Volume 125, Issue 7 e2019JE006159
Research Article
Free Access

Chemical Cycling in the Venusian Atmosphere: A Full Photochemical Model From the Surface to 110 km

C. J. Bierson

Corresponding Author

C. J. Bierson

Department of Earth and Planetary Sciences, UC Santa Cruz, Santa Cruz, CA, USA

Correspondence to:

C. J. Bierson,

[email protected]

Search for more papers by this author
X. Zhang

X. Zhang

Department of Earth and Planetary Sciences, UC Santa Cruz, Santa Cruz, CA, USA

Search for more papers by this author
First published: 08 April 2020
Citations: 39


Venus is an exceptional natural experiment to test our understanding of atmospheric sulfur chemistry. Previous modeling efforts have focused on understanding either the middle or lower atmosphere. In this work, we performed the first full atmosphere analysis of the chemical transport processes on Venus from the surface to 110 km using a 1-D diffusion model with photochemistry. We focused on the cycling of chemical species between the upper and lower atmospheres and interactions between distinct species groups including SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0001, CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0002 + OCS, chlorides, NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0003, O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0004, and S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0005. We tested different eddy diffusivity profiles and investigated their influences on the vertical profiles of important species. We find that the assumed boundary conditions in previous models strongly impacted their simulation results. This has a particularly large effect for SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0006. We find the high SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0007 abundance in the lower atmosphere is readily transported into the middle atmosphere, far exceeding observed values. This implies some yet unknown chemistry or process limiting SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0008 mixing. We summarize outstanding questions raised by this work and note chemical reactions that should be the highest priority for future laboratory studies and ab initio calculations.

Key Points

  • We perform the first thorough analysis of a Venus atmospheric chemistry model that extends from the surface to 110 km
  • Some previously proposed chemical pathways break down when more complete chemistry is used
  • To match observed abundances SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0011 must be inhibited from diffusing through the clouds by some unknown process

Plain Language Summary

Venus's atmosphere can be broadly separated into lower and middle regions, separated by a thick cloud deck. Chemistry in the lower atmosphere is controlled by the high temperatures below the clouds. In the middle atmosphere, photochemistry stimulated by solar UV radiation is dominant. Previous works have modeled either the lower or middle atmosphere to understand these chemical processes. In this work, we create a single model that encompasses both regions to understand how chemical species are cycled. We find that the large abundance of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0009 in the lower atmosphere is transported into the middle atmosphere, far exceeding what is observed. We argue that this suggests some as of yet unknown chemistry or process that is limiting the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0010 flux from the lower atmosphere.

1 Introduction

Venus is a natural laboratory for understanding atmospheric sulfur chemistry. Over the last decade, there have been a wealth of observations of the minor species in Venus's atmosphere from both ground-based observations (Arney et al., 2014; Jessup et al., 2015; Krasnopolsky, 2010a; Marcq et al., 2006; Sandor et al., 2012; Sandor & Clancy, 2017, 2018) and Venus Express (Belyaev et al., 2008, 2012; Marcq et al., 2008). Chemical models have been developed for the lower atmosphere (Krasnopolsky, 2013; Yung et al., 2009) and the middle atmosphere (Jessup et al., 2015; Krasnopolsky, 2012; Zhang et al., 2012). These two regions are separated by the thick cloud layers that extend from roughly 40 to 70 km altitude. In the lower atmosphere, the high temperatures (400–700 K) drive chemical systems to thermochemical equilibrium. In the middle atmosphere, where the temperatures are more similar to Earth's stratosphere (200–300 K), photochemistry is the dominant process.

In this work, we present the first detailed chemical model that couples the lower and middle atmospheres. This is an attempt to understand how the limited domain of previous models impacted their results and determine what open questions can be better understood using a coupled model. Our results are broken up by species groups. We summarize the major outstanding questions (both preexisting and raised by this study) and what future work is needed. The goal of this work is not to present the “best” or most highly tuned Venus atmospheric chemistry model. Instead, we aim to use this model as a tool for exploring the interactions between species and regions of the Venusian atmosphere. First, we briefly review the key observational constraints and previous modeling efforts.

1.1 Overview of Chemical Cycles

In this work, we focus on the chemical cycles for three species groups: sulfur oxides, carbon oxides, and chlorides (shown in Figure 1). Other groups including oxygen, nitrous oxides, and polysulfur species will be discussed briefly. Here, we provide an overview of the interactions between these groups. For more background, see the recent reviews by Mills and Allen (2007) and Marcq et al. (2017). In section 3, these processes will be evaluated and quantified.

Details are in the caption following the image
Cartoon showing the main species groups discussed in this work and their interactions.

By far, the most thoroughly observed (and modeled) species in the lower atmosphere are carbon monoxide (CO) and carbonyl sulfide (OCS). At urn:x-wiley:jgre:media:jgre21306:jgre21306-math-001230 km, OCS is present at the ppm level. The OCS mixing ratio then decreases rapidly with altitude at about 35 km (Arney et al., 2014; Marcq et al., 2006). Conversely, CO has a near surface mixing ratio of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-001320 ppm and a gradual increase with altitude. CO is also observed to increase in mixing ratio by urn:x-wiley:jgre:media:jgre21306:jgre21306-math-00145 ppm within 30° of the poles (Cotton et al., 2012; Marcq et al., 2006). OCS does not have an unambiguous trend with latitude, although Marcq et al. (2006) and Marcq et al. (2008) have suggested a possible decrease in mixing ratio at high latitudes. Several chemical pathways have been proposed for conversion between CO and OCS (Krasnopolsky & Pollack, 1994; Yung et al., 2009) and are discussed in detail in section 3.2.

Photolysis of carbon dioxide (CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0015) in the middle atmosphere produces CO and O, which are slow to directly recombine back to CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0016. This leads to a buildup of oxygen species including O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0017. However, upper limits on the O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0018 column density suggest that there is some catalytic mechanism providing more efficient recombination (section 3.3 and Mills & Allen, 2007).

In the middle atmosphere, the most thoroughly observed minor species is sulfur dioxide (SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0019). Vandaele et al. (2017) provide an excellent review of observations from ground-based and Venus Express measurements. Venus Express terminator measurements allow for well-resolved vertical profiles of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0020 (Belyaev et al., 2012). These observations show that the mixing ratio of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0021 decreases briefly above the clouds before inverting and increasing with altitude from urn:x-wiley:jgre:media:jgre21306:jgre21306-math-002280 up to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0023100 km. It was initially suggested that this could be due to the photolysis of sulfuric acid (H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0024SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0025) at high altitudes (Zhang et al., 2010), but follow-up observations found that not enough H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0026SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0027 was present (Sandor et al., 2012). The source of sulfur driving this inversion is still unclear, and polysulfur aerosols remain a possibility (Zhang et al., 2012).

SO, SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0028, and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0029 are quickly recycled inside the sulfur oxides group above the clouds. The relative abundance of these species is set by the balance between photolysis (creating more SO) and oxidation (producing SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0030 and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0031). SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0032 will readily react with water vapor to form sulfuric acid (H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0033SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0034). This sulfuric acid then condenses to form the bulk of the cloud deck. At the base of the clouds ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-003540 km), temperatures are high enough that the sulfuric acid droplets evaporate and the vapor is then thermally decomposed back into SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0036 and water. This lower atmosphere SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0037 is also converted to the more stable SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0038 and diffused upward to the middle atmosphere completing the sulfur cycle.

The most abundant chlorine species in the Venusian atmosphere is hydrochloric acid (HCl). HCl has been observed in the lower and middle atmospheres with a roughly constant mixing ratio of 100–400 ppm (Arney et al., 2014; Sandor & Clancy, 2017). Recently, chlorine monoxide (ClO) was also observed to have a mixing ratio of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-00392 ppb at 85 km (Sandor & Clancy, 2018). Sandor and Clancy (2018) interpreted this concentration of ClO as requiring an additional significant reservoir of chlorine in the middle atmosphere apart from HCl.

1.2 Previous Models

Previous chemical models of the Venusian atmosphere have focused on either the lower (Krasnopolsky, 2007, 2013) or middle atmosphere (Krasnopolsky, 2012; Zhang et al., 2010, 2012). The model of Krasnopolsky (2007) is not notably different from the updated version in Krasnopolsky (2013), and so, we only discuss the latter. For the same reason, we only review Zhang et al. (2012) and not Zhang et al. (2010). The only model to have a domain including both the lower and middle atmospheres is Yung et al. (2009). However, Yung et al. (2009) primarily show species profiles from a model with the domain restricted to the middle atmosphere. The only species profile they show from their extended model is OCS (Figure 2).

Details are in the caption following the image
Comparison of previous chemical models of the Venusian atmosphere. The curves from Zhang et al. (2012) are their Model A. In Krasnopolsky (2013), SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0040 is only reported for values larger than urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0041.
Table 1. Mixing Ratios of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0042 at the Cloud Level Model Boundary in Different Lower and Middle Atmosphere Models
Reference Domain (km) SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0043 (ppm)
Krasnopolsky (2007) 0–47 130
Krasnopolsky (2013) 0–47 130
Krasnopolsky (2012) 47–112 9.7
Zhang et al. (2012) 58–112 3.5
Yung et al. (2009) 58–112 130
  • Note. For the three middle atmosphere models, this is simply the lower boundary condition. In the lower atmosphere models, this value is calculated but is the same as the surface boundary condition as there is no significant chemistry. Note that the model of Yung et al. (2009) that uses a higher SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0044 concentration also produced an SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0045 abundance two orders of magnitude larger than observations in the middle atmosphere.

Figure 2 shows species profiles from the studies described above. Within a given domain (lower or middle) of the atmosphere, the different models are generally in agreement. The exception to this is the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0046 profile of Yung et al. (2009) that exceeds the observed mixing ratio by nearly two orders of magnitude at 80 km. The middle atmosphere models of Zhang et al. (2012) and Krasnopolsky (2012) reproduce these observations by dramatically lowering their lower boundary condition (Table 1).

The greatest disagreement between previous models is at the boundary between the lower and middle atmospheres. This disagreement is modest for CO and OCS and is most significant for the sulfur oxides. In the lower atmosphere models, SO and (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0047 are often not even included or are included only in a limited capacity. We discuss how this affects the OCS chemistry in section 3.2. Lower atmosphere models do not predict any vertical gradient in SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0048 and as such report constant values around the observed 130 ppm (Arney et al., 2014; Marcq et al., 2006). To match the low mixing ratio observed in the middle atmosphere, models such as Zhang et al. (2012) and Krasnopolsky (2012) use a fixed abundance lower boundary condition of less than 10 ppm. It is these inconsistencies that motivate the need for models like the one presented here that can characterize the flux of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0049 species through the cloud layer (discussed in section 3.1).

2 Model Description

For this study, we use the JPL/Caltech kinetics 1-D photochemistry-diffusion model (Mills, 1998; Yung & Demore, 1982; Zhang et al., 2012). Our model solves a 1-D continuity equation with diffusion and chemical reactions. The atmosphere is assumed to be in hydrostatic equilibrium. We use 2-km grid cells covering 0–112 km. We calculate the diurnally averaged radiation field from 100 to 800 nm using a modified radiative transfer scheme including gas absorption, Rayleigh scattering by molecules, and Mie scattering by aerosols with wavelength-dependent optical properties (Zhang et al., 2012). We also parameterized an additional UV opacity source in the radiative transfer calculation contributed by the unknown UV absorber (see Zhang et al., 2012, for details). Our calculations are set at 45° latitude with fixed solar insolation to approximate the global-mean situation.

We use the Venus International Refrence Atmosphere (Seiff et al., 1985) (Figure 3). Because observational constraints for the eddy diffusivity ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0050) are sparse, a number of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0051 profiles have been used in the literature. For our nominal model configuration, we use the profile from Zhang et al. (2012) linearly extrapolated in log space into the lower atmosphere. For a description of the other urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0052 profiles used and their motivations, see section 2.2.

Details are in the caption following the image
Left panel: temperature profile from Seiff et al. (1985). Right panel: urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0053 profiles used in this study. The Zhang et al. (2012) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0054 profile has been extrapolated into the lower atmosphere (nominal). The other urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0055 profiles are described in section 2.2. The horizontal dashed line at 47 km indicates the boundary between the middle atmosphere model in Krasnopolsky (2012) and the lower atmosphere model in Krasnopolsky (2013). Observational estimates of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0056 are shown in magenta (upper limit, von Zahn et al., 1979; star, Woo & Ishimaru, 1981; and line, Lane & Opstbaum, 1983).

In this work, we calculate the vertical profile of 53 species: O, O( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0057), O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0058, O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0059( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0060), O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0061, H, H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0062, OH, HO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0063, H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0064O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0065, N, NO, NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0066, NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0067, N urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0068O, HNO, HNO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0069, HNO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0070, Cl, Cl urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0071, ClO, HCl, HOCl, ClCO, COCl urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0072, ClCO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0073, CO, CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0074, S, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0075, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0076, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0077, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0078, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0079, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0080, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0081, SO, (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0082, SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0083, SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0084, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0085O, SH, H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0086S, HSO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0087, ClSH, ClS, ClS urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0088, Cl urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0089S, Cl urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0090S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0091, OSCl, ClSO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0092, SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0093Cl urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0094, and OCS. Additionally, we have three species with fixed profiles, N urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0095, H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0096O, and H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0097SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0098 (shown in Figure 4). N urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0099 is fixed at a mixing ratio of 3.4%. N urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0100 acts as a catalyst in some reactions and is a source of N via photolysis. H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0101O and H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0102SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0103 are also held constant as both species are condensible, but condensation is not included in this model. In addition, as noted by Parkinson et al. (2015), in the middle atmosphere, the H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0106O and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0107 abundances are strongly coupled. While the specific bifurcation behavior described by Parkinson et al. (2015) was due to numerical errors, the strong coupling is robust (Shao et al., 2020). By holding the water profile constant (at observed values), we remove this source of sulfur variability. The water profile used is a spline interpolation to the observational results from Bertaux et al. (2007) in the middle atmosphere and Marcq et al. (2006), Barstow et al. (2012), and Arney et al. (2014) in the lower atmosphere (see Figure 4).

The H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0108SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0109 profile is fixed at the saturation vapor pressure assuming 90 wt% H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0110SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0111 above 57 km. This is consistent with observation of cloud acidity (Arney et al., 2014; Barstow et al., 2012). Our model is not very sensitive to this value as H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0112SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0113 is not a significant chemical source in the upper atmosphere. Below 47 km (at the base of the lower cloud), we use the profile calculated by Krasnopolsky (2013). Between 47 and 57 km, the H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0114SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0115 profile is linearly interpolated in log space. The resulting profile shown in Figure 4 is consistent with the upper limits by Sandor et al. (2012) in the middle atmosphere and the Venus Express observations at the base of the clouds (Oschlisniok et al., 2012). A test of the model sensitivity to this profile is presented in supporting information, Text S1. The effect of this profile is discussed in detail in section 3.1.

Details are in the caption following the image
Profiles of all fixed species in the model and corresponding observations. H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0104SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0105 upper limit is obtained from Sandor et al. (2012) and observed abundance from Oschlisniok et al. (2012) and Imamura et al. (2017). Water vapor observations are from Marcq et al. (2006), Bertaux et al. (2007), Krasnopolsky (2010a), Barstow et al. (2012), and Arney et al. (2014).

In order to couple the chemical systems in the middle and lower atmospheres, one needs to include both the photochemistry in the middle atmosphere and the high-temperature thermochemistry with both forward and reverse reactions in the lower atmosphere. In this study we have updated the chemistry of Zhang et al. (2012) to include new experimental results as described in Burkholder et al. (2015) and the lower atmosphere chemistry of Krasnopolsky (2013). The chemical reactions used and their sources are described in Table S1. These updates to the chemistry have a negligible effect on our model compared directly with Zhang et al. (2012). In section 4, we discuss which reactions with unconstrained reaction constants have the largest impact on our results.

2.1 Boundary Conditions

For most species in our model, we use a zero flux (closed box) boundary condition at both the upper and lower boundaries. This is consistent with the standard setting for previous lower and middle atmosphere models (Krasnopolsky, 2007, 2012; Zhang et al., 2012). In a model with a zero flux boundary condition for all species, the resulting steady-state solution is highly sensitive to the initial condition. This is due to the fact that the total abundance of each group is set by the initial condition. To avoid this, at least one species in each group was assigned a fixed mixing ratio at the lower boundary. We preferentially chose species that have observational constraints on their abundance. This allows the abundance of each chemical group to equilibrate to the observed values through surface fluxes. The upper boundary conditions used in this model follow the work of Mills (1998). Upward fluxes of CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0116, O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0117, and O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0118( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0119) are imposed at the upper boundary to account for photolysis taking place above the model boundary. We use the fluxes calculated by Mills (1998). In order to conserve the number of atoms in the domain, downward fluxes of CO and O of the same magnitude as the loss of CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0120 are also imposed at the upper boundary.

Details are in the caption following the image
Chemical loss timescales for different SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0121 species and the transport timescale profile. All profiles are for the nominal model configuration. The local minimum in the OCS lifetime corresponds to the region with excess SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0122 at the base of the clouds (Figure 7).

2.2 Modifications to Our Nominal Case

To discuss the sensitivity of our results to different changes in our nominal model, it is useful to name particular cases we will refer back to. Here, we describe those alternate cases in the sensitivity study. The alternate urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0123 profiles are shown in Figure 3.
  • K urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0124: uses the combined urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0125 profiles from Krasnopolsky (2012) and Krasnopolsky (2013).
  • Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0126: uses our nominal urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0127 profile but reduced by one order of magnitude. We present this profile not because it is a plausible urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0128 profile but because it is instructive for understanding the model sensitivity.
  • Cloud urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0129: It was suggested by Marcq et al. (2017) that the cloud layer may inhibit transport. This profile tests a stable cloud region by using our nominal urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0130 profile modified to a value of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0131 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0132 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0133 between 45 and 65 km.
  • S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0134: changes the upper boundary condition of S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0135 to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0136. This polysulfur flux from the upper atmosphere was proposed by Zhang et al. (2012) to explain the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0137 inversion in the middle atmosphere.
  • urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0138 OCS: uses the reaction rate constant for OCS + SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0139 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0140 CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0141 + (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0142 from Krasnopolsky (2007).

3 Results

3.1 SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0143

Discussing SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0144 in the Venusian atmosphere is nearly equivalent to discussing SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0145. SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0146 is the most abundant, most stable and best observed of the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0147 group everywhere except the highest altitudes. Using our nominal urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0148 profile, the chemical transport timescale ranges from urn:x-wiley:jgre:media:jgre21306:jgre21306-math-014910 years near the surface to months at 90 km (Figure 5). The chemical loss timescale for SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0150 exceeds urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0151100 years until photolysis begins to dominate in the upper middle atmosphere. As such, SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0152 is well mixed from the lower boundary of the model until the base of the clouds.

In contrast to SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0153, SO and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0154 have chemical loss timescales of hours or less throughout the model domain. In the lower atmosphere, where there are no significant sources of these species, this leads to mixing ratios below 1 ppt. Above the cloud deck, their abundance is set by the relative contributions of oxidation and photochemical destruction. The source of oxygen for these reactions is primarily derived from the photolysis of CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0155 and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0156.

Here, we quantitatively examine the sulfur flux exchange between the lower and middle atmospheres through the clouds, as illustrated in Figure 1. The primary reservoirs of sulfur below 35 km are SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0157 ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0158100 ppm) and OCS ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-015930 ppm), both of which have fixed mixing ratios at the surface. There may be a few ppm S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0160 present; however, currently, there is no direct observational evidence for this (see section 3.5). Right above  35 km, the observed rapid decrease of OCS suggests a rapid conversion from OCS to SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0161 or S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0162, which will be discussed in section 3.2. As a result, the primary source of sulfur into the middle atmosphere through the clouds is SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0163. The magnitude of this flux is set by the observed SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0164 mixing ratio and urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0165 profile. In our nominal case, the upward flux of the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0166 through the clouds (59 km) is urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0167 cm s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0168. The value ranges from urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0169 to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0170 cm s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0171 across all of our cases.

Above the cloud, SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0172 is converted to SO and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0173 via photochemistry. The downward fluxes of SO and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0174 to the lower atmosphere are negligible due to their short loss timescales. The main sink of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0175 is sulfuric acid formation (SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0176 + 2H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0177O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0178 H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0179SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0180 + H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0181O). The entire H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0182SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0183 formation process is primarily limited by SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0184 formation, which is itself primarily limited by the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0185 flux. Thus, among our model configurations, the column-integrated production rate of H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0186SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0187 above 59 km shows little variation and tracks with the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0188 flux. Values range from urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0189 to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0190 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0191 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0192, which is in line with previous results by Zhang et al. (2012) ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0193 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0194 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0195) and Krasnopolsky (2015) ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0196 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0197 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0198). The H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0199SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0200 produced in the middle atmosphere rapidly condenses to form the bulk of the Venus clouds. Cloud droplets settle into the lower atmosphere where they evaporate releasing H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0201SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0202 vapor. This vapor further thermally decomposes into SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0203 and H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0204O. This SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0205 is rapidly converted to SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0206 creating a source region of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0207 at the cloud base. Then the long-lived SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0208 from the base of the clouds is diffused away from this source to complete the sulfur cycle.

In our nominal model configuration, the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0209 mixing ratio approaches that of a well-mixed profile with minimal vertical gradient, inconsistent with the observations from Venus Express (Belyaev et al., 2008, 2012). This problem was not observed in previous middle atmosphere models because they could limit the flux by fixing the mixing ratio of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0210 in their lower boundary condition (located in the cloud deck). As shown in Table 1, the models that successfully reproduce the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0211 minima (Krasnopolsky, 2012; Zhang et al., 2012) require a lower boundary condition of less than 10 ppm. The only model with a lower boundary condition consistent with the lower atmosphere observations is Yung et al. (2009). Their model (like ours) does not match the middle atmosphere observations.

There are two way to overcome this issue, decreasing the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0212 transport or increasing the chemical sink in the cloud layer. Two of our urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0213 profiles attempt to limit the transported SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0214: Cloud urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0215 and Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0216. Cloud urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0217 is based on the suggestion from Marcq et al. (2017) that SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0218 transport is inhibited by a stable cloud layer. This profile has a value of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0219 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0220 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0221 between 45 and 65 km (shown in Figure 3). Additionally, we use the Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0222 profile that has a urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0223 lower than our nominal case by an order of magnitude throughout the domain. Both the Cloud urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0224 and Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0225 models come close to matching the observed SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0226. It is worth noting that in both of these cases, the flux of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0227 at 59 km is actually twice that of the nominal case due to the large mixing ratio gradient. This gradient leads to a much lower SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0228 mixing ratio at the cloud tops, which in turn allows the radiation to penetrate deeper into the middle atmosphere. This further lowers the mixing ratio. In other words, it is the mixing ratio of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0229 at the cloud top that is important, not the flux.

Figure 7 shows that the observed SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0230 profile can be matched allowing for any arbitrary amount of transport. This raises an important question: How plausible are these particular urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0231 profiles? Both of these profiles are inconsistent with the radio scintillation estimate as urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0232 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0233 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0234 at 60 km (Woo & Ishimaru, 1981) and urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0235 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0236 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0237 at 45 km (Woo et al., 1982). From the mixing length theory, urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0238 can be estimated as the product of vertical velocity and a length scale. Imamura and Hashimoto (2001) used velocities from both models and the Vega balloons to estimate a cloud region urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0239 of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0240 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0241 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0242. The value seems even larger between 50 and 60 km. Using the turbulent velocity on the order of 1 m s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0243 measured by Vega balloon at 54 km (Blamont et al., 1986) and the mixing length scale of 1 km, the estimated urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0244 is about urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0245 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0246 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0247. Based on the static stability profiles from the Pioneer Venus probes, which are consistent with the VeRa/VEx data (Limaye et al., 2018), McGouldrick and Toon (2007) estimated the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0248 between urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0249 and urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0250 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0251 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0252 in the cloud region. These values have been validated by cloud microphysical models that are sensitive to the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0253 to match the observed cloud droplet size distribution (Gao et al., 2014; McGouldrick & Toon, 2007). These values are higher than our Cloud urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0254 and Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0255 configurations by at least a factor of three. Even using the low end of these plausible values of kzz (i.e., urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0256 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0257 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0258) cannot reproduce the observed SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0259 data, as already shown in our K urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0260 configuration. This implies that the low SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0261 mixing ratio above the clouds is due to some as of yet unknown chemical sink in the cloud region or interactions with the cloud droplets. This also highlights the need for 2-D and 3-D dynamical models to quantify the transport in greater detail, as the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0262 is a global-mean approximation of the 3-D dynamical transport (Zhang & Showman, 2018b, 2018a).

An important caveat in our results is the artificial source of sulfur as a result of the fixed H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0263SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0264 vapor profile. As our model does not include cloud formation, and so the gas H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0265SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0266 abundance cannot be self consistently calculated. Instead, the H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0267SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0268 profile is held to the profile expected from condensation and evaporation (Figure 4). This parameterization can be checked for self-consistency by comparing the net production and loss rates. Figure 6 shows the production and loss reaction rates as well as the net production minus loss for H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0269SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0270. Although the total column formation and destruction rates of H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0271SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0272 are both roughly urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0273 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0274 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0275, the column destruction rate is a bit higher than production by urn:x-wiley:jgre:media:jgre21306:jgre21306-math-02760.2% in all our model configurations. This slight excess in H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0277SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0278 destruction over production impacts the lower atmosphere sulfur budget. As noted above, the evaporation of sulfuric acid and conversion from SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0279 and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0280 at the cloud base provides a SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0281 source, which needs to be diffused away, both upward and downward in our model. The downward diffusion flux goes into the surface to satisfy the lower boundary condition (fixed mixing ratio, Table 2). Because the production and loss of H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0282SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0283 are not exactly equal, our model uses this boundary condition to achieve steady state. Across our model configurations, this downward flux of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0284 ranges from urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0285 to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0286 molecules cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0287 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0288. These are much larger than the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0289 to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0290 molecules cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0291 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0292 suggested as plausible rates for surface reactions (Fegley et al., 1997). These change very little even for large changes in the H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0293SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0294 profile (see Text S1). To properly model the lower atmosphere, flux of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0295 will require a model that can fully conserve sulfur. In any model that does not conserve sulfur, the lower boundary will numerically accommodate the excess.

Details are in the caption following the image
Profiles for the production and loss reaction rates of sulfuric acid for two model configurations. The red curve shows the net production minus loss where dashed lines indicate local H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0337SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0338 production and dots indicate local loss. In both cases, H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0339SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0340 is primarily produced in the zone of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0341 photolysis.
Details are in the caption following the image
Observations and model results of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0296. Observed SO and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0297 are shown in magenta. The middle atmosphere boxes denote the variability in SO and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0298 observed by the SPICAV solar occultation instrument on Venus Express (Belyaev et al., 2012; Vandaele et al., 2017). The individual points show a typical observed profile from Belyaev et al. (2012) to show the shape of the observed SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0299 profile. The one middle atmosphere SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0300 point outside the box is from Krasnopolsky (2010a). Lower atmosphere box shows the variation in SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0301 measured by VIRTIS on Venus Express (Vandaele et al., 2017). Individual lower atmosphere measurement from Arney et al. (2014). Model configurations with varying urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0302 are shown in the various colors. For each urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0303, the run including a flux of S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0304 from the upper boundary is shown with a dashed line. For some of these model configurations, the dashed line is on top of the solid line (hence is not visible) because the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0305 flux from the lower atmosphere dominates. The exception is the Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0306 and Cloud urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0307 configurations. This is because the significantly lower SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0308 mixing ratio allows the high-altitude oxidation of S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0309 species to make a significant impact on the profile.
Table 2. Boundary Conditions Applied to the Model
Species Lower Upper
O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0310 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0311
O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0312 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0313 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0314
O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0315 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0316 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0317
NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0318 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0319
CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0320 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0321
CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0322 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0323 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0324
SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0325 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0326 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0327
OCS urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0328 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0329
  • Note. urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0330 is the flux (cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0331 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0332), urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0333 is the escape velocity (cm s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0334), where urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0335 is the maximum escape diffusion velocity and urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0336 is a fixed mixing ratio. The boundary condition for all other species is zero flux. The sign convention is positive upwards.

In addition to the sulfur cycle discussed above, there might be an additional source of sulfur from the upper atmosphere. Such a source was proposed by Zhang et al., 2010 (2010, 2012) in order to explain the increase of SO and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0342 with altitude above 80 km. In this work, we parameterize this as a polysulfur (S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0343) flux from the model upper boundary as proposed by Zhang et al. (2012). It is only in the runs with this polysulfur flux that we observe any inversion. However, because we have an overabundance of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0344 in the middle atmosphere generally, this model should not be used to quantify the required flux.

We also use our model to examine a recently proposed candidate for the unknown UV absorber in the Venusian atmosphere. Frandsen et al. (2016) suggested that (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0345 could be the unknown UV absorber (Esposito, 1980). In creating their simple model to estimate the (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0346 abundance, they assume either a 12- or 20-ppb mixing ratio at their lower boundary (58 km). In contrast, we find that at 58 km, (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0347 is essentially nonexistent (mixing ratio less than urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0348). The concentration does increase higher in the atmosphere but never exceeds 1 ppb (Figure 8). In the model configuration that best matches the observed SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0349, Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0350 + S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0351, the (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0352 abundance never exceeds even 1 ppt. The mixing ratio also peaks at urn:x-wiley:jgre:media:jgre21306:jgre21306-math-035390 km altitude, far above the cloud tops where the UV absorber is observed. In summary, our results are consistent with the findings of Krasnopolsky (2018) and Marcq et al. (2020) suggesting that (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0354 is not abundant enough to be the main UV absorber.

Details are in the caption following the image
Profiles of (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0378 for the same cases as are shown in Figure 7.

3.2 CO and OCS

This section is an overview of the carbon cycling in Venus's atmosphere. While CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0355 is the most abundant carbon species throughout the atmosphere, the carbon flux is primarily controlled by CO and OCS in conjunction with CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0356.

There are two primary sources of CO, photolysis of CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0357 in the middle atmosphere and conversion from OCS in the lower atmosphere. Figure 9 shows the important production and loss reactions of CO at each altitude in addition to the production and loss profiles. In our nominal case, above 59 km, CO photolysis (CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0358 = CO + O + M) has a column rate of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0359 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0360 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0361 taking place primarily above 85 km. This plus the CO from the upper boundary ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0362 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0363 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0364, Table 2) is transported downward with a nearly constant flux of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0365 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0366 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0367. The difference between this photochemical production and flux is due to slow reoxidation by OH (column rate urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0368 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0369 s urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0370). These rates can vary by an order of magnitude between our different urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0371 profiles. This sensitivity is largely due to the different radiation environment created by the varying SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0372 abundance (Figure 7). This leads to the different CO mixing ratio profiles shown in Figure 10. Regardless of the value, in all of these cases, this downward flux of CO in the middle atmosphere is matched by an upward flux of CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0373, closing the cycle.

In the lowest 30 km of the atmosphere, CO and OCS are cycled between each other. This cycling is mediated by S and S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0374 (Figure 9). In all of our model configurations, OCS is diffused upwards from the lower boundary with a flux between urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0375 and urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0376 molecules cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0377. In each case, this flux is matched by a comparable downward flux of CO towards the surface. As with the middle atmosphere, the absolute value of these fluxes is dependent on the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0379 in the lower atmosphere, which is not observationally constrained. This balance in fluxes implies some surface chemistry moderating CO and producing OCS (Zolotov, 2018, and the references within).

In all our model configurations, CO is converted back to CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0380 in the cloud region via

The rate constant used for this reaction was assumed by Krasnopolsky and Pollack (1994) using analogy to CO + NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0382 = CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0383 + NO. The high abundance of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0384 in the lower cloud region makes this reaction very fast and effectively prevents exchange of CO between the middle and lower atmospheres.

CO may also be rapidly produced in the cloud region by OCS-SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0385 chemistry. The observed decrease in OCS mixing ratio at 35 km is coincident with the bottom of the cloud deck. Because of this, there is a local excess of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0386 from the thermal decomposition of H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0387SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0388. This leads to the natural hypothesis that there is some reaction between SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0389 and OCS driving this reduction. Krasnopolsky and Pollack (1994) calculated the free energy associated with possible products of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0390 + OCS and found that they are all endothermic. The least endothermic was
Krasnopolsky and Pollack (1994) further suggested that the OCS destruction would be enhanced by the secondary reaction

This set of reactions has since been used in follow-up models of Venus's lower atmosphere (Krasnopolsky, 2007, 2013) but has yet to be measured in the lab. This pathway was tested in our urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0393 OCS model configuration.

In the lower atmosphere models of Krasnopolsky (2007) and Krasnopolsky (2013), R2 and R3 are the only reactions that feature (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0394. Because of this, (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0395 will build up in their models until reaction R3 has the same rate as reaction R2. In contrast, our model includes reactions that equilibrate (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0396 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0397 2SO using the rates estimated by Mills (1998). As a result, (SO) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0398 does not build up and reaction R3 is effectively halted. This has two effects: It dramatically reduces the OCS destruction rate and interrupts the pathway converting OCS to CO. Because of this, none of the models using urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0399 OCS fit the observed OCS observations (Figure 10). This could potentially be remedied by dramatically increasing the rate coefficient of reaction R2. However, such a modification would still not allow for the efficient conversion of OCS to CO.

Details are in the caption following the image
CO production and loss for our nominal case. The left two panels show the percentage of the production or loss that is due to a particular reaction. Unfilled (white) space is due to reactions not listed. All reactions that contribution at least 20% at any altitude are shown. The right panel shows the absolute production and loss curves for CO. In the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0400 OCS model configurations (not shown), loss exceeds production at 40 km due to the absence of reaction R5.
Details are in the caption following the image
Profiles of OCS and CO for different model configurations. These profiles are not sensitive to the upper boundary S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0401 flux.
Yung et al. (2009) proposed that the primary OCS destruction pathway is

This reaction rate has been measured in the lab (Lu et al., 2006), yet it is still difficult to test the impact of this reaction on OCS abundances due to the poor knowledge of S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0403 chemistry. As discussed in section 3.5, the S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0404 chemistry is some of the most uncertain in the entire model. In our model, this reaction is insufficient to reproduce the OCS observations because the concentration of atomic sulfur is five orders of magnitude lower than what was proposed by Yung et al. (2009). There is also no a priori reason why atomic sulfur would be concentrated at the base of clouds, leading to the observed drop in OCS concentration at that particular altitude. Indeed, in our model, this reaction is fastest within 30 km of the surface, not at 35 km where the OCS mixing ratio declines (Figure 9). This does not rule out this mechanism, however, as a better understanding of S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0405 chemistry is needed to definitively test it.

In this work, we propose a new pathway for OCS destruction. The reaction
is exothermic in contrast to R2. It is a three-body reaction and so would likely require some unknown intermediate steps. To estimate the rate constant, we propose that it is similar to that of other three-body reactions with urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0407. As shown in Figure 10, this allows us to match the observed OCS profile. The observed value of OCS above the clouds can also be achieved given some tuning of this rate constant and the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0408 profile. While we favor reaction R5, we also note that, as with reaction R2, this pathway has not been verified either by lab studies or ab initio calculations.

Due to reaction R5, our model does produce an increasing CO mixing ratio with altitude in the lower atmosphere but overshoots the observed values. As noted above, the vertical profile of CO is more sensitive to the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0409 profile than many other species, and the chemistry in the cloud region is dominated by reactions with unmeasured rate constants. It is possible that future work could use this sensitivity of CO to constrain the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0410 of Venus's atmosphere; however, better chemical rate constants are first required.

3.3 O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0411

A long-standing problem with middle atmosphere photochemistry models of Venus is the overabundance of O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0412 when compared to observations. In the middle and upper atmosphere, atomic oxygen is primarily produced by the photodissociation of CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0413 and SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0414. The direct recombination of CO and O to form CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0415 is much slower than photolysis, allowing atomic oxygen to accumulate. While there is no resolved vertical profile for any O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0416 species, there are upper limits on the column abundance above the cloud deck of O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0417 that range from urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0418 to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0419 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0420 (Marcq et al., 2017). Our model configurations produce O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0421 column abundance values between 5 and urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0422 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0423. The lower end of these is values ( urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0424 cm urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0425) that correspond to the model configurations that come closest to matching the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0426 observations and are consistent with previous middle atmosphere models.

O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0427 has also been observed in the Venusian atmosphere with a mixing ratio of 0.1–1 ppm at 100 km (Montmessin et al., 2011). In contrast to O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0428, O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0429 mixing ratios in our model configurations are consistently below these observed levels. At 100 km, most model configurations have urn:x-wiley:jgre:media:jgre21306:jgre21306-math-04300.04 ppm of O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0431, consistent with Zhang et al. (2012). The outlier was the K urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0432 configuration that had urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0433 ppm. This is because, with the K urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0434 profile, the O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0435 mixing ratio peaks at urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0436 ppm at 90 km and declines steeply at higher altitudes. This is consistent with the results presented in Krasnopolsky (2012).

For both O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0437 and O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0438, our Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0439 configuration matches previous model results but diverges from the observations. Recent reanalysis of Venus Express observations has found cloud top O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0440 mixing ratios of 10 to 20 ppb concentrated in the polar regions (Marcq et al., 2019). These values were also observed to vary with local time and year to year. This may suggest that understanding the oxygen observations will require 2-D or 3-D models that include transport in the middle and upper atmospheres. Alternatively, the fact that O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0441 is overabundant while O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0442 is under-abundant may imply some important, but yet unknown, oxygen chemistry in the Venusian atmosphere.

3.4 Chlorides

The only chloride species with observed abundances are HCl (Arney et al., 2014; Krasnopolsky, 2010b; Mahieux et al., 2015; Sandor & Clancy, 2012, 2017) and ClO (Sandor & Clancy, 2018). In both the lower and middle atmospheres, HCl is observed to have a mixing ratio of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0445 ppb, although cloud top abundance has been observed as low as 100 ppb (Sandor & Clancy, 2012). This near constant mixing ratio is consistent with our model results (Figure 11). The only inflection seen in our model is above 90 km where photolysis begins to become important. This acts as a source of chlorine in the upper middle atmosphere producing a variety of other species.

Sandor and Clancy (2017) observed a decline in the HCl mixing ratio in the upper middle atmosphere with altitude. The gradient reported by Sandor and Clancy (2017) is steeper than our steady-state solution. Sandor and Clancy (2017) also note that there appear to be secular variations of up to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0446 ppb that occur on timescales of a month and do not appear correlated with local time. The difference between our model and these observations could be due to the temporal variations in HCl in Venus's atmosphere or a discrepancy between the true HCl photolysis rate and that calculated by the models.

The observations of Sandor and Clancy (2017) are in conflict with the Venus Express observations analyzed by Mahieux et al. (2015). Mahieux et al. (2015) found a cloud top value of 100 ppb increasing to 1 ppm at 110 km. They also found that this increasing mixing ratio with altitude was not sensitive to latitude or local time. Such a profile would require some high-altitude chlorine source, which is not included in this work.

The observations of ClO by Sandor and Clancy (2018) have a mixing ratio of urn:x-wiley:jgre:media:jgre21306:jgre21306-math-04472 ppb at 85 km. This is four orders of magnitude higher than our nominal model values. This value is very sensitive to the photolysis rate of HCl. To illustrate this, it is useful to compare our nominal model configuration to the Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0448 + S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0449 configuration. This case has significantly less SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0450 than the nominal model, making it in better agreement with the observed middle atmosphere SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0451 abundance (see section 3.1). The Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0452 + S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0453 has significantly more ClO, implying that the excessive amount of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0454 in our nominal case shields the middle atmosphere from UV radiation and thereby limits the HCl photolysis rate. However, even the higher abundance of ClO in the Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0455 + S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0456 model configuration is a factor of 3 below the observed value. This difference could be due to uncertainties of the model HCl photolysis rate, some unknown Cl source as proposed by Sandor and Clancy (2018), or as-yet unobserved temporal variability in ClO similar to HCl.

Details are in the caption following the image
Model chlorine species results from our nominal and Low urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0443 + S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0444 configurations. Observations for HCl from Krasnopolsky (2010b), Sandor and Clancy (2012) (vertical profile), Arney et al. (2014), and Sandor and Clancy (2017) (boxed region). ClO observation from Sandor and Clancy (2018).

3.5 Other Species

Model profiles for NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0457 species are presented in Figure 12. NO has been observed by Krasnopolsky (2006) and was interpreted to be sourced from lightning in the lower atmosphere. We find that any lower atmosphere NO mixing ratio is readily mixed into the middle atmosphere. This is consistent with Krasnopolsky (2006) but is agnostic to the source of NO in the lower atmosphere. No significant chemistry affects the NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0458 species outside of photochemistry in our current model.

Details are in the caption following the image
Nominal model profiles for the NO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0459 group. Observation of NO from Krasnopolsky (2006).

Figure 13 shows the model profiles for the S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0460 species. Data analysis of the Venera 11 spectra by Krasnopolsky (2013) has provided measurements for S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0461 and S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0462 abundance below the clouds. In that work, an updated lower atmosphere model is also presented that can match these observations by including the photolysis pathway S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0463 urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0464 S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0465 + S. The S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0466 chemistry is some of the least constrained in the entire model. The reactions used in this work come primarily from Moses et al. (2002), although those rate constants still lack experimental or ab initio validation. While the pathway proposed by Krasnopolsky (2013) may be correct, more lab work is needed to properly understand this system.

Details are in the caption following the image
Nominal model profiles of S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0467 and reduced sulfur species. Observation for S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0468 and S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0469 from Krasnopolsky (2013).

4 Summary

In this work, we perform the first detailed analysis of a Venus atmospheric chemistry model that extends from the surface through the middle atmosphere. We find that a large flux of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0470 transported through the cloud layers, independent of the eddy diffusivity profile, prevents the model from matching the low SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0471 observed around 75 km. We suggest that there is either some unknown chemical sink in the cloud region or that interactions with the cloud droplets themselves prevent the transport of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0472. We also propose a new chemical pathway by which SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0473 below the cloud deck could convert OCS to CO. In the preceding sections, we have detailed how each chemical system behaves; however, it is equally important to understand how these systems interact. These systems and interactions are summarized in Figure 1. In this section, we also present a summary of outstanding questions raised by this model and the highest priority laboratory studies that could help address them.

The base of the clouds is one of the most chemically active regions in the Venusian atmosphere. The excess SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0474 released from the thermal decomposition of H urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0475SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0476 is highly reactive as it finds pathways to form the more stable SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0477. In the process, it oxidizes CO to CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0478. Simultaneously, it may be actively producing CO from OCS (section 3.2). However, this important chemical pathway still lacks laboratory measured rate constants adding large uncertainty to this and similar models.

In the middle atmosphere, photochemically produced atomic chlorine is highly reactive. In this model, the abundances of ClO, ClCO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0479, and Cl urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0480SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0481 were not sufficient to be significant reservoirs of CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0482, SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0483, or O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0484. They can, however, act as important chemical catalysts (Mills & Allen, 2007).

There are two key factors that connect all the chemical groups we have discussed: the radiation field and oxygen abundance. When any species becomes too abundant, it can shield other species from photolyzing radiation. From a modeling perspective, this means an overabundance of one phytochemically active species can impact chemically unrelated species (see the example of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0485 and ClO in section 3.4). Given the large temporal variability of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0486 (Encrenaz et al., 2012), future work should also explore the effects such shielding may have in the Venusian atmosphere.

The oxygen abundance of the middle atmosphere sets the relative abundance for many species. Yet, as noted in section 3.3, models consistently overestimate the O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0487 abundance. ClCO(O) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0488 may be an important intermediate species in facilitating CO oxidation to CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0489 in the middle atmosphere (Mills & Allen, 2007). More work is needed to understand the plausibility of these reactions and how such pathways could impact these chemical networks.

As an integration of previous models and observations, we think it is useful to summarize the key outstanding questions in Venus atmospheric chemistry. These questions are all demonstrated in earlier sections and have been discussed to varying degrees by previous authors. They are presented in no particular order.
  • SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0490 flux through clouds: What limits the flux of SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0491 through the cloud deck and why is this process so temporally variable? (section 3.1)
  • OCS destruction at 35 km: What chemical pathway controls the sudden decline in OCS at urn:x-wiley:jgre:media:jgre21306:jgre21306-math-049235 km altitude? (section 3.2)
  • Surface chemistry: What are the surface reactions that buffer CO and OCS and on what timescales can those be maintained? (section 3.2)
  • O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0493 abundance: What process is catalyzing photochemically created O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0494 back to CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0495? (section 3.3)

Addressing these questions will require both a better understanding of physical processes (i.e., transport and condensation) and new laboratory measurements of reaction rates.

In this work, we find that, due to chemical interactions, different species exhibit very different responses to urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0496 (Figures 6 and 10). While our 1-D work suggested that the observed drop in the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0497 mixing ratio cannot be resolved by transport, future 3-D GCM studies including more realistic radiative heating and cooling, microphysical cloud processes, gas chemistry, and hydrodynamics could shed more light on the detailed chemical transport mechanisms in the system and further investigate this problem and the possible solution. The globally averaged urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0498 from these 3-D models should also be complimented by more observational estimates of the urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0499 through revisiting either the methods of previous studies (Lane & Opstbaum, 1983; Woo & Ishimaru, 1981; von Zahn et al., 1979) or new methods.

In our model, urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0500 of the reaction rate coefficients have no lab-measured value and many more rate coefficients are based solely on upper limits or rates at one temperature (Table S1). Given this, it is important to highlight the reactions that have the largest effect in our model and are also lacking any constraints on their reaction rate constants.

The reaction for which lab measurements would have the largest impact on our understanding of lower atmosphere chemistry is SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0501 + OCS urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0502 products. As discussed in section 3.2, understanding this reaction is key to understanding if and how the CO and OCS cycles are linked in Venus's atmosphere. A related process is CO + SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0503 = CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0504 + SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0505, which is critical for re-oxidizing CO at the cloud base and also has no measured rate coefficient.

Another gap in our laboratory measurements is understanding the chemistry and stability of the Cl urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0506O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0507 groups. Species such as ClCO(O) urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0508 and ClS have been proposed to be important in the recycling of O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0509 back to CO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0510 and thus bringing model O urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0511 values in line with observations (section 3.3 and Mills & Allen, 2007).

Finally, the S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0512 chemistry plays a large role in several outstanding questions. S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0513 species have been proposed to be the sulfur source for the SO urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0514 inversion layer (section 6 and Zhang et al., 2012). Our ability to interpret the observed lower atmosphere abundances of S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0515, S urn:x-wiley:jgre:media:jgre21306:jgre21306-math-0516, and OCS is also limited by our understanding of these rates (section 3.2).


Data for species profiles and reactions rates for all models presented in this work are archived at the Harvard Dataverse https://doi.org/10.7910/DVN/QLHLPR. We thank Cheng Li for providing some of our analysis tools, Wencheng Shao for identifying problems in early model results, and Huazhi Ge for suggesting the three-body OCS reaction. We thank Claire Newman and an anonymous reviewer for their helpful comments and suggestions. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant DGE 1339067. X. Z. acknowledges support from NSF Grant AST1740921.