Report

Strong Release of Methane on Mars in Northern Summer 2003

See allHide authors and affiliations

Science  20 Feb 2009:
Vol. 323, Issue 5917, pp. 1041-1045
DOI: 10.1126/science.1165243

Abstract

Living systems produce more than 90% of Earth's atmospheric methane; the balance is of geochemical origin. On Mars, methane could be a signature of either origin. Using high-dispersion infrared spectrometers at three ground-based telescopes, we measured methane and water vapor simultaneously on Mars over several longitude intervals in northern early and late summer in 2003 and near the vernal equinox in 2006. When present, methane occurred in extended plumes, and the maxima of latitudinal profiles imply that the methane was released from discrete regions. In northern midsummer, the principal plume contained ∼19,000 metric tons of methane, and the estimated source strength (≥0.6 kilogram per second) was comparable to that of the massive hydrocarbon seep at Coal Oil Point in Santa Barbara, California.

The atmosphere of Mars is strongly oxidized, composed primarily of carbon dioxide (CO2, 95.3%), along with minor nitrogen (N2, 2.7%), carbon monoxide (CO, 0.07%), oxygen (O2, 0.13%), water vapor (H2O, 0 to 300 parts per million), and radiogenic argon (1.6%); other species and reduced gases such as methane (CH4) are rare. CH4 production by atmospheric chemistry is negligible, and its lifetime against removal by photochemistry is estimated to be several hundred years (13) or shorter if strong oxidants such as peroxides are present in the surface or on airborne dust grains (4). Thus, the presence of substantial amounts of CH4 would require its recent release from subsurface reservoirs; the ultimate origin of this CH4 is uncertain, but it could be either abiotic or biotic (2, 5, 6).

Before 2003, all searches for CH4 were negative (79). Since then, three groups have reported detections of CH4 (1018); see (1924) for discussion. Spectral data from the Mars Express mission contain five unidentified spectral features between 3000 and 3030 cm–1, one of which coincides with the expected position of the CH4 Q branch (15, 18, 25). The data span all seasons and extend over several years, but low S/N ratios require averaging the spectra over two of the three key dimensions (longitude, latitude, and time). Other searches featured low spatial coverage (16) or sparse seasonal coverage (16, 17), and the results (CH4 mixing ratios) are best interpreted as upper limits.

We report measurements of CH4 in northern summer in 2003 and estimate its source strength and its (short) destruction lifetime. Our search covered about 90% of the planet's surface and spanned 3 Mars years (MYs) (7 Earth years). Our results (1014) are based on the simultaneous detection of multiple spectrally resolved lines of CH4, and each observation is spatially resolved, allowing examination of spatial and temporal effects. Our spatial maps reveal local sources and seasonal variations.

To search for CH4 and other gases on Mars, we used the high-dispersion infrared spectrometers at three ground-based telescopes. Here we report data from CSHELL/IRTF (Hawaii) and NIRSPEC/Keck-2 (Hawaii) [supporting online material text 1 (SOM-1)]. Each spectrometer features a long entrance slit that is held to the central meridian of Mars (Fig. 1A) while spectra are taken sequentially in time (fig. S1). Pixelated spectra were acquired simultaneously at contiguous positions along the entire slit length, for each observation, providing 35 spectra at 0.2–arc second (arc sec) intervals (∼195 km at disk center) when Mars' diameter is 7 arc sec (Fig. 1A). We binned these data (in groups of three along the slit) to provide latitudinally resolved spectra, and then in time (longitude) to improve the S/N ratio (SOM-1). Here we focus on three dates in 2003 [universal time (UT) 12 January, 19 March, and 20 March] and one in 2006 (UT 26 February) (Table 1).

Fig. 1.

Detections of CH4 and water vapor on Mars on UT 19 and 20 March 2003. (A) Mars is shown as it appeared at the mean time of the R0 and R1 observations (Table 1). The subsolar (*) and sub-Earth (+) points are marked, along with several prominent features. Contours of constant altitude are shown at intervals of 3 km, and regions beyond the afternoon terminator (the night side) appear in gray. The entrance slit of the spectrometer was oriented north-south (N-S) on Mars along the central meridian and is shown to scale. (B) Spectra taken on 20 March were extracted at 11 equal intervals (0.6 arc sec each) along the slit (ranging from 70°N to 70°S), after binning over longitudes 277° to 323°W. At the sub-Earth position, the binned spectrum samples a footprint 3215 km (east-west) by 586 km (N-S) (see text and SOM-1). Strong lines of terrestrial water and CH4 (labeled) and weak lines of ozone (3036 to 3038 cm–1) appear in a typical spectrum, shown at the top of this panel. See SOM-3 for data reduction. Narrow spectral lines of H2O (three lines, short dashes) and CH4 (the R1 line, long dashes) are seen at the Doppler-shifted positions expected for this date. Corresponding Mars ozone (O3) absorptions for this date are weak and would appear in the southern polar region (not sampled). (C) Spectra taken on 19 March are shown as in (B) but binned over the longitude range from 289° to 335°. Spectral lines of H2O (three lines, short dashes) and CH4 (the R0 line, long dashes) are seen. The longitude range sampled was systematically westward (by 12°) of that sampled for R1 (Table 1), owing to the slower rotation of Mars relative to Earth. The residual spectra shown in (B) and (C) are scaled by a factor of either 10 or 25 (gray background), to make the lines more apparent. (D and E) show residual intensities in a grayscale format, to more easily show the spatial distribution of the gases with latitude. The lines are displayed as measured; that is, corrections for two-way air mass on Mars (Sun to Mars surface and Mars surface to Earth), for Mars local topography, and for terrestrial transmittance have not yet been applied to the residuals (see Fig. 2).

Table 1.

Observational searches for Mars CH4 on selected dates. A complete listing of all dates searched, spanning 7 years, is available from the authors.

View this table:

Our spectra exhibit strong lines of terrestrial H2O (2ν2 band) and CH43) along with weaker lines of O3 (3ν3) seen against the continuum (Fig. 1, B and C, top). We corrected the data for telluric extinction (SOM-2). At 3.3 μm, Mars is seen mainly in reflected sunlight, so the collected spectra also contain Fraunhofer lines (SOM-3). Removing these two components from a composite spectrum exposed the residual Mars atmospheric spectrum (Fig. 1, B and C) (26). One line of CH4 and three distinct lines of H2O are seen in each panel.

CH4 consists of three separate nuclear spin species (A, E, and F) that act as independent spectral entities. The identification of two spectrally resolved CH4 lines (R0 and R1, of the A and F species, respectively) and the good agreement (within measurement error) of the CH4 column densities obtained independently from each line support our detection (systematic uncertainty and stochastic errors are discussed in SOM-2 and SOM-3). At a later season [areocentric longitude of the Sun (Ls) = 220°], we detected the P2 doublet, consisting of lines of the E and F species (SOM-1 and fig. S2). Detections (or upper limits) reported by others are based on the intensity summed over the frequencies of multiple unresolved lines [the Q branch (15, 18)] or on the summed intensity of several undetected individual lines (16, 17).

Column densities and mixing ratios for CH4 obtained from the spectra of Fig. 1 show a broad CH4 plume in late summer when averaged over 46° in longitude (Fig. 2, A and B) (SOM-3); more restricted longitudinal binning shows higher maxima (Fig. 2C, profile d). In early spring (profile a), the CH4 mixing ratio was small at all latitudes and showed only a hint of the marked maximum seen in late summer. By early summer (profiles b and c), CH4 was prominent but the maximum mixing ratio occurred at more northerly latitudes and was somewhat smaller than in late summer. The late summer profile (d) in Fig. 2C differs from the profile shown in Fig. 2B, owing to different longitudinal binning (27). The mixing ratios shown in Fig. 2C (profile d) represent a subset of the data of Fig. 1B, centered at central meridian longitude (CML) 310° (see aspect, Fig. 1A) and binned over only 30 min. Including the slit width, 30-min binning provides a footprint at the sub-Earth point that spans ∼16° (948 km) in longitude and 10° (586 km) in latitude (fig. S1 and Table 1).

Fig. 2.

Absolute abundances, spatial profiles, and seasonal changes of CH4 on Mars. (A) The total CH4 column density (in molecules per square meter) along a two-way path (Sun to Mars surface and Mars surface to Earth) needed to reproduce the measured lines (after correcting for terrestrial transmittance, SOM-2). The R0 (Fig. 1C) and R1 (Fig. 1B) lines were analyzed independently for a range of longitudes that spanned 46° but differed by 12° in mean longitude (312° for R0 versus 300° for R1, Table 1). The apparent differences in CH4 column density seen at low latitudes (30°N to 15°S) reflect (in part) differences in mean topography sampled on the two dates. The confidence limits contain both systematic uncertainty and random error; the systematic uncertainty affects all extracted values in the same way whereas the random error introduces scatter among the individual points (SOM-2 and SOM-3). (B) The local mixing ratio (in parts per billion) of CH4 obtained from the column density (A) in each footprint, after correcting for two-way air mass on Mars and for topography (SOM-3). The mixing ratios derived from R0 and R1 of CH4 agree, within confidence limits. The remaining systematic difference at low latitudes is consistent with stronger weighting of local sources (Fig. 3) on 20 March (R1), when the longitude range sampled was more nearly centered over them. The differences in mixing ratio (R0 versus R1) should then have decreased with increasing distance from the source or sources, as they did (compare values at 60°N, 40°N, and 25°S, 35°S, 47°S, and 62°S). (C) Geographic and temporal variability of Mars CH4. Latitudinal profiles of CH4 mixing ratios for different longitudes and seasons are shown; the width of the color envelope represents the ±1σ confidence envelope. The areocentric seasons (Ls) are early northern spring (a: 17°), early northern summer (b and c: 122°), and late northern summer (d: 155°) (Table 1). These extracts are taken from spectra centered at the indicated longitude (CML), and the sub-Earth footprints span longitude-latitude ranges (Table 1) with these physical dimensions: a, 770 km × 535 km; b and c, 1274 km × 818 km; d, 948 km × 586 km. The mixing ratios shown in profile d are larger than those shown in Fig. 2B, owing to different longitudinal binning, and they reflect the longitudinal maximum of the plume (SOM-1, figs. S1 and S6, and Fig. 3).

A quantitative release rate can be inferred by considering the observed temporal changes (with season) and measured spatial profiles. The summer mixing ratios (profiles b, c, and d) show a clear maximum for each north-south spatial profile (Fig. 2C). Moving southward by about 30° from the latitude of the peak, the mixing ratio decreased by a factor of 2 in each case, and for profile d the northward gradient was similar to the southward one. These latitudinal gradients suggest that there was a local source or sources and the resulting plume or plumes were being dispersed by atmospheric circulation.

We consider the dimension of the hypothesized CH4 plume to be about 60° in latitude [full width at half maximum (FWHM); Fig. 2C, profile d] and assume a similar dimension in longitude. The latter view is weakly supported by profiles b and c, which differ by 27° in central longitude and by a factor of 2 in peak mixing ratio. It is also supported by the profile formed by binning over 46° in longitude (277° to 323°, Fig. 2B), which has a peak mixing ratio [24 parts per billion (ppb)] reduced by a factor of 2 from the peak value (45 ppb) obtained when binning over only 16° of longitude (302° to 318°, Fig. 2C, profile d). The slight increase of profile d near 40° N is consistent with enhanced CH4 (perhaps owing to continued release at that latitude; compare profiles b and c), whereas the slight increase in profile c near 15°S suggests a small contribution from a source to the west (compare peak position, profile d). Together, these profiles suggest that there may be two local source regions, the first centered near 30°N, 260°W and the second near 0°, 310°W. The vapor plume from each is consistent with ∼60° in both latitude and longitude.

The amount of trace gas present in each plume can be estimated from these parameters (SOM-4). In the central plume of profile d (FWHM diameter ∼60°), the mean CH4 mixing ratio is ∼33 ppb (120 mol km–2), and the plume contains ∼1.17 × 109 mol of CH4 (∼1.86 × 107 kg, or ∼19,000 metric tons). If seasonally controlled, the duration of release must be substantially shorter than 0.5 MYs, requiring a mean CH4 release rate of ≥39 mol s–1 (≥0.63 kg s–1). For comparison, the massive hydrocarbon seep field at Coal Oil Point in Santa Barbara, California, releases CH4 at a rate of ∼0.4 to 1.0 kg s–1 (28).

We considered three models for plume formation, to constrain aspects of CH4 release and its migration in latitude and longitude (SOM-5). A model based on release from a central source region coupled with eddy diffusion fits the observed plume parameters. Models of meridional flow using a global circulation model suggest that released gas would move northward by ∼3.3 cm s–1 at this season (29), for a total displacement by not more than ∼170 km from its central source. If the mixing coefficients (Kx and Ky) in zonal and meridional directions are identical (Kh), a steady source would fill the plume (profile d) in 60 days if Kh∼6.4×108 cm2 s–1. For this case, the required source strength would be ∼3.66 kg s–1. The filling time and Kh vary inversely, whereas Kh and source strength vary proportionately. For a filling time of 0.5 MYs (∼344 Earth days), Kh ∼ 1.1×108 cm2 s–1 and the source strength is ∼0.63 kg s–1. A reasonable limit for filling time (<120 days) requires Kh ∼ 3.2 × 108 cm2 s–1 and source strength ∼1.8 kg s–1.

These parameters are consistent with release from a single central source region, followed by efficient eddy mixing (SOM-5). The central source could be activated thermally by warming of a surface zone, or by connecting subpermafrost regions to the atmosphere through seasonally opened pores in scarps or crater walls. The plume would reflect the gross morphology of active release zones (and their intensity), and the peak could suggest a region of enhanced release. For comparison, the subsolar latitude was 24°N at Ls = 122° (compare profiles b and c) and 10°N for Ls = 155° (profile d).

Additional information is obtained from a high-resolution map constructed from our data for midsummer 2003 (Fig. 3 and fig. S1). CH4 appears notably enriched over several localized areas: A (east of Arabia Terra, where we also measure greatly enriched water vapor), B1 (Nili Fossae), and B2 (the southeast quadrant of Syrtis Major). Unusual enrichments in hydrated minerals (phyllosilicates) were identified in Nili Fossae by Mars Express (30, 31) and by the Mars Reconnaissance Orbiter (32) (Fig. 3). The observed morphology and mineralogy (33, 34) of this region suggest that these bedrock outcrops, rich in hydrated minerals, might be connecting with reservoirs of buried material rich in volatile species. The characteristic arcuate ridges in the southeast quadrant of Syrtis Major were interpreted as consistent with catastrophic collapse of that quadrant, from interaction with a volatile-rich substrate (34).

Fig. 3.

Regions where CH4 appears notably localized in northern summer (A, B1, and B2) and their relationship to mineralogical and geomorphological domains. (A) Observations of CH4 near the Syrtis Major volcanic district. (B) Geologic map of Greeley and Guest (45) superimposed on the topographic shaded relief from the Mars Orbiter Laser Altimeter (46). The most ancient terrain units are dissected and etched Noachian plains (Npld and Nple) (∼3.6 to 4.5 billion years old, when Mars was wet) and are overlain by volcanic deposits from Syrtis Major of Hesperian (Hs) age (∼3.1 to 3.6 billion years old).

The low mean abundance measured in early spring 2006 (profile a) provides an important constraint on the CH4 lifetime. The plume seen in March 2003 (19,000 tons, Fig. 2C, profile d) implies a global mean mixing ratio of ∼2 ppb, if later spread uniformly over the planet. The content of the central plume of profile c is similar. Combining data for the entire region mapped during northern summer brings the total CH4 to 42,000 tons, or 6 ppb if spread uniformly over the planet. However, the mean mixing ratio displayed in the early spring equinox in 2006 (profile a, Fig. 2C) was only 3 ppb (SOM-3). If CH4 is not removed by other means, the implied destruction lifetime is ∼4 Earth years if the 2003 event was singular, to as little as ∼0.6 Earth years if the event repeats each MY. In either case, the destruction lifetime for CH4 is much shorter than the time scale (∼350 years) estimated for photochemical destruction (16). Another process thus must dominate removal of atmospheric CH4 on Mars, and it must be more efficient than photochemistry by a factor ≥100.

Heterogeneous (gas-grain) chemistry is a strong candidate. The presence of strong oxidants in the soil was suggested first by the labeled-release experiment on Viking landers, and laboratory simulations suggested that peroxides (such as H2O2) were responsible; the apparent discovery of perchlorate (XClO4) by the Phoenix lander (35, 36) suggests the existence of another family of strong oxidants, although their presence at low latitudes has not been established. The lofting of oxidant-coated soil particles into the atmosphere could permit rapid oxidation of the CH4 that collides with them. H2O2 is also produced photochemically and is a known trace gas in the atmosphere (37) and (being polar) might bind to aerosol surfaces. Electrochemical processes in dust storms may produce additional peroxide efficiently (4, 6, 23, 24) (SOM-6). Peroxide-coated grains might provide an efficient sink for CH4 for many years thereafter if they settle to the surface and are sequestered in the regolith (6, 24). Sequestered oxidants should also efficiently destroy upward-diffusing CH4, reducing the fraction that might escape to the atmosphere.

The most compelling question relates to the origin of CH4 on Mars. The CH4 we detected is of unknown age—its origin could be ancient (38) or perhaps recent. Both geochemical and biological origins have been explored, but no consensus has emerged. Most theses draw on known terrestrial analogs such as production in magma (19, 20) or serpentinization of basalt (39), or production by psychrophilic methanogenic biota in Mars-analog cryoregimes such as permafrost (21). The annual release of CH4 from an arctic tundra landscape on Earth (at 72°N) was measured to be 3.15 g m–2, with midsummer CH4 fluxes of typically 30 mg m–2 day–1 (40). If similar release rates applied to our midsummer plume, the tundra-equivalent area of (assumed uniform) release would be ∼6000 km2 compared with a plume footprint ∼9.7 × 106 km2. If CH4 release were uniform over the plume footprint, the mean release rate could be smaller than the arctic rate by a factor of ∼1600.

Of special interest are the deep biocommunities that reduce sulfate to sulfide using H2 or abiotic CH4 as the electron donor, releasing H2S. These communities thrive at depths of 2 to 3 km in the Witwatersrand Basin of South Africa and have been isolated from the surface (and photosynthesis) for tens of millions of years (41, 42). It might be possible for analogous biota (methanogens or methanotrophs) to survive for eons below the cryosphere boundary on Mars, where water is again liquid, radiolysis can supply energy, and CO2 can provide a carbon source. Gases accumulated in such zones might be released to the atmosphere if pores or fissures open seasonally, connecting these deep zones to the atmosphere at scarps, crater walls, or canyons. The location of CH4 maxima over the Syrtis Major shield volcano and the nearby Nili Fossae district suggests a possible relation to serpentinization and/or to the phyllosilicates discovered there (43, 44).

Supporting Online Material

www.sciencemag.org/cgi/content/full/1165243/DC1

SOM Text

Figs. S1 to S6

References and Notes

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article