Andesites on Mars? Dianne Taylor Winter, 2004 “The Mars Debates”

1. Andesites on Mars? Dianne Taylor Winter, 2004 “The Mars Debates” Mars Science A Surveyor's Chronicles.htm

2. Andesites on Mars? I think not. Dianne Taylor Winter, 2004 “The Mars Debates” Weathered basalt Mars Science A Surveyor's Chronicles.htm

3. What is TES?Thermal Emission Spectroscopy • The TES Instrument has 3 parts: • Michelson Interferometer/spectrometer • Bolometric (broadband) thermal radiance (4.5 to ~100µm) channel • Solar reflectance channel (0.3 to 2.7 µm ) – measures brightness of reflected solar energy. • Covers the 6 to 50 µm (1655 – 200 cm-1) wavelength range • Spectral Resolution: 5 – 10 cm-1 • Spatial Resolution: 3 km from MGS orbit • Entered Mars orbit on board the Mars Global Surveyor (MGS) on 9/11/1997. • Collected 4.8 x 106 spectra of Mars over 510 days during a series of 1284 aerobraking orbits prior to 350 km mapping orbit. • PI: Philip R. Christensen (ASU, grad student at UCLA)

4. MGS - TES • Six sensors in 3 x 2 array • Spatial resolution of 3 km from MGS orbit • About the size of a small microwave • Weighs 32 lbs. • Powered by solar sails on spacecraft – uses 14.5 watts! • Also designed to study composition of atmosphere, atmospheric temp. field and polar energy balance.

5. How does TES determine surface composition? Step 1 - Apply atmospheric correction to raw spectra • Bandfield et al. 2000: • Estimated the spectral shape of the dust and water-ice cloud components. • Spectral shapes were shown to be constant in space and time over a wide range of atmospheric conditions (dust loading and ice content). • Dust spectrum found to contain little, if any, surface spectral characteristic. • Smith et al. 2000 • Developed two different algorithms for surface-atmosphere separation. • Both algorithms gave very similar results. • Derived a suite of surface-only spectra for classic dark region of Cimmeria Terra • The low albedo Dark Regions were looked at first since they are believed to be relatively free of surface dust coating. • Tested the techniques using TES data collected under two very different atmospheric temperature and opacity conditions. Bandfield, J.L., Christensen P.R., Smith, M.D. (2000) Spectral data set factor analysis and end-member recovery: Application to analysis of Martian atmospheric particulates. J. Geophys. Research, 104, 9573-9588. Smith, M.D., Bandfield, J.L., Christensen, P.R. (2000) Separation of atmospheric and surface spectral features in Mars Global Surveyor Thermal Emission Spectrometer (TES) spectra. J. Geophys. Research, 104, 9589-9607.

6. Atmospheric correction Models Two atmospheric correction models: • Radiative transfer model • Deconvolution model Comparison of surface emissivity spectral shapes Comparison of dust and water ice spectral shapes obtained by both models. The agreement is very good. Difference in dust spectral shapes are caused by CO2 hot bands at 1075 cm-1. Smith et al (2000)

7. Radiance and Emissivity 70 F • A blackbody emits energy with a Planck distribution • Vibrating molecules emit and absorb infrared light - depends on bond energies. • Unique spectrum for a given mineral. • Infrared energy measured in terms of radiance – Watts per unit of area • The radiance from a mineral at one temperature will be different from that at another temperature Quartz (From ASU Thermal Emission Spectroscopy website)

8. Radiance and Emissivity, cont. • Need to remove the temperature effect. • Divide the radiance spectrum of selective emitter by blackbody (perfect emitter) at same temperature. • Result: Emissivity spectrum (dimensionless) • Values <1 are wavelengths where molecules absorb energy • For quartz, SiO2 molecules are responsible for absorption.

9. Mixed Spectra • Rocks are a mixture of minerals • Emissivity spectrum from individual components of a mixture add together in a simple linear fashion. • The linearity of the mixed spectrum allows it to be deconvolved.

10. Step 2: Deconvolution of Mixed Spectra Thermal Emission Spectra of Feldspars* • Assemble a spectral library of mineral end members (ASU has an online spectral library) • Write a deconvolution algorithm which adds and subtracts endmembers and calculates the best fit (rms error) • Depth of absorption feature is directly related to abundance (or smoothness – see later) • NB – the mineral mix you come up with depends on the end members you start with! • This is from a suite of feldspars with grain size of 710-1000 μm • Vertical lines call attention to various spectral features which change with composition. *From ASU Thermal Infrared Mineral Spectroscopy Library

11. Two distinct surface types found on Mars The spectral fits were modeled by two groups: Bandfield et al. (2000): used 45 end-members representing igneous, sedimentary and metamorphic minerals (rms 0.00018, 0.0009) Hamilton et al. (2001): used a narrower range of 29 mineral spectra of unweathered basalts and andesites (rms 0.0026, 0.0014) (rms = average error over the entire spectrum) Bandfield, J.L., Hamilton, V.E., Christensen P.R. (2000) A global view of Martian surface composition from NGS-TES. Science 287, 1626-1630. Hamilton V.E., Wyatt M.B., McSween H.Y., Christensen P.R. (2001) Analysis of terrestrial and Martian volcanic compositions using thermal emission spectroscopy 2. Application to Martian surface spectra from the Mars global surveyor thermal emission spectrometer. J. Geophys. Research, 106, 14,733-14,746..

12. Two distinct surface types found on Mars Type 1 - Similar to Basalt Type 2 – Andesite? ( < 52 wt% SiO2) ( 52-63 wt% SiO2) Mostly in southern highlands Mostly in northern lowlands (similar to Pathfinder results) (note the larger percentage of high silica glass is the main diff.)

13. A note on Methods TES data cover 1650-200 cm-1, spectral fitting constrained to 1,280-400 cm-1 CO2 exclusion region (Bandfield et al 2000) 540-800 cm-1 Atmosphere correction excludes high wave number region because of numerous water vapor and minor CO2 features 400-200 cm-1 range excluded due to residual atmospheric water vapour rotational bands in Martian surface spectra and ASU mineral library spectra. Range is restricted so that algorithm won’t attempt to fit water vapor features instead of surface mineralogy.

14. MGS TES Basalt Map Basalt (Type 1 spectra) concentrated in Southern Highlands

15. MGS TES Andesite Map Andesite (type 2 spectra) appears concentrated in Northern Lowlands, but also intermixed with basalt in Southern Highlands.

16. Mars Andesite

17. Mars Basalt

18. Mars Hematite

19. Ways to form Andesite (on Earth) • Converging plate margins (tectonics!): Water in descending oceanic crust promotes melting of mantle wedge above it --- or ---- • Fractional crystallization of basaltic magma (settling out of crystals with less SiO2). • Depending on starting comp., 60%-90% of original basalt magma must crystallize to give andesite. • The andesite should be intimately associated with the basalt, which it is not in the northern hemisphere.

20. Compare to possible “global ocean” Wyatt and McSween (2002, Nature) noticed:. The surface interpreted as andesite (red, Surface Type 2) is concentrated in the northern hemisphere in a large, low region previously interpreted as an ancient ocean on Mars. The white line outlines the location of the possible shoreline. This poses a new question – maybe aqueous alteration?

21. Redo deconvolution without high-silica glass phase. Surface Type 1 Wyatt and McSween (2002, Nature): • Without using any silica glass, they still produced a good fit for basalt for Surface Type 1 • Rms fits • Bandfield et al.: 0.0018 • Hamilton et al.: 0.0026 • Wyatt and McSween: 0.0018 “The low r.m.s. values and similar modeled mineral abundances derived from different end member sets indicate these modes accurately reflect the Surface Type 1 composition.” Wyatt, M.B. and McSween, H.Y. (2002) Spectral evidence for weathered basalt as an alternative to andesite in the northern lowlands of Mars, Nature, 417, 263-266.

22. Redo deconvolution without high-silica glass phase. Surface Type 2 Wyatt and McSween (2002, Nature): • For Surface Type 2, clays replaced high-silica glass in the modeled compositions. • This can be interpreted as weathered basalt instead of andesite. • Rms fits: • Bandfield et al.: 0.0009 • Hamilton et al.: 0.0023 • Wyatt and McSween: 0.0014 Wyatt, M.B., McSween, H.Y. (2002) Spectral evidence for weathered basalt as an alternative to andesite in the northern lowlands of Mars, Nature, 417, 263-266.

23. Spectra of clays vs. high silica glass • The main absorption feature to distinguish clays from high-silica glass is in 500-550 cm-1 wavenumber range. • But, CO2 atmosphere of Mars is opaque in this region. • Otherwise they are similar in overall shape and positions of spectral features. • Clays are Fe-smectite and Ca-montmorillonite.

24. Compare Mars to Columbia River Basalt (CRB) • Wyatt and McSween obtained spectra of fresh-cut and weathered surfaces of CRB. Light colored fresh basalt Darker weathered basalt

25. Compare Mars to Columbia River Basalt (CRB) • Added a blackbody component to account for band-depth particle-size effects. (TES = sand sized particles) • When deconvolved weathered CRB with Bandfield et al. and Hamilton et al. mineral sets, got high plag, pyroxene and silica glass (which is not present in CRB) • Using clays, W&McS modeled plag + alteration minerals + lesser pyroxene (best agreement to actual composition)

26. Implications of Wyatt and McSween modeled mineral content • Maybe there is high-SiO2 glass…. But this could also represent an amorphous high-silica alteration product spectrally similar to high silica glass – not Andesite.

27. Implications of Wyatt and McSween modeled mineral content • Maybe there is high-SiO2 glass…. But this could also represent an amorphous high-silica alteration product spectrally similar to high silica glass – not Andesite. • Or, maybe there are alteration products present (Clays)

28. Implications of Wyatt and McSween modeled mineral content • Maybe there is high-SiO2 glass…. But this could also represent an amorphous high-silica alteration product spectrally similar to high silica glass – not Andesite. • Or, maybe there are alteration products present (Clays) • But, in any case --- the Martian northern lowland plains materials are basalts weathered under submarine conditions and/or sediments derived from weathered basalts and deposited in the northern basin.

29. Infrared stealthy surfaces: Why TES and THEMIS may miss substantial mineral deposits on Mars – Kirkland et al., 2003 • Optically rough surfaces can remain undetected (including regional mineral deposits and rock outcrops at 100% exposure) • Rough materials are called “Infrared Stealthy” • Claimed TES mineral results are based on washed, pure minerals, large samples and smooth surfaces. Kirkland L.E., Herr K.C., Adams P.M. (2003) Infrared stealthy surfaces: Why TES and THEMIS may miss some substantial mineral deposits on Mars and implications for remote sensing of planetary surfaces, J. Geophys. Research, 108, Dec 2003, pp., 11-1. (Themis – Thermal Emission Imaging System on Mars 2001 Odyssey Orbiter, also ancient Greek goddess of justice)

30. Spectral contrast reduction – caused by texture • Comparing spectral contrast variation of: • Calcite hand sample • Calcrete hand sample (intensely lithified) • Airborne spectrum of region covered with calcrete boulders • Field spectrometer of same target region. • Close-up of 11 m region • So, texture can impact the spectral signature of all geological classes.

31. Three cases of “spectral misbehavior” impact Mars mineral interpretations Case 1: Minerals that are actually there are undetected. Case 2: Abundance variations and surface texture variations have same effect on spectrum. Increased smoothness looks same as increased abundance. Case 3: Physical effects actually alter spectral band shape. So, roughness at the grain scale can determine whether or not a mineral is detectable.

32. Future work • Gamma ray spectrometer on board Mars Odyssey spacecraft can measure actual elemental abundances (instead of controversial modal mineralogy) • If type 1 and 2 differ appreciably in Si, then it is more likely that they are basalt and andesite. • “Mini-TES” results from Spirit and Opportunity. • 2008 mission to northern plains of Mars. • Need more modeling to distinguish between clays and high-silica alteration coatings on basalt which are spectrally similar to andesite. • Develop theoretical models describing how global weathering could occur on Mars which is consistent with the small amount of carbonates that are found.

33. SNC Meteorites and TES • Another problem with the TES spectra - does not find any area on Mars that matches the SNC meteorites - calling into question how good a job it is doing identifying the proper minerals • But since they seem to have been ejected by just a few giant impacts, maybe they just happen to come from a few unusual areas that haven’t been imaged, or that do not cover a large enough area to be represented by TES. - Could also be from deeper in the crust. ALH 840001 is thought to be deep.

34. Pathfinder and Andesite • Did Pathfinder (1997) really find andesite at its landing site, i.e. that TES “confirmed” the finding of andesite on Mars? • A design difference between flight APXS and lab APXS resulted in 14% greater alpha intensity in flight instrument. • Preliminary results overestimated the light elements and underestimated the heavy elements. • Also had inconsistencies in the reference library and mixed reference library with polished and rough surfaces. • This calls into question the original interpretation that “Barnacle Bill” was andesitic.

35. References Bandfield, J.L., Hamilton, V.E., Christensen P.R. (2000) A global view of Martian surface composition from NGS-TES. Science 287, 1626-1630. Bandfield, J.L., Christensen P.R., Smith, M.D. (2000) Spectral data set factor analysis and end-member recovery: Application to analysis of Martian atmospheric particulates. J. Geophys. Research, 104, 9573-9588 Christensen P.R., Bandfield, J.L., Smith, M.D., Hamilton V.E., Clark R.N.. (2000) Identification of a basaltic component on the Martian surface from thermal emission spectrometer data. J. Geophys. Research, 105, 9609-9621. Hamilton V.E., Wyatt M.B., McSween H.Y., Christensen P.R. (2001) Analysis of terrestrial and Martian volcanic compositions using thermal emission spectroscopy 2. Application to Martian surface spectra from the Mars global surveyor thermal emission spectrometer. J. Geophys. Research, 106, 14,733-14,746.. Kirkland L.E., Herr K.C., Adams P.M. (2003) Infrared stealthy surfaces: Why TES and THEMIS may miss some substantial mineral deposits on Mars and implications for remote sensing of planetary surfaces, J. Geophys. Research, 108, Dec 2003, pp., 11-1. Minitti M.E., Rutherford M.J., Weitz C.M. (2001) Spectra of Martian andesitic materials, LPSC XXXII, Abstract 1976. Ruff S.W., (2003) Basaltic andesite or weathered basalt: A new assessment, Sixth International Conference on Mars (abstract #3258) Smith, M.D., Bandfield, J.L., Christensen, P.R. (2000) Separation of atmospheric and surface spectral features in Mars Global Surveyor Thermal Emission Spectrometer (TES) spectra. J. Geophys. Research, 104, 9589-9607 Wyatt, M.B., McSween, H.Y. (2001) An alternative hypothesis for basalt and andesite on mars: Global surface compositions from MGS-TES. 64th Annual Meteoritical Society Meeting, Abstract 5392. Wyatt, M.B., McSween, H.Y. (2002) Spectral evidence for weathered basalt as an alternative to andesite in the northern lowlands of Mars, Nature, 417, 263-266.

36. What is a Michelson Interferometer? D = mλ/2 The Michelson interferometer produces interference fringes by splitting a beam of monochromatic light so that one beam strikes a fixed mirror and the other a movable mirror. When the reflected beams are brought back together, an interference pattern results.

37. Why the dust spectrum does not contain surface spectral characteristics • Close similarity between derived dust spectrum and spectra acquired with high dust opacity and/or high emission angles. • Good agreement between the derived atmospheric dust spectrum and TES limb spectra that view only the atmosphere • Consistent surface spectra derived using the dust spectrum for atmospheric dust opacities that varied by a factor of 5-10 • Consistent atmospheric dust spectra in pairs of day/night observations in which the dust is alternatively viewed in transmission and emission.