Venus Cloud Properties from Venus Express VIRTIS Observations
J. Barstow1 , F. Taylor1 , C. Tsang1,2 , C. Wilson1 , P. Irwin1 , P. Drossart3 and G. Piccioni4
( Atmospheric, Oceanic and Planetary Physics, University of Oxford, UK 2 Southwest Research Institute,
Boulder, Colorado, USA 3 LESIA, Observatoire de Paris, Meudon, France 4 INAF-IASF, Rome, Italy)
1
Summary
Near-infrared spectra from the Visible and Infrared
Thermal Imaging Spectrometer (VIRTIS) on Venus
Express have been used to investigate the vertical
structure and global distribution of cloud properties
on Venus. The spectral range covered by VIRTIS is
sensitive on the nightside to absorption by the lower
and middle cloud layers, which are back-lit by radiation from the lower atmosphere and surface. The
cloud model used to interpret the spectra is based
on previous work by Pollack et al. (1993) and others,
and assumes a composition for the cloud particles
of sulfuric acid and water, with acid concentration as
a free parameter to be determined. Other retrieved
parameters are the average size of the particles
and the altitude of the cloud base in the model. The
sensitivity to these variables across the measured
spectral range (1.5 to 2.6 µm) is investigated, and
radiances at suitable pairs of wavelengths are used
in model branch plots to recover each variable independently. Model spectra are generated using
the NEMESIS radiative transfer and retrieval code
(Irwin et al. 2008).
Spatial variation of sulfuric acid concentration in
the cloud particles has been estimated for the first
time. This is then used in the determination of other
cloud properties and gaseous abundances. Key
findings include increased acid concentration and
decreased cloud base altitude in regions of optically thick cloud, a peak in cloud base altitude at
-50◦ , and an increased average particle size near
the pole. These results are being used to develop
better models of the structure and variability of the
clouds, which are needed to understand the chemistry, meteorology and radiative energy balance on
Venus.
Cloud Model
The cloud model used was based on that of Pollack
et al. (1993) and contains four distinct size modes,
assumed to be composed of sulfuric acid with a concentration between 75 and 96 wt %. The properties
of the four modes are listed in Table 1. Mode 2 is
confined to the upper cloud. Mode 20 and mode 3
are confined to the lower cloud (below 57 km).
Table 1: Aerosol particle size distribution as used in
this work. All size distributions are log-normal.
Mode
Mode 1
Mode 2
Mode 20
Mode 3
Effective Radius (µm)
0.30
1.00
1.40
3.65
Variance (µm)
0.44
0.25
0.21
0.25
Method
Figure 1: VIRTIS-M-IR spectrum with sensitivity to
the acid concentration, mode 3 abundance, base altitude and CO and H2 O between 1.6 and 2.6 µm.
The branch plot method uses a comparison between the radiances at two different wavelengths to
study the variation of a parameter. This can be a
property of the cloud that affects the spectral behaviour of infrared absorption or scattering by the
cloud particles, or it could be the abundance of a gas
that has an absorption band at one of the two wavelengths chosen. Synthetic spectra are created for
different amounts of mode 20 . Mode 20 abundance
varies along each branch, and different branches
are produced by changing other model input parameters, such as mode 3 abundance. These branch
plots define a parameter space that may depend on
one or more of the varying input quantities. If the
space depends primarily on a single parameter, the
position of a data point from a real spectrum on the
branch plot allows the variable to be fixed for that
data point. Once this variable is fixed, branch plots
that are sensitive to this variable and one other can
be used to fix a second variable, and so on.
and 2.56 microns, to investigate variation in acid
concentration, fractional abundance of mode 3 particles, cloud base altitude, CO abundance at ∼35
km and H2 O abundance at ∼35 and ∼50 km. A
branch plot for 2.2 µm radiance against 1.74 µm
radiance is sensitive only to the concentration of
the acid in the clouds. Figure 2 shows radiances at
these wavelengths for a series of models with acid
concentrations of 85% and 96%, fractional mode
3 abundances of 1, 1.5 and 2, and cloud base altitudes of 44, 48 and 52 km. It can be seen that
the model branches with different mode 3 fractional
abundances and cloud base altitudes lie on top of
each other, so the dominant mode of variability in
this parameter space is the acid concentration.
The variables of interest in this work are mainly related to the cloud. They are the concentration of the
acid in the cloud particles, the abudance of large
mode 3 particles (which when taken in conjunction
with the abundance of mode 20 particles indicates
the average particle size in the lower cloud) and the
altitude of the cloud base (an indicator of the vertical
distribution of cloud particles). Gaseous absorption
lines for CO and water vapour are also present. CO
has a band at 2.32 µm, sensitive to the abundance
near 35 km altitude. Three water vapour bands are
present, centred at 1.77 µm (sensitive to water near
25 km altitude), 2.4 µm (35 km) and 2.56 µm (50
km). The sensitivities in the spectral region of interest are outlined in Figure 1.
Five other parameters remain to be constrained.
After correction for acid concentration, the ratio between radiances at 2.3 and 1.74 microns may be
used to determine the relative abundance of mode
3 particles. The base altitude can be determined
using the ratio of radiances at 2.53 and 2.4 microns, after correcting for acid concentration and
H2 O abundance at 50 km. The gaseous abundances can be calculated using the ratios of 2.32
to 2.3 µm radiance (CO), 2.4 to 2.3 µm radiance
(H2 O, 35 km) and 2.56 to 2.53 µm radiance (H2 O,
50 km).
Data Processing
The data used have been spatially coadded in boxes
of 20 × 20 pixels, shifted every 10 pixels to ensure Nyquist sampling. This reduces a 256 × 256
pixel image to a 25 × 25 pixel image. Single radiances for 1.74, 2.2, 2.29, 2.32, and 2.4 microns
have been extracted for each coadded spectrum.
The 2.53 µm radiance used has been smoothed
using a triangular filter in order to remove the oddeven effect, which is a discrepancy in the response
of odd and even spectels in the spectral dimension
of VIRTIS-M and is more prounounced at low radiances. This smoothing calculates the 2.53 µm
0
radiance as I2.53
= 0.25×I2.52 + 0.5×I2.53 + 0.25×I2.54 .
The same smoothing is used for the 2.56 µm radiance.
Figure 2: Model branches for 2.2 against 1.74 microns. Dashed and continuous lines represent different acid concentrations, different colours different base altitudes, and different symbols different
amounts of mode 3. The two distinct branches are
formed by varying the acid concentration.
The VIRTIS viewing geometry results in the planet
being observed at a range of emission angles. This
results in variable limb-darkening across the VIRTIS
dataset. A correction for this effect has been applied
to each of the radiances by creating model spectra
for different input emission angles, and deriving a
relationship between emission angle and radiance.
Branch plots have been generated for combinations of radiances at 1.74, 2.2, 2.3, 2.32, 2.4, 2.53
2
Results
923 VIRTIS-M cubes, recorded between July 2006
and October 2008, have been processed using the
branch plot method and the results collated. The
results for each variable have been analysed as a
function of 1.74 µm radiance, latitude and local solar
time. These results provide comprehensive coverage of the nightside southern hemisphere.
species are present in the lower cloud.
A negative correlation between acid concentration
and 1.74 µm radiance can be seen in Figure 3. 1.74
µm radiance is inversely correlated with the cloud
opacity, so in regions of high cloud opacity we see a
corresponding increase in the cloud droplet sulfuric
acid concentration. It is likely that more cloud forms
in regions of high sulfuric acid vapour abundance,
so this seems intuitive. The acid concentration also
increases slightly towards the cold polar collar at
-60◦ , but is mostly constant with latitude.
Figure 4: Variation of base with 1.74 µm radiance.
Base altitude is constant with radiance except for regions of very high or very low cloud opacity, where
the base altitude is correlated with radiance. Grey
points are individual data, red points are averaged
in radiance.
The altitude of the cloud base also varies strongly
with latitude (Figure 5). The altitude remains approximately constant from the equator to -60◦ , with
a slight increase at -50◦ , but drops off rapidly polewards of -60◦ . This may be correlated with the
thicker cloud present in the polar region, and could
also be an interesting indicator of dynamical processes in this poorly-understood area of Venus.
The number of large (mode 3) particles is relatively
high in the polar regions (Figure 6), as found by
Carlson et al. (1993) and Wilson et al. (2008). We
also find that the relative number of large particles
increases in regions of low cloud opacity (high 1.74
µm radiance). This may be due to increased volatility of smaller particles, which enables them both to
condense and to evaporate more easily than their
larger counterparts, meaning the larger particles
are more stable.
Figure 3: Variation of acid concentration with 1.74
µm radiance. Acid concentration decreases as radiance increases. Grey points are individual data, red
points are averaged in radiance.
The altitude of the cloud base decreases sharply in
regions of low radiance, and increases in regions
of high radiance (shown in Figure 4). This indicates that opacity changes are mainly caused by
creation/removal of the lower cloud layer. The base
altitude values for radiances below 0.1 W/m2 /µm/sr
seem low for a sulfuric acid cloud, as it is expected
that sulfuric acid will thermally dissociate below ∼40
km. This may indicate that other as yet unidentified
We also present the first measurement of the water
vapour abundance near 50 km on the nightside, using the 2.56 µm band.
3
Conclusion
Venus Express- VIRTIS data have been used to
study the variation with cloud opacity, latitude, and
local solar time of the acid concentration in the
Venusian cloud, the cloud base altitude, and the
representative particle size. The absolute numerical
values found are model dependent, but the trends
uncovered are likely to be real. These include strong
meso- and planetary-scale variations which suggest
important meteorological processes at work, and
also indicate that the common assumption in modelling and data analysis of a single a priori cloud
structure for the whole planet may be invalid.
Acknowledgements
This work is supported by the Science and Technology Facilities Council (UK), the Centre Nationale
d’Etudes Spatiales (France), the Agenzia Spaziale
Italiana and ESA.
Figure 5: Variation of base altitude with latitude.
Base altitude increases slightly towards the polar
collar, but drops off sharply polewards of -60◦ . Grey
points are individual data, red points are averaged
in radiance.
References
Carlson, R. W., Kamp, L. W., Baines, K. H., Pollack, J. B., Grinspoon, D. H., Encrenaz, T.,
Drossart, P., Taylor, F. W., 1993. Variations in
Venus cloud particle properties: A new view
of Venus’s cloud morphology as observed by
Galileo Near-Infrared Mapping Spectrometer.
Planetary and Space Science 41, 477–485.
Irwin, P. G. J., Teanby, N. A., de Kok, R., Fletcher,
L. N., Howett, C. J. A., Tsang, C. C. C., Wilson, C. F., Calcutt, S. B., Nixon, C. A., Parrish, P. D., 2008. The NEMESIS planetary atmosphere radiative transfer and retrieval tool.
Journal of Quantitative Spectroscopy and Radiative Transfer 109, 1136–1150.
Pollack, J., Dalton, J., Grinspoon, D., Wattson, R.,
Freedman, R., Crisp, D., Allen, D., Bézard, B.,
de Bergh, C., Giver, L., 1993. Near-Infrared
Light from Venus’ Nightside: A Spectroscopic
Analysis. Icarus 103, 1–42.
Wilson, C. F., Guerlet, S., Irwin, P. G. J., Tsang,
C. C. C., Taylor, F. W., Carlson, R. W.,
Drossart, P., Piccioni, G., 2008. Evidence
for anomalous cloud particles at the poles
of Venus. Journal of Geophysical Research
(Planets) 113 (E12).
Figure 6: Variation of mode 3 relative abundance
with latitude. There is a significant increase polewards of -60◦ . Grey points are individual data, red
points are averaged in radiance.
4