Ozone abundance on Mars from infrared heterodyne spectra

Icarus 181 (2006) 419–431 www.elsevier.com/locate/icarus

Ozone abundance on Mars from infrared heterodyne spectra I. Acquisition, retrieval, and anticorrelation with water vapor Kelly Fast a,∗,1 , Theodor Kostiuk a,1 , Fred Espenak a,1 , John Annen a,1 , David Buhl a , Tilak Hewagama a,b , Michael F. A’Hearn b , David Zipoy b,1,2 , Timothy A. Livengood a,c,1 , Guido Sonnabend d,1 , Frank Schmülling e,1 a Planetary Systems Laboratory, Code 693, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA b Department of Astronomy, University of Maryland, College Park, MD 20742-2421, USA c Universities Space Research Association, 1101 17th St. NW, Suite 1004, Washington, DC 20036, USA d National Research Council Associate at Planetary Systems Laboratory, Code 693, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA e I. Physikalisches Institut, Universität zu Köln, Zülpicher Str. 77, 50937 Köln, Germany

Received 11 August 2005; revised 30 November 2005 Available online 25 January 2006

Abstract Observations of ozone on Mars were made using the Goddard Space Flight Center’s Infrared Heterodyne Spectrometer and Heterodyne Instrument for Planetary Wind and Composition at the NASA Infrared Telescope Facility. Ozone is an important observable tracer of martian photochemistry. Infrared heterodyne spectroscopy with spectral resolution
106 is the only technique that directly measures ozone in the martian atmosphere from the surface of the Earth. Ozone column abundances down to the martian surface were acquired in seven data sets taken between 1988 and 2003 at various orbital positions (LS = 40◦ , 74◦ , 102◦ , 115◦ , 202◦ , 208◦ , 291◦ ). Ozone abundances are compared with those retrieved using ultraviolet techniques, showing good agreement. Odd hydrogen (HOX ) chemistry predicts anticorrelation of ozone and water vapor abundances. Retrieved ozone abundances consistently show anticorrelation with corresponding water vapor abundances, providing strong confirmation of odd hydrogen activity. Deviation from strict anticorrelation between the observed total column densities of ozone and water vapor suggests that constituent vertical distribution is an additional, significant factor.  2005 Elsevier Inc. All rights reserved. Keywords: Mars, atmosphere; Infrared observations; Photochemistry; Spectroscopy; Abundances, atmospheres

1. Introduction The planet Mars has gone from a subject of mystery in antiquity to the target of spacecraft at every launch opportunity in recent years. It is one of the more accessible planets via Earth-based and in situ observations. Along with Venus, Mars * Corresponding author. Fax: +1 301 286 0212.

E-mail address: [email protected] (K. Fast). 1 Visiting astronomer at the Infrared Telescope Facility, which is operated by

the University of Hawaii under Cooperative Agreement No. NCC 5-538 with the National Aeronautics and Space Administration, Office of Space Science, Planetary Astronomy Program. 2 Present address: 119 Hibiscus Drive, Punta Gorda, FL 33950, USA. 0019-1035/$ – see front matter  2005 Elsevier Inc. All rights reserved. doi:10.1016/j.icarus.2005.12.001

is important for understanding properties of terrestrial planets in general as well as properties and processes that shape conditions on Earth. Mars is a unique geological and atmospheric environment, raising new questions with each discovery. The atmosphere of Mars has an average surface pressure of ∼6 mbar (<1% that of Earth), and is rich in chemical and dynamical processes. Atmospheric conditions on Mars vary over the course of the year. Mars has an eccentric orbit compared to Earth and Venus, which results in ∼45% greater insolation at perihelion, LS = 251◦ , than at aphelion, LS = 71◦ , where LS is the orbital position relative to 0◦ at the northern spring equinox. Perihelion and aphelion occur near the solstices (Fig. 1), accentuating seasonal variations (e.g., colder southern winter). Atmospheric pressure changes by ∼20% over the course of

420

K. Fast et al. / Icarus 181 (2006) 419–431

Fig. 1. Seasons on Mars. The solar longitude, or LS , is the position of Mars in its orbit around the Sun. The northern spring equinox is at 0◦ , the northern summer solstice is at 90◦ and so on. Dates for observations used in this work are displayed above their corresponding orbital position.

the year due to sublimation and refreezing of CO2 at the polar caps. The existence of southern highlands and northern lowlands results in a range of atmospheric pressures with latitude. Water vapor sublimates mainly from the north polar cap during warmer periods and freezes out of the atmosphere during colder periods. The importance of water vapor to the stability of the 95% CO2 martian atmosphere was recognized by Parkinson and Hunten (1972) and McElroy and Donahue (1972). Solar ultraviolet radiation of wavelength 205 nm dissociates CO2 into CO and O, and direct recombination of CO and O to reform CO2 involves a slow, spin-forbidden, three-body reaction. The three-body reaction that forms O2 from O is orders of magnitude faster, thus a buildup of CO and O2 would be expected up to relatively high equilibrium values. Measurements of martian CO and O2 reveal abundances that are far lower than those expected from this “pure CO2 ” scheme. Water is dissociated through photolysis (243 nm) into OH and H. Subsequent chemistry involving “odd hydrogen” or “HOX ” species (e.g., H, OH, HO2 ) and oxygen species results in reaction schemes that provide OH, which can go on to react with CO to form CO2 at a rate that is orders of magnitude faster than the three-body recombination reaction of CO and O. This CO2 reformation process involving OH is catalytic because additional reactions recycle the OH, and therefore the minute amount of water vapor present on Mars (∼0.02%) is sufficient to support this process. Odd hydrogen chemistry is largely responsible for recycling and stabilizing CO2 , but other processes increase the complexity of the photochemistry. Molecular nitrogen is dissociated and processed in the ionosphere (above ∼120 km), resulting in “odd nitrogen” species like NO, some of which are transported to the lower atmosphere to participate in catalytic recycling of OH. Ionospheric processes result in hydrogen escape to space from H2 dissociation and oxygen escape from exothermic ion chemistry. In the lower atmosphere, heterogeneous chemistry can play a role in processes such as odd hydrogen loss. The extreme seasons and topography of Mars, along with atmospheric circulation, also come into play. Earth-based and spacecraft observations are required to constrain models of the chemistry responsible for CO2 stability, atmospheric constituent abundances, and atmospheric conditions. Although chemistry resulting from water vapor photolysis is the key to the accepted solution of the CO2 stability problem, odd hydrogen species had not been directly observed until the

recent detections of hydrogen peroxide, H2 O2 , by Clancy et al. (2004) and Encrenaz et al. (2004). Ozone (O3 ) is formed by the three-body reaction of O and O2 and is destroyed by odd hydrogen, making it a sensitive tracer of the photochemistry that stabilizes the CO2 atmosphere of Mars. Ozone and atomic oxygen interconvert during the day, therefore photolysis and other reactions that result in atomic oxygen are not permanent ozone loss mechanisms. It is expected that ozone destruction should be controlled by the abundance of the odd hydrogen photochemical parent, water vapor. Ozone/water anticorrelation was first observed by the UVS experiment onboard the Mariner 9 spacecraft, which saw northern hemispheric ozone abundances rise as winter progressed and water vapor froze out of the atmosphere, as well as a similar effect in the southern hemisphere during its winter (Lane et al., 1973). Additional measurements of ozone and water at the same time are important to constrain photochemical models that predict this ozone/water anticorrelation and to understand the behavior of the atmosphere with orbital position. Ultraviolet observations of ozone on Mars must be made outside Earth’s atmosphere. Martian ozone was first measured by Mariner 7 (Barth and Hord, 1971) and Mariner 9 (Lane et al., 1973; Barth et al., 1973; Wehrbein et al., 1979). Hubble Space Telescope observations of ozone using the Faint Object Spectrograph (HST-FOS) have been made on a number of occasions (Clancy et al., 1996b, 1999). Mars Express SPICAM (Bertaux et al., 2000, 2004) is simultaneously studying ozone and water vapor abundance and distribution. Spacecraft have limited instrumentation and lifetimes, however, and cannot always focus on the atmosphere, so measurement of atmospheric constituents must be done from Earth as well. Earth-based observations can be made anywhere on the visible disk and can characterize the atmosphere over long periods, which also provides a historical record to connect periods of in situ and orbital exploration. An indirect method of ozone abundance retrieval from the ground involves measuring the intensity of the O2 (1 ∆) 1.27-µm dayglow emission on Mars (Noxon et al., 1976; Krasnopolsky and Bjoraker, 2000; Novak et al., 2002; Krasnopolsky, 2003). Ozone photolysis produces rotationally excited molecular oxygen in the 1 ∆ metastable state. At altitudes above ∼20 km, the excited oxygen state has a high probability of radiative decay to produce the dayglow emission, which is a tracer of ozone abundance. Dayglow intensities can be compared to those predicted by photochemical models. They can also be converted to ozone column abundances, but assumptions must be made about the quenching rate, the ozone vertical distribution, and the lower boundary of the retrieved ozone. A third method, infrared (IR) heterodyne spectroscopy, directly measures martian ozone absorption features from the surface of the Earth, and was first applied by Espenak et al. (1991) in 1988 measurements. IR heterodyne spectroscopy can achieve the very high spectral resolving power of λ/λ
106 , orders of magnitude greater than that of conventional IR spectrometers. The spectral region between 9 and 10 µm is rich in martian ozone target absorption lines, overwhelmed by their telluric counterparts at conventional resolution. The very high resolution of IR heterodyne spectroscopy provides access from

Ozone abundance on Mars from IR heterodyne spectra

421

in order to explore the anticorrelation predicted by photochemical theory at the various latitudes and orbital positions (LS ) covered by the measurements. IR heterodyne ozone retrievals are compared to retrievals obtained using ultraviolet techniques. The application of IR heterodyne ozone abundances to the testing of photochemical models and the exploration of vertical ozone distribution through the combination with O2 (1 ∆) measurements will be presented in future papers. 2. Observations and data reduction 2.1. Instrumentation

Fig. 2. Telluric atmospheric transmittance between 10.2 and 8.9 µm (upper panel), and at 9.7 µm at heterodyne resolution (lower panel). Martian ozone features can be observed when Doppler-shifted to regions of higher telluric transmittance (arrow).

Fig. 3. Example heterodyne spectrum of Mars. The low-resolution (25 MHz) filter bank samples both a deep CO2 absorption feature (∼1200 MHz) and ozone absorption features (∼400 MHz). The ozone features are simultaneously sampled by the high-resolution (5 MHz) filter bank (inset). Displayed single-sideband spectra were constructed by removing modeled telluric contribution and one Mars sideband from measured double-sideband spectra.

the surface of the Earth to fully resolved martian ozone features when they are Doppler-shifted away from their telluric counterparts. This is illustrated in Fig. 2, comparing Earth’s atmospheric transmittance between 10.2 and 8.9 µm with the bandpass of an IR heterodyne spectrometer at 9.7 µm, resolving individual terrestrial ozone absorption features and exposing regions of higher atmospheric transmittance where martian features are observable. IR heterodyne spectroscopy of martian ozone is sensitive to the total ozone column integrated down to the surface. Absorption lines of martian CO2 are present in the same spectra (Fig. 3) and can be used to model the thermal profile, which is important for obtaining accurate ozone abundance. We present the results of IR heterodyne observations of ozone on Mars made between 1988 and 2003 (Fast, 2005). We detail the IR heterodyne technique and the data acquisition, reduction and analysis schemes. The resulting ozone retrievals are compared to contemporaneous water vapor measurements

Spectroscopic measurements of ozone features on Mars were acquired with the NASA/Goddard Space Flight Center Infrared Heterodyne Spectrometer (IRHS, Abbas et al., 1976) between 1988 to 1999 and with the Heterodyne Instrument for Planetary Wind and Composition (HIPWAC, Schmülling et al., 1999) in 2003 at the NASA Infrared Telescope Facility (IRTF) on Mauna Kea, Hawai’i. The IRHS was mounted at the Coudé focus of the IRTF, succeeded by the Cassegrainmounted HIPWAC in January 2000. Both instruments utilize similar techniques. Source light from the telescope is directed into the instrument front-end. A dichroic mirror transmits visible light to a guide camera for active tracking of the source while reflecting the infrared light further into the instrument. Optical components collimate the source beam and combine it with an infrared CO2 laser local oscillator (LO) beam, focusing them together onto a liquid nitrogen-cooled HgCdTe photomixer, providing a radio-frequency (RF) signal at the difference frequency between the source and the laser frequency. The HgCdTe photodiode mixer generates absolute frequency differences and does not distinguish between frequency differences below and above the frequency of the laser local oscillator. The resultant difference spectrum is a double-sideband (DSB) spectrum comprised of contributions from the lower and the upper sideband folded about the laser frequency. The frequency and the intensity information of the infrared spectrum are preserved in this beat frequency spectrum. In Mars measurements, each sideband contributes a different portion of Mars’ spectrum to the DSB spectrum, with only one sideband containing the target ozone absorption lines. The source signal is chopped with an in-system (“on-the-table”) chopper or the telescope’s wobbling secondary mirror for synchronous detection of the difference between the source and background sky. A local blackbody source provides the system intensity calibration, verified with direct measurements of continuum emission from the Moon. The measured lunar continuum intensity is within 3 K of prediction (Montgomery, C.G., Saari, J.M., Shorthill, R.W., Six Jr., N.F., 1966. Directional characteristics of lunar thermal emission. Technical Note R-213, Boeing Document D1-82-0568). Both IRHS and HIPWAC utilized a back-end of two RF filter banks to analyze the RF beat-frequency signal. The lowresolution (LR) filter bank contains 64 filters that are 25 MHz (0.00083 cm−1 ) in width, providing a 3200-MHz bandpass folded about the laser local oscillator into a 1600-MHz doublesideband spectrum. The detector mixing gain in the lower and

422

K. Fast et al. / Icarus 181 (2006) 419–431

upper 1600-MHz sidebands is assumed equal since mixing is done in a 3.2-GHz frequency band centered around the laser frequency of ∼30 THz (∼10 µm wavelength). Over this small fractional frequency range no significant gain difference between sidebands is expected, and the photomixer does not distinguish between frequencies in the two sidebands. The highresolution (HR) filter bank contains 64 filters that are 5 MHz (0.000166 cm−1 ) in width, tunable by mixing the detector signal with a radio-frequency local oscillator (RFLO) to adjust the frequency positioning of the HR filter bank’s 320 MHz bandpass. The wide LR bandpass samples both martian ozone and CO2 (for temperature retrieval) while the HR filter bank simultaneously samples the ozone features at higher resolution for better ozone mole fraction retrieval (Fig. 3). In one instance (November 2003), an acousto–optic spectrometer (AOS) with a bandwidth of 1500 MHz and a spectral resolution of 1 MHz was used along with the filter banks to analyze the heterodyne signal. The signal-to-noise ratio in a measured spectrum varies with frequency due to a decrease in detector sensitivity with frequency from the laser local oscillator, and due to the changing transmittance of the Earth’s atmosphere with frequency in the 9.7-µm region. Further details of the IR heterodyne technique is given by Kostiuk (1994) and Kostiuk and Mumma (1983). 2.2. Data acquisition and reduction IR heterodyne observations of ozone on Mars were made at a variety of orbital positions (seasons) between 1988 and 2003 (Table 1 and Fig. 1). Heterodyne spectra of ozone on Mars obtained in 1988 by Espenak et al. (1991) have been re-reduced and re-analyzed in this work using updated techniques, revising the retrieved ozone abundances downward by ∼30%. The heterodyne local oscillator used is a frequency-stabilized 16 O12 C16 O gas laser. The laser was tuned to the P branch of the 9.7-µm band of CO2 , to either the P (32) transition (1035.4736 cm−1 ) or the P (36) transition (1031.4774 cm−1 ), depending on the Earth/Mars radial velocity at the time of the observations. The spectral regions around these transitions

contain telluric ozone lines distributed in such a way that corresponding martian ozone lines can be Doppler shifted into regions of higher telluric transmittance. Observations were made at various latitudes on Mars with a beam that covered ∼10◦ of latitude in most cases (∼15◦ at high latitudes). The pointing was maintained over the course ∼1–2 h, and individual 2–4min integrations (scans) were stacked together in order to form the measured spectrum. As a result, observed longitudes change with the rotation of Mars and each spectrum covers ∼20◦ –40◦ of longitude, but the local time on Mars is maintained. Most observations were made toward the limb of Mars (∼50◦ –55◦ zenith angle) in order to increase the available column abundance of ozone (e.g., Figs. 4 and 6). The data are reduced first by dividing individual spectral scans by blackbody calibration scans taken close in time to remove the instrumental intensity roll-off that results from the frequency response of the IR detector. Spectra are assembled by co-adding scans according to their date of observation and position on Mars. The resulting spectra contain features of both Earth and Mars, therefore the telluric components must be removed in order to retrieve martian ozone abundances. Observations were made of the atmosphere of Earth against the Moon in order to characterize the transmittance of Earth’s atmosphere at the time of observation of Mars. The subsequent radiative transfer analysis retrieved model Earth atmospheres to be used for removing the telluric component from spectra of Mars. A detailed description is given by Fast et al. (2004), which also demonstrated the accuracy of retrieved terrestrial ozone profiles through comparison with those retrieved through nearby contemporaneous ozonesonde and lidar measurements. 3. Radiative transfer modeling An IR heterodyne spectrum is a double-sideband spectrum, which is essentially a normal (single-sideband) spectrum folded about the laser local oscillator frequency, combining lower and upper sidebands. A measured spectrum of Mars consists of the thermal continuum emission from the surface modified by the

Table 1 Infrared heterodyne observations of ozone on Mars LS a

Date

Northern martian season

Subsolar latitude

SubEarth latitude

Angular diameterd (arcsec)

Contemporaneous water vapor observation

40◦ 74◦ 102◦ 115◦ 202◦ 208◦ 291◦

February 14–16, 1993b March 17–23, 1995b July 4–7, 1993b March 24–29, 1999b June 10–16, 2003c June 3–7, 1988b November 1–3, 2003c

Mid-spring Late spring Early summer Early summer Early autumn Early autumn Early winter

16.3◦ 24.4◦ 24.8◦ 23.0◦ 8.8◦ 11.9◦ 23.7◦

3.7◦ N 16.8◦ N 24.7◦ N 15.3◦ N 21◦ S 23.2◦ S 23◦ S

11.6 11.6 4.6 13.3 13.8 10.6 14.7

March 9–11, 1993e March 20, 1995e June 3, 1993e March 26–28, 1999f June 16–18, 2003f June 4, 1988g November 2–4, 2003f

a b c d e f g

LS is solar longitude, where 0◦ is northern spring equinox. Instrument: IRHS. Instrument: HIPWAC. Instrument field-of-view 1-arcsec FWHM. Sprague et al. (1996). Smith, M.D., et al. (2001), Smith (2004). Rizk et al. (1991).

N N N N S S S

Ozone abundance on Mars from IR heterodyne spectra

423

Fig. 4. Martian observing scheme and retrieved single-sideband heterodyne spectra from 14 to 16 February 1993, LS = 40◦ (northern mid-spring). The target ozone feature was present in the upper sideband (USB) and the target CO2 feature was present in the lower sideband (LSB) of the observed double-sideband spectrum. In order to display all features in the single-sideband plots here, the portions of the single-sideband spectrum retrieved after modeling that contain the features of interest are plotted together, resulting in a discontinuity. The 1-arcsec instrument field-of-view is displayed at the targeted positions on Mars, and the asterisk indicates the sub-solar point.

transmittance of the atmospheres of both Mars and Earth. The spectral shapes of both transmittances vary across the lower and upper sidebands as a result of the CO2 and multiple ozone absorption features present. Due to this, the contributions from both Earth and Mars to a double-sideband spectrum must be modeled simultaneously in an iterative manner before extracting a Mars-only spectrum. The martian components of the spectra are modeled using the radiative transfer package BEAMINT (Hewagama et al., 1998), developed at NASA’s Goddard Space Flight Center. BEAMINT combines a layer-by-layer radiative transfer modeling engine with an algorithm to combine the contribution from sub-resolution segments of the instrument beam to form the overall spectrum. This can be a noticeable improvement over a single point mean viewing angle model, especially when the viewing geometry is such that the 1 FWHM beam sees contribution from a wide range of planetary longitudes. Due to

the relatively small size of our field-of-view (∼1 ) relative to the disk of Mars (usually 11 –15 ) at the times of observations, a single-element model for the beam was sufficient for this work. BEAMINT accepts planetary parameters, observation circumstances, molecular and thermal height profiles, and a molecular line atlas. Atmospheric and other parameters can be iterated until a best fit of a model spectrum to an observed spectrum is achieved. Uncertainties based on correlation between free parameters and the variance between the observed and model spectra are returned. The uncertainties on ozone abundance therefore reflect uncertainties in other free parameters (e.g., surface and atmospheric temperature) and thus are larger than if other parameters are held fixed. The telluric components of the spectra are modeled through calls by BEAMINT to the terrestrial radiative transfer package GENLN2 (Edwards, D.P., 1992. GENLN2: A General Lineby-Line Atmospheric Transmittance and Radiance Model: Ver-

424

K. Fast et al. / Icarus 181 (2006) 419–431

sion 3.0 Description and Users Guide. Rep. NCAR/TN-367 + STR. Natl. Cent. for Atmos. Res., Boulder, CO). BEAMINT applies the single-sideband GENLN2 transmittance to the single-sideband model spectrum of Mars before folding it about the laser local oscillator frequency and comparing it to an observed double-sideband spectrum. Initial models of telluric transmittance for the martian ozone analysis were taken from the GENLN2 analysis of ozone on Earth described by Fast et al. (2004). The ozone and CO2 abundances for Earth were adjusted to fine-tune the telluric transmittance model for particular spectra of Mars. The martian atmospheric parameters that must be either varied or constrained in the fitting process include a scale factor applied to the initial vertical ozone abundance profile, an additive adjustment to the initial thermal profile, atmospheric pressure at the surface, and surface temperature. The martian CO2 absorption feature present in the instrument bandpass is the primary constraint on all of the parameters except ozone abundance, and is very important for constraining the thermal profile so that the ozone abundance retrieval is accurate. The CO2 line shape is sensitive to the combination of surface temperature, surface pressure, and atmospheric temperature profile. The shape and depth of the line can indicate the presence of large amounts of dust. These measurements at 9.7 µm are not so strongly sensitive to normal amounts of dust as observations at shorter wavelengths, such as ultraviolet measurements. Possible extinction greater than 10–20% was not evident in the CO2 lines shapes and retrieved surface temperatures for any of the observing campaigns. Uncertainties in dust extinction are smaller than derived errors (
20%) in the ozone abundance from the IR heterodyne measurements. The martian surface temperature and initial thermal profile are varied until the best fit to the CO2 line is achieved. The atmospheric surface pressure is externally constrained. Surface vehicle data, such as those from Viking (Zurek et al., 1992), show that surface pressure varies consistently from year to year, barring disruptions such as large-scale dust storms. The pressure model of Tillman et al. (1993) was updated with additional pressure information from the Mars Pathfinder mission and incorporated into software3 developed at NASA’s Goddard Space Flight Center. This software estimates a surface pressure for a particular latitude and longitude based on the corresponding altitude and orbital position (LS ). Altitudes of surface features come from the Mars Global Surveyor (MGS) Mars Orbiter Laser Altimeter (MOLA) (Smith, D.E., et al., 2001). Initial thermal profiles used in the analysis are those retrieved from MGS Thermal Emission Spectrometer (TES) observations (e.g., Fig. 5) (Smith, M.D., et al., 2001). MGS-TES began observing in March of 1999, and the profiles utilized here were taken either around the same time as a heterodyne observation of Mars, or at the same orbital position (LS ) as a pre-MGS observation. MGS samples 14:00 local time and the ozone spectra analyzed here sample roughly between 10:00 and 15:00. The entire MGS thermal profile is modified by adjusting 3 The original code was written by H. Kieffer. Modifications were made by M. Kaelberer, A. Mekelburg, J. Pearl, and E. Winter.

Fig. 5. Example thermal profile from Mars Global Surveyor (Smith, M.D., et al., 2001) and contribution functions (dT /dh · Bν (T ), see text). The left panel shows how various heights are probed by the fully resolved CO2 line at various frequencies from line center. The right panel shows contribution of ozone line center frequencies which probe below ∼20 km.

the value of a temperature increment added to the profile. Mars Exploration Rover mini-TES observations over the course of a day show that this is a reasonable way to modify the profile during the local times considered (Smith et al., 2004). Initial surface temperatures are also taken from MGS-TES observations and are allowed to vary in the fitting process along with the thermal profile additive term in order to best fit the CO2 absorption line shape. The final fit parameters ultimately deviate somewhat from the MGS results because of local time and pointing variations. High altitude non-LTE emission (Mumma et al., 1981; Deming et al., 1983; Livengood et al., 2003) fills the CO2 line core (e.g., Figs. 3 and 6). Channels at the core of the CO2 line are masked and ignored in the fitting process. The CO2 line wing is sensitive to the lower altitude temperatures where the bulk of the ozone contribution to the spectra originates (below ∼10–20 km). IR heterodyne measurements probe ozone column abundance down to the surface, and retrieved column abundances are not strongly affected by reasonable variations to a constantwith-height ozone distribution (Fast, 2005). A constant-withheight ozone mole fraction is therefore assumed in the model, modified by a fitted multiplicative factor. IR heterodyne measurements of ozone and CO2 on Mars resolve the shapes of individual absorption features (Fig. 3). The line shapes define the contribution to the spectrum from the different heights (pressure levels) in the atmosphere. Example contribution functions are shown in Fig. 5. Contribution functions are calculated by multiplying the change in atmospheric transmittance with height, dT /dh, by the Planck function, Bν (T ), at a particular frequency and altitude (i.e., temperature). Contribution functions indicate the region of formation of the modeled absorption feature at frequencies offset from line center along the wing. Higher altitudes are probed by the CO2 line core while the wings probe lower altitudes (higher pressures), as shown in the left panel. The right panel shows that the bulk of the ozone feature is formed at altitudes below ∼20 km for a constant-with-height ozone mole fraction model. Deviations from this distribution

Ozone abundance on Mars from IR heterodyne spectra

425

Fig. 6. Martian observing scheme and retrieved single-sideband heterodyne spectra from 3 to 7 June 1988, LS = 208◦ (northern early autumn). Displayed single-sideband spectra were constructed by removing modeled telluric contribution and one Mars sideband from measured double-sideband spectra.

would result in enhancement or deficit of contribution at particular altitudes, and will be explored in a later paper. An additional parameter, a scale factor applied to the data, also is required. Although the data have been absolutely calibrated, a scale factor is required to account for calibration uncertainty, pointing uncertainty, tracking errors and seeing, as well as the variation in telluric transmittance. By fitting the scale factor, the fully resolved CO2 line shape drives the fit of the atmospheric temperature and surface temperature combination. At each position on Mars, a 25-MHz-resolution spectrum covers 1600 MHz (double sideband) and contains both the martian ozone and CO2 absorption features. A simultaneous 5MHz-resolution spectrum covers 320 MHz (double sideband) across the ozone absorption features, providing denser frequency sampling. The CO2 absorption feature contains the thermal information on the observed region. The 5-MHz-resolution spectrum of ozone is modeled simultaneously with the 25MHz-resolution CO2 spectrum, yielding the ozone abundances

quoted in this work. Fits to the 25-MHz-resolution ozone features are consistent but less precise. A Mars-only spectrum can be constructed from the data and the Mars and Earth models. The portion of the model martian spectrum in the sideband opposite the target martian absorption features is modified by the model telluric transmittance and subtracted from the double-sideband data to create a singlesideband (SSB) spectrum of Mars and Earth. The SSB spectrum is divided by the model telluric transmittance in that sideband, resulting in a single-sideband Mars-only “data” spectrum. Examples of these retrieved Mars spectra and their spatial distribution are presented in Figs. 4 and 6. The 1σ uncertainties on each of the fit parameters that are returned in the radiative transfer analysis take into account the correlations with other parameters as well as the variance between the observed and model spectra. The ozone column abundance uncertainties reflect correlations with the other parameters (atmospheric and surface temperature, scale factor, telluric ozone, and CO2 ) and are therefore more conservative

426

K. Fast et al. / Icarus 181 (2006) 419–431

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 7. Total ozone column densities from this work at the latitudes indicated, as a function of LS . The curve indicates measurement of total water vapor column density (MGS-TES, Smith et al., 2001) for the martian year that ran from mid-2002 to early 2004. The correlation coefficient, r, between the ozone and water vapor column densities is indicated, along with the percent probability that uncorrelated values could produce that particular r value. The correlation coefficients between the total column densities are consistently negative, indicating anticorrelation, but are not always close to −1, indicating the need to consider the distribution of species within the column.

representations of the accuracies of the retrievals than uncertainties in ozone abundance when the other parameters are held fixed. The column abundances themselves are integrated above the local topography on Mars, as are the water vapor column abundances from other investigators that are presented below. 4. Ozone abundances from infrared heterodyne spectra of Mars Ozone column densities from the 5-MHz-resolution heterodyne spectra at latitudes investigated more than once in this work are displayed as functions of LS in Fig. 7, along with water vapor column densities from MGS-TES (Smith, M.D., et al., 2001; Smith, 2004). The ozone data sets sample a variety of orbital positions from pre-aphelion to post-perihelion.

The retrieved total ozone column densities show a general decline between LS = 40◦ –291◦ . A decline in ozone abundance is expected from photochemistry, which predicts anticorrelation with water vapor abundance. During the warmer perihelion seasons (LS ∼ 180◦ –360◦ ), more water vapor is available as a source of odd hydrogen, which destroys ozone. During the colder aphelion seasons (LS ∼ 0◦ –180◦ ), water vapor freezes out of the atmosphere, decreasing odd hydrogen production. The decline in ozone column density between LS = 40◦ and 291◦ confirms in general the anticorrelation of ozone and water vapor predicted by photochemistry. However, a strict anticorrelation of total ozone and water vapor column density is not observed. Correlation coefficients, r, between the ozone and water vapor column densities are noted in Fig. 7, along with

Ozone abundance on Mars from IR heterodyne spectra

the percent probabilities that uncorrelated values could produce those correlation coefficients. The correlation coefficients between the total column densities are consistently negative, indicating anticorrelation. For example, water vapor abundance is high at 40◦ –60◦ N around aphelion due to a cold trap inhibiting transport south of ∼30◦ N, resulting from a low water vapor saturation altitude or “hygropause” (∼10 km) and aphelion Hadley circulation proposed by Clancy et al. (1996a) and evident in MGS-TES observations (Smith, M.D., et al., 2001). IR heterodyne ozone abundances at those latitudes are seen to drop in response (Fig. 7). Not all of the correlation coefficients are close to −1, however, and studying constituent vertical distribution would be an important next step. For instance, in their seasonal photochemical model, Clancy and Nair (1996) note variations in ozone number density at 20 and 40 km altitude that are much greater than the variation in the total ozone column, which they attribute to variations in the hygropause altitude. During cooler aphelion seasons, the hygropause is lower, at 10–15 km, and therefore ozone abundance at the higher altitudes can increase in the absence of large amounts of odd hydrogen. During warmer perihelion seasons, the hygropause rises to ∼40 km, providing a source of odd hydrogen that destroys ozone at the higher altitudes. The three-dimensional photochemical model of Lefèvre et al. (2004) predicts orbital variations in the vertical distribution of ozone. The varying hygropause altitude and model predictions point to the need to measure the vertical distribution of ozone as well in order to understand the lack of strict anticorrelation of total column abundances. Another way to investigate ozone/water anticorrelation is to look at the variation in the two constituents with latitude at particular orbital positions (LS ). Contemporaneous water vapor column abundances (see Table 1) are displayed in Fig. 8 along with the ozone abundances as functions of latitude for the various orbital positions (LS ) and correlation coefficients. Again, the consistently negative correlation coefficients between the total column densities point to an anticorrelation of ozone and water vapor abundance. The correlation coefficients are not consistently close to −1, indicating the need to consider the vertical distribution of the species within the column when interpreting their measurements and investigating their chemistry. The diurnal behavior of ozone was investigated on a few occasions (LS = 74◦ at 60◦ N, LS = 115◦ at 40◦ S, and LS = 208◦ at 60◦ S). The interval of daytime local times ranged from 4 to 7.5 h. Photochemical models predict significant diurnal variation in ozone abundance (Nair et al., 1994; Lefèvre et al., 2004). However, no statistically significant variation in ozone abundance was seen over any of the relatively short daytime local time intervals observed. 5. Infrared heterodyne and ultraviolet ozone retrievals compared 5.1. Ultraviolet spectroscopy of ozone on Mars Ozone on Mars has been studied in the ultraviolet (UV) from outside Earth’s atmosphere by sensing absorption in the 200–

427

330-nm region, or Hartley bands, brought about by electronic excitation of ozone (1 B2 ← X 1 A1 ), resulting in a broad absorption feature centered around 255 nm. Rayleigh scattering of solar photons contributes wavelength-dependent continuum brightness in the spectral region. There is a small surface reflectance contribution, greater over the polar ice caps. Water ice clouds and dust are additional sources of scattering that can add significant opacity (τ ∼ 0.25), decreasing the effective path length through the ozone. The modeling of UV spectra yields ozone column densities, as well as height information in the case of occultations of the Sun or stars by the limb atmosphere. Spectral modeling requires assumptions about the vertical distribution of scattering components, which modify the effective path length of incoming and reflected solar UV radiation through the ozone (e.g., Lindner, 1995). The high-resolution ozone IR spectroscopic measurements in this work are not strongly sensitive to scattering in the martian atmosphere. IR heterodyne spectroscopy samples the entire ozone column because it detects thermal emission from the surface modified by CO2 and ozone opacity. Apart from major dust storm activity, Mars’ atmosphere is relatively clear at 9– 10 µm. The degree of continuum opacity (dust, aerosols) can be modeled from the pressure-broadened 9.6 µm CO2 line shape. IR spectroscopy samples the vertical distribution of ozone and temperature down to the surface. UV band spectroscopy samples the column of ozone traversed by the UV radiation, which may not reflect contribution from the entire ozone column. 5.2. Ozone retrievals from IR heterodyne and UV spectra The analysis in this work provides an excellent opportunity to test abundance retrieval of the same constituent by two very different methods. Contemporaneous measurements of ozone column density by IR heterodyne spectroscopy (this work) and UV spectroscopy (Clancy et al., 1996b, 1999; Clancy, R.T., private communication), as well as those made during similar orbital periods (LS ) in different years, can be compared. Clancy et al. (1996b) detail the modeling of HST-FOS UV spectra. They employ three atmospheric layers, distributing cloud opacity among the lower two layers and constraining water condensation level and Rayleigh scattering distribution according to the thermal profile. They assume a uniform ozone distribution throughout 0–40 km for perihelion measurements and an altitude-increasing mixing ratio (Clancy and Nair, 1996) for aphelion measurements. The atmospheric scattering properties may or may not allow the entire ozone column to be probed by the measurements themselves, but assumed distributions of ozone and scattering allow retrieval of the total ozone column abundance down to the surface. Different assumed distributions of ozone and scattering can result in different total ozone column abundance retrievals. The uncertainty in the accuracy of the ozone retrievals from UV spectra is given by Clancy et al. (1999) as +50%, −20%, shown in Fig. 9. The IR heterodyne retrievals have accuracy uncertainties in the range of ±15–35%, determined from the noise in each individual measurement. The temporal correspondence between the IR and UV measure-

428

K. Fast et al. / Icarus 181 (2006) 419–431

(a)

(b)

(c)

(d)

(e)

(f)

(g) Fig. 8. Ozone abundances (filled diamonds) and 1σ accuracy bars retrieved from 5-MHz-resolution infrared heterodyne spectra are displayed with contemporaneous water vapor column densities (open diamonds and dotted curves, see Table 1). Accompanying numbers indicate day of month of heterodyne observations. The correlation coefficient, r, between the ozone and water vapor column densities is indicated, along with the percent probability that uncorrelated values could produce that particular r value. The correlation coefficients between the total column densities are consistently negative, indicating anticorrelation, but are not all close to −1, indicating the need to consider the distribution of species within the column.

Ozone abundance on Mars from IR heterodyne spectra

429

(a)

(b)

(c)

(d)

(e) Fig. 9. Ozone retrievals from infrared heterodyne spectroscopy (diamonds with 1σ uncertainties) compared to those from ultraviolet measurements made by the Hubble Space Telescope Faint Object Spectrograph (Clancy et al., 1996b, 1999; Clancy, R.T., private communication). The dates and orbital positions (LS ) of the UV measurements (open squares) are detailed in Table 2. (In panel (a) large squares are from the same year as this work, and small squares are from a different year, and the UV measurements were earlier than IR measurements by ∼10◦ in LS .) The error bars on the UV measurements represent the uncertainty in the accuracy as quoted in Clancy et al. (1999), which was +50%, −20%. The error bars on the infrared heterodyne measurements reflect the uncertainty in the accuracy and were calculated for each individual measurement.

Table 2 Hubble Space Telescope observations comparable to this work Fig. 9 panel

Infrared heterodyne

HST-FOS Date

LS

Date

Reference

(a)

74◦

March 17–23, 1995

(b) (c) (d) (e)

102◦ 115◦ 202◦ 208◦

July 4–7, 1993 March 24–29, 1999 June 10–16, 2003 June 3–7, 1988

63◦ 61◦ 104◦ 118◦ 212◦

February 25, 1995 January 4, 1997 January 17, 2001 February 16, 2001 June 28, 2003

Clancy et al., 1996b Clancy et al., 1999 Clancy, R.T., priv. comm., 2004

LS

430

K. Fast et al. / Icarus 181 (2006) 419–431

ments is detailed in Table 2. Retrieved total ozone column abundances are compared in Fig. 9 in those instances in which measurements were available at comparable orbital positions (LS ). Orbital periods from aphelion to approaching perihelion have been sampled by the two methods. Overall, the UV (open squares) and IR heterodyne ozone (filled squares) retrievals in Fig. 9 show consistent behavior. There is generally good agreement in the variation of ozone at low to mid-latitudes, from high around aphelion (Fig. 9a) to low approaching perihelion (Figs. 9d and 9e). Both IR and UV retrievals show much more short-scale variability in late northern spring (Fig. 9a) than at the other periods. The good agreement between the two very different measurement techniques provides mutual validation of the ozone retrieval methods. This is important for expanding the pool of available ozone measurements for testing models of atmospheric photochemistry. The only disagreements seen in the variation with latitude are at 60◦ –80◦ N in late northern spring (Fig. 9a) and in the opposite sense at 20◦ –40◦ S at early northern summer (Fig. 9c). Uncertainties in the thermal profiles retrieved from the IR heterodyne measurements cannot account for these differences. The disagreements may be due to slightly different LS (Table 2) or due to the observing geometry (highest observable latitudes). The UV retrievals are sensitive to dust opacity, resulting in an underestimate of ozone abundance if the assumed opacity is too low. UV spectra primarily sample the upper part of the ozone column and an assumed ozone distribution is used to estimate total column abundances. The IR heterodyne spectra probe total ozone abundance down to the surface, but are subject to additional uncertainty due to telluric effects. Coordinated measurements would help to resolve the few discrepancies. 6. Conclusions Recognizing the role of odd hydrogen chemistry in the atmospheric stability of Mars gave direction to the solution of the CO2 reformation problem. Odd hydrogen resulting from water vapor photolysis participates in catalytic reactions to reform CO2 , and it also destroys ozone. It follows that measurements of ozone abundance can trace the abundance of the odd hydrogen species that have not yet been directly observed in the atmosphere of Mars. Observations of ozone are not possible from the ground by conventional UV and IR means due to telluric ozone. IR heterodyne spectroscopy (λ/λ
106 ), as applied in this work, can measure ozone absorption features on Mars when they are Doppler-shifted away from telluric counterparts, enabling the only direct measurement of martian ozone from the surface of the Earth. This makes it possible to conduct long-term studies of the chemistry and stability of the martian atmosphere through a tracer of the odd hydrogen species thought to participate in the reformation of CO2 . The analysis of IR heterodyne spectra of Mars in this work confirms in general the ozone/water anticorrelation expected for odd hydrogen photochemistry and supports the odd hydrogen solution to the CO2 stability problem. Spatial and seasonal anticorrelation is seen in the ozone and water vapor column

abundances (Figs. 7 and 8). However, the quantitative behavior of total ozone abundance did not show the same degree of variation as the water vapor. This and the variability of the hygropause altitude with orbital position suggest that the vertical distribution of the constituents is important to photochemical processes and the accurate interpretation of measured data by all methods and the optimization of atmospheric models of Mars. Fully resolved IR heterodyne line shapes have the potential to probe the vertical distribution of ozone down to the surface on their own and in combination with O2 (1 ∆) measurements (Fast, 2005). The use of two very different techniques to measure the same constituent is important for testing the accuracy of both methods. Contemporaneous and seasonal/orbital comparisons of ozone retrievals from this work and from Hubble Space Telescope UV observations show very good agreement, from aphelion to near perihelion. This is important for broadening the pool of available data with which to validate photochemical models. The Mars Express instrument SPICAM (Bertaux et al., 2000, 2004) is conducting the first long-term study of ozone from orbit since Mariner 9. IR heterodyne spectroscopy can provide an important test of those results and, as with the HSTFOS observations, discrepancies may highlight observational and retrieval issues for either technique, or issues of chemistry on Mars. IR heterodyne spectroscopic observations of ozone on Mars can be made from the surface of the Earth and are limited only by Doppler shift due to the relative motion of Mars. There is always a need for Earth-based observations to provide longterm monitoring and to validate and complement spacecraft data. Spacecraft have limited instrumentation and lifetimes, and sending spacecraft to Mars is an enormous (and not always successful) undertaking. Ongoing Earth-based observations of atmospheric constituents over longer time spans and with higher spectral resolution such as IR heterodyne measurements of ozone will continue to improve the understanding of the chemistry and stability of the martian atmosphere. Acknowledgments The authors thank the directors and staff of the NASA Infrared Telescope Facility for their support of infrared heterodyne observations of Mars. We thank A.L. Sprague, M.D. Smith, and R.T. Clancy for providing relevant martian data for comparison, V. Krasnopolsky and R. Novak for informative discussion, and J. Delgado for providing technical support. This work was supported by the NASA Planetary Astronomy Program. References Abbas, M.M., Brown, L.W., Buhl, D., Clark, T.A., Hillman, J., Kostiuk, T., Kunde, V., Mumma, M.J., 1976. A 10-micron superheterodyne receiver for spectral line observations. Bull. Am. Astron. Soc. 8, 508. Barth, C.A., Hord, C.W., 1971. Mariner ultraviolet spectrometer: Topography and polar cap. Science 173, 197–201. Barth, C.A., Hord, C.W., Stewart, A.I., Lane, A.L., Disk, M.L., Anderson, G.P., 1973. Mariner 9 ultraviolet spectrometer experiment: Seasonal variations of ozone on Mars. Science 179, 795–796.

Ozone abundance on Mars from IR heterodyne spectra

Bertaux, J.-L., Fonteyn, D., Korablev, O., Chassefière, E., Dimarellis, E., Dubois, J.P., Hauchecorne, A., Cabane, M., Rannou, P., Levasseur-Regourd, A.C., Cernogora, G., Quemerais, E., Hermans, C., Kockarts, G., Lippens, C., De Maziere, M., Moreau, D., Muller, C., Neefs, B., Simon, P.C., Forget, F., Hourdin, F., Talagrand, O., Moroz, V.I., Rodin, A., Sandel, B., Stern, A., 2000. The study of the martian atmosphere from top to bottom with SPICAM light on Mars Express. Planet. Space Sci. 48, 1303–1320. Bertaux, J.-L., Korablev, O., Quemerais, E., Perrier, S., Fedorova, A., Muller, C., and the SPICAM Team, 2004. SPICAM on Mars Express: First results and first observations of water ice at south pole. Lunar Planet. Sci. 35. Abstract 2178. Clancy, R.T., Nair, H., 1996. Annual (perihelion–aphelion) cycles in the photochemical behavior of the global Mars atmosphere. J. Geophys. Res. 101, 12785–12790. Clancy, R.T., Grossman, A.W., Wolff, M.J., James, P.B., Rudy, D.J., Billawala, Y.N., Sandor, B.J., Lee, S.W., Muhleman, D.O., 1996a. Water vapor saturation at low altitudes around Mars aphelion: A key to Mars climate? Icarus 122, 36–62. Clancy, R.T., Wolff, M.J., James, P.B., Smith, E., Billawala, Y.N., Lee, S.W., Callan, M., 1996b. Mars ozone measurements near the 1995 aphelion: Hubble Space Telescope ultraviolet spectroscopy with the Faint Object Spectrograph. J. Geophys. Res. 101, 12777–12783. Clancy, R.T., Wolff, M.J., James, P.B., 1999. Minimal aerosol loading and global increases in atmospheric ozone during the 1996–1997 martian northern spring season. Icarus 138, 49–63. Clancy, R.T., Sandor, B.J., Moriarty-Schieven, G.H., 2004. A measurement of the 362-GHz absorption line of Mars atmospheric H2 O2 . Icarus 168, 116– 121. Deming, E., Espenak, F., Jennings, D., Kostiuk, T., Mumma, M., Zipoy, D., 1983. Observations of the 10-micron natural laser emission from the mesospheres of Mars and Venus. Icarus 55, 347–355. Encrenaz, Th., Bézard, B., Greathouse, T.K., Richter, M.J., Lacy, J.H., Atreya, S.K., Wong, A.S., Lebonnois, S., Lefèvre, F., Forget, F., 2004. Hydrogen peroxide on Mars: Evidence for spatial and seasonal variations. Icarus 170, 424–429. Espenak, F., Mumma, M.J., Kostiuk, T., Zipoy, D., 1991. Ground-based infrared measurements of the global distribution of ozone in the atmosphere of Mars. Icarus 92, 252–262. Fast, K.E., 2005. Mars ozone abundances from infrared heterodyne spectra and their application to study of the stability of the martian atmosphere. Ph.D. dissertation, University of Maryland. Fast, K.E., Kostiuk, T., Espenak, F., Livengood, T.A., Hewagama, T., A’Hearn, M.F., 2004. Stratospheric ozone profiles from Mauna Kea, Hawai’i (19.8◦ N, 155.5◦ W) using infrared heterodyne spectroscopy, 1988–2003. Geophys. Res. Lett. 31, doi:10.1029/2004GL019443. L08109. Hewagama, T., Goldstein, J., Buhl, D., Espenak, F., Fast, K., Kostiuk, T., Livengood, T.A., 1998. Spectral line analysis for planetary atmospheric dynamics retrieval. Bull. Am. Astron. Soc. 30. Abstract 41.P11. Kostiuk, T., 1994. Physics and chemistry of upper atmospheres of planets from infrared observations. Infrared Phys. Technol. 35, 243–266. Kostiuk, T., Mumma, M.J., 1983. Remote sensing by IR heterodyne spectroscopy. Appl. Opt. 17, 2644–2654. Krasnopolsky, V.A., 2003. Mapping of Mars O2 1.27-µm dayglow at four seasonal points. Icarus 165, 315–325. Krasnopolsky, V.A., Bjoraker, G.L., 2000. Mapping of Mars O2 (1 ∆) dayglow. J. Gephys. Res. 105, 20179–20188. Lane, A.L., Barth, C.A., Hord, C.W., Stewart, A.I., 1973. Mariner 9 ultraviolet spectrometer experiment: Observations of ozone on Mars. Icarus 18, 102– 108.

431

Lefèvre, F., Lebonnois, S., Montmessin, F., Forget, F., 2004. Three-dimensional modeling of ozone on Mars. J. Geophys. Res. 109, doi:10.1029/ 2004JE002268. E07004. Lindner, B.L., 1995. Mars ozone: Mariner 9 revisited. Icarus 113, 213–216. Livengood, T.A., Kostiuk, T., Fast, K.E., Annen, J.N., Sonnabend, G., Hewagama, T., 2003. Meridional mapping of mesospheric temperatures from CO2 emission along the MGS ground track. Bull. Am. Astron. Soc. 35. Abstract 03.04. McElroy, M.B., Donahue, T.M., 1972. Stability of the martian atmosphere. Science 177, 986–988. Mumma, M., Buhl, D., Chin, G., Deming, D., Espenak, F., Kostiuk, T., Zipoy, D., 1981. Discovery of natural gain amplification in the 10-micrometer carbon dioxide laser bands on Mars—A natural laser. Science 212, 45–49. Nair, H., Allen, M., Anbar, A.D., Yung, Y.L., Clancy, R.T., 1994. A photochemical model of the martian atmosphere. Icarus 111, 124–150. Novak, R.E., Mumma, M.J., DiSanti, M.A., Dello Russo, N., 2002. Mapping of ozone and water in the atmosphere of Mars near the 1997 aphelion. Icarus 158, 14–23. Noxon, J.F., Traub, W.A., Carleton, N.P., Connes, P., 1976. Detection of O2 dayglow emission from Mars and the martian ozone abundance. Astrophys. J. 207, 1025–1035. Parkinson, T.M., Hunten, D.M., 1972. Spectroscopy and aeronomy of O2 on Mars. J. Atmos. Sci. 29, 1380–1390. Rizk, B., Wells, W.K., Hunten, D.M., Stoker, C.R., Freedman, R.S., Rouch, T., Pollack, J.B., Haberle, R.M., 1991. Meridional martian water abundance profiles during the 1988–1989 season. Icarus 90, 205–213. Schmülling, F., Kostiuk, T., Buhl, D., Rozmarynowski, P., Segal, K., Livengood, T., Hewagama, T., 1999. A new infrared heterodyne instrument for measurements of planetary wind and composition. Bull. Am. Astron. Soc. 31. Abstract 08.03. Smith, D.E., Zuber, M.T., Frey, H.V., Garvin, J.B., Head, J.W., Muhleman, D.O., Pettengill, G.H., Phillips, R.J., Solomon, S.C., Zwally, H.J., Banerdt, W.B., Duxbury, T.C., Golombek, M.P., Lemoine, F.G., Neumann, G.A., Rowlands, D.D., Aharonson, O., Ford, P.G., Ivanov, A.B., McGovern, P.J., Abshire, J.B., Afzal, R.S., Sun, X., 2001. Mars Orbiter Laser Altimeter (MOLA): Experiment summary after the first year of global mapping of Mars. J. Geophys. Res. 106, 23689–23722. Smith, M.D., 2004. Interannual variability in TES atmospheric observations of Mars during 1999–2003. Icarus 167, 148–165. Smith, M.D., Pearl, J.C., Conrath, B.J., Christensen, P.R., 2001. Thermal Emission Spectrometer results: Mars atmospheric thermal structure and aerosol distribution. J. Geophys. Res. 106, 23929–23945. Smith, M.D., Wolff, M.J., Lemmon, M.T., Spanovich, N., Banfield, D., Budney, C.J., Clancy, R.T., Ghosh, A., Landis, G.A., Smith, P., Whitney, B., Christensen, P.R., Squyres, S.W., 2004. First atmospheric science results from the Mars exploration rovers mini-TES. Science 306, 1750–1753. Sprague, A.L., Hunten, D.M., Hill, R.E., Rizk, B., Wells, W.K., 1996. Martian water vapor, 1988–1995. J. Geophys. Res. 101, 23229–23241. Tillman, J.E., Johnson, N.C., Guttorp, P., Percival, D.B., 1993. The martian annual atmospheric pressure cycle—Years without great dust storms. J. Geophys. Res. 98, 10963–10971. Wehrbein, W.M., Hord, C.W., Barth, C.A., 1979. Mariner 9 ultraviolet spectrometer experiment: Vertical distribution of ozone on Mars. Icarus 38, 188–199. Zurek, R.W., Barnes, J.R., Haberle, R.M., Pollack, J.B., Tillman, J.E., Leovy, C.B., 1992. Dynamics of the atmosphere of Mars. In: Kieffer, H.H., Jakosky, B.M., Snyder, C.W., Matthews, M.S. (Eds.), Mars. Univ. of Arizona Press, Tucson, p. 839.