16 July 1971, Volume 173, Number 3993 SCIENCE

Disk Spectra

Mariner Ultraviolet Spectrometer:Topography and Polar Cap

Ultraviolet measurements reveal the topography of Marsand show that ozone may be adsorbed on the polar cap.

Charles A. Barth and Charles W. Hord

Ultraviolet spectrum observationsfrom the 1969 Mariner flyby missionsyielded an unexpected abundance ofmeaningful data about the lower atmo-sphere and surface of Mars. The ultra-violet spectrometer experiment was de-signed primarily to determine the com-position and structure of the Mars upperatmosphere (1). The experiment wassuccessful in determining the propertiesof the lower atmosphere and surfacebecause of two characteristics of theMars surface. First, the ultraviolet re-flectivity of the Mars surface in thedesert region is low; and, second, thereflectivity of the polar cap in the ultra-violet is high. The first property meansthat any ultraviolet light reflected by theplanet over desert regions is the resultof scattering by atmospheric constitu-ents; thus, ultraviolet intensity is ameasure of the number of atmosphericscatterers. By interpretation of thesemeasurements, the local atmosphericpressure may be determined; andvariations in the local atmosphericpressure may be used to measure thetopography of the planet. The secondproperty (the high reflectivity of thepolar cap) indicates that any ultravioletabsorber that lies on or above the polarcap may be detected by measuring thereflected ultraviolet light. An ultravioletabsorber, ozone, is present on the polar

The authors are with the department of astro-geophysics and the Laboratory for Atmosphericand Space Physics of the University of Colorado,Boulder 80302.

16 JULY 1971

cap, and its presence may be indicativeof a general property of a carbon di-oxide polar cap-namely, the ability toadsorb minor constituents out of theatmosphere.

Instrument and Spacecraft

The ultraviolet spectrometer, alongwith the television cameras and infraredinstruments on the scan platform ofthe spacecraft, was aimed at selectedregions of the planet as the two Mari-ners flew by Mars at 7 kilometers persecond. Mariner 6 viewed several lightand dark areas in the equatorial region;Mariner 7 flew farther south and ob-tained observations of the south polarcap and the bright desert region Hellas.The field of view of the ultraviolet

spectrometer through its 250-millimeterfocal length telescope is 0.23 X 2.3 de-grees. When the spacecraft was closestto the planet, the instrument viewed an

area on the surface that was 10 by 100kilometers. A spectral scan was com-

pleted every 3 seconds, during whichtime the projected field of view moved21 kilometers in the direction of itsnarrow dimension. Although the spec-

tral range of the 250-millimeter focallength scanning spectrometer is from1100 to 4300 angstroms, the results re-

ported here were obtained in the wave-

length region between 2100 and 3600angstrom at a resolution of 20 ang-

stroms (1).

Several hundred spectra were ob-tained with the ultraviolet spectrome-ters pointed at various locations on thedisk of Mars. Lighting conditions atthe points of observation varied fromthe sun nearly overhead to near the ho-rizon. These measurements were madeat phase angles of 46, 63, and 90degrees.

Three examples of the spectra ob.served over desert regions are shownin Fig. 1, where each spectrum is thesum of several individual spectra. Theupper spectrum from the Candor regionis an example of an observation madewith a nearly overhead sun. The middlespectrum, which has the angles of in-cidence and emission nearly equal, wasrecorded near the western edge of Meri-diani Sinus, where the radio occulta-tion experiment measured the surfacepressure (2). The lower spectrum wasobtained in Deucalionis Regio, whichwas near the terminator during thisobservation.

Ultraviolet Reflectance

An example of the reflectance of thedisk of Mars is shown in Fig. 2, whichwas obtained by dividing the spectrumfrom the Meridiani Sinus region by thespectrum of the sun that was measuredin rocket experiments by the Naval Re-search Laboratory (3). The bulk of vari-ation has been removed, and thus allthe detail in the disk spectra can beseen to be due to the Fraunhofer struc-ture in the solar spectrum. In fact, thevariation that remains may be causedby an inexact comparison between theMars spectra and the solar spectrum,which were obtained by two differenttechniques with different types of in-strumentation. The reflectance of theMars disk in the Meridiani Sinus re-gion increases with decreasing wave-length. The monotonic increase in thereflectance at the shorter wavelengths issuggestive of atmospheric scatterers withan optical depth of T = 0.1 at 3050angstroms. The same spectrum shape isfound in most of the several hundred

197

CZ

.1000

2000 2500 3000 3500 4000Wavelength (A)

Fig. 1. Mars spectra observed over desert regions. The upper spectrum was obtainedwhile the ultraviolet spectrometer was observing Mars at 40N,70°W, with a solarzenith angle o0 of 200, an emission zenith angle 0 of 820, and a phase angle of 630.The middle spectrum was obtained at 1 OS,9 °W, with 0, = 470, = 420, and = 62 0.

The lower spectrum was obtained at 180S,350°W, with O0 = 650, 0 = 270, and 0 = 90°.

ultraviolet spectra taken over the Marsdesert regions. There is no evidence ofozone, which absorbs strongly between2000 and 3000 angstroms, in any of thespectra obtained over the desert regionsof Mars.The magnitude of the reflectance at-

tributed to !atmospheric scattering isabout three times that expected from a

Rayleigh scattering atmosphere of pure

carbon dioxide with a surface pressure

of 6.6 millibars (2). Excess scattering isalso observed in Earth's atmosphere,where it is attributed to the existence ofaerosols. Results from both the Marinertelevision and Mariner infrared spec-

trometer indicated the existence of po-

tential nonmolecular scatterers. The tel-evision experiments (4) showed a limbhaze at altitudes between 15 and 40kilometers, and the infrared spectrome-ter (5) measured a spectral reflectionfeature attributed to condensate of at-mospheric carbon dioxide at the brightlimb. The infrared spectrometer also ob-served an atmospheric spectrum featureat the dark limb, which is attributed toa silicate material. The particulate mat-ter measured by these experiments mayor may not be the same material thatis producing the ultraviolet scattering.

Surface Pressure

If the ultraviolet reflectance of thedesert regions of Mars is identified as

originating primarily from atmosphericscattering, it becomes possible to de-termine the atmospheric pressure at thesurface from the ultraviolet measure-

ments (6). The method requires theassumption that small particles also

198

contribute to the atmospheric scatteringand that these particles are uniformlydistributed in the atmosphere. This may

or may not be a valid assumption. How-ever, if the particles are small enough toproduce the observed increased reflec-tivity with decreasing wavelengh, theywould remain suspended in the atmo-sphere for very long times (7). Sincethe spectroscopic character of the datashows that the ground is essentiallyblack in the ultraviolet and since themagnitude of the intensity shows thatthe scattering is optically thin, the ultra-violet reflectance should be directlyproportional to the number of scatterersobserved by the instrument. Since we

have assumed that the total number ofscatterers is proportional to the numberof molecules and since the total num-

ber of molecules in a column is directlyproportional to the pressure when theviewing geometry is taken into account,we may normalize the intensity to thepressure at one location and expect theultraviolet intensity measured elsewherealong the observation track to be a di-rect measure of pressure.The normalization point chosen was

.06-

4, .03a)

.I-qCK

2600

f l xI l I I lI L

3100 3600

Wavelength (A)

Fig. 2. Ultraviolet reflectance of a desertregion. This reflectance was derived fromthe middle spectrum in Fig. 1.

centered at latitude 1°S and longitude10°W near the Meridiani Sinus region,where the radio occultation experimentmeasured the surface pressure as 6.6millibars (2). When this method wasapplied to the ultraviolet reflectance at3050 angstroms, the surface pressurewas found to vary along the observa-tion track, with the lowest pressurebeing slightly greater than 4 millibarsand the highest, 8 millibars.

Topography

When the variation of surface pres-sure along the observation track is com-pared with the surface features re-vealed by the television pictures fromthe Mariner spacecraft (4), there is acorrelation between pressure variationsand topographical features on both asmall and a large scale. The most plau-sible explanation of this correlation isthat the variation in the measured pres-sure is produced by elevation differenceson the surface of Mars. The atmosphericpressure over high elevations is low, andthe highest pressures are over the lowestelevations. A simple atmospheric mod-el, in which the scale height near thesurface is constant both as a function ofaltitude and location, was used to con-vert the pressure variations into altitudevariations. A scale height of 10 kilo-meters was used, which is representativeof the value measured by the radio oc-cultation experiment at the Mariner 6entry point (2). The elevation varia-tion found from the Mariner observa-tions has a total excursion of 7 kilo-meters, with the 8-millibar pressure cor-responding to an elevation that is 3kilometers lower than the MeridianiSinus region, and the 4-millibar pressureto an elevation that is almost 4-kilo-meters higher than this reference level.A detailed comparison between the

elevation measurements and the topo-graphic features is shown in Fig. 3. Amap prepared by the Army TopographicService shows, in addition to the classi-cal names and markings, topographicdetail derived from the Mariner 6 andMariner 7 television pictures. The ob-servation track of the ultraviolet spec-trometer, shown by elongated areas en-closed by heavy lines, for the most partfalls within the areas where near-en-counter television pictures were ob-tained. These are shown in the rectan-gular areas enclosed by light lines. Tofacilitate comparison with the map, lon-gitude is used as the parameter for thealtitude data given in the lower part of

SCIENCE, VOL. 173

v l I I

the figure. The uppermost of the threegraphs corresponds to the Mariner 6track at the most northerly latitudes;the middle graph corresponds to theMariner 6 track at middle latitudes;and the lowermost graph correspondsto the Mariner 7 tracks at the mostsoutherly latitudes.

Detailed correlations may be recog-nized. Begin with the Mariner 6 trackat 0°,70°W; the elevation starts high,drops rapidly in moving to the right, andthen rises abruptly at 60°W. At this lo-cation, the map shows what appears to

be a high crater wall being illuminatedfrom the left. To the right, the eleva-tion continues to fall and reaches aminimum at the edge of a crater wallat 48°W. The uppermost track at lati-tudes between 13°N and 2°S is moreor less level on a large scale, but showsdetails that may be matched with indi-vidual craters. At 22°W, for example,a crater shows as a depression on thealtitude scale. On the Mariner 6 trackbetween 100 and 20°S, there is a gen-eral depression (when viewed from leftto right) which reaches a minimum near

18°W and then rises farther to theright. The altitude variation from 430to 18°W is 4 kilometers. The rise to theeast from 180 to 350°W regains this 4kilometers. Once again, there is a de-tailed correlation with individual cra-ters-for example, at longitudes 110and 13°W. The Mariner 7 track, whichmoves in a southeasterly direction start-ing at 35°S,50°W, shows a rapid riseof 4 kilometers, which reaches a peakat 440S,41 °W and then falls rapidlyfarther to the southeast. The large cra-ter at 50°S,34°W sthows up as a

6064 ~ o Latitude2 - , .,, 200(N) to 10C(S)_

-2 I

-6 I I l I I I I I I _61.._ 1 -t I I I | I I _

4 L e Latitude -

o2 * ....1st(S) to 20(S)S

-2

-6

80 60 40 20 0

Longitude340 320 300 280

Fig. 3. Elevations of regions on Mars observed from Mariner 6 and Mariner 7. The areas enclosed by heavy lines on the map showthe areas observed by the ultraviolet spectrometer. The upper graph gives altitude corresponding to the Mariner 6 observation tracksbetween 20°N and 10°S. The middle graph corresponds to the Mariner 6 tracks between 10° and 20'S, and the lower graph cor-responds to the Mariner 7 tracks between 200 and 60'S.16 JULY 1971 199

depression. Starting at 20°S,355°W,the Mariner 7 track shows the Noachisregion to be a high plateau with in-dividual crater depressions. At 318°W,the elevation drops abruptly eastwardinto Hellas, a drop 5 kilometers ina horizontal distance of 500 kilometers.Hellas, which is devoid of craters, risesslowly in elevation as the observationtrack moves eastward.The topography of Mars has also

been mapped in part by ground-basedradar, ground-based infrared observa-tions!(8), and the infrared spectrometeron Mariner 6 and Mariner 7 (9).Where it has been possible to com-pare the topography determined bythe ultraviolet spectrometer, there isqualitative agreement. When the in-frared and ultraviolet spectrometersviewed the same areas on Mars, theoverall agreement was quantitative.Comparison on a detailed level revealssmall-scale differences, notably in Hellas,where the ultraviolet measurements in-dicate a relatively smooth surface ascompared with large variations in alti-tude observed by the infrared spectrom-eter. However, the overall agreementis strong circumstantial evidence thatthe basic concepts used in the two meth-ods are correct.

o<

-.

4)

Cs

Polar Cap

A series of ultraviolet spectrum obser-vations were made across the southerndesert regions and onto the polar cap at60°S. The striking difference betweenthe ultraviolet spectra of the desert re-gions and the polar cap is illustrated inFig. 4. The lower spectrum was obtainedover the Mare Australe region (55OS,27°W), just north of the edge of thepolar cap, and the upper spectrum wasrecorded well onto the polar cap at 65°S,316°W. The polar cap spectrum is muchmore intense, a factor of 3 at 2900angstroms; however, it falls off morerapidly toward shorter wavelengths; theratio is about 2 at 2550 angstroms. Inthe polar cap spectra, as in the rest ofthe disk spectra, all of the spectrumdetail that appears with 20-angstromresolution is due to Fraunhofer featuresin the solar spectrum. Both the polarcap and desert spectra extend shortwardto 2085 angstroms, where the incidentsolar spectrum has a discontinuity thatdrops sharply in intensity at shorterwavelengths.The ultraviolet reflectance of the polar

cap in the 2200- to 2900-angstromwavelength region was obtained bydividing the Mars spectrum by the solar

2000 2500

Wavelength (A)

Fig. 4. Comparison of Mars spectrum observed over the polar cap with one observedover desert region. The upper spectrum was obtained while the ultraviolet spectrometerwas observing the polar cap at 67°S,5°W, with a solar zenith angle o0 of 410, anemission zenith angle 0 of 58°, and a phase angle of 55°. The lower spectrum wasobtained while the spectrometer was observing the desert at 54°S,29°W, with o0 = 510,0 = 500, and P = 55°.

200

spectrum. Many spectrum fluctuationsremain in the reflectance obtained bythis method caused by the shortcom-ings of the available solar spectrumshortward of 2635 angstroms. A bettermethod to determine the spectrum char-acteristics of the polar cap is to dividethe polar cap spectrum by a desert spec-trum obtained at similar lighting angles.The resulting ratio (plotted in Fig. 5)shows a broad absorption feature cen-tered at 2550 angstroms. The primecandidate to explain this absorption isozone, which has an absorption con-tinuum between 2000 and 3000 ang-stroms, centered at 2550 angstroms.Each of the 45 polar cap spectra hasthe spectrum absorption shape expectedfrom ozone. Ozone is likely to be pro-duced in the photochemistry of a carbondioxide atmosphere (10).Where might this absorber be located

to produce the characteristics that ap-pear in the polar cap spectra? First,the observed absorption could be aproperty of the polar cap itself; sec-ond, the absorber could be in the at-mosphere above the polar cap but notover the rest of the planet; and, third,the absorber could exist in the atmo-sphere over the entire planet but onlybe seen when viewed with the brightcap as background. We will consider thethird possibility first: for a planet-wideabsorber to be visible only over thepolar cap, it must have a small scaleheight. Over the desert regions, wherethe ground albedo is very small, the ab-sorber would go undetected because itwould lie beneath the bulk of the at-mospheric scatterers. Over the polarcap, where the ground albedo is high,the absorber would be easily detectablesince the observed radiation would havetraversed the absorbing layer twice.

If the absorber is in the atmosphere,an amount of ozone equivalent to anoptical depth of 0.3 at 2550 angstromsis needed to match the data. This isequivalent to 1 X 10-3 centimeter-at-mosphere of ozone or 3 x 1 016 mole-cules per square centimeter. If thisnumber of ozone molecules is dis-tributed uniformly in the lower at-mosphere, the volume density at thesurface would be 3 X 1010 moleculesper cubic centimeter, or a mixing ratioslightly greater than one part in 107.This amount of ozone falls within therange calculated by carbon dioxide pho-tochemistry (JO). A solution to the mys-tery of why the desert spectra do notshow the ozone absorption has beensuggested by Sagan (11). There may be

SCIENCE, VOL. 173

more particulate aerosols above thedeserts than above the polar caps. Thedust above the desert regions wouldshield the ozone from observation.An ingenious hypothesis suggested by

Broida is that the ozone is trapped inthe solid carbon dioxide of the polarcap and the observed spectrum is thatof the trapped ozone. Laboratory mea-surements show that the reflectance offreshly precipitated solid carbon diox-ide is large and constant with wave-length between 2000 and 3000 ang-stroms. Broida and his collaborators (12)have shown that solid carbon dioxidewith ozone trapped in it shows the ultra-violet absorption that is characteristicof ozone. These same laboratory ex-periments have also shown that ozoneis adsorbed when it is brought in con-tact with solid carbon dioxide. Thisresult suggests that the carbon dioxidepolar cap on Mars serves as a sink fornonvolatile gases, as well as for thevolatiles that condense out at the tem-perature of the cap. Gases such asoxygen, methane, and ammonia, if theyare present in the Mars atmosphere atall, may be adsorbed or trapped on thepolar cap along with the water that willcondense there. When the cap sub-limes, the trapped gases includingozone would be released and would pro-duce a local source in the atmosphere.

Landing Sites

In the fall of 1971, it is anticipatedthat a Mariner spacecraft will orbitMars and, with its complement of scien-tific instruments, will make extensiveobservations over the entire planet. Oneof the instruments is an ultraviolet spec-trometer, which will be able to carryout the topographical mapping overnearly all the desert regions and willalso be able to study the behavior ofozone on the polar cap for an extendedperiod (6). In addition to the intrinsicimportance of these observations, theywill aid directly in plans to select land-ing sites for unmanned spacecraft thatare planned for future years.

There are several reasons why it maybe desirable to land in a region of lowelevation and high pressure (13). If thelanding vehicle uses aerodynamic brak-ing as part of the landing technique, sucha system could be made simpler if itwere designed to be operated where thepressure was 8 millibars rather than 4millibars. If it is considered desirableto land in a region where there is the16 JULY 1971

5-

4-

0 3-

2-

I-

Summary

.4

n4 I I I

2000 2500Wavelength (A)

Fig. 5. Ratio of reflectance of polar capto reflectance of a desert region. Thisratio was obtained by dividing the upperspectrum of Fig. 4 by the middle spec-trum of Fig. 1.

greatest probability of finding liquidwater, then a low elevation is desirable.Water vapor has been measured in theMars atmosphere from ground-basedtelescopes, but, because of the low at-mospheric pressure, liquid water may

not be present on the surface (14).Since one objective of a spacecraft land-ing on Mars may be to determine wheth-er or not liquid water is present, theplausible place to search is at low ele-vations where the atmospheric pressure

is greater than 6 millibars, the pressure

corresponding to the triple point ofwater (14). The amount and form ofthe water retained by Mars are of inter-est in understanding the formation andevolution of the planet. Biological ex-

periments are most likely to meet withearly success if they are performed ata location where water is available inliquid form. The pressure mapping bythe ultraviolet spectrometer experimenton the 1971 Mariner spacecraft shouldhelp locate such places on Mars.

Another criterion in selecting a land-ing site may be the desire to enhancethe probability of detecting minor con-

stituents in the lower atmosphere thatmight be present as the result of bio-logical or geological activity. The in-terpretation of the ultraviolet spec-

trometer data that the polar cap ad-sorbs ozone suggests that the solid car-

bon dioxide cap might be an efficienttrap for other gases as well. A landingsite chosen within the polar cap bound-ary would allow analytical instrumen-tation on a Mars lander to take advan-tage of the concentrating effect of thisadsorption trap. The chemical historyof the previous season may be foundthere. During the 1971 Mariner exper-

iments, the idea that the polar cap trapsminor constituents may be verified andits temporal behavior determined.

Mars, the red planet, reflects sun-light in the ultraviolet, but it is theatmosphere, not the surface, that isresponsible for the reflected light. Eventhough there are atmospheric scatterersin addition to the molecular scatterers,it is possible to relate the intensity ofthe scattered radiation with the atmo-spheric pressure. The variation of pres-sure over the planet reveals the topog-raphy to vary over 7 kilometers inheight and to be correlated with visiblefeatures. The carbon dioxide polar cap,in addition to being a cold trap for vola-tile gases in the atmosphere, may alsobe a very efficient adsorption trap fornonvolatiles. This last property maymake the cap a repository for gases pro-duced by geological or biological ac-tivity (15).

References and Notes

1. C. A. Barth, Appl. Opt. 8, 1295 (1969);, W. G. Fastie, C. W. Hord, J. B.

Pearce, K. K. Kelly, A. I. Stewart, G. E.Thomas, G. P. Anderson, 0. F. Raper, Sci-ence 165, 1004 (1969); C. A. Barth, C. W.Hord, J. B. Pearce, K. K. Kelly, G. P. An-derson, A. I. Stewart, J. Geophys. Res. 76,2213 (1971); J. B. Pearce, K. A. Gause, E.F. Mackey, K. K. Kelly, W. G. Fastie, C. A.Barth, Appl. Opt. 10, 805 (1971).

2. A. Kliore, G. Fjeldbo, B. L. Seidel, S. I.Rasool, Science 166, 1393 (1969); S. I. Rasool,J. S. Hogan, R. W. Stewart, L. H. Russell,J. Atmos. Sci. 27, 841 (1970).

3. N. Wilson, R. Tousey, J. D. Purcell, F. S.Johnson, C. E. Moore, Astrophys. J. 119, 590(1954); R. Tousey, Space Sci. Rev. 2, 3 (1963).

4. R. B. Leighton, N. H. Horowitz, B. C. Mur-ray, R. P. Sharp, A. H. Herriman, A. T.Yoimg, B. A. Smith, M. E. Davies, C. B.Leovy, Science 166, 49 (1969); C. B. Leovy,B. A. Smith, A. T. Young, R. B. Leighton,J. Geophys. Res. 76, 297 (1971).

5. K. C. Herr and G. C. Pimentel, Science 167,47 (1970).

6. C. W. Hord, C. A. Barth, J. B. Pearce,Icarus 12, 63 (1970).

7. J. B. Pollack and C. Sagan, Space Sci. Rev.9, 243 (1969).

8. G. H. Pettengill, C. C. Counselman, L. P.Rainville, I. I. Shapiro, Astron. J. 74, 461(1969); M. J. S. Belton and D. M. Hunten,Science 166, 225 (1969).

9. K. C. Herr, D. Horn, J. M. McAfee, G. C.Pimentel, Astron. J. 75, 883 (1970). The in-terpretation that the ultraviolet spectrometerwas actually measuring topography received astrong impetus from the early publication ofthe topography results of the infrared spec-trometer. When a comparison is made be-tween the detailed results of the two experi-ments, the differences as well as the agree-ments between the two techniques shouldreveal still more information about the Marslower atmosphere and surface.

10. M. B. McElroy and D. M. Hunten, J. Geo-phys. Res. 75, 1188 (1970); D. M. Huntenand M. B. McElroy, ibid., p. 5989.

11. C. Sagan, personal communication.12. H. P. Broida, 0. R. Lundell, H. I. Schiff,

R. D. Ketcheson, Science 170, 1402 (1970).13. C. Sagan and J. B. Pollack, J. Geophys. Res.

73, 1373 (1968).14. T. Owen and H. P. Mason, Science 165, 893

(1969); A. P. Ingersou, ibid. 168, 972 (1970);E. S. Barker, R. A. Schorn, A. Woszczyk,A. G. Tull, S. J. Little, ibid. 170, 1308 (1970).

15. For some years now, the Stanford Univer-sity biologist Joshua Lederberg has advocateda polar landing site for Viking.

16. This research was sponsored by the Na-tional Aeronautics and Space Administration.

201