Icarus 272 (2016) 48–59
Contents lists available at ScienceDirect
Icarus
journal homepage: www.elsevier.com/locate/icarus
Carbon monoxide observed in Venus’ atmosphere with SOIR/VEx
A.C. Vandaele a,∗, A. Mahieux a,b, S. Chamberlain a, B. Ristic a, S. Robert a, I.R. Thomas a,
L. Trompet a, V. Wilquet a, J.L. Bertaux c
a
Planetary Aeronomy, Belgian Institute for Space Aeronomy, 3 av. Circulaire, 1180 Brussels, Belgium
Fonds National de la Recherche Scientifique, 5 rue d’Egmont, 1000 Brussels, Belgium
c
LATMOS, 11 Bd d’Alembert, 78280 Guyancourt, France
b
a r t i c l e
i n f o
Article history:
Received 3 November 2015
Revised 6 February 2016
Accepted 12 February 2016
Available online 27 February 2016
Keywords:
Atmosphere composition
Venus
Infrared observations
a b s t r a c t
The SOIR instrument on board the ESA Venus Express mission has been operational during the complete
duration of the mission, from April 2006 up to December 2014. Spectra were recorded in the IR spectral
region (2.2–4.3 μm) using the solar occultation geometry, giving access to a vast number of ro-vibrational
lines and bands of several key species of the atmosphere of Venus. Here we present the complete set of
vertical profiles of carbon monoxide (CO) densities and volume mixing ratios (vmr) obtained during the
mission. These profiles are spanning the 65–150 km altitude range. We discuss the variability which is
observed on the short term, but also the long term trend as well as variation of CO with solar local time
and latitude. Short term variations can reach one order of magnitude on less than one month periods.
SOIR does not observe a marked long term trend, except perhaps at the beginning of the mission where
an increase of CO density and vmr has been observed. Evening abundances are systematically higher
than morning values at altitudes above 105 km, but the reverse is observed at lower altitudes. Higher
abundances are observed at the equator than at the poles for altitude higher than 105 km, but again the
reverse is seen at altitudes lower than 90 km. This illustrates the complexity of the 90–100 km region of
the Venus’ atmosphere where different wind regimes are at play.
© 2016 Elsevier Inc. All rights reserved.
1. Introduction
The earliest observation of CO above the clouds was performed
by Connes et al. (1968) who measured a disk-average mixing ratio of 45 ± 10 ppm at 64 km. Observations of an average value
of 70 ± 10 ppm at 60° latitude (8 AM local solar time for altitudes between 68 and 71 km) were reported (Krasnopolsky, 2008),
using the ground-based CSHELL instrument at NASA’s Infrared
Telescope Facility (IRTF), and complemented two years later by
measurements at different local solar times (Krasnopolsky, 2010).
Irwin et al. (2008) reported latitudinal distribution of CO above
the cloud level on the nightside using VIRTIS-M observations,
on board Venus Express (VEx). They found an average value of
40 ± 10 ppm (at 65–70 km) with little variation in the middle latitudes. Iwagami et al. (2010) observed a nearly uniform distribution of CO above the clouds on the day side consistent with the
findings of Krasnopolsky (2008). Recently, Marcq et al. (2015) used
high-resolution observations acquired by CSHELL/IRTF at 4.5 μm
(corresponding to soundings at 70–76 km altitude) to search for
∗
Corresponding author. Tel.: +32 23730367.
E-mail address: [email protected] (A.C. Vandaele).
http://dx.doi.org/10.1016/j.icarus.2016.02.025
0019-1035/© 2016 Elsevier Inc. All rights reserved.
latitudinal variations of CO on both day and night sides of Venus,
which were found to be weak.
Observations at millimetre and sub-millimetre wavelengths are
sensitive to the CO abundance located in the 80–105 km region.
They allow the determination of CO vertical profiles on the nightside and dayside of the planet, which are of great interest to
understand the CO behaviour in the Venus’ atmosphere. They
moreover showed that the CO abundance is very variable at these
altitudes (Clancy and Muhleman, 1985b; Clancy and Muhleman,
1991; Clancy et al., 2008,2012b; Gurwell et al., 1995; Lellouch
et al., 1994; Wilson et al., 1981). CO densities at higher altitudes
(100–150 km) were obtained using the 4.7 μm non-LTE emission
bands of CO obtained during daytime limb observation performed
by VIRTIS-H/VEx (Gilli et al., 2015).
CO has also been observed below the clouds by ground-based
instruments (Bézard et al., 1990; Crovisier et al., 2006; Marcq
et al., 2005; Marcq et al., 2006; Pollack et al., 1993) and VIRTIS/VEx (Marcq et al., 2008; Tsang et al., 2008; Tsang et al., 2009).
On the nightside, Tsang et al. (2008) observed tropospheric CO
at altitudes of 35 km. They reported higher CO values at dusk
compared to dawn. They also saw an increase of the CO abundance from the equator to the poles, with a maximum at latitudes
around 60°. In fact, the first hint of a possible variation of CO with
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
latitude was reported by Collard et al. (1993) based on measurements recorded by the Near-Infrared Mapping Spectrometer
(NIMS) on board Galileo during its 1990 fly-by of Venus.
SOIR, which was part of the SPICAV/SOIR instruments’ suite on
board VEx, routinely observed CO during the VEx mission, providing vertical profiles extending from 65 to 150 km. CO observations
performed by SOIR were already reported (Vandaele et al., 2015)
but the authors limited the discussion to the short term variations.
They focused their analysis on four groups of measurements out of
the complete data set, to show the CO density and vmr variation
on periods of days. In this paper, we will present the complete CO
data set obtained by SOIR. We will further investigate the variability observed in terms of different time scales, and the latitudinal
variations seen at different altitude regions, as well as the local
solar time dependence.
2. Instrument description
The SOIR (Solar Occultation in the IR) instrument has been designed to measure spectra of the Venus’ atmosphere in the IR region (2.2–4.3 μm) using the solar occultation technique (Nevejans
et al., 2006). This method derives unique information on the vertical composition and structure of the mesosphere and lower thermosphere of Venus, probing the altitude range from 65 to 170 km
(Fedorova et al., 2008; Mahieux et al., 2015a; Mahieux et al., 2010;
Mahieux et al., 2015b; Mahieux et al., 2012; Mahieux et al., 2015c;
Vandaele et al., 2008; Vandaele et al., 2015). SOIR was an extension mounted on top of the SPICAV instrument (Bertaux et al.,
2007). SPICAV/SOIR was one of the seven instruments on board
VEx, a planetary mission of the European Space Agency (ESA) that
was launched in November 2005 and inserted into orbit around
Venus in April 2006 (Titov, 2006). The SOIR instrument is unique
in terms of spectral coverage and spectral resolution (0.15 cm−1 ),
and is ideally designed to probe the Venus’ atmosphere, above the
cloud deck, for CO2 as well as trace gases. CO is particularly well
covered since the (2-0) vibrational band strongly absorbs in the
390 0–440 0 cm−1 range, which is well inside the sensitivity range
of the SOIR instrument.
The instrument has already been extensively described elsewhere (Bertaux et al., 2007; Mahieux et al., 2008; Nevejans
et al., 2006) and will only be briefly outlined here. SOIR is an
Echelle grating spectrometer operating in the IR, combined with an
acousto-optic tunable filter (AOTF) for the selection of the recorded
wavenumber interval. The wavenumber range covered by the instrument extends from 2200 to 4370 cm−1 (2.2–4.3 μm) and is divided into 94 diffraction orders (from 101 to 194). The definition
and limits of these diffraction orders are presented in Vandaele
et al. (2013a). The bandwidth of the AOTF was originally designed
to be 20 cm−1 , as measured in the lab before launch (Nevejans
et al., 2006), to allow light from only one order into the spectrometer. However, the measured bandwidth of the AOTF filter is
∼24 cm−1 (Mahieux et al., 2008), creating some order leakage on
the detector, since it is wider than one diffraction order. This effect
is called superposition of orders hereafter.
The detector has 320 pixels along the spectral axis and 256
rows. The instrument entrance slit is projected on 32 rows. Due
to telemetry limitations, only the equivalent of 8 rows per second
can be downlinked to Earth. In most of the observations 4 different
values of the AOTF frequency are chosen to record spectra in 4 different spectral intervals per second, hence increasing the number
of species potentially detectable simultaneously. In that case, only
two spectra per AOTF frequency can be downlinked, corresponding
to two ‘bins’ of rows on the detector. It was noticed early on that
the outermost rows of the slit projection on the detector were not
fully illuminated, so only the 24 central rows were considered after
orbit 332.1 (March 19, 2007), leading to the definition of two bins
49
of 12 rows each (compared to the 2 bins of 16 rows that were considered before orbit 332.1). As a result, four different wavenumber
regions (orders) are measured during each occultation, and each
order measurements are obtained for the two bins of the detector
array, resulting in eight independent series of spectra.
The resolution of the SOIR instrument varies slightly from the
first to the last order, from 0.12 to 0.20 cm−1 , as does the spectral sampling interval which varies from 0.030 cm−1 in diffraction
order 101 to 0.055 cm−1 in order 194, increasing with the pixel
number and the diffraction order. Signal to noise ratio on the SOIR
transmittances varies between 250 and 50 0 0 but is typically of the
order of 20 0 0. These SNR values are deduced from the recorded
spectra as explained in Vandaele et al. (2013b). The spectra used
in this work are PSA Level 3 calibrated data, which have been intensively described in Vandaele et al. (2013b).
Because of the geometry of the solar occultation observations,
all observations are performed at the terminator, either on the day
side (local solar time, LST 6 AM) or night side (LST 6 PM). They
cover all latitudes on both hemispheres. The instrument probes
the mesosphere (70–110 km) and the lower thermosphere (above
110 km). Those were up to now very poorly known regions of the
Venus’ atmosphere. Indeed, few measurements were performed in
that altitude range at all latitudes prior to SOIR. Depending on the
species, SOIR delivers vertical profiles extending from the top of
the clouds (60–70 km) up to 170 km. The vertical resolution and
sampling of the profiles depend essentially on the distance separating VEx and the planet. Therefore they both vary significantly
since the orbit of the spacecraft was very elliptic, with a periapsis of about 300 km and apoapsis of 45,0 0 0–60,0 0 0 km. The resolution ranges from a few hundreds of meters (typically 200 m at
the North Pole) up to some kilometres (typically 5 km at the South
Pole). All vertical profiles resulting from the retrieval procedure are
interpolated on an altitude grid of 1 km from 60 to 200 km.
3. Retrieval technique and data set
The retrieval method has already been described in detail elsewhere for CO2 and temperature (Mahieux et al., 2015a, 2010, 2012)
and specifically for trace gases (Vandaele et al., 2015). The method
determines the number densities, temperature and aerosol extinction profiles using an iterative procedure. The algorithm takes the
saturation of atmospheric lines into account, and only spectra with
unsaturated lines of the weakest simulated vibrational bands are
considered (Mahieux et al., 2015a). The a-priori number densities are taken from a modified VIRA (Venus International Reference Atmosphere model) model based on Hedin et al. (1983) above
100 km and Zasova et al. (2007) below. The covariance on the densities is set to 25% of the a priori number density logarithm. The
vertical resolution, which can vary a lot with latitude, is accounted
for by considering a multiple ray tracing across the SOIR field of
view. Typically a 24 point grid spanning the FOV is considered as
shown in Fig. 1 in Mahieux et al. (2015a). The measurement at
one given tangent height is then calculated as the algebraic average of the contribution of all 24 grid points (corresponding to 24
slightly different tangent heights distributed around the main tangent point).
The temperature profile is not fitted directly, but is retrieved
from the CO2 density, assuming hydrostatic equilibrium. The retrieved temperatures profile affects the modelled absorption cross
sections, which are temperature-dependent in the infrared domain.
At the end of each iteration of the fitting procedure, the profile of one particular species is built by interpolating the individual
number density values (weighted by their uncertainties) obtained
from the different wavenumber regions (orders) and bins of the
detector (Mahieux et al., 2012). The retrieved profile is indeed defined by a series of (altitudes, local densities) couples where the
50
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
Fig. 1. Location of the CO observations, in terms of orbit number versus latitude. The circles represent the observations at the morning side of the terminator and the
triangles the observations at the evening side of the terminator. The numbers at the centre of the picture are the occultation season numbers.
altitudes correspond to all tangent heights of the spectra containing spectral information of the targeted species. This profile is interpolated on a final grid corresponding to a 1 km altitude step
scale, in the following way: at a given altitude, z, all couples lying
within [z- α SH, z + α SH], with SH the scale height and α a user
defined parameter, are considered and fitted to a linear expression
(log(density) = A x Altitude + B), all points weighted by their uncertainty. Finally this polynomial is interpolated at z, defining the final ‘interpolated’ density at that altitude. During the retrieval process, the vertical box width used for the interpolation is equal to
± 2 scale heights (α = 2). This method ensures the convergence by
removing local oscillations which might appear from one iteration
to the other that could hamper the smooth and rapid convergence.
The scale height which is considered here is the one obtained from
the mean atmosphere mass density, and not the scale height of CO
only.
The width of this box (α ) can be chosen by the user; by default
it is set to 2 (the value used during the fitting procedure). In the
following investigation, we have used a scaling factor of α = 0.2.
The complete procedure is described in detail in the first paper
dealing with CO retrieval (Vandaele et al., 2015).
This ‘smoothing/interpolating’ method is in no case performed
to reduce the error on the fits, but to account for the fact that for a
random occultation, depending on the numbers of orders in which
a given species is measured, on the inclination of the slit relative
to the limb and on the latitude, the pattern of the measurement
altitudes can be very different from an occultation to another. In
fact, for some occultations, the measurement altitudes will be distributed evenly within the altitude range, while for others there
will be large altitude gaps between groups of points. This is intrinsic to the solar occultation method used here, coupled to the instrument characteristics (measurement in different orders on two
detector bins) and the VEx orbit shape (very elliptical orbit impacting the vertical sampling).
This two steps procedure, i.e. the determination of the number
density profiles followed by the determination of the hydrostatic
temperature profile, is done in an iterative scheme. Convergence is
achieved when all the density and temperature profiles stabilize,
i.e. remain within the uncertainty of the previous iteration. It usually occurs in three to four steps.
The primary results deduced from the analysis of SOIR spectra
are densities. Conversion to volume mixing ratio (vmr) requires the
knowledge of the total density, or that of CO2 , if a CO2 vmr is assumed:
vmrTraceGas =
nTraceGas × vmrCO2
nCO2
(1)
If the CO2 density was retrieved simultaneously during the
same occultation, then it is directly used to obtain the vmr using Eq. (1). If no information on CO2 was available from the observation itself, the CO2 density will be derived from the VAST
(Venus Atmosphere from SOIR data at the Terminator) data model
(Mahieux et al., 2015a, 2012). In both cases, an assumption is made
on the CO2 mixing ratio vmrCO2 , whose values are taken from the
VIRA model: data from Seiff et al. (1985) for altitudes up to 100 km
and from Hedin et al. (1983) above that altitude have been used.
Vmr vertical profiles averaged in the different latitudinal bins are
also available for the VIRA compilation.
4. Results and comparison with the literature
CO has been observed regularly with SOIR during the VEx mission. The individual observations are shown in Fig. 1, as a function
of orbit number and latitude. In this figure, circles denote morning
side measurements and triangles the evening ones. Table 1 gives
the number of observations for the different latitudinal bins considered in this study (0°–30°, 30°–60°, 60°–70°, 70°–80°, 80°–90°).
Morning/evening difference has been considered, for both North
and South hemispheres. The equatorial and mid-latitude bins considered in this study are larger than the more polar ones. This is
done to compensate the poorer coverage at these latitudes due to
the very elliptical orbit of the spacecraft.
Individual profiles of CO densities corresponding to the 218 observations reported in Table 1 are shown in Fig. 2. Data for the
North and South hemispheres have been separated, and profiles
have been gathered in the different latitudinal bins listed above,
while morning and evening observations are shown with different
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
51
Fig. 2. CO density profiles obtained by SOIR during the Venus Express mission for the different latitudinal bins considered in this work and for the Northern (Left panels)
and Southern (Right panels) hemisphere. For each bin, morning (blue) and evening (orange) observations are shown, with the VIRA profiles corresponding to the day (solid
purple) and night (dashed purple) conditions. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
52
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
Table 1
Statistics of the CO observations for the different latitude bins and on each
side of the terminator considering a hemispherical symmetry.
0°–30°
30°–60°
60°–70°
70°–80°
80°–90°
Total
North hemisphere
South hemisphere
Number of
observations
AM
Number of
observations
PM
Number of
observations
AM
Number of
observations
PM
8
3
3
12
38
64
9
3
7
13
26
58
11
16
6
4
6
43
15
20
6
5
7
53
colours. Errors are also plotted indicating the level of quality of
the retrieval procedure. Higher error levels are encountered at the
bases of the profiles, corresponding to the less intense spectra.
The density profiles all show a distinctive change of slope occurring at an altitude usually lying between 100 and 110 km, which
is also present in the VIRA model, but at a lower altitude. The VIRA
model for CO is based on observations reported by Keating et al.
(1985) above 100 km. Below 100 km, the values are those reported
by von Zahn and Moroz (1985) later updated by de Bergh et al.
(2006) and Zasova et al. (2006) (see a full description of the VIRA
models in Table 1 of Vandaele et al. (2016)). This slope is not seen
in the CO densities retrieved from VIRTIS limb emission measurements (Gilli et al., 2015). However, it is quite well reproduced in
the profiles simulated using the Venus Thermospheric General Circulation Model (VTGCM) which are shown in Fig. 12 of Gilli et al.
(2015), although at higher altitudes (between 110 and 120 km). Already from Fig. 2, we can infer some general trends which will
be investigated in more detail in the following sections. In general
we observe that at altitudes above 100–110 km, evening values are
larger than morning values, and that at the lower altitudes, the
reverse is seen, i.e. evening abundances are lower than at morning terminator. The latter has been described by many observers
(Clancy et al., 2012b; Gurwell et al., 1995; Krasnopolsky, 2010; Lellouch et al., 1994). For example, Krasnopolsky (2010) obtained CO
abundances of 40 ± 4 ppm and 52 ± 4 ppm near the evening and
morning sides of the terminator respectively (at 68 km). However
the behaviour observed at higher altitudes has never been reported
before. It can be hinted in some of the profiles reported in Clancy
et al. (2012b) at the highest altitude sounded by their instrument
(105 km).
Fig. 3 illustrates the comparison of SOIR CO profiles with data
from the literature. SOIR data have been grouped into the latitudinal bins defined in Table 1. For each latitude bin, we have
plotted the average profile and the minimum and maximum values measured at each altitude. This gives an idea of the variability observed in the SOIR dataset. Densities and vmr are plotted
in different panels of the figure. The consistency between SOIR
values and the literature can be readily appreciated in both panels. Connes et al. (1968) reported a CO abundance of 45 ppm at
50–70 km altitude; Irwin et al. (2008) found an average value of
40 ± 10 ppm at 65–70 km; Krasnopolsky (2008) observed a vmr of
70 ± 8 ppm at 68–71 km; Marcq et al. (2015) reported values between 25 and 45 ppm at 70 km; Lellouch et al. (1994) could derive a CO profile with CO vmr going from 18 ppm at 75 km up to
2100 ppm at 115 km. A series of recent CO profiles obtained by the
JCMT spectrometer have been provided (Clancy, private communication). These spectra have been obtained with the same instrument and in similar conditions, whilst not in the same year, as the
data described and discussed in Clancy et al. (2012b). They all correspond to dayside observations with local time varying between 8
and 11.30 am. Krasnopolsky (2012)’s profile is in very good agreement from 65 km up to 110 km. The vertical profile obtained by
Marcq et al. (2005) from observations performed at the NASA Infrared Telescope Facility (IRTF) in Hawaii with the SpeX imaging
spectrometer is also plotted on the figure. Only Gilli et al. (2015)
reported CO density profiles at higher altitudes from the analysis
of the 4.7 μm non-LTE emission band of CO observed by VIRTIS-H.
Different profiles corresponding to different local solar times and
different latitude bins are plotted in Panel B.
Scale heights of CO have been deduced from the complete data
set presented here, see Fig. 4. These values were calculated directly from the CO density vertical profiles considering layers of
5 km, and not from the temperature profiles (it may thus be different from the scale height used in the retrieval procedure described in Section 3). Up to 90 km, the scale height is almost constant (4.8 km, std. dev. = 1.5 km). This is in very good agreement
with the value found by Marcq et al. (2015) who derived the
value from ground-based observations probing the 70-76 km altitude range (5 ± 0.5 km). Above 90 km, the scale height increases
with most values ranging from 5 to 15 km, with an average value
of 10.8 km (std. dev. = 3.5 km) at 115 km. Scale heights deduced
from the measurements of Lellouch et al. (1994) have been added
on the plots and also show an increase with increasing altitudes.
Above 115 km, the scale height decreases to reach values again
comprised between 5 and 10 km, with an average value of 6.9 km
(std. dev. = 1.3 km) at 142.5 km. From the various density profiles
presented in Gilli et al. (2015) we have calculated scale heights
from 107.5 to 142.5 km. The mean values at different altitudes are
plotted in the figure, along with their standard deviation. These
values comfort the trend seen by SOIR, i.e. increase up to 115 km
then decrease above. We have investigated a possible dependence
on latitude, but it was not significant. Evening observations seem
to correspond to higher scale heights above 100 km, with an almost constant difference of 1.8 km. Moreover the spread of values
at altitudes between 90 and 120 km is larger for evening observations than for morning ones. This could reflect a higher variability
in evening CO compared to morning’s.
In the following discussion we will consider broader latitude bins in order to improve the statistics of the observations
when calculating average profiles and deviation standard. The bins
extend from 0° to 60° (equatorial region), 60–80° (high-latitude
region) and 80–90° (polar region). To illustrate the values of the
standard deviation around the averages, average profiles have been
plotted in Fig. 5 considering two different scenarios: first, averaging on all evening values from both hemispheres (left panel), then
averaging on all morning and evening data from only the Northern hemisphere (right panel). We see that in both cases, above
110 km, the average profiles are quite well separated falling outside of their respective deviation standards. Below 90 km, all deviation standards overlap each other. This will be important when
discussing the pertinence of the differences observed between average profiles in the following sections.
4.1. Local solar time variations
The local solar time difference observed by SOIR is highlighted
in Fig. 6 where average profiles corresponding to morning and
evening have been plotted for the three latitudinal bins covering
the equatorial region (0–60°), the high-latitudes (60–80°), and the
polar regions (80–90°). Here both hemispheres have been considered together. These profiles summarize perfectly what SOIR is observing: above 105 km, evening values are in average higher than
morning values for all three latitudinal bins. The impact seems the
lowest at the poles. Below 95 km, although the number of observations is lower than above, we clearly see the reverse pattern.
The critical region where the inversion occurs lies between 90 and
110 km (p ∼ 3.0-0.001 mbar). This region corresponds to a very intricate dynamical regime, combining zonal and meridional flows,
as well as remnants of the subsolar to antisolar (SSAS) circulation.
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
53
Fig. 3. Comparison of SOIR CO vmr (Top) and density (Bottom) profiles with data from the literature. SOIR average profiles, with minimum and maximum values, have been
plotted for the five latitudinal bins defined in Table 1.
54
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
Fig. 4. CO scale heights for all orbits considered in this work. Orange dots correspond to evening observations, while blue dots are morning observations. The solid lines
correspond to the average values for the morning (blue), evening (orange) and all combined (black). The dashed lines are the corresponding standard deviations. Values from
(Gilli et al., 2015; Lellouch et al., 1994; Marcq et al., 2015) are also mentioned on the figure in blue, dark green and light green respectively. The horizontal bars represent
the standard deviation on the scale heights determined from (Gilli et al., 2015). (For interpretation of the references to colour in this figure legend, the reader is referred to
the web version of this article.)
Fig. 5. Average profiles and their standard deviation for three latitudinal bins considering all values at the evening terminator and both hemispheres (Left panel) and all
values at morning and evening but only in the Northern hemisphere (Right panel).
4.2. Short-term variations and long-term trend
Short term variations of CO in the Venus atmosphere as seen by
SOIR have already been addressed and discussed in Vandaele et al.
(2015). However, here, we bring new insights to the discussion by
considering the complete data set obtained during the whole VEx
mission. In our previous study, we concluded that high variability
was observed on the short term. This is confirmed and illustrated
in Fig. 7 where CO vmr have been plotted versus time for the full
data set. Each dot represents one observation at a given pressure
level with the colour indicating the latitude at which the observation was performed. The six pressure levels considered here span
the region sounded by the instrument. They correspond approximately to the altitudes of 129, 120, 111, 102, 93, and 84 km.
Day to day variations of one to two orders of magnitude are
observed all along the mission and at all altitudes. On the long
term, however, no clear trend seems visible, except at the beginning of the mission when, at all pressure levels, a steep rise in
abundance can be noticed, ending at around orbit 500. This increase is in fact seen in other species, for example in CO2 or even
in the aerosol optical depths (Vandaele et al., 2013b).There seems
to be no clear trend even considering the latitudinal dependence
(i.e. isolating one colour). However the detailed investigation of
this long term trend will be soon performed considering all the
variables retrieved by SOIR (CO2 and trace gas abundances, as well
as temperature), searching in particular for specific periods and
frequencies.
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
55
Fig. 6. CO vmr for morning (solid line with dots) and evening (dashed lines with x) observations and for different latitudinal bands. The latter have been chosen large
enough to be statistically meaningful. Hemispheric symmetry has been considered here.
Fig. 7. Evolution of the CO vmr during the entire course of the Venus Express mission at different pressure levels (the corresponding altitude is indicated in brackets). Circles
and triangles represent observations at the morning and evening sides of the terminator respectively, while colour shows the latitude (blue for equator to red for the poles).
Each dot corresponds to a single observation at a given pressure level. The plots have been done for six different levels spanning the region sounded by the instrument. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
4.3. Latitudinal variations
Above 100 km altitude, a steadily decrease of CO abundance (either density or vmr) is observed from the equator to the poles.
This is seen for morning and evening observations, as illustrated
in Figs. 8 and 6. Fig. 8 shows the averaged profiles for the Northern and Southern hemispheres for the three latitudinal bands al-
ready introduced, and corresponding to the equatorial region, the
high-latitudes and the polar regions (0–60°, 60–80°, and 80–90°,
respectively). Morning and evening data are considered together
to improve the statistics in each latitudinal box. We see that, at
high altitudes, the trend is very clear: higher CO vmr are observed
in the equatorial zone compared to the poles (factor of 5 to 7).
Values obtained at high-latitudes are very similar in both hemi-
56
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
Fig. 8. CO vmr for the Northern (solid line with dots) and Southern (solid lines with x) hemispheres and for different latitudinal bands. The latter have been chosen large
enough to be statistically meaningful. Morning and evening observations are not discriminated.
spheres while the difference between polar and equatorial values
with the high-latitude values appear lower in the Southern hemisphere than in the Northern. However we cannot rule out the impact of the statistics on the average due to the disparate number of
observations in each box (see Table 1, for the exact numbers). Similar pole-to-equator behaviour was observed by Gilli et al. (2015)
who investigated the CO abundance detected by limb measurements performed by the VIRTIS-H/VEx instrument. A decrease of
a factor of 2 was detected between high latitudes and equator for
altitudes higher than 130 km, which is very similar to what is seen
by SOIR (factor of 2.2 between polar and high-latitude values at
1.0 × 10−5 mbar, approx. 130 km, for both hemispheres).
At lower altitudes, a reverse trend is observed. Although suffering from a definite under-sampling at low altitudes, SOIR detects more CO at high-latitude than at the poles and equator. This
is clear for the Northern hemisphere, less obvious for the Southern data, remembering that the statistics of the latter are less
favourable (see Table 1: only 21 profiles have been obtained for
the 60–80°S bin, compared to 35 for the same band in the Northern hemisphere). This is also observed at lower altitudes, even if
the connection with the SOIR observations is not straightforward.
For example Tsang et al. (2008), analysing VIRTIS-M/VEx data in
the 2.18–2.50 μm, showed that at an altitude of 35 km, CO vmr
was increasing from equator to poles going from 23 ± 2 ppm to
30 ± 2 ppm, with marked maxima at 60° and 70° latitudes in the
Southern and Northern hemispheres respectively. A very similar
evolution was found based on the analysis of the VIRTIS-H/VEx
data at 2.3 μm (Marcq et al., 2008).
Fig. 9 helps to visualize the CO latitudinal distribution and how
it changes with altitude. We have selected six different pressure
levels (corresponding to altitudes approximately every 10 km from
80 to 130 km) at which we have plotted the CO vmr with respect
to the latitude. The abundance shows a systematic pronounced decrease at Northern latitude above 60°, which seems not so marked
or even absent at the Southern pole. However, when fitting a polynomial of degree 2 through all individual data at the different pressure levels, the results show that indeed the hemispheric sym-
metry is observed. For all levels above 90 km, we see that the fit
shows higher abundances at the equator, decreasing smoothly towards the poles. However the trend is clearly reverse at 1.0 mbar
(84 km), although it should be reminded that the statistics at these
low altitudes is poorer.
4.4. Discussion
Carbon monoxide is produced by photo-dissociation of CO2
on the day side in the upper atmosphere of Venus by solar UV
radiation. The 3-body CO + O recombination reaction into CO2 is
very slow. However, faster recombination processes through catalytic cycles involving chlorine (ClOx ) or hydrogen (HOx ) oxides
have been proposed to reform CO2 (Krasnopolsky and Parshev,
1983; Krasnopolsky, 2012; Yung and DeMore, 1982) as well as catalytic CO oxidization reactions within the clouds (Mills and Allen,
20 07; Mills et al., 20 07). Hence, mesospheric and lower thermospheric CO abundances are strongly impacted by dynamical transport which makes it a very good tracer for the dynamics at play.
Observations of tropospheric CO (Tsang et al., 20 08, 20 09),
which indicate an increase of CO abundance from equator to poles,
with a maximum around 60° latitude, were interpreted as a proof
of the existence of a Hadley cell circulation type. The uplifting
of air occurs at the equator, and the descending branches of the
Hadley cell have been shown to be located at latitudes close to 60°
by GCM modelling (see for example Lee et al. (2007)), supporting
the interpretation of the measurements of CO below the clouds.
The following mechanism was then suggested (Marcq et al., 2015;
Taylor, 1995; Taylor and Grinspoon, 2009; Tsang et al., 2008): CO is
produced through the photolysis of CO2 occurring at high altitudes
above the clouds, the CO-rich air above the cloud deck is entrained
below the clouds by the descending branch of the cell, leading to
the observed bulge at 60° latitude in both hemispheres. Looking at
Fig. 8 or Fig. 6, we see that the profiles corresponding to the highlatitudes, i.e. the location of the descending branch, cross the ones
from the equatorial regions at an altitude of 100 km. This would
represent the maximum altitude at which, or at least below which
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
57
Fig. 9. Latitudinal variations of the CO vmr in function of the latitude for different pressure levels (the corresponding altitude is indicated in brackets). Blue circles and
orange triangles represent observations at the morning and evening sides of the terminator respectively. Each dot corresponds to a single observation at a given pressure
level. The plots have been done for six different pressure levels spanning the region sounded by the instrument. The black solid lines represent the polynomial of degree 2
which best fits the data at each level. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
the meridional transport occurs, i.e. the location of the top zonal
branch of the Hadley cell. This supports the mechanism (Taylor,
1995; Taylor and Grinspoon, 2009) in which the deep Hadley circulation on Venus extends from well above the clouds to the surface
and from the equator to the edge of the polar vortex.
At high altitudes (well above the clouds), CO is produced by the
photolysis of CO2 by UV radiation. Since this process is driven by
the Sun illumination, CO production would then be the highest at
the subsolar point. Gilli et al. (2015) observed maximum CO abundances around noon decreasing towards the morning and evening,
with no significant difference between morning and afternoon values. SOIR observations, on the contrary, indicate that above 100 km
altitude, CO is slightly more abundant at the evening terminator
than in the morning.
CO is thus produced at high altitudes where the SSAS circulation prevails (Bougher et al., 1997). CO is then transported from the
subsolar point (or near the subsolar point) in the equatorial region
to the higher latitudes, undergoing chemical interaction during
the transport (Clancy and Muhleman, 1985a; Krasnopolsky, 2012),
leading to a decrease of its abundance. This was observed already
by Gilli et al. (2015) who saw a clear decrease of CO density from
equator to pole at high altitudes (above 130 km). Below that altitude, the gradient was still present in the available data but lying
within the noise. SOIR observations confirm this latitudinal trend:
CO abundances at the equator are higher than high-latitude (60–
80°) and polar (80–90°) abundances by a factor of 2 and 4 respectively at an altitude of 130 km, when considering all data (North
and South hemispheres, morning and evening). Below 100 km, as
we pointed out earlier, the trend is reversed. In fact we observe a
convergence of all profiles around 100 km, see for example Figs. 6
and 8, in which however the inversions in the trends relative to
morning/evening and equator/high-latitude/pole values are visible
below that altitude. This is compatible with several previous ob-
servations which reported weak (Irwin et al., 2008; Marcq et al.,
2015) or non-significant (Iwagami et al., 2010; Krasnopolsky, 2008)
latitudinal variations in different altitude ranges spanning the 65–
76 km region. When a weak trend was mentioned, it corresponded
to a slight increase from equator to poles.
The altitude region sounded by SOIR corresponds to a region where different circulation patterns coexist: the SSAS
circulation driven by strong diurnal temperature gradients (mostly
active above 120 km) and the retrograde zonal circulation (mostly
active at and below 70 km). It has been suggested that gravity
waves, generated in the unstable cloud region, could force the retrograde zonal flow well above 70 km extending up to higher altitudes (Bougher et al., 2006). Wind measurements (Lellouch et al.,
1994) suggested that a meridional component to the circulation
might be present in the lower thermosphere, which is consistent
with our observations. Indeed, we showed that this region is also
probably where the upper branch of the Hadley cell circulation
occurs. However, the understanding of all these processes is still
patchy, and the circulation of the Venus’ meso- and thermosphere
is thus far from being fully known (Limaye and Rengel, 2013).
The high variability of the Venus atmosphere and in particular
that of the trace gases such as CO, is also not yet fully understood.
Models of the atmosphere are still unable to reproduce any such
variability. Simultaneous measurements of CO, temperature and
wind measurements in the lower thermosphere of Venus (Clancy
et al., 2012a, b) showed that although the circulation in that region exhibits large-scale instability, there was no clear association
between the amplitudes of the SSAS and zonal winds and the observed temperature and CO spatial and temporal variability. They
suggested that a distinctive circulation pattern would force largescale air masses downward, creating dynamic increases in temperature and CO mixing ratio. Interestingly O2 airglow observations
around 96 km showed small scale-features (∼100 km) embedded
58
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
in larger scale structure (∼10 0 0–30 0 0 km) (Hueso et al., 2008).
Those small scale features seemed to be correlated to regions of
strong subsidence associated with downward flows increasing the
volume concentration of O atoms and inducing adiabatic warming,
while the larger structures had their own motion characterized by
timescale of a few hours with day-to-day changes. Clearly, more
observations of the atmosphere of Venus are needed to even try
to grasp its complexity. With the end of the VEx mission, we now
have to rely on ground-based observations, which unfortunately do
not probe the complete range of altitudes required to have a global
view of the different processes at play.
5. Conclusions
Carbon monoxide densities have been measured by the SOIR
instrument on board Venus Express on a regular basis since the
beginning of the mission and up to its end in December 2014. Observations cover both hemispheres from 90S to 90 N latitudes, and
are all obtained at the terminator (LST 6 AM and 6 PM). Here we
have presented the complete set of vertical profiles of CO densities and volume mixing ratios (vmr) obtained by SOIR. These profiles are spanning the 65-150 km altitude range. The high variability observed on the short term, which was investigated in a
previous paper (Vandaele et al., 2015), can reach one order of magnitude on less than 1 month periods, even from one day to the
other. SOIR does not observe a marked long term trend, except
perhaps at the beginning of the mission where an increase of CO
density was observed. Similar increases have been noticed in other
species. Evening abundances are systematically higher than morning values at altitudes above 105 km, but the reverse is observed
at lower altitudes. Higher abundances are observed at the equator than at the poles for altitude higher than 105 km, but again
the reverse is seen at altitudes lower than 90 km. This illustrates
the complexity of the 90-100 km region of the Venus atmosphere
where different wind regimes are at play.
Acknowledgments
The research program was supported by the Belgian Federal Science Policy Office and the European Space Agency (ESA – PRODEX
program – contracts C 90323, 90113, and C40 0 0107727). The research was performed as part of the “Interuniversity Attraction
Poles” programme financed by the Belgian government (PlanetTOPERS) and a BRAIN research grant BR/143/A2/SCOOP. The authors
also recognize the support from the FP7 EuroVenus project (G.A.
606798) as well as that of the HiResMIR International Scientific
Coordination Network (GDRI, CNRS). AM would like to thank the
FNRS for his chargé de recherches position.
References
Bertaux, J.L., Nevejans, D., Korablev, O., et al., 2007. SPICAV on Venus Express: Three
spectrometers to study the global structure and composition of the Venus atmosphere. Planet. Space Sci. 55, 1673–1700.
Bézard, B., de Bergh, C., Crisp, D., et al., 1990. The deep atmosphere of Venus revealed by high-resolution nightside spectra. Nature 345, 508–511.
Bougher, S.W., Alexander, M.J., Mayr, H.G., 1997. Upper Atmosphere Dynamics:
Global Circulation and Gravity Waves, Venus II. The University of Arizona Press,
Tucson, pp. 259–291.
Bougher, S.W., Rafkin, S., Drossart, P., 2006. Dynamics of the Venus upper atmosphere: Outstanding problems and new constraints expected from Venus express. Planet. Space Sci. 54, 1371–1380.
Clancy, R.T., Muhleman, D.O., 1985. Chemical-dynamical models of the Venus mesosphere based upon diurnal microwave CO variations. Icarus 64, 183–204.
Clancy, R.T., Muhleman, D.O., 1985. Diurnal CO variations in the Venus mesophere
from CO microwave spectra. Icarus 64, 157–182.
Clancy, R.T., Muhleman, D.O., 1991. Long-term (1979-1990) changes in the thermal,
dynamical and compositional structure of the Venus mesosphere as inferre from
microwave spectral line observations of 12 CO, 13 CO, and C18 O. Icarus 89, 129–
146.
Clancy, R.T., Sandor, B., Moriarty-Schieven, G., 2008. Venus upper atmospheric
co, temperature, and winds across the afternoon/evening terminator from
june 2007 JCMT sub-millimeter line observations. Planet. Space Sci. 56, 1344–
1354.
Clancy, R.T., Sandor, B., Moriarty-Schieven, G., 2012. Circulation of the Venus upper mesosphere/lower thermosphere: Doppler wind measurements from 2001–
2009 inferior conjunction, sub-millimeter CO absorption line observations.
Icarus 217, 794–812.
Clancy, R.T., Sandor, B., Moriarty-Schieven, G., 2012. Thermal structure and CO distribution for the Venus mesosphere/lower thermosphere: 20 01–20 09 inferior conjunction sub-millimeter CO absorption line observations. Icarus 217, 779–793.
Collard, Taylor, F., Calcutt, S., Carlson, R., et al., 1993. Latitudinal distribution of carbon monoxide in the deep atmosphere of Venus. Planet. Space Sci. 41, 487–494.
Connes, P., Connes, J., Kaplan, L., et al., 1968. Carbon monoxide in the Venus atmosphere. Astrophys. J. 152, 731–743.
Crovisier, J., Lellouch, E., de Bergh, C., et al., 2006. Carbon monoxide emissions at
4.7 μm from Venus’ atmosphere Planet. Space Sci. 54, 1398–1414.
de Bergh, C., Moroz, V.I., Taylor, F.W., et al., 2006. The composition of the atmosphere of Venus below 100 km altitude: An overview. Planet. Space Sci. 54,
1389–1397.
Fedorova, A., Korablev, O., Vandaele, A.C., et al., 2008. HDO and H2 O vertical distributions and isotopic ratio in the Venus mesosphere by Solar Occultation
at Infrared spectrometer onboard Venus Express. J. Geophys. Res 113, E00B22.
doi:10.1029/20 08JE0 03146 .
Gilli, G., Lopez-Valverde, M.A., Peralta, J., et al., 2015. Carbon monoxide and temperature in the upper atmosphere of Venus from VIRTIS/Venus Express non-LTE
limb measurements. Icarus 248, 478–498.
Gurwell, M., Muhleman, D.O., Shah, K., et al., 1995. Observations of the CO bulge on
Venus and implications for mesospheric winds. Icarus 115, 141–158.
Hedin, A.E., Niemann, H.B., Kasprzak, W.T., 1983. Global empirical model of the
Venus thermosphere. J. Geophys. Res. 88, 73–83.
Hueso, R., Sanchez-Lavega, A., Piccioni, G., et al., 2008. Morphology and dynamics of Venus oxygen airglow from Venus Express/Visible and Infrared Thermal
Imaging Spectrometer observations. J. Geophys. Res. 113, E00B02. doi:10.1029/
20 08JE0 03081.
Irwin, P.G.J., de Kok, K., Negrao, A., et al., 2008. Spatial variability of carbon monoxide in Venus’ mesosphere from Venus Express/VIRTIS measurements. J. Geophys. Res. 113, E00B01. doi:10.1029/2008JE003093.
Iwagami, N., Yamaji, T., Ohtsuki, S., et al., 2010. Hemispherical distribution of CO
above the Venus’ clouds by ground-based 2.3 μm spectroscopy. Icarus 207, 558–
563.
Keating, G., Bertaux, J.L., Bougher, S.W., et al., 1985. Models of Venus neutral upper
atmosphere: structure and composition. Adv. Space Res. 5, 117–171.
Krasnopolsky, V., Parshev, V.A., Hunten, D.M., Colin, L., Donahue, T.M., Moroz, L.V.,
1983. Photochemistry of the Venus atmosphere. Venus. University of Arizona
Press, Tucson, pp. 431–458.
Krasnopolsky, V.A., 2008. High-resolution spectroscopy of Venus: Detection of OCS,
upper limit to H2S, and latitudinal variations of CO and HF in the upper cloud
layer. Icarus 197, 377–385.
Krasnopolsky, V.A., 2010. Spatially-Resolved High-Resolution Spectroscopy of Venus.
1. Variations of CO2, CO, HF, and HCl at the Cloud Tops. Icarus 208, 539–547.
Krasnopolsky, V.A., 2012. A photochemical Model for the Venus Atmosphere at 47112 km. Icarus 218, 230–246.
Lee, C., Lewis, S.R., Read, P.L., 2007. Super-rotation in a Venus General Circulation
Model. J. Geophys. Res. 112, E04S11.
Lellouch, E., Goldstein, J., Rosenqvist, J., et al., 1994. Global circulation, thermal
structure and Carbon Monoxide distribution in Venus’ mesosphere in 1991.
Icarus 110, 315–339.
Limaye, S., Rengel, M., Bengtsson, L., Bonnet, R.-M., Grinspoon, D., Koumoutsaris, S.,
Lebonnois, S., Titov, D., 2013. Atmospheric ciculation and dynamics. Towards
Understanding the Climate of Venus. Applications of Terrestrial Models to Our
Sister Planet. Springer, New York, pp. 55–70.
Mahieux, A., Berkenbosch, S., Clairquin, R., et al., 2008. In-Flight performance and
calibration of SPICAV SOIR on board Venus Express. Appl. Optics 47, 2252–2265.
Mahieux, A., Vandaele, A.C., Bougher, S.W., et al., 2015. Update of the Venus density
and temperature profiles at high altitude measured by SOIR on board Venus
Express. Planet. Space Sci. 113-114, 309–320.
Mahieux, A., Vandaele, A.C., Drummond, R., et al., 2010. Densities and temperatures
in the Venus mesosphere and lower thermosphere retrieved from SOIR onboard
Venus Express: Retrieval technique. J. Geophys. Res. 115, E12014. doi:10.1029/
2010JE003589.
Mahieux, A., Vandaele, A.C., Robert, S., et al., 2015. Venus mesospheric sulfur dioxide
measurement retrieved from SOIR on board Venus express. Planet. Space Sci.
113-114, 193–204.
Mahieux, A., Vandaele, A.C., Robert, S., et al., 2012. Densities and temperatures in
the Venus mesosphere and lower thermosphere retrieved from SOIR on board
Venus Express: Carbon dioxide measurements at the Venus terminator. J. Geophys. Res. 117, E07001. doi:10.1029/2012JE004058.
Mahieux, A., Wilquet, V., Vandaele, A.C., et al., 2015. Hydrogen Halides measurements in the Venus upper atmosphere retrieved from SOIR on board Venus Express. Planet. Space Sci. 113-114, 264–274.
Marcq, E., Bézard, B., Drossart, P., et al., 2008. A latitudinal survey of CO, OCS, H2 O,
and SO2 in the lower atmosphere of Venus: Spectroscopic studies using VIRTISH. J. Geophys. Res. 113, E00B07. doi:10.1029/2008JE003074.
Marcq, E., Bézard, B., Encrenaz, T., et al., 2005. Latitudinal variations of CO and OCS
in the lower atmosphere of Venus from near-infrared nightside spectro-imaging.
Icarus 179, 375–386.
A.C. Vandaele et al. / Icarus 272 (2016) 48–59
Marcq, E., Encrenaz, T., Bézard, B., et al., 2006. Remote sensing of Venus’ lower atmosphere from ground-based IR spectroscopy: Latitudinal and vertical distribution of minor species. Planet. Space Sci. 54, 1360–1370.
Marcq, E., Lellouch, E., Encrenaz, T., et al., 2015. Search for horizontal and vertical
variations of CO in the day and night side lower mesosphere of Venus from
CSHELL/IRTF 4.53 μm observations. Planet. Space Sci. 113-114, 256–263.
Mills, F.P., Allen, M., 2007. A review of selected issues concerning the chemistry in
Venus’ middle atmosphere. Planet. Space Sci. 55, 1729–1740.
Mills, F.P., Esposito, L.W., Yung, Y.L., 2007. Atmospheric composition, chemistry and
clouds. Geophys. Monogr. Series 176, 73–100.
Nevejans, D., Neefs, E., Van Ransbeeck, E., et al., 2006. Compact high-resolution
space-borne echelle grating spectrometer with AOTF based on order sorting for
the infrared domain from 2.2 to 4.3 micrometer. Appl. Optics 45, 5191–5206.
Pollack, J.B., Dalton, J.B., Grinspoon, D.H., et al., 1993. Near-infrared light from Venus’
nightside: A spectroscopic analysis. Icarus 103, 1–42.
Seiff, A., Schofield, J.T., Kliore, A., et al., 1985. Models of the structure of the atmosphere of Venus from the surface to 100 kilometers altitude. Adv. Space Res. 5,
3–58.
Taylor, F.W., 1995. Carbon monoxide in the deep atmosphere of Venus. Adv. Space
Res. 16, 81–88.
Taylor, F.W., Grinspoon, D.H., 2009. Climate evolution of Venus. J. Geophys. Res. 114,
E00B40. doi:10.1029/2008JE003316.
Titov, D.V., Koschny, H., Hoofs, D., et al., 2006. Venus Express science planning.
Planet. Space Sci. 54, 1279–1297.
Tsang, C.C.C., Irwin, P.G.J., Wilson, C.F., et al., 2008. Tropospheric carbon monoxide
concentrations and variability on Venus from Venus Express/VIRTIS-M observations. J. Geophys. Res. 113, E00B08.
59
Tsang, C.C.C., Taylor, F.W., Wilson, C.F., et al., 2009. Variability of CO concentrations
in the Venus troposphere from Venus Express/VIRTIS using a Band Ratio Technique. Icarus 201, 432–443.
Vandaele, A.C., Chamberlain, S., Mahieux, A., et al., 2016. Contribution from
SOIR/VEX to the updated Venus International Reference Atmosphere (VIRA).
Adv. Space Res. 57, 443–458.
Vandaele, A.C., De Mazière, M., Drummond, R., et al., 2008. Composition of the
Venus mesosphere measured by SOIR on board Venus Express. J. Geophys. Res.
113, E00B23. doi:10.1029/2008JE003140.
Vandaele, A.C., Drummond, R., Mahieux, A., et al., 2013. Trace Gases in the Mesosphere and Lower Thermosphere of Venus from SOIR/VEX Observations. International Venus Workshop, Catania, Sicily, pp. 10–14 June.
Vandaele, A.C., Mahieux, A., Robert, S., et al., 2013. Improved Calibration of
SOIR/Venus Express Spectra. Optics Express 21, 21148–21161.
Vandaele, A.C., Mahieux, A., Robert, S., et al., 2015. Carbon monoxide short term
variability observed on Venus with SOIR/VEX. Planet. Space Sci. 113-114, 237–
255.
von Zahn, U., Moroz, V.I., 1985. Composition of the Venus atmosphere below 100 km
altitude. Adv. Space Res. 5, 173–195.
Wilson, W.J., Klein, M.J., Kahar, R.K., et al., 1981. Venus I. Carbon monoxide distribution and molecular-line searches. Icarus 45, 624–637.
Yung, Y.L., DeMore, W.B., 1982. Photochemistry of the stratosphere of Venus: Implications for atmospheric evolution. Icarus 51, 199–247.
Zasova, L.V., Ignatiev, N.I., Khatuntsev, I., et al., 2007. Structure of the Venus atmosphere. Planet. Space Sci. 55, 1712–1728.
Zasova, L.V., Moroz, V.I., Linkin, V.M., et al., 2006. Structure of the Venusian Atmosphere from Surface up to 100 km. Cosmic Res. 44, 364–383.