The chemical composition and evolution of the Martian upper crust and near surface environment
v Table of Contents
List of Figures vii List of Tables ix Acknowledgements x
Chapter 1: Introduction 1 Overview of Chapter 2 6 Overview of Chapter 3 12 Overview of Chapter 4 17 Overview of Chapter 5 21
Chapter 2: Sediments and the Chemical Composition of the Martian Upper Crust
33 Introduction 33 Sediments as proxy for bulk upper crust 39 Selection and refinement of chemical averages and upper crustal chemistry estimates 44 The bulk chemical composition of the Martian upper crust 49 Homogeneous Martian sediments vs. diverse terrestrial compositions 57 Upper Crust vs. Total Crust 58 Summary 60
vi Chapter 3: Mars Odyssey Gamma-Ray Spectrometer Elemental Abundances and Apparent Relative Surface Age: Implications for Martian Crustal Evolution 74 Introduction 74 Methods 77 Results 88 Discussion 92 Summary 107
Chapter 4: Martian Surface Heat Production and Crustal Heat Flow from Mars Odyssey Gamma-Ray Spectrometry 116 Introduction 116 Surface Heat Production 118 Crustal Heat Flow 119 Discussion 123
Chapter 5: Regional Martian Crustal Heat Flow from Mars Odyssey Gamma-Ray Spectrometry 132 Introduction 132 Heat flow variations 140 Regional crustal heat flow for selected geologic provinces and features 146 Discussion and Conclusions 154
vii List of Figures
Chapter 1 Figure 1 4 Figure 2 9 Figure 3 10 Figure 4 14 Figure 5 16 Figure 6 20 Figure 7 22
Chapter 2 Figure 1 46 Figure 2 51 Figure 3 56
Chapter 3 Figure 1 79 Figure 2 83 Figure 3 90 Figure 4 91 Figure 5 93 Figure 6 94 Figure 7 96 Figure 8 101 Figure 9 105
viii Chapter 4 Figure 1 120 Figure 2 121 Figure 3 122 Figure 4 126
Chapter 5 Figure 1 134 Figure 2 142 Figure 3 145 Figure 4 155
ix List of Tables
Chapter 1 Table 1 25
Chapter 2 Table 1 48 Table 2 50
Supplementary Table 1 71
Chapter 3 Table 1 80
Chapter 5 Table 1 137 Table 2 141
First and foremost, I would like to thank my advisor during this dissertation work, Scott McLennan…and there is much for which I owe thanks. Scott has been a superlative advisor in all matters scientific and I always value his insights. He has helped teach me to think like a scientist. He has also provided me the opportunity to work on both the Mars Exploration Rovers and the Mars Odyssey GRS science teams – a great honor, and the fulfillment of the dream of working on a space mission. I am deeply indebted to Scott and know I will work with him for many years to come. He has also been a great friend, and I wish him and his family the best. I thank my dissertation committee for their time and effort in reviewing the work presented here. If anyone is actually reading these acknowledgements, it means I passed…so I’ve got that going for me…which is nice. Thank you: Dan Davis, Lianxing Wen, Tim Glotch, and Brad Jolliff. A special thanks to Brad for travelling a distance from St. Louis for my defense. Studying under Scott has been rewarding in itself, but I’ve also been very fortunate to work with the wonderful people that have made up our research group over the past several years. Special thanks must go to Nick Tosca and Joel Hurowitz, who have been great friends for the last several years. They have made the stresses of graduate school much less of a burden. I expect (and hope) we shall be working together on various projects in the future. I’d also like to thank the other members of our lab group
through the years: Shannon Arlaukas; Scott Perl; Lauren Beavon; Yuyan Zhao; and Suniti Karunatillake. And, while not an official member of our group, I extend a special thanks to Alex Smirnov, too. Best wishes to all. Additionally, I must thank Steve Squyres and William Boynton, the leaders of the MER Athena Science Team and Odyssey GRS Science Team, respectively. Participation in these missions has been a wonderful and gratifying experience. I thank all the faculty, staff, and my fellow students of the Department of Geosciences at Stony Brook University for support and resources. Special thanks to: Owen Evans, for keeping the ship running smoothly and silently; Loretta Budd, who I hope enjoys her well-deserved retirement; and Diane Isgro and Yvonne Barbour, for much appreciated help. I thank Zantac ® , Tagamet ® and Prilosec ® for support through stressful times. I also must extend thanks to several people and collaborators. Jeff Taylor has been one of the more interesting people I’ve come to know and work with; he’s a great scientist, but I suspect he’d make an even better late night talk show host. Ross Taylor has been a source of wisdom, insight, and good humor. Also, special thanks to Hap McSween and Jeff Moersch for providing a place of employment after graduation. Last, but certainly not least, my family has been a great source of encouragement and support over the past several years while I have worked to complete this dissertation. Thank you, Charles and Bridget Hahn and Jeanette Hahn for your love and support.
1 Chapter 1: Introduction
The Martian crust is an important geochemical reservoir. It is the only portion of the planet surveyed by broad-scale remote sensing observations or in situ lander and rover experiments. Mars experienced a unique mechanism for crustal emplacement, resurfacing, and recycling [Nimmo and Tanaka, 2005] – different from the broadly continuous plate tectonic regime of Earth or the highly episodic, widespread basalt flooding of Venus. The various Martian crustal age provinces are the only geologic records available that preserve evidence of the secular chemical variations produced by the planet’s development and evolution [Hahn et al., 2007]. If Mars was ever a haven for life, it would have survived on or in the upper crust, and any evidence of its existence would be preserved in the geologic record [Knoll et al., 2005]. Accordingly, determining the bulk chemical composition of the Martian crust, constraining the crustal thermal regime, and understanding their relationships to the evolution of Mars are valuable and necessary scientific endeavors. Further, it is well established that crusts of the inner terrestrial planets as well as many dwarf planets and large rocky satellites become substantially enriched in a suite of geochemically important elements during early differentiation and further evolution [Taylor and McLennan, 20009; Taylor, 2001]. Therefore, all planetary crusts have an enhanced chemical significance disproportional to their small relative volume with
2 respect to the parent planet. The Martian crust is especially significant in this respect since, compared to most other inner planets, the Martian crust is volumetrically larger compared to Mars as a whole (>4%) and is the primary repository for many incompatible trace elements, including the geochemically important heat producing elements (K, Th, and U). The Martian crust therefore serves as an important constraint on the chemical composition of the entire planet and on models of planetary differentiation and crust- mantle evolution. Of the data used for the studies collected in this dissertation, of particular importance are the in situ lander measurements of soil and rock chemistry (especially the Spirit and Opportunity rovers) and orbital gamma-ray spectrometer measurements of regional and global elemental abundances. Several successful missions, both orbital remote-sensing spacecraft and surface landers, have returned a great deal of information about Martian crustal chemistry. Several landed missions have returned geographically disparate measurements of surface chemistry. However, the two Viking missions consisted of stationary landers and the Sojourner rover used by the Pathfinder mission had limited mobility and operating lifetime [Soffen, 1977; Golombek et al., 1999]. The successful – and surprisingly long-lived – Mars Exploration Rovers (MER), Spirit and Opportunity, provided many more detailed chemical and mineralogical analyses of soils and rocks from geographically diverse locales on opposite sides of the planet [Squyres et al., 2004a,b] and significantly expanded the catalog of chemical analyses. Apart from several notable exceptions, in situ lander measurements of Martian sediments show very little chemical variability across landing sites. Aside from discrete, local components that influence proximate sediment chemistries, most soils and dust analyses are chemically
3 indistinguishable. For example, Figure 1 plots average elemental abundances for the soils of Meridiani Planum measured by the Opportunity rover, versus average soil abundances for Gusev Crater as measured by the Spirit rover on a log-log plot. The diagonal line indicates a 1:1 equivalence between landing sites, and most elements plot very close to or on this line – indicating near-identical chemistries. In orbit, the Gamma- Ray Spectrometer (GRS) instrument suite on board the Mars Odyssey spacecraft has mapped the surface distribution of several elemental abundances – the first robust determination of global surface chemistry [Boynton et al., 2007]. Chemical abundances for the suite of GRS detectable elements for the specific pixels that incorporate the Spirit and Opportunity landing sites are also plotted on Figure 1 and agree well with lander values. The Alpha-Particle X-ray Spectrometers (APXS) aboard the MER vehicles are the primary instruments for determining elemental compositions [Gellert et al., 2006; Rieder et al., 2004]. In conjunction with the Rock Abrasion Tool (RAT) (which can brush away dust coatings or grind away alteration rinds), chemical analyses have been performed on undisturbed rock surfaces, brushed rock surfaces, rock interiors, trenched subsurfaces, as well as disturbed and undisturbed soils. For a given target, the APXS provides reliable analysis for all the major elements and several trace elements (notably: Zn, Cr, Br, and Ni). However, even with MER’s enhanced mobility, landed missions have only explored a miniscule fraction of the planet’s surface in any detail. Although limited to a narrower suite of elements than surface in situ observations (H, Fe, Si, Cl, K, Th, Al, and Ca), the GRS can reliably analyze most of the planet from
4 Figure 1. Log-Log plot of average soil compositions measured by the Opportunity rover at Meridiani Planum versus the average soil composition measured by the Spirit rover at Gusev Crater [Gellert et al., 2006; Rieder et al., 2004]. The diagonal line represents a 1:1 correlation of equivalent abundances. Generally, Martian soils are chemically indistinguishable across multiple landing sites. Also plotted, the GRS abundances for a suite of measured elements for the GRS pixels that cover the 2 landing sites. GRS abundances agree well with measured landing site averages.
5 orbit (high abundances of surface or sub-surface polar ice limit resolution at higher latitudes) [Boynton et al., 2007]. Penetration depths are mostly dependant on the density of the underlying matrix and the elements measured, but are on the order of tens of centimeters. This deeper penetration depth compared to most remote sensing techniques allows the GRS to be less sensitive to thin coatings of air-fall dust that obscure most other remote sensing data (e.g., TES, THEMIS, and OMEGA) [Christensen et al., 2001, 2003; Bibring et al., 2005]. However, the Odyssey GRS has an inherently low spatial resolving capacity with a large footprint and thus cannot be used effectively for fine-scale regional to local analysis. Additionally, element abundance maps are subject to considerable smoothing and auto-correlation effects due to element correction factors, the large footprint and necessary data processing. This too, limits GRS analysis to areally-large regional or global studies. Using these data sources, the research described in this dissertation was undertaken in order to further understanding of the elemental composition of the upper Martian crust, its evolution through time, and the thermal behavior of the crust as an important reservoir of the incompatible heat producing elements. The results of this research are described in Chapters 2-5 of the dissertation. A brief summary of the major goals and findings for each chapter is provided below.
6 Overview of Chapter 2: The purpose of this study is to use the MER APXS soil composition data and orbital GRS Odyssey global abundance maps to constrain the bulk chemical composition of the upper Martian crust. For Earth, many past studies have used sediment compositions to help constrain estimates of the composition the terrestrial upper continental crust. Weathering, sedimentary transport, and deposition naturally sample a wide array of source rocks with the resultant chemistry being an efficient mixture of source terrains [Taylor and McLennan, 1985, 1995, 2009; Condie, 1993; Plank and Langmuir, 1998; McLennan, 2001; Rudnick and Gao, 2003; McLennan et al., 2006]. Therefore, knowledge of the total sedimentary rock budget (i.e., shales, carbonates, sandstones, and evaporites) provides an excellent proxy for the average bulk chemistry of the upper continental crust from which it is derived. On average, the bulk sedimentary mass appears to be slightly more mafic than the terrestrial upper crust as a whole, but this is probably due to recycling of ancient more mafic crust into the sedimentary record [McLennan et al., 2006; Ronov, 1983; Veizer, 1979; Veizer and Mackenzie, 2003]. However, on the Earth, there is significant partitioning of the major elements among distinct sedimentary lithologies through various aqueous processes. Accordingly, sediment chemistry alone is not the most reliable means of determining bulk major element chemistry of the upper continental crust and major element abundances are thus determined by using weighted averages of major rock provinces. Averages of sedimentary rock chemistry do provide a good proxy for the relatively insoluble trace element abundances found in the continental crust (e.g., REE, Th, Sc). By volume, marine shales dominate the sedimentary rock budget for Earth, and shales also tend to have high abundances of most trace elements. As such, chemical averages of sedimentary
7 products are dominated by an average shale chemistry that is broadly similar across the planet. Thus, the insoluble trace element abundances derived from an average of shale chemistry (with slight corrections for ‘dilution’ effects) best represents that of the upper terrestrial continental crust. It is not yet clear whether or not Mars has a significant sedimentary component in the form of shale. Although phylosilicates have been observed by orbital remote sensing in Noachian layered sequences, the basic lithology of such deposits is not known [Bibring and Langevin, 2008; Bishop et al., 2008]. Sandstones have been observed and studied, such as those found at Meridiani Planum in the Burns Formation and in layered sequences in the Columbia Hills [Grotzinger et al., 2005; McLennan et al., 2005; Arvidson et al., 2008; Lewis et al., 2008]. Additionally, assortments of evaporites, clays and, most recently, carbonates have been detected in varying amounts at various locations around the planet through remote sensing and in situ rover observations. However, the dominant sedimentary products that have been recognized and characterized to date are the Martian dust and basaltic loose regolith or soils. Although different chemically and mineralogically, and resultant from different weathering processes, like terrestrial shales, Martian soils are derived from a large array of source materials from geographically large regions and thus chemical averages of these soils should represent a good proxy for the upper crust provinces from which they are produced. The major conclusions drawn from this study are: • Most Martian soils are chemically similar representing an average “global” soil composition. Soils that show specific, chemically distinct signatures are explained
8 through physical mixing between the average global sediment and local, chemically distinct lithological sources (Figure 2), such as the Fe-enriched hematite concretions at Meridiani Planum or the P-enriched Wishstone class outcrops in the Columbia Hills at Gusev Crater. • Like certain terrestrial sediments, Martian soils are derived from large source terrains and are the best chemical proxy for the composition of the upper Martian crust. Unlike terrestrial sediments, which can only provide insoluble trace element compositions of the upper continental crust, Martian soils are produced primarily through physical weathering processes and, more similar to terrestrial glacial sediments, also preserve the major elemental chemical abundances of the source terrains from which they were derived. • This study uses carefully screened sediment chemical averages from all available lander mission data as well as orbital chemical abundances determined by the GRS instrument to derive the most comprehensive estimate for the bulk chemistry of the Martian crust to date. Average Martian crust is basaltic in composition and modestly enriched in the incompatible heat producing and most likely the light rare earth elements. • The diversity of terrestrial sediments results from an environment dominated by aqueous alteration in a water-rich environment and most notably the presence of an ocean which fractionates sediments into diverse chemical compositions. In contrast, the primary sedimentary products of the Martian weathering environment over at least the past 3 billion years are the Martian soils and dust that are relatively
Figure 2. Ternary diagram illustrating average “global” soil composition versus mixtures with select local components. Most soils are chemically similar and represent an average “global” soil composition (circled). Specific, chemically distinct local sources create mixing trends whereas soils proximate to the local source will reflect mixing between the local source and the average composition. The apices are chosen to simply to illustrate chemically distinct locales and do not represent any particular geological process. For example, average soil mixed with end-member Meridiani Planum hematite concretions forms a trend toward the FeO apex [Jolliff et al., 2007]. At Gusev Crater, soils proximate to the P-rich Wishstone outcrop show enrichments in phosphate, and soils located near silica deposits near the Eastern Valley location show Si-enrichment [Hurowitz et al., 2006; Squyres et al., 2008]. Although not well illustrated here, there are also soils in the Eastern Valley characterized by extremely high sulfur content, also likely related to hydrothermal processes that operated in the vicinity [Wang et al., 2008]
10 Figure 3. (a; top) Total alkalis versus silica petrologic diagram of Martian sediments and APXS landing site rock analyses. Calculated average upper crust plots in the basalt field just below the subalkalic fractionation line. Martian soils plot within the cluster of Martian rock measurements and SNC analyses indicating they represent a physical mixture of primary sources without extreme chemical fractionation. (b; bottom) Plot of terrestrial average continental compositions, relative volumes of major terrestrial sedimentary rocks [Taylor and McLennan, 2009], and compositions of those specific sedimentary products with respect to bulk crustal chemistry. Unlike the Martian soils, terrestrial sediments are chemically diverse and represent products of aqueous chemical alteration resulting in considerable chemical fractionation.
12 unaltered, homogeneous and chemically indistinct (Figure 3). This lack of compositional diversity strongly suggests an arid weathering environment with little long-term surface water and with sedimentary processes dominated by physical mixing. Global sedimentary chemistry, derived from either the in situ lander experiments or GRS global elemental mapping appear to provide no support for the presence of a Martian ocean during this time. • Martian sediments only sample the upper crust and, as such, chemical composition estimates may not be applicable to the total Martian crust. However, unlike the terrestrial continental crust, the Martian crust is unlikely to be nearly as chemically heterogeneous vertically and there is no definitive evidence of the geologic processes that lead to widespread continental crustal differentiation on Earth. As such, the estimate for the chemical composition of the upper Martian crust likely also represents a reasonable first-order approximation to the chemical composition of the bulk total crust.
Overview of Chapter 3: This chapter reports the first results of correlations between element abundances determined by GRS and the apparent relative surface age of the Martian upper crust as determined by existing geologic mapping and explores some of the potential implications for better understanding Martian crustal evolution and global scale surficial processes.
13 Secular changes in element abundances, as recorded in age-specific provinces, reveal significant information about planetary crustal formation and evolution. Therefore, relating the age of a planetary crust to its chemical composition is of great geologic importance. On Earth, it is a relatively straightforward task to sample geographically wide-spread crustal materials, perform detailed chemical analyses of major and trace element abundances, and reliably date samples through a variety of field and laboratory methods. The relationship of age to chemistry has been used for decades to better understand the nature and evolution of the continental and oceanic crust [Armstrong, 1981; Taylor and McLennan, 1985], the reservoirs of heat producing elements throughout the crust-mantle system [Taylor, 2001], magmatic resurfacing histories [Jerram and Widdowson, 2005], sedimentary processes and recycling [Veizer and Jansen, 1985; Taylor and McLennan, 1985], plate tectonics [Richter et al., 1992], and a wide array of other important geologic questions. Likewise, lunar samples returned by the Apollo missions, while more limited in sampling diversity, have been well characterized for chemistry and age and consequently, lunar science has made great progress (see reviews in Taylor, 1982, 2001). However, for planetary bodies outside the immediate terrestrial neighborhood, these data are difficult to obtain and highly limited. Currently, the only means of constraining Martian surface age is through the analysis of crater count statistics, cross-cutting relationships, and stratigraphic position. Since the first orbital images of the Martian surface were returned, the USGS and other groups have been reliably mapping the geology of Mars. We have adapted a suite of these maps [Scott and Tanaka, 1986; Greeley and Guest, 1987; Tanaka and Scott, 1987] to create a global apparent relative surface age dataset of the Martian surface for
14 Figure 4. Global map of apparent surface age used in this study. Ages were adapted from the USGS Martian Geologic Investigations Series I-1802 (ABC) [Scott and Tanaka, 1986; Greeley and Guest, 1987; Tanaka and Scott, 1987], and binned into a 1ºx1º resolution grid. Three primary age categories have been assigned (N-Noachian, H-Hesperian, and A-Amazonian); as well as two, less areally extensive transitional age categories (N/H-Noachian-Hesperian and H/A-Hesperian-Amazonian). A gray-scale MOLA topography base-map and labels of prominent geologic features and landing sites have been provided for reference. For comparison to GRS element abundance datasets, we further re-binned age data to a 5ºx5º resolution.
15 comparison to GRS determined elemental abundances (Figure 4). Discussion continues about the uncertainties attached to relative ages derived by these methods; however, while absolute surface ages may not be well-constrained, relative ages between geologic provinces at regional scales are likely quite robust [Dohm et al., 2001]. We examine chemical variations with respect to the three primary Martian age epochs from the formal stratigraphic systems devised by Scott and Carr, 1978: Noachian, Hesperian and Amazonian. The data reported and analyzed here support several important conclusions about the nature of the Martian crust: • K and Th abundances both show significant decreases with younger apparent surface age. Although weathering and broad-scale transport in an acidic environment could cause some of the variation in these elemental abundances, the age relationship suggests an igneous origin. Younger resurfacing magmas were likely derived from a more evolved, depleted mantle source when compared to the older terrains derived from a more primitive mantle. All abundance averages for all age categories are higher than those determined for the SNC meteorites suggesting that, in this respect, the SNCs are not representative of the bulk Martian crust (Figure 5). • Fe abundances show an increase with younger apparent surface age. However, the cause of this increase cannot be uniquely interpreted as igneous, alteration and weathering, or a combination of both. In a generally acidic Martian surface environment, Fe would be far more mobile than in typical terrestrial weathering scenarios. Alternately, a relatively modest change in primary igneous magma composition could lead to the abundance increase seen in the GRS data.
Figure 5. Plots of K and Th abundances with the calculated standard errors of the means represented by the thickness of the lines as well as elemental abundances of the Martian meteorites with reliable published dating and chemical analyses (see compilation in Meyers, 2003). Note that irrespective of the ages of the meteorites, all (save the Los Angeles meteorite with a non-characteristic high Th value) show significantly lower K and Th abundances than the broad-scale averages reported by the GRS instrument (age boundaries from Hartmann, 2005).
17 • Cl abundances increase with younger surface age, possibly the result of volcanic activity in younger provinces and the leaching of Cl from older, weathered terrains. The involvement of Cl in the Martian hydrologic cycle is supported by the close correlation between GRS determined Cl and H abundances. The global Cl averages determined for all age categories are lower than average soil compositions sampled at landing sites. However, given the site selection bias of landed missions and the disparate sampling depths of GRS versus in situ lander observations, higher Cl in individual measured soils is not unexpected. • No statistically or geologically significant correlation was found between elemental abundance and apparent surface age for Si or H, nor was a correlation expected as Si varies little over the planet and H re-distribution occurs at faster timescales.
Overview of Chapter 4: This chapter presents the results of models calculating the crustal component of Martian crustal surface heat production and crustal heat flow as derived from GRS derived K and Th global distributions. Martian thermal state and evolution depend principally on the heat-producing element (HPE) distributions in the planet’s crust and mantle, specifically the incompatible radiogenic isotopes of K, Th, and U. Normally these elements are preferentially sequestered into a planet’s crust during differentiation [Taylor and McLennan, 2009], and this is especially true for Mars, which possesses a thick and mostly ancient crust that is proportionally large with respect to the planet’s total volume. The GRS can detect all three of these elements and has been used to map the K and Th
18 abundances across nearly the entire Martian surface [Boynton et al., 2007]. As the crust is a repository for approximately 50% of the radiogenic elements on Mars, these models provide important, directly measurable constraints on Martian thermal behavior. These results are valuable for better understanding Martian geodynamics, crust-mantle evolution, the cryosphere, formation and history of geologic provinces, and many other varied applications. Our calculations show considerable geographic and temporal variations in crustal heat flow, and demonstrate the existence of anomalous heat flow provinces. Using smoothed GRS global K and Th maps where the data have been binned into 5°x5° pixels, we determined the radiogenic 40 K and 232 Th surface abundances for each GRS pixel based on well-determined isotopic fractions. Currently, 232 Th is 100% of total Th abundance with a heat release constant of 2.64x10 -5 W·kg -1 ; 235 U and 238 U are 0.7204% and 99.2742% of total U abundance with heat release constants of 5.69x10 -4
W·kg -1 and 9.46x10 -5 W·kg -1 , respectively; and 40 K is 0.012% of total K abundance with a heat release constant of 2.92x10 -5 W·kg -1 [Turcotte and Schubert, 2001]. Uranium abundances ( 235 U and 238 U) were calculated using an assumed Th/U ratio of 3.8; a canonical cosmochemical value thought to be representative of most planetary bodies and that also agrees with analyses of most Martian meteorites unaffected by terrestrial weathering processes [Meyers, 2006]. The GRS instrument measures elemental abundances in the top-most tens of centimeters of the Martian surface, and accordingly is strongly influenced by near-surface soils, ice and dust deposits. These sediments broadly represent the bulk chemistry of the Martian upper crust when renormalized to a volatile- free basis [Taylor and McLennan, 2009] and as such, K and Th values must be