Day-night differences in Mars methane suggest nighttime containment at Gale crater

We report new measurements of atmospheric methane by the Curiosity rover’s Tunable Laser Spectrometer that is part of the Sample Analysis at Mars suite (TLS-SAM), finding nondetections during two daytime measurements of average value 0.05 ± 0.22 ppbv (95% confidence interval CI). These are in marked contrast with nighttime background levels of 0.52± 0.10 (95% CI) from four measurements taken during the same season of northern summer. This large day-night difference suggests that methane accumulates while contained near the surface at night, but drops below TLS-SAM detection limits during the day, consistent with the daytime nondetection by instruments on board the ExoMars Trace Gas Orbiter. With no evidence for methane production by the rover itself, we propose that the source is one of planetary micro-seepage. Dynamical modeling indicates that such methane release is contained within the collapsed planetary boundary layer (PBL) at night due to a combination of nocturnal inversion and convergent downslope flow winds that confine the methane inside the crater close to the point where it is released. The methane abundance is then diluted during the day through increased vertical mixing associated with a higher altitude PBL and divergent upslope flow that advects methane out of the crater region. We also report detection of a large spike of methane in June 2019 with a mean in situ value over a two-hour ingest of 20.5± 4 ppbv (95% CI). If near-surface production is occurring widely across Mars, it must be accompanied by a fast methane destruction or sequestration mechanism, or both.


Introduction
The quantity, distribution, and behavior of methane in the atmosphere of Mars are of great interest to the planetary science community and the general public since the gas is recognized as a potential biosignature, due in part to its predominantly biological origin on Earth (Pachauri et al. 2014). Over the past two decades, Mars measurements from a wide variety of platforms (ground-based telescopes, orbiters, and rover) have reported values from zero to ∼45 parts per billion by volume (ppbv) that have fueled controversy over the measurements and their subsequent interpretation (Zahnle 2015). Reported measurement examples include ground-based telescope observations from the Canada-France-Hawaii Telescope (CFHT) in 1999 (global average value of 10 ± 3 ppbv (Krasnopolsky et al. 2004) and from the NASA Infrared Telescope Facility (IRTF) in 2003 (Villanueva et al. 2013) (∼45 ppbv near the equator), although subsequent ground-based measurements in 2006, 2009, and 2010 found nondetections of methane (Villanueva et al. 2013); and orbital measurements by the Planetary Fourier Spectrometer (PFS) on the Mars Express (MEX) in 2004 and 2010 (Geminale et al. 2011) (global average value of 15 ± 5 ppbv) and in 2019 (Giuranna et al. 2019) (spike of 25 ± 5 ppbv).
The Sample Analysis at Mars (SAM) instrument suite of the Curiosity rover (Mahaffy et al. 2012) has operated successfully on the surface of Mars in Gale crater for over eight years. During this time, the Tunable Laser Spectrometer (TLS-SAM) instrument of the SAM has reported occasional spikes in atmospheric methane (Webster et al. 2015(Webster et al. , 2018 up to 20 ppbv (Moores et al. 2019a) detected above a low persistent background level. By enhancing the TLS-SAM sensitivity by a factor of ∼25 using the gas processing and atmospheric enrichment of the SAM (Mahaffy et al. 2012;Webster et al. 2018), TLS-SAM discovered that over three Martian years, the low background levels suggest seasonal variation from 0.25 to 0.65 ppbv with a mean value of 0.41 ± 0.16 (95% CI) ppbv for the near-surface methane abundance (Webster et al. 2018). However, Gillen et al. (2020) used statistical analysis to question whether a seasonal cycle was present in the data. Up until December 2019, all the TLS-SAM enrichment measurements were made close to midnight local time on Mars. A&A 650, A166 (2021)  Notes. (a) Corresponding to the northern hemisphere (spring equinox at L S 0 • ); Curiosity landing site (4.6 • S, 137.4 • E) is just south of the equator.
Since April 2018, the ESA ExoMars Trace Gas Orbiter (TGO) has been using its Nadir and Occultation for Mars Discovery (NOMAD) (Liuzzi et al. 2019) and Atmospheric Chemistry Suite (ACS)  instruments to search for atmospheric methane, but has found none . A comprehensive investigation of Mars methane and organics with NOMAD includes more recent TGO results of continuing nondetection (Knutsen et al. 2021). While both instruments have excellent sensitivity in solar occultation at altitudes above ∼3 km, the nadir capability unique to NOMAD has a sensitivity of only ∼5 ppbv in the column average, so it would have difficulties to detect even large spikes of methane like the TLS-SAM ∼20 ppbv if the spike occurred only within the first kilometer of the surface. TGO has reported  a robust nondetection upper limit of ∼50 parts per trillion by volume (pptv) of methane in the Martian atmosphere above a few kilometers from the surface over a wide range of latitudes sampled during northern autumn (solar longitudes L S = 160-240 • ). If methane is long-lived, and even ignoring the occasional spikes reported by TLS-SAM, it is difficult to reconcile the TGO low upper limit of 50 pptv with the pervasive background measurements of TLS-SAM that average 410 pptv. The concern is one of excess methane buildup over time. Korablev et al. (2019) calculated that if Gale crater were continuously providing 410 pptv that mixed out globally, then this flux of background emission from Gale crater could only have been going on for at most 24 Martian years before the TGO detection limits would be reached. Taking into account the larger spikes of methane reported by TLS-SAM, this 24-yr time frame would be even shorter. It is highly unlikely that methane emissions have been occurring only over the recent decades, and so we are left with an apparent discrepancy between the rover and orbiter data sets.
Using the TGO ACS spectral retrievals near 3.3 µm, Olsen et al. (2020) recently reported detection of 50-200 ppbv ozone in the Martian atmosphere, and suggested that prior methane detections reported by orbiter (Planetary Fourier Spectrometer, PFS), ground-based (telescope) and rover (Curiosity) may all be affected by or attributed to the presence of ozone, not methane, in the Martian atmosphere. In response to this suggestion, Webster et al. (2020) showed that TLS-SAM is of a much higher spectral resolution than all other instruments, including ACS, and in the recorded Martian spectra readily distinguishes methane from ozone lines, the latter not being detected after loss in the instrument's enrichment cell.
Following the first observations of plumes in Martian methane by Mumma et al. (2009), transient spikes in Martian methane have since been reported by several observations (Table 1) since Curiosity landed in 2012. From the Curiosity rover, TLS-SAM recorded three spikes of 5-10 ppbv occurring over the first seven years of operation, the first one on June 15, 2013, of 5.8 ± 2.3 ppbv (Webster et al. 2018). From Mars orbit, using a staring-mode enhancement, the PFS instrument onboard the Mars Express orbiter recently reported (Giuranna et al. 2019) the detection of 15.5 ± 2.5 ppbv on the same date as the first Curiosity spike in June 2013, and over the Gale crater region, the potential source region identified as east of the crater. TLS-SAM reported (Webster et al. 2015(Webster et al. , 2018 two additional spikes in the period 2013-16 distinguished over the numerous other measurements at low background levels. More recently, from the Earth, using the NASA IRTF at Maunakea, Hawaii, Novak et al. (2019) reported detection of Mars methane at the 25 ppbv level during January 2017 and January 2018. These detections were consistent with this same group's earlier detections (Novak et al. 2019) of plumes up to 35 ppbv in similar geographic regions.

New measurements of the diurnal and seasonal variability
The TLS-SAM enrichment measurements and experiment protocol have been described in detail in earlier publications (Webster et al. 2015(Webster et al. , 2018 and their supplementary material. The five new measurements presented here were made using exactly the same protocols and instrument run scripts on Mars as earlier measurements. To summarize: After evacuation of the sample (Herriott) cell, the Martian atmosphere is ingested across a molecular sieve material to preferentially remove carbon dioxide and effectively enrich the methane amount (by ∼ × 25) in the ∼5 mbar cell pressure reached after the two-hour ingest. Then, with the cell closed, the laser scans over the three strong methane lines every second, and on board, TLS captures average spectra over sequential A166, page 2 of 14 C. R. Webster et al.: Day-night differences in Mars methane suggest nighttime containment at Gale crater Notes. SEM, standard error of the mean; L S , solar longitude; CI, confidence interval; LMST, local mean standard time. The global pressure multiplier is derived from in situ REMS pressure measurements and is the number by which the original measured in situ values of methane were multiplied to correct the results to the global mean annual mixing ratio given in the last two columns. The in situ values can be found by dividing the last two columns by the global pressure multiplier, and are also given in the appendix tables. The quoted SEM value is calculated from the square root of the sum of the squares of the errors on the individual methane lines (see appendix). Earth date and ingest time refer to the time at which the two-hour gas ingest was started. The enrichment results plotted in Fig. 1 and day/night discussion do not include the large spike observed on sol 2442.16 that is described later. 2.7-min periods that are downloaded. Analysis of each of these 2.7-min spectra produces a measurement "point", the process repeated for each of 26 full cell measurements. The cell is then pumped out and an additional 26 "empty cell" spectra are downloaded for analysis (the 52 individual data point values are given in the appendix). All enrichment measurements given in Table 2 result from differencing full and empty-cell measurement mean values. Table 2 includes a large spike of ∼20 ppbv observed on June 2019 that is described in the next section. Figure 1 panel A plots all our enrichment results to date (over a 70-month period), except for the high spike of ∼20 ppbv seen in June 2019 that is not considered representative of the background values. Figure 1 panel B expands the northern summer subset of the data to better illustrate the large day-night differences. We highlight the recent MY35 sequence of three runs (day-night-day) taken over a 30-sol A&A 650, A166 (2021) Fig. 2. Spectra and data points for the June 19, 2020, spike in methane. Left: actual recorded average for the full cell following the two-hour ingest (black) and a sequential empty cell (red) for comparison. TLS records both direct and second harmonic spectra, the latter providing a better signal-to-noise ratio from a method of high-frequency laser modulation that discriminates against noise and broad interference fringes that are seen in the lower direct absorption spectrum. We note that the direct absorption spectrum analysis produced the same abundance value as the second harmonic result within the quoted errors. Right: plots of retrieved in situ CH 4 abundances for each data collection point (26 each for full-and empty-cell runs) as measured in the Herriott cell using 2f spectra comparison with HITRAN 2016 (Gordon et al. 2017). Mean values and standard errors (±1sem) are shown. The reported 20.5 ppbv in situ spike results from taking the difference (full-empty), and dividing by the enrichment factor of 25. Detailed data tables are presented in the appendix.
period. The daytime measurements are the only nondetections that have ever been observed when using the enrichment protocol, while the middle nighttime measurement of 0.52 ppbv falls as expected on the nighttime values that for this season embrace three Mars years.

Large spike in TLS-SAM atmospheric methane
We report here a new detection of a spike in Mars methane of 20.5 ± 4 (95% CI) ppbv measured in situ in June 2019 that is the highest ever recorded by the Curiosity rover TLS-SAM. When a subsequent measurement was made four days later, this spike had disappeared. It diminished by a factor of 100 to a low background value of 0.22 ppbv. Figure 2 shows the full-and empty-cell averaged spectra recorded for the June 19, 2019, enrichment run and the individual associated data points. For the spike, the strongest methane spectral line is ∼0.2% deep at line center, which HITRAN identifies at our cell temperature and pressure as due to a methane mixing ratio of ∼563 ppbv in the cell. Subtracting our emptycell equivalent value (∼50 ppbv) and then dividing the result by our enrichment factor (EF) of 25 (established in test-bed experiments, described in Webster et al. 2018 supplemental material) produces an atmospheric in situ methane value of 20.5 ± 4 ppbv (95% CI), driven predominantly by the error on the EF of 25 ± 4 (95% CI).
The ∼20-ppbv spike (Table 1) was seen by TLS-SAM in the pre-dawn hours and is the average value measured during a two-hour gas ingest as the Martian atmospheric pressure drives the flow across the enrichment cell that continuously removes most of the atmospheric carbon dioxide. We did not monitor the methane abundance during the two-hour ingest, and so we cannot tell the temporal shape of the spike, such as whether it was a steady ∼20 ppbv value, or included much higher and lower fluctuations during the ingest period. The methane source could have been below the rover and venting upward, or transported horizontally from a near or far location. If transported horizontally, with typical horizontal wind speeds at Gale crater of ∼3 m sec −1 (Pla-Garcia et al. 2019), a homogeneous ∼20-ppbv cloud of methane passing our inlet would therefore be ∼20 km in extent for the two-hour ingest. Because we do not resolve the methane abundance versus time over the ingest period, the horizontal extent of a source significantly above background levels could be much smaller or much larger than this.

Absence of evidence for a rover source of methane
Regarding the assertion by Zahnle (2015) that the rover itself is the source of the Curiosity methane, we agree that in this case, one might expect to see seasonal or diurnal variations that follow the temperature of the rover and its environment. However, working with the MSL team, we have exhausted all efforts to find such a source, as first detailed in the supplemental material in Webster et al. (2018), and to date, no one is able to identify or suggest one. A huge effort by the SAM team to use the extensive housekeeping and other data to verify the pumping and filling protocols and vacuum integrity has found no issues. No dependence is found on the rover inlet orientation with respect to the wind direction, or on the local geology or topography.
The only known rover reservoir of methane is the relatively small amount of terrestrial methane trapped in the TLS foreoptics chamber that amounts to ∼10 15 molecules total. From repeated measurements of the foreoptics pressure and methane content (empty-cell values), and investigating correlations using housekeeping and other data over the eight years of operation, there is no evidence of any significant leakage or change in the foreoptics methane amount (except on very few occasions when we deliberately pumped out the chamber to lower pressures). Over the eight years of operation on Mars, and up to the latest measurement reported here, the foreoptics methane amount remains between ∼1-2 nanomoles as measured for the data presented here, which is identical to the last value reported in the data plots in the supplemental material of Webster et al. (2018).
There is not enough methane in the foreoptics to explain the large amounts seen during spike observations, even if all the methane were to be available. For example, if we consider a 1m diameter cross-section of Martian air moving across our external inlet during the two-hour ingest period, even at a lower Mars wind speed of 1 m sec −1 , it would be 7.2 km long, and sweep out a cylindrical volume of ∼5600 m 3 . For the large spike reported here, with an average volume mixing ratio of 20 ppbv at 7 mbar pressure, this would contain ∼2 × 10 19 molecules of methane, or about 20 000 times the total number of methane molecules contained in the foreoptics chamber. By comparison, background levels of ∼0.4 ppbv would also exceed the available foreoptics content by a factor of ∼400. In the absence of any other identified or suggested rover source, we therefore identify this ∼20 ppbv spike and all other TLS-SAM measurements of methane as Martian in origin and not generated by the rover.

Discussion and conclusions
A critical issue is to understand to what extent the nighttime measurements at Gale crater are representative of the methane flux over the whole day at Gale crater and other near-surface locations because day-night differences in the TLS-SAM measurements have the potential to identify the mechanisms at play and to reduce the discrepancy between the Curiosity and ExoMars TGO data sets (Moores et al. 2019a). Numerous mechanisms for Martian methane production and emission have been proposed to date (e.g., Yung et al. 2018, and references therein). However, methane origin by these processes does not explain either the suggested seasonal variation of methane or the methane spikes detected by Curiosity.
Seasonal variation in the methane background levels suggested by the TLS-SAM background data (Webster et al. 2018) has been reproduced to some extent by the one-dimensional numerical model of Moores et al. (2019b) based on temperaturedependent emissions, and by Viúdez-Moreiras et al. (2019) based on wind-dependent emissions. In the first scenario, the model of Moores et al. (2019b) is based on methane adsorption onto and diffusion through the regolith, although methane destruction had a timescale that was a free parameter and methane thermodynamic absorption differed from laboratory-based parameters. With these constraints, the regolith was assumed to be impregnated with methane from earlier plume events or supplied from the underlying surface by microseepage (Etiope & Oehler 2019). Since that study, the very low upper limits on higher altitude methane provided by TGO provide tighter constraints on the calculated magnitude of microseepage that are included in updated modeling calculations by the Moores' group (Moores et al. 2019a) that demonstrates that diurnal variation needs to be considered. This group proposed that the TLS-SAM measurements could be partly reconciled with those of TGO by the inhibition of mixing near the surface overnight whereby methane emitted from the subsurface accumulates within meters of the surface before being mixed below detection limits shortly after dawn. This nighttime accumulation and inhibition of mixing to higher altitudes is enabled by the collapse of the planetary boundary layer (PBL) from many kilometers during the day to only meters at night (Guzewich et al. 2017).
Pla  Outside of the L S 270 • season, mixing between the crater air mass and the external crater air was interpreted to be more subdued. L S 270 • was selected as being representative of "most of the year" for mixing experiments. All MRAMS simulations produce a strong diurnal cycle in the modeled methane abundance, with variations spanning an order of magnitude or more on the order of several hours, increasing during the evening and night, and decreasing during the daytime, as shown in the figures. The methane abundance tends to increase overnight when the PBL is shallow and the downslope, convergent circulations locally confine the surface release. The methane abundance tends to decrease during the day when the PBL grows (Moores et al. 2019b) and divergent, upslope circulations transport methane away from the release area (Pla-Garcia et al. 2019). Animations of the diurnally varying circulation and methane abundance from MRAMS are provided in the appendix. The MRAMS model emphasizes the importance of location of release and horizontal transport in addition to PBL vertical mixing.
Most of the time, the average value is well below the peak (i.e., the peaks are relatively short-lived). The mean values and range of values show that the modeled values depend on location of release, time of day, and season and represent the importance of vertical PBL mixing and horizontal transport, both of which act to determine the local time and magnitude of the peak methane abundance. A release location to the NE produces the highest values, while a location to SE (from Aeolis Mons) produces the lowest. The dependence on direction is due to the prevailing wind. If we assume that micro-seepage is responsible for the methane detected at Gale crater, the Moores et al. (2019a) model allows micro-seepage fluxes at Gale to be derived, consistent with a constant 1.5 × 10 −10 kg m −2 sol −1 (5.4 × 10 −5 tonnes km −2 yr −1 ) source at depth that is an order of magnitude lower than that required to produce the Curiosity measurements if the methane abundance does not exhibit diurnal variation. However, the constraints of the TGO results ) combined with the assumption that methane retains its ∼300-yr lifetime implied that a surface area of only ∼1.5 times that of the Gale crater region could be emitting methane, unless a fast destruction mechanism exists. Under this scenario, the model of Moores et al. (2019a) predicted that daytime TLS-SAM measurements should be close to zero, as reported in this paper.
The release of methane to the atmosphere by means of seepage should also be strongly affected by winds and pressure fluctuations. Model simulations suggest that advective fluxes produced by winds have strong relevance on regolith emissions under Martian conditions in highly permeable soils such as fracture media (Viúdez-Moreiras et al. 2020). A potential seasonal variation in the methane background levels is also indicated by a numerical model considering wind-dependent methane emissions in Aeolis Mons and in other crater locations (Viúdez-Moreiras et al. 2020). MSL data showed evidence of a correlation between the seasonal cycle of surface wind speeds as measured by REMS and the seasonal variation of methane abundance detected by TLS-SAM, pointing to a local source of methane responsible for the background seasonal variation reported by SAM. The temperature-and wind-dependent emission mechanisms are compatible and could be regulating methane emissions constructively.
Because there exist numerous other areas of similar geological features to Gale crater, Etiope & Oehler (2019) argued that a small emitting area of Mars is unrealistic, and that a fast destruction or sequestration mechanism is necessary to avoid the problem of excess methane building up in the Martian atmosphere above the levels observed by TGO. By effectively decreasing the lifetime of methane in the Martian atmosphere, a fast destruction (e.g., Atreya et al. 2011;Delory et al. 2006) or sequestration process (e.g., Jensen et al. 2014) would allow a substantially larger area to be emitting methane than that calculated by Moores et al. (2019a).
The TLS-SAM results show that a fast destruction or sequestration of methane is required Etiope & Oehler 2019), unless Gale crater is the only source of methane (Moores et al. 2019a), which is unlikely. Several mechanisms for fast destruction of methane have been proposed in the past, including energetic electrons resulting from triboelectric process during convective dust events , and, globally, CH 4 sequestration on airborne dust (Jensen et al. 2014), and superoxides from hydrogen peroxide in the surface/subsurface Atreya et al. 2006Atreya et al. , 2011Atreya et al. , 2019. The hypotheses of fast destruction of methane require further modeling and laboratory studies under appropriate environmental and geochemical conditions to assess their quantitative validity for Mars.
The daytime absence of methane recorded by TLS-SAM, and the new additional nighttime values reported here, are a critical update to the Curiosity data set to constrain the possible mechanisms of methane production and removal on Mars and are a step toward reconciling the apparent differences between this data set and that of the TGO. Building upon the models of Moores et al. (2019b,a) (2019), the TLS-SAM data provide evidence for methane production from a near-surface source, most likely from continuous micro-seepage, that is temporarily contained in the near-surface atmosphere during the night due to the low planetary boundary layer, reduced atmospheric mixing, and horizontal transport. Increased atmospheric dynamics during the daytime mixes the gas accumulated in the night into the global atmosphere where it is diluted to the very low levels constrained by the TGO nondetection. We consider it highly unlikely that Gale crater is the only source area of micro-seepage, and therefore for the scenario implied by the Curiosity measurements, a fast methane destruction or sequestration mechanism must be occurring in the lower atmosphere of Mars as discussed above. MRAMS animations are available online. Gale crater plan view diurnal cycle animation of winds and methane abundance evolution in the lowest model layer for a steady-state methane release at L S 90 • ∼ 1 grid point (2.96 km) NW of the Curiosity rover location on sol 305. White arrows represent the wind speed and direction. Black contours represent the topography. "CH 4 detected" means "Modeled CH 4 abundance". Hours in LMST are shown in the upper right corner and methane values are given in ppbv. White arrows represent the wind speed and direction. The x-y axis labels distance in grid points (each of 2.96 km).      Plots of individual data points from Table B.1 The y-axis units are in situ CH 4 average abundance Wefg in ppbv. The x-axis identifies each of the 26 sequential measurement  points. Empty-cell measurements always immediately follow full-cell measurements.