Plain Language Summary
Water vapor plays an important role in the weather and climate on Mars, even though little of it remains today. The behavior of water vapor has been studied for decades, yet how water vapor varies with altitude, especially close to the surface, remains an open question. In this study, we use measurements from two instruments on the Mars Express satellite to learn about the near-surface water vapor. By combining measurements from the SPICAM and PFS spectrometers, a composite full-year climatology is assembled. We measure the total amount of water vapor with great accuracy, and also obtain information about the vertical distribution. The north polar cap is studied in detail during early summer, when part of the polar ice cap sublimates into water vapor and is transported south. The results are compared to model data from the Mars Climate Database, and significant differences between the observations and the model are identified. The total water content is found to be smaller than model estimates, while observations indicate that more water than expected is confined near the surface. This suggests that some aspects of the atmospheric transport processes are not currently fully understood.
1 Introduction
Water vapor on Mars was first detected in 1963 with the use of a ground-based telescope which observed eleven near-infrared absorption lines (Spinrad et al., 1963). Since then, numerous observatories, ground-based, Earth-orbiting, Mars-orbiting, landers and rovers, have observed the highly volatile trace gas. Even as a minor atmospheric constituent, water vapor plays a major role in shaping the climate on Mars (along with the CO2 and dust cycles). Water controls the stability of the atmosphere, as H2O photolysis supplies hydroxyl radicals, the main oxidant of the Martian photochemical cycle (e.g. McElroy and Donahue (1972)), and impacts the radiative equilibrium through cloud formation (Madeleine et al., 2012).
The Mars Atmospheric Water Detector (MAWD) instruments on the Viking orbiters provided evidence that the Northern polar cap is the primary source of atmospheric water, and also indicated a strong north–south asymmetry in the atmospheric water abundance (Farmer et al., 1976; Jakosky & Farmer, 1982). The most complete climatology, upon which modern Martian water climatology is based, was obtained by the Mars Global Surveyor mission and its Thermal Emission Spectrometer (Smith, 2002, 2004). A revised retrieval scheme on TES observations provide an annual reference water vapor cycle with column abundance maximum at high latitudes during midsummer in both hemispheres, reaching a peak of ∼60 pr-μm on average in the north, and ∼25 pr-μm in the south (Pankine et al., 2010). Low water abundances are observed during fall and winter at middle and high latitudes of both hemispheres. General circulation models along with TES observations indicate that water from the southern summer maximum is transported to the Northern Hemisphere (NH) more efficiently than the reverse process (Montmessin et al., 2004).
One of the main objectives of the Mars Express (MEX) orbiter is to study the water cycle on Mars. Three spectrometers onboard MEX can measure the water vapor abundance in different spectral bands: The Planetary Fourier Spectrometer (PFS), The Observatoire pour la Minéralogie, l’Eau, les Glaces, et l’Activité (OMEGA) and SPectroscopy for the Investigation of the Characteristics of the Atmosphere of Mars (SPICAM). For the purpose of this study, PFS was selected for its coverage of water vapor diagnostic features in the thermal infrared (TIR) domain, while SPICAM was chosen over OMEGA to cover the near infrared (NIR) due to its higher spectral resolution, and the presence of CO2 bands near the 2.6 µm water feature for OMEGA.
1.1 Water vapor vertical distribution
Until recently, knowledge of the near-surface H2O profile on Mars mostly relied on general circulation models. The vertical distribution of water vapor has been inferred from nadir measurements (Fouchet et al., 2007; Pankine & Tamppari, 2015) and measured directly by solar occultation viewing geometry with SPICAM on Mars-Express since 2004, and with the ExoMars Trace gas Orbiter (TGO) and its infrared spectrometers NOMAD and ACS since 2018. SPICAM occultation campaigns were not a primary focus of the spacecraft and are therefore not performed very often, whereas TGO, with its orbit adapted for occultation measurements with good vertical and temporal resolution, allows the study of dynamical behavior of water distribution including escape processes in great detail. With this technique, new knowledge has been obtained on the vertical distribution of water in the upper atmosphere as a result of supersaturation above the hygropause, and the occurrence of high altitude water during dust storms (e.g. (Aoki et al., 2019; Fedorova et al., 2020)). SPICAM solar occultations were also used to produce a climatology of vertical distribution covering the Martian years (MY) 27 to 34, that encompassed two global dust events (Fedorova et al., 2021). With solar occultation measurements one can obtain very fine vertical resolution, nevertheless, measurements below 10 km are relatively sparse as aerosol loading in the lower atmosphere leads to high opacities which reduces the transmittance significantly. The lower limit for observation is typically 5-10 km for dust-free conditions, and as high as 20-30 km during the dusty perihelion season (e.g., Aoki et al., (2019)). Only under very clear conditions will solar occultation observations be able to probe below 10 km, however such conditions mostly occur at high latitudes. As a result, information about the low-atmosphere water vapor profile remains exceptional.
Below 10 km, surface-atmosphere interactions such as convection, frost sublimation and deposition are expected to be the main forcers on the vertical distribution, while above 10 km water ice clouds are thought to be dominant (Montmessin et al., 2004; Richardson, 2002). Below the saturation level, controversy exists regarding whether water vapor is well mixed with CO2, or distributed in a more complex manner. Davies (1979) used Viking Orbiter 1 data to directly probe the location of water vapor with altitude for the first time. He found that H2O vertical distribution was indistinguishable from the dust vertical distribution and was well mixed up to about 10 km, which has been commonly assumed since. It is also argued that water is either confined to, or reduced in the lower atmosphere, depending on season and location. This ties in with to which extent there is exchange of water vapor between the atmosphere and the regolith.
Adsorption of CO2 by the Martian regolith was first suggested by Davis et al., (1969), and the theory was later expanded upon to include water vapor by Fanale and Cannon (1971), whose adsorption isotherm expression has been widely used since (although found to require modification by Savijärvi and Harri (2021)). Using data from Viking 1 and 2, Jakosky et al., (1997) showed a nocturnal depletion of atmospheric water vapor, suggesting a diurnal exchange cycle between the porous regolith and the atmosphere. Similar results were obtained with the thermal and electrical conductivity probe on the Phoenix lander by Zent et al., (2010) and Fischer et al.,(2019). It was found that the layer that experiences a diurnal exchange of water with the surface was 0.5–1 km deep (Tamppari et al., 2010). This phenomenon was again confirmed by Harri et al., (2014) and Martínez et al., (2017), who used the REMS-H device on Curiosity rover to derive water vapor volume mixing ratios. Savijärvi and Harri (2021) found that regolith exchange is largely indifferent to surface properties, and that diurnal adsorption/desorption generates approximately 1% variation in the column abundance, which matches Earth analogue measurements very well. The results of Fouchet et al., (2007) indicate that the vertical distribution is controlled by an intermediate state where the water is controlled by atmospheric saturation on one hand, and confined to a surface layer on the other, pointing to significant regolith-atmosphere exchange processes. This result is inferred by investigating the correlation of water columns and pressure, and was not observed directly. Maltagliati et al., (2011) and Trokhimovskiy et al,. (2015) also attempted to discern a diurnal exchange process between atmosphere and regolith, but found no evidence of local time variation in H2O abundances. Thus, the extent of exchange between regolith and atmosphere remains an open question.
1.2 Spectral synergy
When observing an atmosphere in nadir viewing geometry, the outcome is normally a column abundance value of the target species. However, it is possible to obtain information about the vertical distribution of the species by combining multiple spectral domains in the retrieval process. This approach is commonly referred to as a spectral synergy, and was developed for Earth observation by Pan et al., (1995, 1998), who predicted higher sensitivity to near-surface layers of CO if near and thermal infrared spectral bands were combined. This was later confirmed by Edwards et al., (2009), who demonstrated that combining NIR and TIR measurements in a common retrieval allowed for a significantly higher sensitivity in the troposphere. The method has also been used to increase near-surface sensitivity to other gasses such as CO2 (Christi & Stephens, 2004), O3(Landgraf & Hasekamp, 2007) and CH4 (Razavi et al., 2009).
TIR measurements are mostly sensitive to the middle atmosphere (at the origin of the photon emission) where the temperature contrast of the atmosphere with respect to the surface is high. NIR measurements on the other hand are sensitive to any molecule present along the column as the technique relies on solar photons traversing the entire atmosphere back and forth. Although Trokhimovskiy et al., (2015) indicate the NIR technique is mostly sensitive to the atmosphere below 30 km, it is only true from a mixing ratio perspective, which favors the denser layers of the atmosphere. In other words, any given change in H2O mixing ratio will be easier to sense in the bottom of the profile as pressure and number density is continuously increasing towards the surface. If seen from a number of molecules perspective, the NIR inversion technique has no preference to a particular position of the column, unless this portion concentrates more water molecules at a specific location. One must note however that dust modulates this assertion. At high dust opacity, part of the incoming flux does not reach the surface and is sent back to space without sampling the entire column. Only in such cases will the NIR technique become altitude dependent.
This difference in sensitivity of NIR and TIR can be viewed as a difference in the shape and peak altitude of the weighting function of water vapor retrieval in a particular wavelength domain, and has been advocated to explain the dispersion of H2O column abundance values retrieved by the various instruments of MEX (Tschimmel et al., 2008). On the other hand, the difference in sensitivity can also be considered a way to offer simultaneous access to different regions of the atmosphere, leading to the derivation of more than a single parameter representative of the whole column, as is usually the case with instruments that study water vapor using nadir observations. In fact, combining two spectral domains increases the degree of freedom of the signal (DOF). The DOF gives an estimate of the number of independent bits of information in an atmospheric measurement (Rodgers, 2000), and a DOF higher than 1 indicates the presence of some amount of profile shape information.
If attempting to retrieve vertical information with only one instrument, one could argue that as the single instrument is primarily sensitive to a specific altitude region, the obtained vertical confinement is not a “real” partitioning. Instead, the obtained partitioning might be a product of a lack of sensitivity to other, and perhaps wetter, altitude regions, thus producing an artificial vertical partitioning. This problem is avoided with the use of a spectral synergy, as each wavelength interval is susceptible to emission/absorption signatures in separate regions, and therefore obtains information from different altitudes.
This consideration led Montmessin and Ferron (2019) to investigate the potential for a synergistic retrieval of water vapor in the Martian atmosphere using MEX, as the spacecraft constitutes the only asset at Mars observing water in both NIR (SPICAM, OMEGA, PFS) and TIR (PFS) spectral intervals. Despite their differences in field-of-view, sampling and coverage, SPICAM (NIR) and PFS (TIR) were selected for this study as the two have the most extensive records of water vapor retrievals on Mars among the MEX instruments (Fedorova et al., 2006; Fouchet et al., 2007; Giuranna et al., 2019; Trokhimovskiy et al., 2015). As Montmessin and Ferron (2019) concluded on the promising potential for a synergistic retrieval of water vapor on Mars with MEX, this work is intended to follow-up on this earlier study and present the analysis of a multi-annual dataset covering the period from MY 27 to 34.
The first part of the manuscript provides an overview of the instruments used in this study (Section 2), and continues in Section 3 with an outline of the synergistic retrieval method, including a description of the selection of measurements within the dataset. The results are presented in Section 4, where in 4.1 a complete synergistic column abundance climatology is presented, followed by a comparison of the column abundance between the synergy, the model and the single spectral domain retrievals are made, before the vertical and spatial distribution is elaborated upon. A discussion of the results and how they compare to previous works follow in Section 5, and Section 6 concludes the findings of this study.
2 Instruments
The Mars Express mission was launched in June 2003, and began nominal science operations in mid-January 2004 (Chicarro et al., 2004), corresponding to the very end of MY 26. From a quasi-polar orbit with a period of 7.5 hours, MEX has a particularly detailed view of the polar caps at the sublimation onset. With three instruments able to measure the atmospheric water vapor content (OMEGA, PFS, SPICAM), either in the solar reflected or in the thermal component, MEX has delivered a vast amount of valuable data with complete global and seasonal coverage. The PFS and SPICAM instruments cover the thermal and near-infrared domains, respectively, within which water vapor possesses diagnostic emission/absorption signatures. As each spectral interval provides a distinct sensitivity along the vertical, observations of the same species in separate wavelength regions provide constraints on the vertical distribution.
The measurements used in the following analysis were retrieved from nadir observations, and were selected according to a number of criteria to ensure satisfactory quality of every individual measurement, sufficient geographical and seasonal coverages, and a minimum error of radiative transfer modeling due to surface inhomogeneity (Montmessin & Ferron, 2019). For a detailed description on the selection and averaging processes used for the creation of a dataset compatible with a synergistic extraction of water vapor, the reader is referred to Montmessin and Ferron (2019).
2.1 Mars Express PFS
The Planetary Fourier Spectrometer is an infrared spectrometer with two wavelength channels optimized for atmospheric sensing. The short wavelength channel covers the range 1700-8200 cm-1with a full width at half maximum (FWHM) of the instantaneous field of view (FOV) of 1.6°, while the long wavelength channel spans the 250-1700 cm-1 (5.88-40 µm) with a FWHM FOV of 2.8°, which at an altitude of 250 km corresponds to a 12 km diameter surface footprint. Both channels have a spectral resolution of 1.3 cm-1. Only the long wavelength channel was utilized for this work. For further details, see Formisano et al., (2005) and Giuranna et al., (2005).
For the synergistic approach, several windows in the long wavelength channel were selected. The windows from 8-10 µm and 19-25 µm were used to obtain surface temperature and dust model properties, the region at 12-19 µm is dominated by the absorption of the 15 µm CO2vibrational transition which was used to retrieve atmospheric temperature profiles, while the 20-35 µm thermal emission band was used to retrieve the water vapor abundance, henceforth referred to as TIR.
The PFS spectrum is used to retrieve several parameters, such as surface temperature, dust properties and water vapor column abundance. Because of this, a high signal-to-noise ratio (SNR) is required, and one individual spectrum obtained with PFS is not satisfactory. Therefore, the retrievals were performed on the average of nine consecutive spectra. The total time passed between the acquisition of the first spectrum to the last of the nine to be averaged is 108 seconds, as it takes 4.5 seconds to acquire a single PFS interferogram and the repetition time is 8.5 seconds (Fouchet et al., 2007).
After years of operation, an issue with PFS caused the interferogram peak to not always be centered. The instrument line-shape used here (a sinc function with 1.3 cm-1 FWHM) is then not optimal, and could lead to biased water vapor retrievals, with a tendency of being too low. This issue started around orbit 6000 (MY 29), became particularly relevant after orbit 7500 (MY 30), but data obtained in MY 32 and after are less affected. In an effort to largely avoid this problem, we exclude all measurements during MY 30 and MY 31 from further analysis.
2.2 Mars Express SPICAM
The SPICAM UV-IR instrument (Spectroscopy for the Investigation of the Characteristics of the Atmosphere of Mars) is a dual-channel spectrometer designed to study the Martian atmosphere from top to bottom (Bertaux et al., 2006). In this study, only the IR channel was utilized working in the spectral range of 1-1.7 µm with a spectral resolution of 3.5-4.0 cm-1, a complete description of which can be found in Korablev et al., (2006).
In nadir viewing geometry, the IR channel has an instantaneous FOV of 1°, corresponding to a 4 km footprint on the surface when the spacecraft is near the 300 km pericenter of its orbit. The incoming flux is separated into two detectors, where detector 1 was used for this work as it provides significantly higher performance in nadir. The wavelength interval 1.34-1.43 µm is defined as the NIR range for the synergy, as it covers the strong water absorption band at 1.38 µm. Averages of ten SPICAM-IR spectra are demonstrated to have a SNR sufficient for reliable retrievals of water vapor column abundances (Fedorova et al., 2006; Trokhimovskiy et al., 2015). For the sake of the synergy, the SPICAM observation closest in time to the center PFS spectrum is selected, and averaged together with the seven previous and following spectra. The 15 spectrum average has a FOV similar to that of the nine PFS spectrum average. Together, the SPICAM and PFS average spectra constitute a co-located observation.
3 Data set and retrieval
The first time a synergistic retrieval method was tested on a planet other than Earth, atmospheric H2O on Mars was retrieved by combining measurements from PFS and SPICAM on Mars Express, probing the TIR and NIR spectral intervals respectively (Montmessin & Ferron, 2019). In a nadir viewing geometry, retrievals have traditionally returned a single item of information regarding the target species: either the column-integrated abundance (CIA) in the case of NIR, or the middle atmosphere concentration in the case of TIR, leaving open the question of how the species is distributed along the vertical and in particular whether and how it might interact with the surface. By taking advantage of multiple spectral regions, it is possible to increase the DOF for the signal, and thereby resolve the vertical partitioning of water vapor on Mars.
In the earlier demonstration of the synergy method applied to Martian water vapor, a subset of 449 co-located observations from 133 orbits distributed through MY 27 were presented (Montmessin & Ferron, 2019), showcasing that the synergy brings additional robustness to the retrieval of water vapor column abundance, and provides insight into the vertical distribution of water vapor. In this study, we expand on those findings, and conduct a comprehensive analysis of the complete synergistic dataset available from MEX, which contains nearly 200 000 measurements.
The dataset presented here consists of co-located observations taken over 1379 individual orbits distributed across seven Mars years from Ls 334° of MY 26 to Ls 297° of MY 34, with no measurements from MY 30-31. The geographical and seasonal coverage is highly variable from year to year, several being quite sparsely covered. Some sparsity is due to operational constraints, as not all instruments can be concurrently active, while most is due to the requirement of co-located measurements from both SPICAM and PFS.
3.1 Synergistic retrieval routine
The synergistic approach requires a set of co-located PFS and SPICAM observations on which to apply the retrieval method. To obtain a satisfying PFS SNR for the fitting of multiple parameters, nine consecutive spectra are averaged together. The SPICAM observation closest in time to the central PFS spectrum is then selected and averaged with the seven observations prior to it and the seven after it, resulting in a combined FOV similar to that of the nine combined PFS observations. A screening process is conducted on this set of co-located observations, the details of which can be found in Montmessin and Ferron (2019). The simultaneous inversion of H2O follows the approach outlined in Montmessin and Ferron (2019), and will only be briefly described here.
A priori values of the water vapor and temperature profiles are extracted from the Mars Climate Database (MCD) based on the general circulation model developed at the Laboratoire de Météorologie Dynamique (LMD GCM) (Forget et al., 1999; Millour et al., 2018) with an uncertainty of the water equal to the abundance values. MCD version 5.3 is used. For each year the corresponding scenario is chosen, except for MY 34, which is not yet included (the version used was last updated on 11/01/2019). A composite scenario was therefore built for MY 34 by combining the scenario of MY 33 with the standard MCD dust storm scenario 4 and the warm and dusty scenario 7 (for the intervals Ls=180°-200° and Ls=200°-220° respectively).
Temperature and aerosol parameters are retrieved individually from the PFS average spectra, which are then injected into the synergistic routine. The overall spectral fitting procedure uses the HITRAN 2012 spectroscopic database (Rothman et al., 2013) as a baseline for the computation of absorption coefficients of H2O and CO2, and then relies on a Bayesian approach that consists in maximizing the probability that a given retrieval satisfies both the observed averaged spectra and falls within a range of plausible values specified by prior assumptions on the value and its dispersion. The weight of the prior assumption in the retrieval is dictated by its prior uncertainty, which is set equal to the prior water vapor column value.
Water vapor is inferred from the set of combined NIR and TIR spectra, by a simultaneous inversion from both spectral domains. In practice, the algorithm adjusts the water vapor abundance along the vertical profile at nine altitude points separated by 2.5 km from ground to 10 km, and by 5 km from 10 to 30 km. All points are correlated with a Gaussian kernel, such that the points are less strongly correlated when the distance between them is increasing. The results include a posterior covariance matrix, from which the DOF can be calculated from the sum of the trace of the matrix. The DOF normally fluctuates around 1 when the retrieval includes a single spectral domain (NIR or TIR), which implies only one independent parameter can be inferred from a water vapor measurement (e.g. the CIA), while with a higher DOF some information of the water vapor vertical distribution can be obtained.
Some example spectra are shown in Figure 1, where the selected NIR and TIR spectral intervals include strong diagnostic features of water vapor. The co-located observations are from early summer of MY 27, at high latitudes. The corresponding vertical profile obtained from the synergistic retrieval performed on the spectra is also shown and compared to the MCD prior profile, along with the averaging kernel and the posterior-to-prior error ratio by altitude for the synergy and each single spectral domain retrieval. The post-to-prior profile shows at which altitudes added information is coming from, indicating that the synergy is more sensitive to the lower atmosphere than both PFS and SPICAM.
We quantify the amount of information added by the synergy at each altitude level by comparing the synergistically retrieved error profiles to the MCD prior error profiles, shown in the bottom right panel of Figure 1. In this way, we demonstrate that the synergy does not simply reproduce the prior when calculating vertical profiles, and that for the lower atmosphere, the synergy brings more information than the single spectral domain retrievals. The MCD prior and the retrieved vertical mixing ratio profiles are close to identical above 15 km, but start to deviate below this, where the synergy provides a significant amount of added information.