant NNX07AI63G to the California Institute of Technology. MCL was supported by an NSC grant to Academia Sinica. 12

13 References Arakawa, A., Lamb, V. R., Computational design of the basic dynamical processes of the UCLA general circulation model. Methods in Computational Physics 17, Andrews, D. G., Holton, J. R., Leovy, C. B., Middle atmosphere dynamics. Academic, San Diego, California. Cullen, M. J. P., The Unified Forecast Climate Model. Meteo. Mag. 122, Hall, T. M., Waugh, D. W., Timescales for the stratospheric circulation derived from tracers. J. Geophys. Res. 102, Jiang, X., et al., Quasi-biennial oscillation and quasi-biennial oscillation-annual beat in the tropical total column ozone: A two-dimensional model simulation. J. Geophys. Res. 109, doi: /2003jd Krasnopolsky, V. A., Pollack, J. B., H2O-H2SO4 system in Venus clouds and OCS, CO, and H2SO4 profiles in Venus troposphere. Icarus 109, Krasnopolsky, V. A., Chemical kinetic model for the lower atmosphere of Venus. Icarus, doi: /j.icarus Liang, M. C., et al., Meridional transport in the stratosphere of Jupiter. Astrophysical Journal Letters 635, Lee, C., Lewis, S. R., Read, P. L., Superrotation in a Venus general circulation model. J. Geophys. Res. 112, doi: /2006je Lewis, J. S., Venus: atmospheric and lithospheric composition. Earth Planet. Sci. Lett. 10, Lu, C. W., Wu, Y. J., Lee, Y. P., Zhu, R. S., Lin, M. C., 2006, Experimental and theoretical investigation of rate coefficients of the reaction S(P-3)+OCS in the temperature range of K, J. Chem. Phys. 125 (16): Art. No Marcq, E., Bezard, B., Encrenaz, T., Birlan, M., Latitudinal variations of CO and OCS in the lower atmosphere of Venus from near-infrared nightside spectro-imaging. Icarus 179,

14 Marcq, E., Encrenaz, T., Bezard, B., Birlan, M., Remote sensing of Venus The lower atmosphere from ground-based IR spectroscopy: Latitudinal and vertical distribution of minor species. Plan. Space Sci. 54, Mills, F. P., Allen, M., A review of selected issues concerning the chemistry in Venus middle atmosphere. Plan. Space Sci. doi: /j.pss Mills, F. P., L.W. Esposito, Y. L. Yung, Atmospheric composition, chemistry, and clouds, in Exploring Venus as a Terrestrial Planet, eds. Esposito, L.W., E. Stofan, T. Cravens, American Geophysical Union, Washington, DC. Pp Mills, F. P., I. Observations and Photochemical Modeling of the Venus Middle Atmosphere. II. Thermal Infrared Spectroscopy of Europa and Callisto. Ph.D. thesis, California Institute of Technology, Pasadena, CA. Molina, L.T., Lamb, J. J., Molina, M. J., 1981, Temperature-Dependent UV Absorption Cross-Sections for Carbonyl Sulfide, Geophys. Res. Lett. 9, Moses, J.I., Zolotov, M.Y., Fegley, B., Photochemistry of A Volcanically Driven Atmosphere on Io: Sulfur And Oxygen Species From A Pele-Type Eruption, ICARUS 156 (1): Pernice, H., et al., Laboratory evidence for a key intermediate in the Venus atmosphere: Peroxychloroformyl radical. Proc. Natl. Acad. Sci. U. S. A. 101, Pollack, J. B., Dalton, J. B., Grinspoon, D., Wattson, R. B., Freedman, R., Crisp, D., Allen, D. A., Bezard, B., Debergh, C., Giver, L. P., Ma, Q., Tipping, R. Nearinfrared light from Venus' nightside A spectroscopic analysis. Icarus 103,1-42. Prinn, R. G., Venus: chemical and dynamical processes in the stratosphere and mesosphere. J. Atmos. Sci. 32, Prinn, R. G., Venus: chemistry of the lower atmosphere prior to the Pioneer Venus mission. Geophys. Res. Lett. 5, Prinn, R. G., On the possible roles of gaseous sulfur and sulfanes in the atmosphere of Venus. Geophys. Res. Lett. 6, Sander, S. P., et al., 2006, Chemical Kinetics and Photochemical Data for Use in Stratospheric Modeling Evaluation Number 15. Jet Propul. Lab. Pub. 06-2, Pasadena, Calif. ( 14

15 Seinfeld J. H. and Pandis S. N., Atmospheric Chemistry and Physics: From Air Pollution to Global Change, John Wiley and Sons, New York, Toon, O. B., R. P. Turco, J. B. Pollack (1982), The ultraviolet absorber on Venus: amorphous sulfur, Icarus, 51, 358. Von Zahn, U., Kumar, S., Niemann, H., Prinn, R. G., Composition of the Venus Atmosphere. in Venus, edited by D.M. Hunten, L. Colin, T.M. Donahue and V.I. Moroz. Univ. of Arizona Press, Tucson. pp Yung, Y. L., Allen, M. Pinto, J. P., Photochemistry of the Atmosphere of Titan - Comparison Between Model and Observations, Astrophys. J. Suppl. S. 55 (3): Yung, Y. L., Demore, W. B., Photochemistry of the Stratosphere of Venus - Implications for Atmospheric Evolution. Icarus 51,

16 Table 1. List of reactions related to the chemistry of S n as shown in Figure 6. The photolytic coefficients are given for a diurnally averaged model at 30 N at the top of the atmosphere (s -1 ). Two-body and three-body rate coefficients are given in units of cm 3 s -1 and cm 6 s -1, respectively. Reaction Rate Coefficient Reference R1 OCS + hv CO + S J 1 = (a) R2 S 2 + hv S + S J 2 = (b) R3 S 3 + hv S 2 + S J 3 = 1.15 (b) R4 S 4 + hv S 2 + S 2 J 4 = (b,c) R5 S 4 + hv S 3 + S J 5 = (b,c) R7 OCS + S CO + S 2 k 7 = T 2.57 e -1180/T (d) R8 S 2 + S + M S 3 + M k 8 (0) = (e) k 8 ( ) = R9 S 2 + S 2 + M S 4 + M k 9 (0) = (e) k 9 ( ) = R10 S 4 + S 4 + M S 8 + M k 10 (0) = (e) k 10 ( ) = R11 S 3 + S S 2 + S 2 k 11 = = (e) R12 S 4 + S S 3 + S 2 k 11 = = (e) R13 CO + S + M OCS + M k 13 (0) = e -1940/T (e,f) k 8 ( ) = (a) Molina et al. [1981] (b) Moses et al. [2002] (c) Branching ratio estimated (d) Lu et al. [2006] (e) Mills [1998] (f) Estimated based on analogy with O + CO + M CO 2 + M 16

17 Figure Captions Figure 1: Stream function ψ (top) and horizontal eddy mixing coefficient K yy (bottom) calculated from Lee et al. (2007). Units are 10 6 cm 2 s -1 for ψ and 10 9 cm 2 s -1 for K yy. Minimal K yy is set to be cm 2 s -1. Figure 2: Age of air (Earth year) derived from the circulation shown in Figure 1. Figure 3: Vertical profiles of OCS of global mean (solid curves) and that at 65 N (dashed curves) in our 2-D model. Left: Loss by photolysis of OCS alone. Right: Additional sink is included. See text. Figure 4: Modeled 2-D profile of OCS (in ppmv). Additional sink is included. See text. Figure 5: Modeled OCS distribution at 33 km. The Venus Express (red) and groundbased telescope IRTF (black) are shown by symbols. Dotted line represents the cosine function of latitude. Figure 6: Schematic diagram illustrating the S n chemistry in the lower atmosphere of Venus. 17

18 Figure 1: 18

19 Figure 2: 19

20 Figure 3 20

21 Figure 4: 21

22 Figure 5: 22

23 Figure 6: S S 2 hv S 3 S 4 S hv CO OCS S 2 hv S hv hv S 3 hv S 4 S 4 S 8 S 23