Poster Content

AbstractDiagram of Venusian atmosphere + processes considered (nucleation, condensation, coagulation)Description of CARMA (including what schemes it's using for all the processes)Describe meteoric infallInclude plots of temperature, pressure, wind and eddy diffusion profiles, and mention their sourcesAlso include plots of gas, condensation nuclei flux vs. altitude. CARMA results, compared with data (separate plots for particle-height distribution, gas-height distribution, and particle-size distribution; the latter could be split into many plots for any height levels); also try wind vs. no wind

Plots of results with different gas and condensation nuclei fluxMention possibilities of ice crystals in the upper haze instead of liquid droplets, as we've assumed. References

Bimodal Distribution of H2SO4 Aerosols in the Upper Atmosphere of VenusPeter Gao1, Xi Zhang1, David Crisp2, Charles G. Bardeen3, and Yuk L. Yung1

1California Institute of Technology, Pasadena, CA, USA, 2Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA,3National Center for Atmospheric Research, Boulder, CO, USA

Email: pgao@caltech.edu

AbstractThe vertical and size distributions of Venus' atmospheric

aerosols – major factors in determining their effects on the planet's climate – were measured in situ by the Pioneer

Venus atmospheric Probes [1]. The sulfuric acid cloud deck, sandwiched between thinner haze layers, have been studied

using 1D atmospheric models (e.g. [2], [3], and [4]). Recently, the SPICAV/SOIR instruments on the Venus Express space-

craft were used to study the Upper Haze via solar occultation experiments. From the data, Wilquet et al. [5] was able to

retrieve a bi-modal size distribution in the haze aerosols. We propose that the larger mode results from vertical diffusion

of cloud aerosols from the cloud deck, while the smaller mode is generated from the nucleation of infalling meteoric dust.

We test this hypothesis by using CARMA 3.0 [6, 7], with which we simulate both the Upper haze and the cloud deck. Our re-

sults show good agreement between Pioneer and Venus Express data in both vertical and size distributions, though a vertical wind is necessary to loft the cloud aerosols to the observed

heights in the Upper Haze.

CARMA 3.0The Community Aerosol and Radiation Model for Atmo-spheres (CARMA) code was first written by Turco et al. [6], then subsequently updated by Bardeen et al. [7]. Version 3.0 is written in Fortran 90, and is used mainlyas a 1D microphysical model capable of vertical trans-port. It is currently designed to handle both water and sulfuric acid as condensates. It has been modified for

this project to include eddy diffusion and gas transport.

Gas/CN ProductionWe use Imamura and Hashi-moto's production profiles for sulfuric acid vapour andsulfur condensation nuclei[3] (which we take to be sul-fate, and assume similar pro-duction rates), as shown to the left. In their paper, they propose that H2SO4 is creat-ed by the photolysis of SO2 and subsequent reaction with water:

SO2 + hv SO + O SO + hv S + O

SO2 + O + M SO3 + M SO3 + H2O H2SO4

In our model runs, we “man-ually” add in the necessaryamount of gas and CN at thebeginning of every time step.

Meteoric InfallInfalling meteors ablate upon contact with the at-mosphere. The ablated material, usually in the form of metallic vapour, then condenses out form-ing tiny dust particles. We use a gaussian production profile (shown to the right) for these dust particles modeled after that of Kala-shnikova et al. [8]. The par-ticles considered here are 1.3 nm in size.

NucleationWe use CARMA 3.0's dropletactivation scheme for nuclea-ting CNs. Droplets activate when the supersaturation ex-ceeds the critical value, given by:

Where Akel and Bsol are the Kel-vin curvature and solute effectterms, respectively.

Future WorkSummary

Model SetupThe temperature and pressureprofiles for our model Venus at-mosphere is based on the

Wind = HashimotoEddy diffusion = Hashimoto +Venus bookInitial + boundary conditionsSpherical particles (solid CN, liquid droplets)

Results

References[1] Knollenberg R. G. and Hunten D. M. (1980) J. Geophys. Res., 85, 8039-8058. [2] Yamamoto M. and Takahashi M. (2006) J. Geophys. Res., 111, E08002, 12pp. [3] Imamura T. and Hashimoto G. L. (2001) J. Atmos. Sci., 58, 3597-3612. [4] James E. P. et al. (1997) Icarus, 129, 147-171. [5] Wilquet V. et al. (2009) J. Geophys. Res., 114, E00B42, 13pp. [6] Turco R. P. et al. (1979) J. Atmos. Sci., 36, 699-717.[7] Bardeen C. et al. (2011) The CARMA 3.0 microphysics package in CESM, Whole Atmosphere Working GroupMeeting.[8] Kalashnikova O. et al. (2000) Geophys. Res. Lett., 27, 3293-3296.[9] de Pater I. and Lissauer J. J. (2001) In: Planetary Sciences. Cambridge, UK: Cambridge University Press. [10] Toon O. B. et al. (1989) J. Geophys. Res., 94, 11359-11380.

40 km

50

60

70

80

90

100 km

Condensation& EvaporationCARMA 3.0 uses the growth al-gorithm given in [10], which calculates the change in mass mwith time t:

Where Psat is the saturation va-pour pressure, S is the supersat-uration, Aks is the combined Kel-vin and solute factors, and g0 and g1 are growth kernels.

Advection

Sedimentation

Vertical TransportSedimentation in CARMA 3.0 for Reynold's number Re

< 1 is given by Stoke's flow, augmented by the slip corr-ection factor. For Re > 1, Re is recalculated after consi-dering turbulence, then the fall velocity is calculated

from it. A vertical wind is added by adding the wind ve-locity to the fall velocity. The code has been modified

to include eddy diffusion, given by [9]:

Where vi is the velocity of species i, K is the eddy diffu-sion coefficient, N is the atmospheric number density,Ni is the number density of species i, and z is the verti-

cal scale.

Coagulationfff

Scrit= 4 Akel27Bsol

vi = −K NN i

∂ N i/N∂ z

dmdt

=Psatg0S1−Aks

1g0g1P sat

Bimodal Distribution of H2SO4 Aerosols in the Upper Atmosphere of VenusPeter Gao1, Xi Zhang1, David Crisp2, Charles G. Bardeen3, and Yuk L. Yung1

1California Institute of Technology, Pasadena, CA, USA, 2Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA,3National Center for Atmospheric Research, Boulder, CO, USA

Email: pgao@caltech.edu

AbstractThe vertical and size distributions of Venus' atmospheric aerosols – key factors in determining the planet's climate

– were measured in situ by the Pioneer Venus atmospheric Probes [1]. The sulfuric acid cloud deck, sandwiched between thinner haze layers, have been studied using 1D atmospheric

models (e.g. [2] and [3]). Recently, the SPICAV/SOIR instru-ments on the Venus Express spacecraft were used to study the upper haze via solar occultation experiments. From the data, Wilquet et al. [4] were able to retrieve a bi-modal size distri-

bution in the haze particles. We propose that the larger mode results from vertical transport of cloud aerosols from the cloud deck, while the smaller mode is generated from the

nucleation of infalling meteoric dust. We test this hypothesis by using CARMA 3.0 [5, 6], with which we simulate both the upper haze and the cloud deck. Our results show good agree-

ment between Pioneer height data and simulated height profile but there are discrepancies in the size distributions and

the upper haze height profile. A vertical wind also appears to be necessary to loft the cloud aerosols to the observed heights in the upper haze – diffusion alone is not enough.

CARMA 3.0The Community Aerosol and Radiation Model for Atmo-spheres (CARMA) code was first written by Turco et al. [5], then subsequently updated by Bardeen et al. [6]. Version 3.0 is written in Fortran 90, and is used mainlyas a 1D microphysical model capable of vertical trans-port. It is currently designed to handle both water and sulfuric acid as condensates. It has been modified for

this project to include eddy diffusion and gas transport.

Gas/CN ProductionWe use Imamura and Hashi-moto's production rate pro-files for sulfuric acid vapour and sulfur condensation nu-clei (CNs) [2] (which we take to be sulfate, and assume si-milar production rates), as shown to the right, since theyare a good match to past models and fit well to data. In their paper, they propose that H2SO4 is created by the photolysis of SO2 and subse-quent reaction with water:

SO2 + hv SO + O SO + hv S + O

SO2 + O + M SO3 + M SO3 + H2O H2SO4

Meteoric InfallInfalling meteors ablate upon contact with the atmosphere. The ablated material, usually in the form of metallic vapour, then condenses out forming tiny dust particles. We use a gaussian production profile (shown to the right) for these dust particles modeled after that of Kalashnikova et al. [7]. The particles considered here are 1.3 nm in size. Though theprofile in [7] is for Earth, we will assume a similar profile for Venus due to the similari-ties between the 2 planets.

NucleationWe use CARMA 3.0's dropletactivation scheme for nuclea-ting CNs. Droplets activate when the supersaturation ex-ceeds the critical value, given by:

Where Akel and Bsol are the Kel-vin curvature and solute effectterms, respectively.

Future WorkOur immediate work will focus on resolving the vertical trans-

port issues that may be resulting in the disagreements betweendata and model. Our long term plans are to apply CARMA to

the atmospheres of exoplanets in order to investigate the roleaerosols may play in them, and specifically how they affect

these planets' radiative transfer processes.

SummaryWe have modeled the aerosol distribution of the Venus clouds

and upper haze using CARMA 3.0 and compared our results with data. We have shown that the number density distribu-

tion with altitude of the clouds can be modeled by CARMA to a good accuracy, while in the upper haze the agreement with

data is less, possibly due to issues of vertical transport. We have also seen that two modes can be generated in the upper haze from meteoric dust nucleation and cloud particle upwelling,

though the two modes are difficult to separate from each other. Finally, we saw some agreement between data and the gas con-

centration, though issues with boundary conditions and SVP scheme may have prevented better agreement. In all, there is

clear order-of-magnitude agreement between most of the simulation and data, but more work needs to be done to tease

out the true nature of the disagreements.

Model SetupThe temperature and pressure pro-files for our model Venus atmosph-ere are based on those of [11]. The wind profile is based on those of [2], with extrapolation to above 70 km. The eddy diffusion profile below 70 km is also based on [2], while above 70 km we use a gaussian profile cen-tered at 100 km to approximate the observed profile as reported in [12]. Initial concentrations of H2SO4 gas, sulfate CNs, and meteoric dust CNs are 4 ppm, 40/cm3 from 40 to 70km(and 0 elsewhere), and 40/cm3 from 70 to 100km (and 0 elsewhere), res-pectively. For the bottom boundary we assume 40 sulfate CNs In the smallest bin, and 4ppm of H2SO4 gas. The upper boundary is kept open to allow for an upward flux.

ResultsFigures 1, 3, 5, and 6 are derived from a 6.41x106 s run that covered the 40-100 km range, while figures 2 and 4 are from a 107 s run

that covered the 40-70 km range. We see in figure 2 that the CARMA-derived number density vs. height distribution matches the data [1] well, reproducing the bump at 61 km due to CN and gas production, as well as the convective region between 50 and 55 km. The bad fit below 50 km is possibly due to horizontal winds, which CARMA does not cover. Figure 4 shows the progression of

having only 1 mode ( CNs) at high altitude to two modes lower down (CNs from evaporated particles and larger, sedimenting drop-lets that are beyond the Kelvin barrier). However, comparison to data at 54.2 km [1] show disagreement on either side of the peak;

this could be due to issues involving the coupling of the coagulation scheme to our eddy diffusion scheme. Figures 1 and 3 com-pare our results with analysis from Wilquet et al. [4]. We see that the size distribution is not unimodal, but mode 2 is not complete-ly independent of mode 1; the mean radii of the two modes are also slightly larger than those proposed by [4]. Meanwhile, figure 1

shows that the mode 2 height profile fits the data well, but mode 1 fits rather poorly. One possible cause for this is the upward wind, which reaches ~10 m/s in this region; this could overwhelm all other effects, causing the slopes of all 3 height profiles to be

the same, and the size distributions at various heights to be self-similar. It should be noted that there is a bump in the data at 84 km similar to the bump in the simulation at 87 km; this could constrain the site of meteorite dust production on Venus. Figures 5 and

6 show the gas-height distribution. We see that the gas is saturated for most of the atmospheric column, as expected, except for 1). above 80 km, where the saturation vapour pressure (SVP) becomes too small and numerical effects cause CARMA to “stall” at a gas concentration of ~10-15 ppm, 2). a supersaturation at 61 km, where gas is produced, and 3). below 47 km where we fit to data fromKolodner and Steffes [13]. We see that we overestimate the gas concentration below the SVP, which could be caused by the boun-

dary concentration being too large. Above 47 km, the data lies above our SVP, indicating a possible issue with our SVP scheme.

References[1] Knollenberg R. G. and Hunten D. M. (1980) J. Geophys. Res., 85, [7] Kalashnikova O. et al. (2000) Geophys. Res. Lett., 27, 3293-3296. 8039-8058. [8] de Pater I. and Lissauer J. J. (2001) In: Planetary Sciences. Cam- [2] Imamura T. and Hashimoto G. L. (2001) J. Atmos. Sci., 58, 3597- bridge, UK: Cambridge University Press. 3612. [9] Toon O. B. et al. (1989) J. Geophys. Res., 94, 11359-11380. [3] James E. P. et al. (1997) Icarus, 129, 147-171. [10] Jacobson M. Z. et al. (1994) Atmos. Environ., 28, 1327-1338. [4] Wilquet V. et al. (2009) J. Geophys. Res., 114, E00B42, 13pp. [11] Seiff A. et al. (1985) Adv. Space Res., 5, 3-58.[5] Turco R. P. et al. (1979) J. Atmos. Sci., 36, 699-717. [12] Hunten D. M. et al. (1983) In: Venus. Tucson, USA: The Univer-[6] Bardeen C. et al. (2011) The CARMA 3.0 microphysics package sity of Arizona Press. in CESM, Whole Atmosphere Working Group Meeting. [13] Kolodner M. A. and Steffes P. G. (1998) Icarus, 132, 151-169.

40 km

50

60

70

80

90

100 km

Condensation& EvaporationCARMA 3.0 uses the growth al-gorithm given in [9], which calculates the change in mass mwith time t:

Where Psat is the saturation va-pour pressure, S is the supersat-uration, Aks is the combined Kel-vin and solute factors, and g0 and g1 are growth kernels.

Advection

Sedimentation

Vertical TransportSedimentation in CARMA 3.0 for Reynold's number Re

< 1 is given by Stoke's flow, augmented by the slip corr-ection factor. For Re > 1, Re is recalculated after consi-dering turbulence, then the fall velocity is calculated

from it. A vertical wind is added by adding the wind ve-locity to the fall velocity. The code has been modified

to include eddy diffusion, given by [8]:

Where vi is the velocity of species i, K is the eddy diffu-sion coefficient, N is the atmospheric number density,Ni is the number density of species i, and z is the verti-

cal scale.

CoagulationCARMA 3.0 uses Jacobson et al.'smethod [10] for calculating loss and growth rates of particles by coagulation. CARMA's coagula-tion kernel takes into account VanDer Waal's forces, brownian mo-tion, convective diffusion, and gravitational collection.

Scrit= 4Akel27Bsol

vi = −K NN i

∂ N i/N∂ z

dmdt

=Psatg0S1−Aks

1g0g1P sat

70

Mode 1

Mode 2?

63 km58 km54 km51 km48 km

93 km88 km84 km81 km78 km Gas

Conc.SVP Curve

Gas Conc.SVP Curve

TotalMode 1Mode 2

1 3 5

642

LCPS

LCPS RadioOccultation