Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Planetary Science. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 22 June 2021

# Composition and Chemistry of the Neutral Atmosphere of Venus

• Ann Carine VandaeleAnn Carine VandaeleBelgian Institute for Space Aeronomy

### Summary

The atmosphere of Venus is quite different from that of Earth: it is much hotter and denser. The temperature and pressure at the surface are 740 K and 92 atmospheres respectively. Its atmosphere is primarily composed of carbon dioxide (96.5%) and nitrogen (3.5%), the rest being trace gases such as carbon monoxide (CO), water vapor (H2O), halides (HF, HCl), sulfur-bearing species (SO2, SO, OCS, H2S), and noble gases. Sulfur compounds are extremely important in understanding the formation of the Venusian clouds which are believed to be composed of sulfuric acid (H2SO4) droplets. These clouds completely enshroud the planet in a series of layers, extending from 50 to 70 km altitude, and are composed of particles of different sizes and different H2SO4/H2O compositions. These act as a very effective separator between the atmospheres below and above the clouds, which show very distinctive characteristics.

### Subjects

• Planetary Atmospheres and Oceans

Comparative planetology helps investigate the global processes occurring within different atmospheres and provides a wider perspective for the understanding of life and habitability. By comparing a wide range of compositions, chemistries, and climates, driving mechanisms can be better understood. Planetary atmospheric dynamics and chemistry are governed by the same physical laws in all atmospheres, only the values and ranges of the parameters—composition, thermal and radiation balance, internal structure, existence of a magnetic field, and so on—vary from one planet to another.

Venus is a wonderful natural laboratory to investigate the way in which quite different outcomes can be reached, although starting from similar building blocks. Understanding how the atmosphere of Venus evolved will help us to understand and better place in perspective the atmospheric evolution and climate of our own planet.

The atmosphere of Venus, which is mostly composed of $CO2$ (96.5%) and $N2$ (3.5%), with other chemical species present in trace amounts, such as $CO,H2O,NO,OCS,SO2,HCl$, and $HF$ (see Table 1 for a detailed inventory of the atmosphere), is much hotter and denser than that of Earth, with temperature and pressure reaching 740 K and 92 atmospheres respectively at its surface. Cloud structures (Titov, Ignatiev, McGouldrick, Wilquet, & Wilson, 2018), made of sulfuric acid droplets ($H2O·H2SO4$), completely enshroud the planet, physically separating the lower part of the atmosphere from the layers above.

#### Table 1: Chemical Composition of the Atmosphere of Venus

Venus and Earth formed from the same interstellar gas and dust, and therefore their initial composition should be very similar (Grinspoon, 2013). However liquid water has always been present at the surface of Earth, while Venus evolved towards a dry and hot planet. Some have proposed that Venus could have been formed in a drier region of the Solar nebulae, which would explain the current depletion in water (Lewis, 1974). But observations clearly indicate that, once in its history, Venus had more water than in the present epoch. Indeed, Venus might have harbored enough water to form an ocean. However, all this water was lost over time (Bullock & Grinspoon, 2013; Kasting, 1988). Water is photodissociated by ultraviolet (UV) radiation emitted by the Sun and reaching the upper layers of the atmosphere, generating hydrogen $(H)$ and oxygen $(O)$ atoms. The hydrogen atoms can escape the planet’s gravitational field under favorable conditions, i.e. when their thermal energy is high enough. This escape is confirmed by measurements of present-day D/H ratios, which demonstrate an overabundance of deuterium (heavy hydrogen, D) relative to hydrogen. The free oxygen produced by photo-dissociation reacted to create both carbon dioxide and sulfur dioxide.

Thermal escape of hydrogen is a slow process; at least hundreds of millions of years were required for Venus to lose its ocean (Kasting, 1988). During that time water vapor evaporating from the ocean and trapped in the atmosphere contributed to the increase of the atmosphere temperature, through the greenhouse effect. Water vapor is an even more potent greenhouse gas than carbon dioxide. Increasing the temperature also accelerated the evaporation of the ocean, leading to a situation where all the water entered the atmosphere, leading to eventual loss by escape of H following photodissociation. At some point, the surface of Venus became so hot that the carbon trapped in rocks sublimated into the atmosphere and combined with oxygen to form even more carbon dioxide, creating the dense atmosphere observed. The accumulation of $CO2$ was boosted by the lack of any efficient $CO2$ cycle, transferring $CO2$ back to the crust, as is happening on Earth: the $CO2$ emitted into the atmosphere, remained in the atmosphere.

### Chemistry of the Venusian Atmosphere

The chemistry in Venus’s atmosphere is controlled by ion–neutral and ion–ion reactions in the ionosphere, by photochemistry within and above the cloud global layer, and by thermal equilibrium chemistry, which prevails near the surface (Mills, Esposito, & Yung, 2007).

Above ~100 km altitude, the atmosphere of Venus is less dense and overlaps with the ionosphere: ion–neutral and ion–ion reactions along with photodissociation processes control the chemistry. In the middle atmosphere (~60–100 km), above the cloud layer, UV radiation from the Sun drives the processes through photochemistry. Little UV radiation reaches the lower atmosphere below 60 km, while the temperature gradually increases towards the surface. The chemistry is controlled by thermal processes, called “thermodynamic equilibrium” chemistry, or thermochemistry. Venus’s clouds clearly define the transition between the lower and middle atmospheres, i.e. they separate the photochemistry-dominant and the thermochemistry-dominant regions (Esposito, Bertaux, Krasnopolsky, Moroz, & Zasova, 1997). The cloud and haze region extends from 30 to 90 km in altitude, with the main cloud deck located at 45–70 km. Within this regions, heterogeneous chemistry on aerosols and cloud particles may also play an important role (Mills, Sundaram, Slanger, Allen, & Yung, 2006).

The first studies of the chemistry of the lower atmosphere were based on the assumption that thermochemical equilibrium is reached at each altitude (Esposito et al., 1997). This assumption is valid if the time to reach equilibrium is less than the mixing time, which is of the order of a few years in that region. Krasnopolsky and Parshev (1983) and Krasnopolsky and Pollack (1994) suggested that thermochemical equilibrium could serve as an approximation near the surface, but is generally invalid in the lower atmosphere. Gas-phase thermal equilibrium can only be reached below 1 km above the surface; above that level the atmosphere is more oxidizing than predicted by thermodynamics only. Fegley, Zolotov, Lodders, and Widemann, (1997) combined thermodynamics and chemical kinetics to study the lowest 10 km layer of the atmosphere. Finally, a self-consistent chemical kinetic scheme was developed for the lower atmosphere that does not apply the approximation of thermochemical equilibrium (Krasnopolsky, 2007, 2013b). The near-surface layer is dominated by surface–atmosphere exchange and interaction.

Very few observations of the atmosphere below 7 km above the surface exist; no precise measurements of N2 abundance have been reported (Von Zahn, Kumar, Niemann, & Prinn, 1983). Only the VEGA-2 probe has delivered a temperature profile (VEGA Balloon Science Team, & Seiff, 1987; Zasova, Moroz, Linkin, Khatountsev, & Maiorov, 2006), which indicated a very unstable vertical gradient near the surface. Lebonnois and Schubert (2017) proposed that this could be explained by a change in the atmospheric composition, with the presence of a non-homogeneous layer close to the surface, in which N2 abundance gradually decreases to near-zero at the surface. They further suggested that the atmosphere would not be in a gaseous phase anymore, but rather a super-critical $CO2/N2$ fluid mixture, i.e. the atmosphere is neither a gas, nor a liquid, but, as a mixture of gas and supercritical fluid, has some properties of both.

Two dominant chemical cycles have been identified in the Venus atmosphere (Krasnopolsky & Lefèvre, 2013; Mills & Allen, 2007; Mills et al., 2007)—the carbon dioxide and sulfur cycles.

The $CO2$ cycle includes the photodissociation by UV radiation of $CO2$ on the dayside

$Display mathematics$

followed by the transport of most of the $CO$ and $O$ produced to the nightside, where recombination of $O$ atoms produces excited oxygen $O2(a1Δg)$ (Gérard et al., 2017).

$Display mathematics$

The de-excitation of oxygen to its ground state can occur through quenching

$Display mathematics$

or through

$Display mathematics$

during which radiation, $O2$ airglow emission is emitted at its characteristic wavelength of 1.7 μ‎m.

The oxygen nightglow, which is illustrated in Figure 1, has been reported by a series of ground-based observations and by VIRTIS on board Venus Express. The investigation of the VIRTI-M spectra (Gérard, Soret, Piccioni, & Drossart, 2014; Soret, Gérard, Montmessin, Piccioni, Drossart, & Bertaux, 2012) indicated that the emission peaks at an altitude around 97–99 km (Piccioni et al., 2009; Soret, Gérard, Montmessin, et al., 2012). The $O2$ airglow is characterized by a bright area of maximum emission centered near the antisolar point. These observations proved not only that the proposed chemistry was correct but also that the circulation model at those altitudes were correctly assuming a solar–antisolar circulation.

The direct recombination

$Display mathematics$

is spin-forbidden and is much slower than the $O$ recombination to form $O2$. Photolysis would completely remove $CO2$ above the clouds in ~14,000 years and all $CO2$ from the whole atmosphere in ~5 million years. Moreover this scheme should produce observable amounts of $O2$, which is not the case. Only upper limits of $O2$ have been detected, which would correspond to a uniform mixing ration of 0.3–2 ppm above 60 km (Krasnopolsky, 2006b; Mills, 1999; Trauger & Lunine, 1983). This points to the probable production of $CO2$ through catalytic reactions. Several processes have been proposed, one of the most promising being a cycle of reactions involving the chlorine family (Krasnopolsky & Lefèvre, 2013; Krasnopolsky & Parshev, 1983; Mills & Allen, 2007):

$Display mathematics$
$Display mathematics$
$Display mathematics$
$Display mathematics$
$Display mathematics$

This mechanism has been confirmed in the laboratory (Pernice et al., 2004), by showing that the peroxychloroformyl radical $(CIC(O)OO)$ is thermally and photolytically stable under the conditions prevailing in Venus’s mesosphere. None of the intermediate products have been observed in the atmosphere. However, Sandor and Clancy (2018) detected $ClO$ (the final product of the cycle when $X=Cl$) in the nightside atmosphere. This observation would corroborate the possibility of the chlorine catalytic cycle. $ClO$ would then combine with atomic oxygen to produce $O2$ and $Cl$,

$Display mathematics$

This reaction combined with the above cycle would lead to the net reaction:

$Display mathematics$

However, even considering these reactions, models still predict too high $O2$ abundances above the clouds (by a factor 10). One of the most critical parameters of the models to which the abundance of $O2$ is very sensitive, is the thermal stability of $ClCO$. Recent assessment of this parameter would indicate that the proposed catalytic cycle might not be as efficient as first thought (Marcq, Mills, Parkinson, & Vandaele, 2017; Mills et al., 2007).

Another mechanism has been proposed (Mills et al., 2006): the oxidization of $CO$ through reactions on or within aerosols or cloud particles. This would reduce the modeled value of $O2$ abundance while, however, also reducing the abundance of $CO$ in the lower mesosphere to levels significantly lower than those observed (Mills & Allen, 2007).

The sulfur oxidation cycle starts with upward transport of $SO2$ and its oxidation to form $SO3$ subsequently forming $H2SO4$. Models have shown that $H2SO4$ is formed in a thin layer of the atmosphere, centered at 66 km (Krasnopolsky & Lefèvre, 2013). Condensation of $H2SO4$ and $H2O$ forms the clouds and haze.

$Display mathematics$
$Display mathematics$
$Display mathematics$
$Display mathematics$

The sulfuric acid in the form of cloud droplets is transported by the meridional circulation to the poles, where it is transported downward (Imamura & Hashimoto, 1998). Eventually the droplets evaporate and $H2SO4$ is decomposed to regenerate $SO2$.

$Display mathematics$
$Display mathematics$

This cycle of oxidation of $SO2$ to $H2SO4$ followed by condensation, transport, evaporation and decomposition has been called the “fast atmospheric sulfur cycle” (Von Zahn et al., 1983) (see Figure 2).

Another cycle, the “slow sulfur cycle,” involving photochemical and thermochemical processes with reduced sulfur-bearing species $(OCS, H2S)$ has been proposed by Prinn (1978), based on the assumption that these species, and in particular $OCS$, were the dominant sulfur-bearing species in the atmosphere of Venus. According to this cycle, the following photochemical reactions would occur in the middle atmosphere:

$Display mathematics$
$Display mathematics$
$Display mathematics$

while in the lower atmosphere, below the clouds, $H2SO4$ thermally decomposes:

$Display mathematics$
$Display mathematics$
$Display mathematics$

Observations have demonstrated that $OCS$ and $H2S$ are not the major sulfur species, but rather $SO2$. Revised schemes have been proposed (Krasnopolsky, 2007, 2013b, 2016; Mills & Allen, 2007; Yung et al., 2009).

A variation of the slow cycle has been proposed which involves the production of polysulfur $Sx$. Upward transport of $SO2$ and $OCS$ is followed by photodissocation, generating $S$ atoms

$Display mathematics$
$Display mathematics$

which can react via

$Display mathematics$

to produce $S2$. Production of higher polysulfurs is possible through successive addition reactions such as

$Display mathematics$

Usually the term polysulfur refers to $Sx$ with $x>3$. Polysulfurs absorb strongly in the UV, and have been proposed as being the still-unidentified UV absorber in the upper atmosphere of Venus (Mills et al., 2007; Titov et al., 2018). Krasnopolsky (2017) showed, however, that the production of polysulfurs is negligible above the clouds, and therefore can contribute only partially to near-UV absorption; they cannot explain it by themselves.

The polysulfur branch of the slow sulfur cycle is completed in the lower atmosphere by

$Display mathematics$

The fast and slow cycles both imply that there must be a significant downward flux of $CO$ from the middle to the lower atmosphere, balanced by an upward flux of $CO2$. Observations, however, favor conversion of $CO$ to $OCS$ at 30–45 km altitude. The model of Krasnopolsky (2013b) is a slightly modified version of the reaction set involved in the conversion of $CO$ to $OCS$ and requires a smaller flux of $CO$ from the middle atmosphere. Nevertheless the exchange of constituents through the cloud layer requires more study both in modeling and observations.

There is a third sulfur cycle, involving surface–atmosphere interaction—the geological cycle. This cycle starts with the outgassing of reduced gases, $COS$ and $H2S$, from the crust. These weathering reactions are slow compared to the other cycles, and there are still no accurate measurements of the mineralogy of the surface of the planet. Different surface compositions have been proposed, including carbonate ($CaCO3$), wollastonite ($CaSiO3$), anhydrite ($CaSO4$), magnetite ($Fe3O4$) or pyrite ($FeS2$). The following reactions have been proposed (Fegley & Treiman, 1992; Von Zahn et al., 1983):

$Display mathematics$
$Display mathematics$

In the case of these reactions, $OCS$ and $H2S$ are formed, which will evolve into $SO2$ through photochemical reactions. When $SO2$ abundances exceed the thermochemical equilibrium relative to the surface minerals, $SO2$ will react, through thermochemical reactions, with carbonate rocks to form anhydrite, which will be transformed back to pyrite, closing the cycle.

$Display mathematics$
$Display mathematics$

### Observations of the Venusian Atmosphere

Lomonosov was the first to suspect the existence of an atmosphere around Venus, thanks to his observations of the 1761 Venus transit (Lomonosov, 1955): several observers reported the appearance of a halo when Venus entered and exited the Sun’s disk, which was attributed to the presence of an atmosphere on Venus. The systematic observation of Venus started in the early 20th century: the clouds’ structure and UV markings on the clouds were discovered (Dolfus, 1953); $CO2$ was detected (Adams & Dunham, 1932) and was shown to be the main constituent of the atmosphere (Belton, Hunten, & Goody, 1968; Connes & Connes, 1966); the greenhouse effect from $CO2$ combined with the role of the aerosols (clouds and hazes) was proposed to explain the high surface temperature, an idea which was further developed by Sagan (1960) in the early 1960s and confirmed by microwave observations from Earth (Drake, 1962; Mayer, McCullough, & Sloanaker, 1957) and a fly-by of Venus by Mariner 2 (Barath, Barrett, Copeland, Jones, & Lilley, 1964). This was the beginning of a long series of observations of the planet by different methods and instruments.

What we know of the atmosphere of Venus has been inferred from observations performed from Earth-based telescopes, including the Hubble Space Telescope, from space missions—the Venera and Mariner missions, Pioneer Venus, Magellan (see Fegley, 2004, for a review of these missions), and more recently Venus Express, and Akatsuki, and from entry probes and balloons—Venera missions, Pioneer Venus, and the VEGA probes.

Our knowledge of the composition of Venus’s atmosphere below the clouds comes essentially from observations of the thermal radiation emitted by the surface and different layers of the atmosphere, escaping through several near-infrared (IR) spectral atmospheric windows. These observations are only possible during nighttime, when there is no contribution from solar radiation. Thermal radiation emitted by the surface and lower atmosphere escapes towards space in a series of spectral atmospheric windows located between strong absorption features from $CO2$ and $H2O$. Below 2.4 μ‎m, $H2SO4$ cloud droplets scatter light conservatively, allowing light to escape, while above 2.5 μ‎m, the refractive index of $H2SO4$ changes and cloud droplets start to absorb strongly, preventing any longer-wavelength radiation escaping through the clouds. Following their discovery by Allen and Crawford (1984), several transparency windows have been identified between 1.0 and 2.4 μ‎m. These emissions can only be observed during the night, since their magnitude is about four times weaker than the reflected solar contribution. Radiation in each of the windows is emitted at a particular altitude range: radiation around 1.01 μ‎m is emitted from the surface; 1.10 and 1.18 μ‎m from the atmosphere between the surface and an altitude of 15 km), 1.27 μ‎m from 15–30 km altitude; 1.31 μ‎m from 30–50 km altitude, 1.74 μ‎m from 15–30 km altitude, and 2.3 μ‎m from 26–45 km. By combining abundances retrieved from different windows, vertical profiles can be reconstructed.

To sound the upper layers of the atmosphere (above the clouds) different techniques are used. Ground-based observation in the UV, IR, sub-millimeter, and millimeter ranges, delivers global views and variations of abundances with latitude and longitude, seasons, and local times. These also include observations from the Hubble Space Telescope. Ground-based observatories provide monitoring of the atmosphere composition over long periods of time, in particular in between space missions. Spaceborne observations sound the atmosphere both in nadir geometry and solar/stellar occultation or limb viewing. They provide respectively integrated abundances and high vertically resolved profiles of abundances. Recently, two missions were dedicated to atmospheric observations, Venus Express (VEX) from ESA, the European Space Agency (Svedhem et al., 2007, Svedhem, Titov, Taylor, & Witasse, 2009; Taylor, Svedhem, & Head, 2018), and Akatsuki from JAXA, the Japanese Space Agency (Nakamura et al., 2007, 2016).

On VEX, different instruments worked together to sound the atmospheric composition from the surface up to the upper layers: VIRTIS (Piccioni et al., 2007) performed both nadir and limb observations in the infrared; SPICAV/SOIR (Bertaux, Nevejans, et al., 2007; Nevejans et al., 2006) probed the atmosphere using nadir and solar/ stellar observations in the infrared and the UV; The VeRA (Pätzold et al., 2007) experiment used radio occultation to sound the atmosphere.

On board Akatsuki, several cameras were sensitive to different spectral ranges which allowed a detailed investigation of the dynamics of the atmosphere (Nakamura et al., 2017). The RS (Radio Science) experiment was sensitive to temperature and $H2SO4$ vapor between 35 and 100 km (Imamura et al., 2017). The images provided by the UVI camera showed the spatial distribution of $SO2$ and the unknown absorber at the cloud tops, and characterized the cloud-top morphologies and haze properties

Observations in the UV provide information on $SO$ and $SO2$ above the clouds, since several absorption bands of both molecules are located in this spectral range: electronic transition bands $B3Σ−−X3Σ−$ (190–240 nm) and $A3Π−−X3Σ−$ (240–260 nm) of the monoxide, and analogue bands $C˜1B1−X˜1A1$ (190–235 nm) and $B˜1B1−X˜1A1$ (250–340 nm) of the dioxide. The strongest absorption bands are located in the 190–235 nm interval for both gases.

Ground-based observations at radio wavelengths in the millimeter and sub-millimeter ranges provide information on abundances of trace gases (Clancy, Sandor, & Moriarty-Schieven, 2012; Encrenaz et al., 2016; Encrenaz, Moreno, Moullet, Lellouch, & Fouchet, 2015; Sandor & Clancy, 2005, 2017; Sandor et al., 2010). Sub-millimeter spectra correspond to spectrally resolved absorption in the Venus mesosphere superposed on the spectrally featureless, continuum emission originating from the deeper and warmer atmosphere. Such spectra are pressure-broadened and are therefore fundamentally sensitive to mixing ratio and not to number density. Moreover the altitude distribution can be retrieved based on the shape of the observed absorption lines.

### Composition of the Venusian Atmosphere

Venus’s atmosphere is dominantly $CO2$ and $N2$ with smaller amounts of other chemical species. In the following, trace gas abundances are expressed in number density or volume mixing ratio (VMR). VMR is a dimensionless quantity defined by the ratio of the partial density (pressure) of the species and the total density (pressure) of the atmosphere. VMR are given in percentages for the main constituents, and parts per million (ppm, 10−6), parts per billion (ppb, 10−9), or even parts per trillion (ppt, 10−12) for trace gases.

Several attempts to tabulate the composition of Venus’s atmosphere have been published in the past, each trying to summarize and reconcile observational data of their epochs.

The Venus International Reference Atmosphere (VIRA) was compiled by Kliore Moroz and Keating (1985) presenting a synthesis of the best data on the neutral atmosphere and the ionosphere of the planet available at that time. This was the first attempt to summarize all the knowledge about Venus, providing a common standard reference through tables and averages. The original VIRA chapter dealing with the structure of the atmosphere from the surface to 100 km of altitude (Seiff et al., 1985) included tables of the vertical temperature and density considering no day–night, latitudinal, or synoptic variations in the lower atmosphere, but allowing for latitudinal variations in the middle layers (33–100 km). One chapter (Keating et al., 1985) was devoted to the structure and composition of the upper atmosphere. It gathered tables with temperature, total density, and densities of $CO2$, $O$, $CO$, $He$, $N$, and $N2$ for altitudes from 100 km to 250 km, considering no latitudinal variations. Data from 100 to 150 km were given only for noon and midnight conditions, while those for 150 to 250 km were provided for different local times. Von Zahn and Moroz (1985) summarized the composition of Venus’s atmosphere below 100 km altitude as known at that time.

Since then, many missions have yielded new and valuable information. A first attempt to update the VIRA model (VIRA 2) was put forward by Moroz and Zasova (1997). They considered new data provided by the VEGA 1 and 2 UV spectrometers (Bertaux, Widemann, Hauchecorne, Moroz, & Ekonomov, 1996; Linkin et al., 1986), Venera 15 and 16 infrared spectrometers and radio occultation experiments (Moroz et al., 1986; Oertel et al., 1985; Zasova, 1995; Zasova et al., 1996; Yakovlev, Matyugov, & Gubenko, 1991), as well as Near-Infrared Mapping Spectrometer (NIMS) observations obtained during the Galileo fly-by of Venus in 1990 (Carlson et al., 1993; Collard et al., 1993; Grinspoon, 1993; Roos et al., 1993; Taylor, 1995). They also included re-analysis of previous data (from previous Venera missions and Pioneer Venus) and some ground-based observations for abundances of trace gases (Bézard, de Bergh, Crisp, & Maillard, 1990; de Bergh et al., 1995; Pollack et al., 1993). The structure of the Venusian atmosphere was updated by Zasova, Moroz, and Linkin (2006) taking into account measurements of the vertical temperature profile by the VEGA spacecraft and balloons, the radio occultation measurements of Magellan, Venera 15, and Venera 16, as well as the temperature profiles derived from the Venera15 IR spectrometer. They proposed a model of the atmosphere in the altitude range from 55 to 100 km consisting of tabulated temperature dependent on local time for five latitudinal regions (<35°, 35°–55°, 50°–70°, 70°–80°, 85°). Nothing similar had ever done for the composition of the Venusian atmosphere except for an updated version of the initial Table 5-1 of von Zahn and Moroz (1985) given in Moroz and Zasova (1997). However, the data corresponding to the higher altitudes (above 100 km) that were investigated in Keating et al. (1985) have never been updated.

Since then, several inventories have been compiled incorporating new observations (Bézard & de Bergh, 2007; de Bergh et al., 2006; Taylor, Crisp, & Bézard, 1997). The Venus Express mission and recent ground-based observations, however, yielded a wealth of new information on Venus’s atmosphere, from the surface up to the highest layers of the atmosphere (Gérard et al., 2017; Marcq et al., 2017).

Table 1 summarizes our current knowledge of Venus’s atmosphere. It provides the VIRA2 accepted abundances (Moroz & Zasova, 1997; von Zahn & Moroz, 1985), and the recommended values proposed by Taylor et al. (1997) and then lists more recent observations. These will be described in the following sections, each of them focused on one species or one species family (“Water Vapor,” “CO and OCS,” “Sulfur-Bearing Species,” “Halides (HBr, HF, HCl),” “Other Species (O2, OH, O3, ClO)”).

Most of these constituents have spatially and temporally variable abundances. The high variability of the atmosphere of Venus, and in particular that of the trace gases, is not yet fully understood. Models of the atmosphere are still unable to reproduce any such variability.

#### Water Vapor

The primary reservoir for $H$ on Venus lies in $H2O$ and in the $H2O·H2SO4$ cloud layers. Abundances of water vapor have been measured by numerous instruments, and it is now accepted that its abundance in the lower atmosphere is about 30 ppm (Bullock & Grinspoon, 2013; Marcq et al., 2017; Pollack et al., 1993; Taylor et al., 1997). This has been confirmed by more recent observations, both from Earth and from space. Ground-based observations of the lower atmosphere started after the discovery of the transparency windows by Allen and Crawford (1984). Chamberlain, Bailey, Crisp, and Meadows (2013) retrieved water abundance from spectra recorded by the Anglo-Australian Telescope between 0 and 15 km (31 ± 9 ppm). More recently, Arney et al. (2014) performed measurements using the Apache Point Observatory probing the atmosphere at different altitudes. These values were also confirmed by the observations by VIRTIS (Marcq et al., 2008; Tsang et al., 2010) and SPICAV-IR (Fedorova, Bézard, Bertaux, Korablev, & Wilson, 2015) both on board Venus Express.

The abundance of water has, up to now, been very poorly constrained above the clouds, indeed only few observations have been performed. The first reliable measurements were performed by Fink Larson, Kuiper, and Poppen (1972). VIRTIS-H/VEX carried out dayside observations which provided water abundance above the clouds (Cottini et al., 2012). They found that the abundance at cloud top was 3 ± 1 ppm between −40° and +40° latitude, increasing to 5 ppm at high latitudes. Observations of both $H2O$ and its heavy isotopologue HDO by SPICAV-SOIR were performed routinely during the whole Venus Express mission. First results of SOIR (Bertaux, Vandaele, et al., 2007; Fedorova et al., 2008; Vandaele, Chamberlain, et al., 2016) showed a depletion around 85 km in both $H2O$ and HDO, an increase of HDO/$H2O$ above the clouds, and no noticeable temporal variability. Measurements using SPICAV-IR (Fedorova et al., 2016) indicated abundances of $H2O$ of 2 to 11 ppm within the clouds and near the cloud top altitude (59–66 km).

Ground-based observations in the millimeter and sub-millimeter range showed that the VMR of water in the mesosphere (70–100 km) is constant with altitude ranging between 0 and 3.5 ppm (Sandor & Clancy, 2005). These measurements revealed strong temporal and spatial variations. Observations from the SWAS satellite performed between 2012 and 2014 (Gurwell, Melnick, Tolls, Bergin, & Patten, 2007) also confirmed that temporal variability of water in the mesosphere is important (variations of a factor of 50 over two days), and is dominant over the diurnal span. They suggested that these fluctuations might be due to moderate variations of the mesospheric temperature (by10–15 K) inducing water condensation. This is not confirmed by sub-millimeter observations using the ALMA telescope (Encrenaz et al., 2015) which reported a uniform vertical mixing of 2.5 ppm, with a potential maximum in the late afternoon (by a factor of 2 to 3), nor by IR observations by TEXES (Encrenaz et al., 2013, 2016, 2012) or by CSHELL (Krasnopolsky, Belyaev, Gordon, Li, & Rothman, 2013). Note that some of these observations rely on the measurement of HDO, and a conversion to $H2O$ abundances using a fixed D/H value. However all studies did not use the same values of D/H.

#### The D/H Ratio

The D/H isotopic ratio is highly variable among the different objects of the Solar System (Clayton, 2003) and is considered to be key in determining if Venus once had abundant water. Most of the processes that enable hydrogen isotopes to escape from the atmosphere strongly discriminate against the loss of deuterium, leading to an enrichment in the heavier isotope. The D/H ratio is usually derived from simultaneous measurements of $H2O$ and its heavy isotopologue, HDO, although it has also been derived from measurements of DCl/HCl and DF/HF. The bulk of Venus’s atmosphere exhibits a very high D/H ratio, more than 100 times the VSMOW value (Vienna Standard Mean Ocean Water) representative of Earth’s isotopic ratio. The first measurement of the D/H ratio was done by the neutral mass spectrometer on the Large Pioneer Venus Probe during its descent through the atmosphere, providing the value of 157 ± 30 (Donahue, Hoffman, Hodges, & Watson, 1982). De Bergh et al. (1991) reported a value of 120 ± 40 on the night side and lower atmosphere, which is compatible with the value of 135 ± 20 found by Marcq, Encrenaz, Bézard, and Birlan (2006) in the 30–40 km range. Reanalysis of Pioneer Venus data (Donahue, Grinspoon, Hartle, & Hodges, 1997) and ground-based dayside observations (Bjoraker, Larson, Mumma, Timmermann, & Montani, 1992) obtained similar results. Krasnopolsky et al. (2013) derived D/H values from observations of $HDO/H2O$, DCl/HCl and DF/HF reporting 95 ± 15, 190 ± 50, and 420 ± 200 respectively. The higher value found in the case of DCl/HCl was explained by the photochemistry of HCl which tends to enrich D in HCl in the mesosphere. The still higher D/H value derived from DF/HF may be explained by the fact that only DF was measured in Krasnopolsky et al. (2013) who used the averaged abundance of HF reported elsewhere (Krasnopolsky, 2010c; Vandaele et al., 2008). Using SOIR data, Fedorova et al. (2008) reported a value of 240 ± 25 from $H2O/HDO$ observations, while Vandaele et al. (2014) reported a value of 208 ± 155 from DF/HF measurements. The large error on this number is due to the difficulty in measuring DF abundance.

#### CO and OCS

Carbon monoxide $(CO)$ is the second most abundant carbon-bearing gas in Venus’s atmosphere. Its abundance (VMR) is altitude-dependent, decreasing towards the surface in the lower atmosphere, increasing with altitude above the clouds.

Carbonyl sulphide $(OCS)$ is the most abundant reduced sulfur gas in the sub-cloud atmosphere of Venus. Its abundance increases with decreasing altitude in the 25–50 km range by almost two orders of magnitude. Extrapolation of this gradient to the surface indicates that OCS abundance would reach tens of ppm at the surface. OCS above the clouds has been observed using the IRTF/CSHELL instrument (Krasnopolsky, 2010d), indicating a highly variable abundance (1–8 ppb) at 65 km altitude, with a scale height of 2.5 km.

$CO$ and $OCS$ have been observed below the clouds by ground-based instruments, see for example Bézard et al. (1990) for the first detection of $OCS$ and Pollack et al. (1993) for the first detailed interpretation of the nightside observations of $CO$ and $OCS$. Using the high-resolution spectra of VIRTIS-H, Marcq et al. (2008) showed that, at 36 km, $CO$ increased with latitude (24 ± 3 ppm to 31 ± 2 ppm from equator to 60°), while $OCS$ showed the reverse behavior. On the nightside, Tsang et al. (2008) observed tropospheric $CO$ and $OCS$ (2.5–4 ppm) at an altitude 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° (23 ± 2 ppm to 32 ± 2 ppm from equator to 60°). In fact, the first hint of a possible variation of $CO$ with 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. Ground-based observations by Arney et al. (2014) provided measurements of $CO$ and $OCS$ below the clouds that confirmed the previous results. A recent re-analysis of all VIRTIS-H data (Haus, Kappel, & Arnold, 2015) definitively confirmed the current picture: $CO$ abundances near 35 km increase from 23 ± 1 ppm near 10° S up to a maximum value (31 ± 2 ppm) in the 65–70° S latitude band, then slightly decrease poleward to 29.5 ± 2.5 ppm at 80° S.

Observations at millimeter and sub-millimeter wavelengths have shown that $CO$ abundance is very variable at high altitudes (around 100 km) (Clancy & Muhleman, 1985a, 1991; Clancy, Sandor, & Moriarty-Schieven, 2008, 2012; Gurwell et al., 1995). These measurements also identified diurnal variations in $CO$ VMR above 90 km altitude. The amplitude of these variations differed from year to year. They also revealed long-term variations in the $CO$ profiles between 75 and 105 km. It was suggested that the observed interannual variations of $CO$ are related to interannual variations in the circulation of Venus’s mesosphere. Gurwell et al. (1995) reported the existence of a CO bulge at 90–100 km on the nightside (with a maximum at 100 km and at 2 a.m. local time). This was explained (Clancy & Muhleman, 1991) by considering the general subsolar-to-antisolar circulation in the mesosphere and a zonal retrograde flow below, in addition to inefficient destruction of $CO$ on the nightside. This can lead to a convergence of the flow in the vicinity of the antisolar point, where the bulge is created by subsidence. The return flow occurs lower in the atmosphere.

Irwin et al. (2008), using VIRTIS-M/VEX observations, reported latitudinal distribution of $CO$ above the cloud level. They found an average value of 40 ± 10 ppm (at 65–70 km) with little variation in the middle latitudes. $CO$ densities at higher altitudes (100–150 km) were obtained by Gilli (2012); Gilli et al. (2015), using the 4.7 μ‎m non-LTE emission band of $CO$ measured by VIRTIS-H. Iwagami Yamaji, Ohtsuki, and Hashimoto (2010), observed a nearly uniform distribution of $CO$ above the clouds on the dayside, consistent with the findings of Krasnopolsky (2008). Using the $CO$ dayglow at 4.7 μ‎m, Krasnopolsky (2014) retrieved abundances of 560 ± 100 ppm near 104 km altitude and 40 ± 5 ppm near 74 km at low and middle latitude (within ± 50°). 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 latitudinal variations of CO on both day- and nightsides of Venus: $CO$ abundance was seen to increase from 35 ± 10 ppm at low latitude to 45 ppm past 30° N. Using VIRTIS-M data, Grassi et al. (2014) found that $CO$ exhibited a similar increase in the southern hemisphere between 40° S and 60° S (peaking at 70 ± 10 ppm), followed by a decrease poleward (60 ± 5 ppm at 70° S). SOIR on board Venus Express measured the CO profiles in the 65–150 km range of altitudes (Vandaele, Mahieux, et al., 2016; Vandaele et al., 2015). Evening abundances were found to be systematically higher than morning values at altitudes above 105 km, but the reverse was observed at lower altitudes. Higher abundances were observed at the equator than at the poles for altitude higher than 105 km (by a factor of 5 to 7), but again the reverse was seen at altitudes lower than 90 km (Figure 3.A). This illustrates the complexity of the 90–100 km region of the Venus’s atmosphere, where different wind regimes are at play.

Clancy et al. (2012) reported high variability of temperature and CO mixing ratio, both spatially and temporally. They showed periods where up to 24 K and 133% spatial variations of the temperature and the $CO$ VMR respectively were observed between 95 and 100 km. They also mentioned several cases of high temporal variability, with variations up to 100–200%. They definitively saw a correlation between a temperature increase above 95 km and an increase of the $CO$ abundance. They associated this correlation to the strong downward vertical advection, drawing the larger $CO$ VMR present at higher altitudes downward, leading to a strong compressional adiabatic heating. In Tsang et al. (2009), later confirmed by Tsang and McGouldrick (2017), $CO$ abundances obtained from VIRTIS observations are reported at an altitude of ~35 km and show unexpectedly high variability in space and time. At that time, the accepted scenario was that a Hadley-type cell would bring enriched $CO$ air from higher altitudes towards the lower layers of the atmosphere. However, the variability seen at 35 km led the authors to speculate that such a simple view was not consistent with the observations. They concluded that the meridional circulation could be more asymmetric than one single Hadley cell. This might imply that the strength of the downwelling might be variable. Such a complex circulation would also explain the high spatial and temporal variability seen in the SOIR data (Vandaele, Mahieux, et al., 2016; Vandaele et al., 2015).

Observations of tropospheric $CO$ (Tsang et al., 2008; Tsang & McGouldrick, 2017; Tsang et al., 2009), 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 General Circulation Modeling (see, e.g., Lee, Lewis, & Read, 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 & 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 (Figure 3.B). Cotton, Bailey, Crisp, and Meadows (2012) performed ground-based observation of CO below the clouds, which confirmed the proposed scheme. However, contrary to previous studies, they considered that the observed variability did not originate at the level of sensitivity (around 36 km), but from above. SOIR observations led to the conclusion (Vandaele, Mahieux, et al., 2016) that the maximum altitude at which, or at least below which the meridional transport occurs, i.e. the location of the top zonal branch of the Hadley cell, would correspond to an altitude of 100 km. This supports the mechanism (Taylor, 1995; Taylor & 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.

$CO$ is produced above the clouds by the photolysis of $CO2$ by UV radiation (see for example Figure 2 in Krasnopolsky (2012), showing that $CO2$ photolysis is the main source of $CO$ above 70 km altitude). 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 (Vandaele, Mahieux, et al., 2016), on the contrary, indicated that above 100 km altitude, $CO$ is slightly more abundant at the evening terminator than in the morning, while the reverse is true below 100 km.

$CO$ is produced at high altitudes where the subsolar–antisolar circulation prevails (Bougher, Alexander, & Mayr, 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 & Muhleman, 1985b; Krasnopolsky, 2012), leading to a decrease of its abundance. This was observed 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 (Vandaele, Mahieux, et al., 2016): $CO$ 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. Below 100 km, the trend is reversed. In fact SOIR observes a convergence of all profiles around an altitude of 100 km, at which the inversions in the trends relative to morning/ evening and equator/ high-latitude/ pole values occur (Figure 3A). This is compatible with several previous observations which reported weak (Irwin et al., 2008; Marcq et al., 2015) or non-significant (Iwagami et al., 2010; Krasnopolsky, 2008, 2010a) 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.

#### Sulfur-Bearing Species

Several sulfur-bearing species have been unambiguously identified in Venus’s atmosphere: $SO2$, $SO$, $OCS$, $S3$, and $H2SO4$. $H2S$ was reported below 20 km by Pioneer Venus, but this has never been confirmed. Moreover, the reported abundance was at least one order of magnitude larger than expected by thermodynamic equilibrium chemistry (Fegley, Zolotov, et al., 1997).

During the Venera 11 probe descent, spectra of the scattered sunlight were recorded from which abundances of $S3$ and $S4$ were derived. A first investigation (Maiorov et al., 2005) yielded abundances of $S3$ ranging from 0.03 ppb at 3 km to 0.1 ppb at 19 km. A more recent re-analysis of these same spectra (Krasnopolsky, 2013b), provided revised numbers (11 ± 3 ppt at 3–10 km; 18 ± 3 ppt at 10–19 km) but also abundances for $S4$ (4 ± 4 ppt at 3–10 km; 6 ± 2 ppt at 10–19 km).

Abundances of $H2SO4$ below the clouds were provided by ground-based observations (Jenkins, Kolodner, Butler, Suleiman, & Steffes, 2002; Sandor, Clancy, & Moriarty-Schieven, 2012), and by radio occultations from the Pioneer Venus Orbiter (Jenkins & Steffes, 1991), from the Magellan spacecraft (Jenkins, Steffes, Hinson, Twicken, & Tyler, 1994; Kolodner & Steffes, 1998), and more recently by the VeRA experiment on board Venus Express (Oschlisniok et al., 2012). These revealed complicated meridional distribution and local-time dependence of the mixing ratio, which are not reproduced in numerical models. Indeed, abundances of about 1–2 ppm were found between 0° S and 70° S latitude in the altitude range from 50 to about 52 km, sometimes increasing to values of about 3 ppm on the dayside and 5 ppm on the night side near 50 km. The abundance at polar latitudes (>70° S) did not exceed 1 ppm within the considered altitude range.

$H2SO4$ saturation vapor depends highly on temperature and concentration, leading to relatively abundant concentration (a few ppm) at the lower cloud boundary (48 km), and to low concentrations, below detection limit of about 1 ppm higher in the atmosphere. This was confirmed by recent observations carried out by the UVI camera on board Akatsuki (Imamura et al., 2017), which provided $H2SO4$ vapor profiles within and above the clouds. They found maximum values of 10 ppm at 48 km, then decreasing above to lower values. They found that the $H2SO4$ vapor VMR roughly follows the saturation curve at cloud heights, suggesting equilibrium with the cloud particles. Sandor et al. (2012) performed sub-millimeter observations of $SO$, $SO2$, and $H2SO4$ with the JCMT, in the 70–100 km altitude range, with a maximum sensitivity at 80–85 km. They could only derive upper limits (1 ± 2 ppb) considering all observations together, while upper limits from single spectra ranged from 3 to 44 ppb. The observed $H2SO4$ vapor distribution suggests that its abundance is controlled by condensation and evaporation in the clouds, thermal decomposition in the lower atmosphere, and global-scale circulation (Imamura & Hashimoto, 1998, 2001; Krasnopolsky, 2015).

Sulfur dioxide $(SO2)$ is the most abundant sulfur-bearing species. Its abundance varies with altitude and is highly spatially and temporally variable.

The abundance of $SO2$ below the clouds was measured by the Venera 12 (Gel’man et al., 1979) and Pioneer Venus (Oyama et al., 1980) gas chromatographs. It can also be inferred from nightside measurements in the IR (2.4 μ‎m). Space observations with VIRTIS-H (Marcq et al., 2008) and ground-based instruments (Arney et al., 2014; Bézard et al., 1993) confirmed the value recommended by Taylor et al. (1997) (130 ± 40 ppm). Microwave observations are also sensitive to the $SO2$ absorption, although their results strongly depend on the assumption made on the temperature profile. Such observations suggested low $SO2$ mixing ratio below the clouds with values lower than 100 ppm at low latitude and lower than 50 ppm in polar regions (Butler, Steffes, Suleiman, Kolodner, & Jenkins, 2001; Jenkins et al., 2002). So far, the number of the available measurements and their accuracy did not allow the detection of any variations in $SO2$ below the clouds. Arney et al. (2014), using ground-based spectra in the 2.3 µm window, reported a hemispheric dichotomy with slightly more $SO2$ in the northern hemisphere, although the difference lies within the retrievals error.

VEGA 1 and 2 probes provided $SO2$ profiles from the cloud top down to the surface (Bertaux et al., 1996), indicating the presence of peak structures at altitudes between 42 and 52 km with abundances reaching 210 ppm. Below 40 km both profiles exhibited the same decrease of $SO2$ with decreasing altitude. It has been proposed that VEGA $SO2$ profiles are decreasing further towards the surface, finally reaching thermochemical equilibrium. Extrapolating the value at the surface (25 ppm), the thermochemical lifetime of $SO2$ would be 320,000 years (Fegley, Klingelhöfer, Lodders, & Widemann, 1997). Clouds would eventually disappear over geological timescales, due to lack of $SO2$ to replenish the sulfuric acid. Detailed thermochemical modeling indicates that volcanic outgassing is the most probable primary source of sulfur dioxide in Venus’s lower atmosphere (Bullock & Grinspoon, 2001), an idea which has been supported by VEX observations of the surface thermal emission which strongly suggest recent or current volcanic activity (Shalygin et al., 2015; Smrekar et al., 2010).

The exchange of $SO2$ from the lower to the upper atmosphere is not fully understood, but most probably involves convective transport, presumably in conjunction with Hadley cell circulation. Secular variation of this circulation may alter the $SO2$ supply from the lower atmosphere being the origin of the high variability seen above the clouds. Photochemical oxidation efficiently destroys $SO2$ from Venus’s upper atmosphere leading to the formation of the sulfuric acid droplets making up the clouds and haze enshrouding the planet. Therefore, the $SO2$ abundance above the clouds drops to ppb levels. Krasnopolsky (2012) suggested that the origin of the high variability of $SO2$ observed above the clouds could be explained by the photolysis of $SO2$ followed by the formation of $H2SO4$ near the cloud top being very sensitive to Eddy diffusion.

Vandaele et al. (2017a, 2017b) investigated VEX-era observations of $SO2$ above the clouds. This obviously involved measurements from instruments on board VEX, but also ground-based observations in the sub-millimeter, using JCMT (Sandor et al., 2012; Sandor et al., 2010) and ALMA (Encrenaz et al., 2015), and in the IR with the TEXES/IRTF instrument (Encrenaz et al., 2013, 2016, 2012), as well as from the Hubble Space Telescope with the UV STIS (Space Telescope Imaging Spectrograph) instrument (Jessup et al., 2015). On VEX, both SPICAV-UV and SOIR (Belyaev et al., 2008; Mahieux, Vandaele, Robert, et al., 2015) observed $SO2$ above the clouds. SPICAV-UV performed measurements during solar occultation (Belyaev et al., 2012), stellar occultations (Belyaev et al., 2017) and in nadir mode (Marcq et al., 2011; Marcq, Bertaux, Montmessin, & Belyaev, 2013).

$SO2$ was first detected in the Venusian upper atmosphere from Earth-based ultraviolet observations with a mixing ratio of 0.02–0.5 ppm at the UV cloud top (Barker, 1979). Space-based identifications of $SO2$ UV absorption followed soon after, with the International Ultraviolet Explorer (IUE) (Conway, McCoy, Barth, & Lane, 1979) and the Pioneer Venus Orbiter (PVO) Ultraviolet Spectrometer (UVS) (Stewart, Anderson, Esposito, & Barth, 1979). The PVO/UVS observations showed a steady decline of the UV cloud top $SO2$ content from 100 ppb down to 10 ppb on a period of 10 years (Esposito et al., 1988). This decline was rather fast over the first year, and much slower later on. Esposito et al. (1988) interpreted this behavior as the result of a massive “injection of $SO2$ into the Venus middle atmosphere by a volcanic explosion.” IUE observations also confirmed this steep decline of $SO2$ abundance, reporting a decrease from 380 ± 70 ppb in 1979 down to 50 ± 20 ppb in 1988 (Na, Esposito, & Skinner, 1990). In 1995 a first measurement from the Hubble Space Telescope was performed, yielding a value of 20 ± 10 ppb (Na & Esposito, 1995) thus confirming the decrease in $SO2$ abundance with time. The VEX observations extended the long-term investigation of the evolution of $SO2$ abundance above the clouds (see right-hand panel of Figure 4): a clear increase in $SO2$ abundance by a factor of 2 was observed at the beginning of the mission from 2006 to 2007, followed by a decrease until 2012 by a factor of 5 to 10 (Marcq et al., 2013). These trends were also observed by SOIR (Mahieux, Vandaele, Robert, et al., 2015) and by SPICAV-UV (Belyaev et al., 2012) solar occultations, at the lowest altitudes sounded by the instrument (69–73 km): from 30 ppb (2007) to a maximum of 100 ppb (2009), decreasing again to 40 ppb (2013). However, the amplitude of the long-term trend was smaller than the short-term variations. Long-term variations were also reported by Encrenaz et al. (2016), showing that the disk-integrated $SO2$ VMR varied by a factor of 10 between 2012 and 2016 with a minimum in 2014 (30 ppb) and a maximum in 2016 (300 ppb).

Short spatial and temporal variability is best observed from Earth-based instruments which can provide global snapshots of the visible part of the planet on relatively short time periods. The three left-hand panels of Figure 4 illustrate the short term variations observed in $SO2$ abundance above the clouds. These maps were obtained with the TEXES spectrometer during three consecutive nights (Encrenaz et al., 2012). The spatial distribution of $SO2$ and the localization of the hot spots were different each night. This was confirmed by several observations (Encrenaz et al., 2013, 2016, 2012, 2015; Jessup et al., 2015; Sandor et al., 2012; Sandor, Clancy, Moriarty-Schieven, & Mills, 2010). Significant variability occurs within a few hours, consistent with a very fast photochemical or microphysical loss mechanism. Supported by ALMA observations, Encrenaz et al. (2015) showed that the short-scale variations were not correlated to day/ night or latitude variations.

#### Halides (HBr, HF, HCl)

Hydrogen halides are active species, involved in all the main chemical cycles governing Venus’s atmosphere, i.e. the $CO2$, sulfur oxidation and the polysulfur cycles (Krasnopolsky & Lefèvre, 2013; Mills et al., 2007).

Krasnopolsky and Belyaev (2017) carried out a search for HBr using CSHELL instrument at the NASA Infrared Telescope Facility in 2015. They derived an upper limit of 1 ppb at cloud top (78 km) and of 20–70 ppb below 60 km. Krasnopolsky and Belyaev (2017) further investigated the impact of such upper limits on the bromine chemistry. They concluded that the bromine chemistry might be effective on Venus, but, if the Cl/Br ratio in the atmosphere would be similar to that in the Solar System, then HBr would only reach 1 ppb in the lower atmosphere, rendering the bromine chemistry completely insignificant.

No new measurements of HF below the clouds were reported since Taylor et al. (1997) and the recommended value of 5 ± 2 ppb is still valid. Krasnopolsky (2010c) showed that, above the clouds, HF is constant at 3.5 ± 0.2 ppb (at 68 km) in both morning and afternoon observations and in the latitude range ± 60°. The SOIR/VEX measurements showed that HF VMR vary from 2.5–10 ppb at 80 km increasing to 30–70 ppb at 103 km (Mahieux, Wilquet, et al., 2015).This would require a source at 103 km and a sink near 80 km.

A recent survey using IR ground-based spectrometry sounded the 15–25 km altitude range (Arney et al., 2014) providing $HCl$ abundances in agreement with previous observations (0.5 ppm). Moreover, these showed no evidence of any horizontal variability. Iwagami et al. (2008) confirmed this value below the clouds and provided also measurements of HCl above the clouds at 60–66 km altitude (0.76 ± 0.1 ppm). They suggested the existence of a production mechanism of $HCl$ within the clouds. However both Krasnopolsky (2010c) and Sandor and Clancy (2012) reported lower abundances at cloud top (0.40 ppm). The SOIR instrument on board VEX provided $HCL$ vertical profiles (Mahieux, Wilquet, et al., 2015). These values at 70 km disagree by an order of magnitude with previous data obtained at cloud top or below and require an unidentified sink near 70 km and a source near 105 km.

Both $HCl$ and $HF$ of SOIR (Mahieux, Wilquet, et al., 2015) showed high short-term variability. $HCl$ values found at the equator and mid-latitudes are higher than at the poles. Sandor and Clancy (2012) did not see clear evidence of any local time variations.

#### Other Species (O2, OH, O3, ClO)

Only upper limits of the $O2$ abundance have been obtained (Mills, 1999; Trauger & Lunine, 1983). Krasnopolsky (2006a) proposed an even more restrictive interpretation of the measurements of Trauger and Lunine (1983). However, the existence of $O2$ has been acknowledged by several studies of its airglow emissions, which have intensively investigated the vertical, temporal, and geographic distribution emitted by excited $O2$ decaying radiatively to the ground-state (Gérard et al., 2017).

Hydroxyl radical was detected in Venus’s upper atmosphere through its nightside airglow emission (Piccioni et al., 2008) using VIRTIS instrument on board Venus Express. The $OH$ emission bands were unambiguously identified in the range 1.40–1.49 μ‎m (Δν‎ = 2 sequence of the $OH$ Meinel band) and 2.6–3.14 μ‎m (Δν‎ = 1 sequence). $OH$ was observed between 85 and 110 km, peaking at an altitude of 96 ± 2 km (Gérard, Soret, Saglam, Piccioni, & Drossart, 2010; Migliorini, Piccioni, Moinelo, Cardesi, & Drossart, 2011; Soret, Gérard, Piccioni, & Drossart, 2012). Krasnopolsky (2010d) detected airglow $OH$ lines using ground-based observations. Gérard, Soret, Piccioni, and Drossart (2012) showed that the $OH$ and $O2$ emissions are highly spatially correlated, indicating the role of $O$ atoms as a precursor of both emissions.

SPICAV-UV detected ozone $(O3)$ in the atmosphere of Venus (Montmessin et al., 2011), showing that it is vertically confined in layers in the thermosphere (between 90 and 120 km, with a mean altitude of 99 km), with densities of the order of 107–108 molecules cm-3.

Sandor and Clancy (2018) performed the first ever measurement of chlorine monoxide $(ClO)$ in the mesosphere. Their observations are compatible with a constant ClO VMR of 2.6 ± 0.5 ppb within a layer located above 85 ± 2 km and extending up to 90–100 km. Temporal variation of a factor of 2 was clearly observed. This altitude distribution of $ClO$ supports the observations of $O3$ and the interpretation proposed by Montmessin et al. (2011), i.e. the observed $O3$ and $ClO$ profiles are consistent with a chlorine-catalyzed ozone-destruction scheme above 90 km. The model of Krasnopolsky (2010d) predicted an ozone layer at 94 km but with 200 times more O3 than observed. This led to a revision of the nighttime photochemical model considering fluxes of $O,N,H$, and $Cl$ from the dayside (Krasnopolsky, 2013a). This model better simulates the observed ozone layer, but predicts 48 ppb of $ClO$ at 88 km, which is in contradiction with the observations.

High-resolution spectra of Venus of the nitric oxide $(NO)$ fundamental band at 5.3 μ‎m were acquired using the TEXES/IRTF instrument, (Krasnopolsky, 2006b). A simple photochemical model for $NO$ and $N$ in the 50–112 km range was coupled to a radiative transfer code to simulate the observed absorption features of the $NO$ and some $CO2$ lines, providing a $NO$ VMR of 5.5 ± 1.5 ppb below 60 km. Krasnopolsky (2006b) assumed that lightning is the only known source of $NO$ in the lower atmosphere of Venus. NO nightside airglow was observed by the Pioneer Venus OUVS spectrometer (Stewart, Gérard, Rusch, & Bougher, 1980). SPICAV/VEX performed observations in the limb and nadir directions (Stiepen, Gérard, Dumont, Cox, & Bertaux, 2013; Stiepen, Soret, Gérard, Cox, & Bertaux, 2012), and the 1.224 μ‎m $NO$ transition was unambiguously detected by VIRTIS on two occasions (Garcia-Munoz, Mills, Piccioni, & Drossart, 2009).

### Conclusions

The exploration of the atmosphere of Venus started in the early 20th century through ground-based observations, and benefited from the early Venera and Pioneer Venus missions. Results of the Venus Express and Akatsuki missions, combined with recent ground-based observations have dramatically improved our knowledge of the atmosphere of Venus. Abundances of most of the trace gases have been investigated from the surface up to the upper layers of the atmosphere. The atmospheres below and above the clouds differ in many respects, having quite different circulation patterns, different chemical composition, as well as different (photo)chemical processes linking the atmospheric constituents. As an example, $SO2,OCS$, and $H2O$ show a drastic change in their mixing ratio of at least one order of magnitude (even more for $SO2$) between the regions below and above the clouds.

Another aspect which is still to be understood is the high variability of the atmosphere of Venus, seen both in terms of spatial and temporal variability. This is observed in the abundances of the neutral atmosphere, as has been shown in this article, but also in the airglow emissions occurring in the upper layers (Gérard et al., 2017), in the clouds’ structures (Titov et al., 2018) and in the thermal structure of the atmosphere (Limaye et al., 2018). Observations indicate that the variation of the incident solar flux plays only a minor role in the rapid changes, which suggests the existence of a non-steady transport. Gravity waves have been proposed to be a potential source of variability, but up to now, there is no evidence that they would provide the required amplitude. Nevertheless, the different waves—gravity, Kelvin, Rossby, and tidal waves—exist on Venus and are thought to play an important role in the general circulation of the atmosphere via their contribution to momentum transport, which varies with altitude and latitude (Sánchez-Lavega, Lebonnois, Imamura, Read, & Luz, 2017).

Although the main cycles driving the composition of the atmosphere have been identified, much work remains to confirm them with observations and to improve models. New chemical schemes are required to fully understand the existing observations and models should be able to use photochemistry, dynamics and microphysics of the clouds in a coherent approach and should encompass the whole atmosphere from ground to the upper atmosphere.

As for observations of the composition of the Venus’s atmosphere, new missions should focus on observations that would enable is to better understand the different chemical cycles, in particular that of sulfur, and the exchanges from the surface to the lower atmosphere and then to the middle and upper atmosphere. This encompasses searches for potential sources at the surface, like volcanisms and hot spots (Smrekar et al., 2010) through measurements of surface emissivity, and better characterization of the compositional gradients and variability below and above the clouds, especially at cloud top. This would enable us to disentangle the different processes involved in the sulfur cycle—photochemistry, dynamics, and microphysics—and provide insights on the possible mechanisms that drive the observed variability (climate variability, circulation variations, active volcanism, etc.). Sensing, at the same time, the surface and sub-surface, their changes in composition and structure, would provide insights on surface–atmosphere interaction. This is in fact the objective of the EnVision mission, a M5 ESA mission, scheduled for launch in 2032 (Ghail et al., 2017). This mission will combine observations by radars and spectroscopic instruments (operating in the UV to the IR) to provide a global view of the planetary surface and interior and their relationship with the atmosphere.

### Glossary

Nadir observations—The instrument on board a spacecraft orbiting the planet is looking down towards the surface of the planet. Measurements are sensitive to the thermal radiation emitted by the surface (the only contribution during night observations) and the solar radiation reflected by the surface and/or scattered by the atmosphere.

Limb observations—Atmospheric remote-sounding technique involving observing radiation emitted or scattered from the limb, which is the portion of a planetary (or stellar) atmosphere at the outer boundary of the disk, viewed “edge on.” Limb viewing provides a much longer path through the atmosphere, and looking through a larger mass of air improves the chances of observing sparsely distributed substances (Livesey, 2014).

Stellar/solar occultation—An atmospheric remote sounding technique involving observing radiation emitted (or reflected) by a distant body (solar, stellar, lunar, or an orbiting satellite), transmitted along a limb path through an absorbing and/or scattering planetary atmosphere, and detected by a remote observer (Livesey, 2014). The stellar/ solar occultation technique is a powerful method of gaining information on the vertical structure of atmospheres. At sunset, the recording of spectra starts well before the occultation occurs (the solar spectrum outside the atmosphere is used for referencing), and continues until the line of sight crosses the planet. At sunrise, the recording of spectra continues well above the atmosphere to provide the corresponding reference. Transmittances are obtained by dividing the spectra measured through the atmosphere by the reference spectrum recorded outside the atmosphere. In this way, transmittances become independent of instrumental characteristics, such as the absolute response or the ageing of the instrument and in particular of the detector. Such observations provide high vertical resolution.

Venus Express—The first European mission to Venus (Svedhem et al., 2007). Its overarching science objectives covered the atmosphere, plasma environment, and surface properties. The payload consisted essentially of instruments inherited from the Mars Express and Rosetta missions and comprised a combination of spectrometers, spectro-imagers and imagers covering a wavelength range from ultraviolet to thermal infrared, a plasma analyzer and a magnetometer. Launched in November 2005, it arrived at Venus in April 2006 and directly started sending back science data until 16 December 2014, when ESA announced that the Venus Express mission had ended. Venus Express dived into the planet’s upper atmosphere to altitudes only 165 km above the surface during a series of low passes in the period 2008–2013, in order to measure the density of the upper polar atmosphere. The campaign showed that the upper layers of the atmosphere were a surprising 60% thinner than predicted and showed high density variability.

Akatsuki—A Japanese mission to Venus. The Planet-C mission was approved in 2001 and was launched in May 2010 (Nakamura et al., 2007). At that time, as usual for Japanese missions, Planet-C was renamed Akatsuki, “Morning Star.” After a first failed orbital insertion attempt, Akatsuki was placed in a solar orbit with a period slightly shorter than the orbital period of Venus. Later, Akatsuki successfully entered Venus orbit on its second attempt in December 2015 (Nakamura et al., 2016). Its current orbit is highly elliptical, near-equatorial.

Akatsuki was designed to observe the atmospheric dynamics of Venus and to better understand the atmospheric super-rotation. To address its scientific objectives, five cameras cover different spectral ranges: IR1 (InfraRed 1 μ‎m camera) (Iwagami et al., 2018), IR2 (InfraRed 2 μ‎m camera) (Satoh et al., 2017), UVI (UltraViolet Imager) (Yamazaki et al., 2018), LIR (Longwave InfraRed camera), and LAC (Lightning and Airglow Camera). Combining these cameras’ observations, allows detection of atmospheric motions at different altitudes. Radio occultation (RS, Radio Science) is sensitive to temperature and H2SO4 vapor between 35 and 100 km (Imamura et al., 2017). Akatsuki has already provided many significant scientific results, one of the most compelling being the discovery of large-scale stationary gravity waves in the atmosphere (Fukuhara et al., 2017).

• Ya Marov, M., & Grinspoon, D. H. (1998). The planet Venus. Yale University Press.

This comprehensive book provides a detailed synopsis of Venus, with particular attention to the history of its observation and exploration, the planet’s formation, and the development of the runaway greenhouse effect.

• Hunten, D. M., Colin, L., Donahue, T. M., & Moroz, V. I. (Eds.). The series Venus. University of Arizona Press, 1983; Bougher, S. W., Hunten, D. M., & Phillips, R. J. (Eds.). (1997). Venus II: Geology, geophysics, atmosphere, and solar wind environment. University of Arizona Press; and Bézard, B., Russell, C. T., Satoh, T., Smrekar, S. E., & Wilson, C. F. (Eds.). (2018). Venus III, Space Science Reviews. Springer.

These publications are comprehensive sources of information on different aspects of the planet. Each of them summarizes the knowledge of their time. The third volume reports on the achievements gained in the era of the Venus Express mission.

• Mackwell, S., Simon-Miller, A., Harder, J., & Bullock, M. (2013). Comparative climatology of terrestrial planets. University of Arizona Press.

This compendium gives a wide overview of the current understanding of atmospheric formation and climate evolution. Particular emphasis is given to surface–atmosphere interactions, mantle processes, photochemistry, and interactions with the interplanetary environment, all of which influence the climatology of terrestrial planets.