sign in sign up
<<
Home , Adirondack (Mars), Aeolis Palus, Bounce Rock, Gale (crater), Gusev (Martian crater), Home Plate (Mars)

Geoscience Faculty Publications Geoscience

5-31-2018

Formation of Evolved Rocks at Crater by Crystal Fractionation and Implications for Crustal Composition

Arya Udry University of Nevada, Las Vegas, [email protected]

Esteban Gazel Cornell University

Harry Y. McSween Jr. University of Tennessee

Follow this and additional works at: https://digitalscholarship.unlv.edu/geo_fac_articles

Part of the Atmospheric Sciences Commons, and the Geology Commons

Repository Citation Udry, A., Gazel, E., McSween, H. Y. (2018). Formation of Evolved Rocks at Gale Crater by Crystal Fractionation and Implications for Mars Crustal Composition. Journal of Geophysical Research, 123(6), 1525-1540. http://dx.doi.org/10.1029/2018JE005602

This Article is protected by copyright and/or related rights. It has been brought to you by Digital Scholarship@UNLV with permission from the rights-holder(s). You are free to use this Article in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s) directly, unless additional rights are indicated by a Creative Commons license in the record and/ or on the work itself.

This Article has been accepted for inclusion in Geoscience Faculty Publications by an authorized administrator of Digital Scholarship@UNLV. For more information, please contact [email protected]. Journal of Geophysical Research:

RESEARCH ARTICLE Formation of Evolved Rocks at Gale Crater by Crystal 10.1029/2018JE005602 Fractionation and Implications for Mars

Key Points: Crustal Composition • Recent in situ data changed our understanding of the crustal Arya Udry1 , Esteban Gazel2 , and Harry Y. McSween Jr.3 • Alkaline and felsic rocks can also be 1Department of Geoscience, University of Nevada, Las Vegas, Las Vegas, NV, USA, 2Department of and Atmospheric reproduced through different degrees 3 of fractional crystallization Sciences, Cornell University, Ithaca, NY, USA, Department of Earth and Planetary Sciences, University of Tennessee, • Continental crust, as defined by large Knoxville, Knoxville, TN, USA amounts of quartz-bearing felsic rocks, is unique to Earth Abstract The recent discovery of some ancient evolved rocks in Gale crater by the rover has

Supporting Information: prompted the hypothesis that continental crust formed in early history. Here we present • Supporting Information S1 petrological modeling that attempts to explain this lithological diversity by magma fractionation. Using the • Figure S1 thermodynamical software MELTS, we model fractional crystallization of different Martian starting • Table S1 • Table S2 compositions that might generate felsic igneous compositions like those analyzed at Gale crater using • Table S3 different variables, such as pressure, oxygen fugacities, and water content. We show that similar chemical and • Table S4 mineralogical compositions observed in Gale crater felsic rocks can readily be obtained through different • Table S5 • Table S6 degrees of fractional crystallization of basaltic compositions measured on the . The results • Table S7 suggest that Gale crater rocks may not represent true liquids as they possibly accumulated and/or • Table S8 fractionated feldspars as well as other phases. In terms of major element compositions and mineralogy, we found • Table S9 that the Gale crater felsic compositions are more similar to fractionated magmas produced in Earth’sintraplate

Correspondence to: volcanoes than to terrestrial felsic continental crust as represented by tonalite-trondhjemite-granodiorite A. Udry and E. Gazel, suites. We conclude that the felsic rocks in Gale crater do not represent continental crust, as it is defined [email protected]; on Earth. [email protected]

1. Introduction Citation: Udry, A., Gazel, E., & McSween, H. Y., Jr. Recent discoveries of diverse felsic (evolved) rock compositions found on the surface of Mars have ques- (2018). Formation of evolved rocks at Gale crater by crystal fractionation and tioned our understanding of the Martian crustal composition. Findings by the implications for Mars crustal (MSL) Curiosity rover at Gale crater, located near the northern edge of the highly cratered highlands composition. Journal of Geophysical (Figure 1) and formed by an impact near the Noachian- boundary at ~3.61 Ga (Le Deit et al., 2013; Research: Planets, 123, 1525–1540. https://doi.org/10.1029/2018JE005602 Thomson et al., 2011), include evidence of ancient diversity in igneous compositions. The large variations in major and minor/trace element compositions likely indicate a wide range of magmatic Received 5 MAR 2018 sources (e.g., Cousin et al., 2017; Grotzinger et al., 2015; Treiman et al., 2016). Most of the igneous rocks ana- Accepted 21 MAY 2018 Accepted article online 31 MAY 2018 lyzed in Gale crater are basaltic (Edwards et al., 2017); However, before (Martian day) 800, felsic, including Published online 22 JUN 2018 alkaline rocks, were encountered (Cousin et al., 2017; Sautter et al., 2015). Felsic and alkaline compositions potentially originated from the Gale crater walls and were transported by the river (Cousin

et al., 2017). These felsic float compositions (up to ~68 wt % SiO2 and ~14 wt % total alkalis) in Gale crater have abundant , K-feldspar, and silica (e.g., Cousin et al., 2017; Edwards et al., 2017; Mangold et al., 2015; Ollila et al., 2014; Payré et al., 2017; Sautter et al., 2014, 2015; Treiman et al., 2016; Vasavada

et al., 2014; Yingst et al., 2013). In addition, alkali enrichments, particularly in K2O, were suggested in igneous rocks and , possibly indicating a metasomatized source (Le Deit et al., 2016; Payré et al., 2017; et al., 2014; Stolper et al., 2013; Treiman et al., 2016), and sedimentary rocks may have had an alkali-feldspar-rich protolith (Siebach et al., 2017). An example of a Gale crater alkaline rock is Jake_M, a float rock that has a basaltic trachyandesitic composition (an evolved alkali-rich silica-undersaturated rock compo- sition), which had not been previously observed on Mars (Stolper et al., 2013). Although subsequently proposed to be a sedimentary rock (Cousin et al., 2017; Grotzinger et al., 2015), Jake_M’s composition suggests that its potential igneous component could have formed by extensive fractional crystallization of a metasomatized alkaline-rich source (Schmidt et al., 2014; Stolper et al., 2013; Udry, Balta, & McSween, 2014).

©2018. American Geophysical Union. Felsic compositions have been observed or inferred prior to the MSL mission in ancient locations outside of All Rights Reserved. Gale crater at the Martian surface and in . Dreibus and Wänke (1985, 1987) first suggested the

UDRY ET AL. 1525 Journal of Geophysical Research: Planets 10.1029/2018JE005602

presence of alkaline magmatism on Mars. Whitaker et al. (2005) and McCubbin et al. (2008) proposed that magmatic diversity was possible on Mars and was due to intraplate magmatism fractional crystallization. Mildly alkaline rocks such as Backstay, Irvine, and Wishstone were observed in the in crater by the rover (McSween et al., 2006). Udry, Balta, and McSween (2014) subsequently showed that alkaline compositions, such as Backstay, can be formed by polybaric fractionation of Gusev crater Fastball-like primary magma. Furthermore, the parental magma of Chassigny (a dunitic ) is alkalic and could be similar to the Gusev crater target Backstay (Filiberto, 2008; Nekvasil et al., 2007, 2009). Figure 1. Location of Gale and Gusev crater spacecraft missions where in situ In addition, the polymict Northwest Africa—NWA—7034 (and rover analyses of igneous crustal rocks have been collected on Mars (image paired meteorites, e.g., NWA 7533; Agee et al., 2013; Bellucci et al., from NASA/GPL/GSFC/Arizona State using Google Mars platform). Major, younger volcanic constructs are also labeled for reference. 2015; Humayun et al., 2013; Yin et al., 2014) revealed clasts with compo- sitions not previously observed among Martian meteorites. Note that many of the igneous components of the NWA 7034 formed at about 4.4 Ga (Humayun et al., 2013; McCubbin, Boyce, Novák-Szabó, et al., 2016; Nyquist et al., 2016), but the breccia itself is thought to have been lithified during a metamorphic event during the (Bellucci et al., 2015; Cartwright et al., 2014; McCubbin, Boyce, Novák-Szabó, et al., 2016). The trachyandesite and trachyte clasts in NWA 7034 are similar to those of evolved rocks found in Gusev and Gale crater (Agee et al., 2013; Cartwright et al., 2014; Gattacceca et al., 2014; Humayun et al., 2013; Nemchin et al., 2014; Santos et al., 2015; Udry, Lunning, et al.,

2014). Gale crater rocks show higher K2OandNa2O than trachyandesite and trachyte clasts in NWA 7034. The intrusive NWA 6963 meteorite is the only shergottite having small pockets of granitic composition (Filiberto et al., 2014, 2018; Gross & Filiberto, 2014). These authors suggested that felsic compositions likely formed through fractional crystallization of the parental magma to NWA 6963. Christensen et al. (2005) also suggested felsic (dacitic) compositions in the Syrtis Major caldera from thermal spectra, and orbital near-infrared observations suggest that anorthosites and/or granitic compositions containing abundant plagi- oclase and/or quartz and K-feldspar may exist at the Martian surface (Carter & Poulet, 2013; Wray et al., 2013). However, new thermal infrared measurements indicate that these lithologies might also simply be feldspar-rich (Rogers & Nekvasil, 2015). The global abundance of silica-rich rocks on Mars is limited by the absence of mappable silica-rich units anywhere on Mars from orbital Gamma Ray Spectrometer data (Boynton et al., 2007). The new MSL discoveries of felsic compositions in Gale crater, especially during the first 20 months of the mission (Cousin et al., 2017), led Sautter et al. (2015) to propose that continental crust is present on Mars and may resemble the Archean tonalite-trondhjemite-granodiorite (TTG, the felsic building blocks of continental crust on Earth) suites. This poses a unique to better understand the formation of felsic rocks in planetary evolution. Here we present thermodynamic models that explore the lithological diversity of Mars by magma fractiona- tion processes using different variables. Note that fractional crystallization was already explored for the for- mation of felsic lithologies (e.g., Edwards et al., 2017; Sautter et al., 2015; Udry, Balta, & McSween, 2014). However, in this study, we investigated the role of different conditions (water content, oxygen fugacity—

ƒO2, and pressure) using various starting compositions in the formation of rock from various groups in Gale crater defined by Sautter et al. (2015) and Cousin et al. (2017) in order to test whether these magmas can be produced by simple fractional crystallization processes or represent early continental crust generation as defined on Earth. For this, we offer a direct statistical comparison of Earth’s Archean continental crust and intraplate volcanoes, with a range of basaltic to felsic compositions from Martian meteorites and missions, with the goal of understanding the generation of felsic magmas on early Mars.

2. Data and Methods 2.1. Martian and Terrestrial Compositions

In this manuscript, we refer to felsic rocks as those containing >55% wt % SiO2 regardless of alkali content. In addition, when mentioned, we define alkaline rocks throughout this manuscript as silica-saturated alkaline rocks falling above the Irvine and Baragar (1971) subalkaline-alkaline boundary in the total alkalies versus

UDRY ET AL. 1526 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Table 1 Starting Martian Compositions and Normative Mineralogies (Normalized to Volatile-Free Basis) Clast 59 Fastball Backstay Champagne Esperanza Home Platea

Sols N/A 750 511 357 1055 1211 SiO2 48.84 48.71 50.70 46.37 49.50 49.13 TiO2 2.16 0.76 0.95 2.79 1.09 1.02 Al2O3 7.18 9.04 13.43 15.25 8.68 10.26 FeO 19.39 19.11 13.53 13.18 20.88 17.53 MnO 0.50 0.46 0.26 0.25 0.39 0.36 MgO 12.62 11.41 8.51 4.83 8.68 9.93 CaO 5.16 6.40 6.15 7.97 5.79 6.75 Na2O 2.01 2.91 4.05 5.05 3.51 3.29 K2O 0.14 0.30 1.05 0.55 0.54 0.51 P2O5 2.01 0.89 1.37 3.76 0.94 1.21 Total oxides 100.00 100.00 100.00 100.00 100.00 100.00 Mg#b 53.92 55.50 54.18 40.89 43.99 53.38

CIPW norm Quartz 0 0 0 0 0 0 Plagioclase 33 42 56 65 44 47 Albite 36 29 30 25 17 29 Orthoclase 1 2 7 4 4 4 Diopside 2 12 5 0 13 11 Hypersthene 51 16 9 0.8 14 14 Nepheline 0 0 0 0 0 0 5 24 19 18 22 21 Ilmenite 3 1 1 3 1 1 Apatite 5 2 3 8 2 3 Corundum 0 0 0 0.7 0 0 Density 3.3 3.2 3.0 3.0 3.2 3.1 aHome Plate_June Emerson composition. bUsing all FeOt as Fe2+ for comparison with other rocks.

silica diagram. We also define mildly alkaline rocks falling on the Irvine and Baragar (1971) classification scheme. This definition is similar to previous studies, such as McSween et al. (2006) and Nekvasil et al. (2007). Note that Edwards et al. (2017) defined alkaline compositions as silica-undersaturated nepheline normative. To better understand the processes that resulted in felsic and alkaline magmas on Mars and the implication for our understanding on continental crust formation, we first compiled a database of Martian igneous compositions, available in the supporting information (Table S1; e.g., Filiberto, 2017; Gellert et al., 2006; Ming et al., 2008; Santos et al., 2015; Sautter et al., 2015; Stolper et al., 2013; Zipfel et al., 2011). The complete database was used to statistically compare felsic Martian and terrestrial compositions and elucidate petrogenetic processes. For crystallization modeling calculations, we used six different starting Martian parental magma composi- tions: Clast 59 from breccia NWA 7034 (Santos et al., 2015) and five Spirit rover-analyzed rocks—Fastball, Backstay, Esperanza, Home Plate_June Emerson, and Champagne (Gellert et al., 2006; McSween et al., 2006; Ming et al., 2008; Squyres et al., 2007)—that were measured using the Alpha Particle X-ray Spectrometer instrument (Table 1). The starting compositions were selected because they represent a wide range of basaltic compositions, avoiding end members that were reported in the literature as altered, with the goal to show that Gale crater compositions can be obtained from fractionation of various parental magma compositions. Clast 59 is a relatively primitive basaltic clast (compared to the other clasts) with granulitic texture, and it shows the lowest total alkalis content (2.15 wt %) of the NWA 7034 clasts analyzed by Santos et al. (2015). The five investigated Gusev crater compositions all represent analyses of (RAT)-abraded surfaces. Fastball, which was analyzed in , was suggested to be near primary in composition (e.g., Filiberto et al., 2010) and was investigated as a parental magma for alkaline compositions by Udry, Balta, and McSween (2014). Backstay was found in (McSween et al., 2006) and has a mildly alkaline com-

position (Na2O = 4.05 wt % and K2O = 21.05 wt %). Esperanza, a member of the Irvine group, was analyzed at

UDRY ET AL. 1527 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Table 2 Felsic and Alkaline Martian Compositions From Sautter et al. (2015) and Normative Mineralogies (Normalized to Volatile- Free Basis) GROUP 2 GROUP 3 GROUP 5

Harrison C1 Harrison C2 Sledgers Chakonipau Becraft Sparkle Angmaat

Sols 512 512 379 338 421 514 337 SiO2SiO2 57.0 57.1 63.9 61.9 60.1 59.4 62.9 TiO2TiO2 0.62 0.49 0.29 0.83 1.31 0.37 0.91 Al2O3Al2O3 19.4 16.8 17.0 16.3 14.8 17.3 16.9 FeO 6.59 7.89 5.04 6.62 9.93 5.40 10.1 MgO 1.96 1.97 1.81 0.92 1.31 1.40 1.31 CaO 5.66 8.88 2.19 0.92 1.69 11.9 3.02 Na2ONa2O 6.08 3.95 5.14 5.43 4.22 3.45 2.52 K2OK2O 2.78 2.96 4.57 7.08 6.65 0.74 2.41 Total oxides 100 100 100 100 100 100 100 Mg#a 34.6 30.8 39.0 19.8 19.1 31.6 18.8

CIPW norm Quartz 0 0 6 0 0 12 25 Plagioclase 63 57 55 45 40 62 39 Albite 28 35 18 0 5 49 40 Orthoclase 18 20 29 45 43 5 16 Diopside 7 18 1 3 5 21 0 Hypersthene 0 6 10 0 8 0 15 Nepheline 5 0 0 0 0 0 0 Olivine 6 0 0 8 3 0 0 Ilmenite 1 1 0 2 1 0 1 Corundum 0 0 0 0 0 0 3 Density 2.5 2.6 2.5 2.5 2.5 2.8 3.0 aUsing all FeOt as Fe2+ for comparison with other rocks.

the summit of Columbia Hills and is enriched in FeO (20.20 wt %; Ming et al., 2008). Champagne, which was encountered on the Cumberland Ridge, shows the highest Na, Al, and Ca concentrations measured in Gusev crater (Gellert et al., 2006). In addition, as shown by Usui et al. (2008), this target has an igneous composition

and has a high merrillite abundance (and high P2O5 content) indicating a possible CO2-rich source. Finally, the Home plate_June Emerson composition (Ming et al., 2008) displays an intermediate composition compared to the other starting compositions. The selected samples differ in major elemental composition,

including Al2O3 between 8.68 wt % (Esperanza) and 14.73 wt % (Champagne), FeO between 12.73 wt % (Champagne) and 20.20 wt % (Esperanza), and K2O between 0.14 wt % (Clast 59) and 1.05 wt % (Backstay; see Table 1 for compositions and normative mineralogies). Sautter et al. (2015) and Cousin et al. (2017) identified 59 igneous rocks in Gale crater analyzed from sol 20 to

sol 800 using SiO2, TiO2,Al2O3, MgO, CaO, Na2O, and K2O compositions. Most of these rocks were found close to the Peace Vallis deposits. These rocks were identified using the ChemCam laser-induced break- down spectrometer (LIBS) and Remote Micro-Imager RMI and other imaging instruments (MAHLI—Mars Hand Lens Image—and MastCam) to investigate textural analyses and geological context. The igneous origin of these compositions was confirmed by Mangold et al. (2017), using a classification scheme based on textural and chemical analyses to distinguish igneous and sedimentary rocks in Gale crater. ChemCam ana- lyzes chemistries, measured using a laser beam size of 300–500 microns varying with distance of ana- lyses (see Maurice et al., 2016 and Cousin et al., 2017 for instrumentations and calibration details), and that can be combined to produce bulk-rock compositions. The Gale crater felsic compositions from Sautter et al. (2015) and Cousin et al. (2017) were divided into five groups according to target rock textures and colors (see Table 2 for compositions and normative mineralogies). Here we consider three groups defined by Cousin et al. (2017; Groups 2, 3, and 5; Table 2), which represent alkaline and felsic compositions. Harrison compositions (Group 2) are porphyritic, containing large light-colored euhedral phenocrysts

of feldspar in dark gray, fine-grained mesostasis, with SiO2 around 57 wt % and total alkalies (Na2O+K2O) between 6.91 and 8.86 wt %. Chakonipau, Sledger, and Becraft (Group 3) are vesiculated rocks that show

UDRY ET AL. 1528 Journal of Geophysical Research: Planets 10.1029/2018JE005602

alkaline compositions with SiO2 contents >60 wt %. Payré et al. (2017) mentions that the major (Na and K) and trace (Ba and Sr) element compositions suggest that Group 3 rocks have an alkali feldspar component. Black Trout (Group 4) has a microgabbroic texture. Sparkle and Aangmat (Group 5) are both subalkaline

and felsic (SiO2 >59 wt %) and are coarse grained (see Sautter et al., 2015 and Cousin et al., 2017 for detailed textural descriptions for each group). These rocks are all normative silica saturated, except Harrison C1. Finally, for comparison with Earth, data for terrestrial Archean TTG were taken from Moyen (2011), and intraplate volcanoes data were compiled from the GEOROC website (http://georoc.mpch-mainz.gwdg.de/ georoc/).

2.2. Rhyolite-MELTS Fractional Crystallization Modeling We performed fractional crystallization models from Martian starting parental compositions (described below) using the open-source software MELTS (Ghiorso & Sack, 1995) that runs the rhyolite-MELTS calibration (Gualda et al., 2012, http://melts.ofm-research.org). Using a set of thermodynamic parameters, the MELTS family of algorithms can model phase stabilities along defined crystallization paths. The rhyolite-MELTS calibration differs from others (MELTS and pMELTS), as it is appropriate for fluid-bearing silicic magmas

(>55-wt % SiO2). Rhyolite-MELTS also includes better calibrations for quartz and K-feldspar saturation as a function of pressure and should be used for pressures below 1 GPa (Gualda et al., 2012). The MELTS algorithm has been widely and successfully used for Martian magmas (Balta & McSween, 2013; McSween et al., 2006; Monders et al., 2007; Peslier et al., 2010; Sarbadhikari et al., 2011; Symes et al., 2008; Taylor et al., 2002; Udry, Balta, & McSween, 2014; Udry, Lunning, et al., 2014). We conducted isobaric fractional crystallization modeling at decreasing temperature intervals of 10 °C, start- ing at the liquidus temperature and ending at 600 °C representing liquidus of highly felsic compositions, or in some cases at higher temperature if the melt was completely exhausted. To evaluate the effect of pressure, we explored models with pressures from 4 to 8 kbar as alkaline magmas are formed at higher pressures (<4 kbar; Green, 1970). Our models follow the modeling procedures reported in previous studies on alkaline compositions (Nekvasil et al., 2007; Stolper et al., 2013),

Three different ƒO2 were applied to our models, relative to the fayalite--quartz buffer—FMQ À 1, FMQ, and FMQ + 1—which correspond to the higher range of the Martian oxygen fugacities recorded using all types of Martian meteorites and surface rocks (McCubbin, Boyce, Novák-Szabó, et al., 2016; Santos et al., 2015; Schmidt et al., 2013; Tuff et al., 2013). For these calculations, examples of liquid lines of descent at FMQ + 1 from the six starting compositions are plotted in Figure 2, where every discontinuity in the lines indi-

cates a new crystallizing mineral phase. An example of the effect of ƒO2 in the liquid lines of descent using Backstay as a starting composition is also provided in Figure 2 (more examples are in supporting information Figure S1). Various water contents have been reported in the parental magmas for Martian meteorites, from completely dry up to 2 wt % (e.g., Dann et al., 2001; Giesting et al., 2015; McCubbin et al., 2010, 2012; Médard & Grove, 2006; Nekvasil et al., 2007; Usui et al., 2012). Most recent papers indicate from 150 up to 750 ppm water in the Martian meteorite mantle parental magmas (Filiberto et al., 2016; McCubbin, Boyce, Srinivasan, et al., 2016). The Martian crust is slightly more water-rich with 1,410 ppm of water (McCubbin, Boyce, Srinivasan, et al., 2016). However, water is required to form felsic magmas on Earth (Moyen, 2011). For this purpose, we produced dif-

ferent models at 0.5 to 1 wt % % H2O added to the Martian parental compositions in our calculations, as also suggested by Nekvasil et al. (2007).

We compared the effects of the different modeling parameters (water contents, ƒO2, and pressures) used in our models of liquid lines of descents from the selected starting compositions with the Gale crater felsic and alkaline compositions from Sautter et al. (2015) and Cousin et al. (2017), using a best-fit (least square) method. In order to quantitatively compare our thermodynamic fractionation models to Gale crater compositions, we use a “best-fit” index. The best-fit index uses the “sum of squares best-fit” method, in which the mean of

major published elemental compositions (using SiO2, TiO2,Al2O3,Fe2O3, MgO, CaO, Na2O, and K2O) for our models and Gale crater compositions are squared and summed (Wheater & Cook, 2006). We obtained best-fit indices for each bulk composition for all the different modeled liquid lines of descent (results reported in Tables 3 and S1). The lowest best-fit index corresponds to the best match between our modeled bulk com- positions and the Gale crater felsic rocks target compositions. Once we identified compositions with lowest

UDRY ET AL. 1529 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Figure 2. Diagrams of (a) CaO versus MgO, (b) FeO versus MgO, (c) SiO2 versus MgO, (d) Al2O3 versus MgO, (e) Na2O+K2O versus SiO2, and, (f) Na2O/K2O versus SiO2 (wt %) showing liquid lines of descent from the seven starting compositions at 4 kbars, 0.5-wt % H2O, and FMQ + 1. Data sources: Gale crater felsic rocks (Sautter et al., 2015), Jake_M (Schmidt et al., 2014), other rover analyses from Gusev crater (Gellert et al., 2006; Ming et al., 2008) NWA 7034 meteorite (Santos et al., 2015), and TTG data (Moyen, 2011). See supporting information for data references. FMQ = fayalite-magnetite-quartz; SNC = shergottites, , and chassignites.

best-fit indices, we compared normative mineralogies calculated using the CIPW norm of both modeled and Gale crater compositions (Table 4).

3. MELTS Modeling Results Overall, the liquid lines of descent calculated for each starting composition using the different water contents are similar (Data in the supporting information Table S2). In addition, the range of oxygen fugacities used in this study does not have a large effect on the different liquid lines of descent, with the exception of higher

alkali content for the same SiO2 content at higher ƒO2 (see Backstay example in Figure 2). The difference in pressures shows the largest effect for the various liquid lines of descent (Figures 2 and S1). For the same

MgO, we observe higher overall CaO, FeO, and SiO2 contents and delayed Al2O3 turnover (representing delayed onset of plagioclase crystallization) at higher pressure for each starting composition. In addition,

we observe a constant Na2O/K2O with increasing SiO2 (Figure 2f). When comparing liquid lines of descent originating from the different starting compositions at the same conditions (Figure 2), almost all evolved toward the trachyte and rhyolite fields with a relatively constant

UDRY ET AL. 1530 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Table 3 Best-fit Indices Calculated by Statistically Comparing the Melt Compositions Formed From the Five Different Martian Parental Melt and the Gale Crater Compositions Clast 59 Fastball Backstay Champagne Esperanza Home plate % of crystallization T (°C)

Harrison C1 81 40 5 8 71 49 65 1,119 Harrison C2 54 31 13 31 52 33 32 1,138 Sledgers 49 31 11 25 41 45 61 1,029 Chakonipau 80 58 31 46 63 66 64 984 Becraft 60 55 33 55 52 53 65 983 Sparkle 66 66 52 61 87 80 39 1,128 Angmaat 23 31 38 47 33 38 61 1,038

alkali ratio, whereas Clast 59 evolved toward the dacite field. The evolution of these trends are likely due to the fact that Clast 59 contains lower alkali content compared to the other starting compositions. Liquid lines of descent are similar for the different starting compositions for each oxide. Only the starting compositions of

Backstay and Champagne reached the Al2O3 and CaO values of the Gale felsic rocks, as they started with higher values of those oxides; nevertheless, most of the Al2O3 and CaO compositions are still too high (Figure 2). Backstay starting composition lead to better fits for Na/K ratio to obtain Gale crater compositions.

Figure 3 illustrates the fractionated phases under the same conditions (4 kbar, FMQ + 1, and 0.5-wt % H2O) as the liquid lines of descent in Figure 2. As spinel (here Cr-spinel) is the modeled liquidus phase but, similar to MELTS and pMELTS calculations from previous studies (Balta & McSween, 2013), spinel stability seems to be overestimated in rhyolite-MELTS as pressure increases. Note that the modal abundances of crystallized

Cr-rich spinel are too low to affect the liquid line of descent (except for Cr2O3). If spinel is ignored, orthopyr- oxene and whitlockite are the first phases to crystallize, followed by clinopyroxene and feldspar. Note that at high degree of fractionation, MELTS models can generate unrealistic assemblages (i.e., olivine and quartz briefly coexist in models using Clast 59 and Home Plate starting compositions). For a liquid with 65 wt %

SiO2 for each parental magma, feldspar is the dominant phase, with varying amounts of clinopyroxene and orthopyroxene but no quartz. Parental magmas with highest MgO (Clast 59) fractionate more orthopyroxene,

whereas parental magmas containing more Al2O3 (Champagne and Backstay) fractionate more feldspar. To better understand if the Gale crater felsic and alkaline rocks formed from fractionation from similar com- positions to our starting parental magmas, we compared the modeled results from the different Martian starting compositions with the Gale crater felsic and alkaline rocks using best-fit indices (Table 3). Table 3 represents the lowest best-fit indices of liquid lines of descent from starting compositions with Gale crater

evolved compositions. The conditions (including pressure, water content, and ƒO2) for the lowest best-fit indices are variable for the different Gale crater compositions (Table S3). Compositions of Gale crater felsic rocks are more closely reproduced by models that start with Backstay composition as highlighted in Table 3. This is also observed on Figure 2. Group 5 felsic rocks (Sparkle and Aangmat) show the lowest-

Table 4 CIPW Normative Mineralogies of Gale Crater Felsic and Alkaline (wt % in orange) Compared to CIPW Normative Mineralogies of Best-fit Modeled Composition (wt % in ) Group 2 Group 3 Group 5

CIPW norm Harrison C1 Harrison C2 Sparkle Becraft Chakonipau Angmaat Sledgers

Quartz 0 0 0 0 12 0 0 0 0 0 25 14 6 1 Plagioclase 63 74 57 66 62 67 40 58 45 63 39 64 55 71 Orthoclase 18 11 20 10 5 11 43 19 45 19 16 3 29 17 Diopside 7 6 18 13 21 13 5 13 3 11 0 7 1 7 Hypersthene 0 0 6 0 0 0 8 0 0 0 15 11 10 3 Nepheline 5 1 0 2 0 7 0 4 0 1 0 0 0 0 Olivine 6 7 0 8 0 0 3 5 8 4 0 0 0 0 Ilmenite 1 1 1 1 0 1 1 0 2 0 1 1 0 1

Na2SiO3 00000001020000

UDRY ET AL. 1531 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Figure 3. Left side: Modal abundances of the different phases for liquid line of descent calculations at ~65 wt % of SiO2 using the following starting compositions: (a) NWA 7034 Clast 59, (b) Esperanza, (c) Fastball, (d) Home Plate, (e) Champagne, and (f) Backstay. Right side: Modal abundances of different phases versus Mg# for the seven different starting compositions. All these models were calculated with starting conditions of FMQ + 1, 4 kbars, and 0.5 wt % H2O.

best-fit indices from parental compositions of Backstay and NWA 7034 Clast 59, respectively, with degrees of

fractionation of 61%. The calculated melts show a lower Al2O3 content compared to the Gale crater composi- tions. Group 3 (Chakonipau, Sledger, and Becraft) show good fits from a parental melt with Backstay compo- sition, with melt fractionation between 39% and 65%, reaching the point of critical crystallinity (Table 4). The calculated melts show a higher CaO content compared to the Gale felsic compositions. Group 2 (Harrison C1 and C2) might have formed from a parental melt with a Backstay composition but at different degrees of frac- tionation (between 32% and 65%). These results suggest that Gale alkaline and felsic rocks could have formed

within the whole range of pressure, water content, and ƒO2 conditions found in the Martian crust (McCubbin et al., 2010; McCubbin, Boyce, Srinivasan, et al., 2016; Santos et al., 2015; Schmidt et al., 2014; Usui et al., 2012). For some of these compositions, the evolved compositions are obtained after high degrees of fractionation (>60%; Table 3), implying that although these compositions can be widespread, their volumes are not likely to be large on the Martian surface. The modeled bulk compositions are similar to the Gale crater com-

positions. However, as shown in Figure 2, the Al2O3 contents of the Gale felsic data are lower and CaO con- tents are higher in our model. This observation suggests that the Gale crater rocks do not represent primary melts, and they possibly fractionated and feldspar (especially plagioclase) was removed or accumulated dur- ing crystallization. Although the bulk compositions in our models and the Gale crater compositions are similar, the modeled fractionated phases have some differences (Table 4 and Figure 4). Table 4 represents modal abundances of best-fit models from the six different Martian compositions compared to modal abundances of Gale crater evolved compositions. For the conditions corresponding to the lowest best-fit indices (melt compositions

between 52 and 63 wt % SiO2), quartz is systematically absent, except for Group 5. Our models also

UDRY ET AL. 1532 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Figure 4. Comparison of modal abundances of phases calculated from ChemCam data (blue) versus the modal abundances of calculated phases corresponding to the best-fit modeled data (red) for each for felsic and alkaline compositions used in this study.

consistently show higher plagioclase abundances compared to the normative Gale modal compositions but lower orthoclase abundances, which is consistent with higher CaO in our model. Small amounts of nepheline are present in the fractionation models for Groups 2 and 3. However, when comparing the normative mineralogy of the best-fit modeled compositions to the Gale crater normative mineralogy, the overall phases are consistent (Tables 2 and 4). The differences between normative and modeled mineralogy for the modeled compositions are difficult to explain but could be due to parameterization errors for spinel in rhyolite-MELTS (Balta & McSween, 2013; Gualda et al., 2012).

4. Discussion 4.1. Formation Processes of Felsic Rocks in Gale Crater According to our best-fit analyses, it is possible to generate felsic compositions through fractional crystalliza- tion with similar chemical and mineralogical compositions to those reported at Gale crater rocks, using var-

ious Martian conditions (pressure, ƒO2, and water content) and a wide range of starting Martian compositions measured on the Martian surface. It is important to note that MELTS calculations do not represent the abso- lute true compositions. As described in Balta and McSween (2013), MELTS algorithm was not calibrated based

UDRY ET AL. 1533 Journal of Geophysical Research: Planets 10.1029/2018JE005602

on experiments but on thermodymanical data, and thus errors, such as overestimation of Cr-spinel liquidus temperatures and an estimated error of 1 wt % in oxides, arise. However, MELTS is a powerful tool to un- derstand conditions of formation of igneous rocks, if various liquid lines of descent are calculated under different parameters, as conducted in this study, to in a time-efficient manner (Balta & McSween, 2013). Similar to our results, using fractionation calculation, Edwards et al. (2017) also suggested that fractionation of olivine from Gusev-like Humphrey basaltic composition can yield the trachy- (Mg# = 27) composition measured by ChemCam at Gale crater. Our results show that, from the different starting melt compositions, fractio- nation of orthopyroxene, plagioclase, and diopside as well as some minor olivine can lead to the felsic compositions (Mg# = 19–38) found in Gale crater during the first 800 sols (Table 4). The modeling pre- sented in this study also agrees with studies on the NWA 6963 meteor- ite, the only shergottite yielding granitic compositions consisting of quartz, plagioclase, and K-feldspar (Filiberto et al., 2014; Gross & Filiberto, 2014). These authors suggested that these highly felsic com- positions likely formed in pockets through fractional crystallization of a basaltic magma with NWA 6963 composition. Additionally, mildly alkaline rocks in the Columbia Hills in Gusev crater such as Backstay, Irvine, and Wishstone (McSween et al., 2006) also likely resulted from fractional crystallization of a Martian primary magmas under polybaric conditions (Udry, Balta, & McSween, 2014).

4.2. Comparison of Mars Felsic Rocks With Evolved Terrestrial Rocks On Earth, evolved compositions are ubiquitous, but most of these rocks are found in the continental crust with some formed in intraplate vol- canism. Earth is the only planetary body with a reported continental crust (Taylor & McLennan, 2009). However, due to the diversity and compositions found at Gale crater, Sautter et al. (2015) suggested that Mars might also have a continental crust, similar to the Archean crust on Earth. However, the definition of continental crust is relatively broad. On Earth, continental crust has an average intermediate (57-

to 64-wt % SiO2) composition and comprises 40% of the Earth’s surface (Rudnick, 1995). A large portion of the continental crust (60%) was Figure 5. Statistical comparison of elemental ratios between felsic rocks formed during the Archean through the generation of TTG suites, from Gale crater (Sautter et al., 2015) and examples from Earth (with which have similar composition to the post-Archean crust (Rudnick & SiO2 >50 wt % and Na2O+K2O >3 wt %, Canaries n = 706; n = 42; Hawaii n = 87; Iceland n = 754; and TTG n = 1343). The radii of the circles on the Gao, 2003). right panel represent the mean of the standard error colored by locality. The The new Gale crater felsic and alkaline composition data collectively names of the localities highlighted in red in the left panel indicate that the mean (represented by larger symbols) is within the ±2σ of the standard error of reveal a more complex picture of Mars crustal diversity. We, and others the mean of the felsic rocks from Gale crater. TTG = tonalite-trondhjemite- (e.g., Edwards et al., 2017; Filiberto, 2008; McCubbin et al., 2008; granodiorite. McSween et al., 2006; Nekvasil et al., 2007, 2009; Sautter et al., 2015; Udry, Balta, & McSween, 2014; Whitaker et al., 2005), have demon- strated that the evolved compositions can form by fractionation of Martian magmas. The question remains as to whether these felsic compositions qualify as continental crust formed in early Martian history (Sautter et al., 2015, 2016). To better understand the formation of felsic Martian compositions, we statistically and analytically compare Gale crater compositions to terrestrial alkaline and felsic compositions from Archean continental crust (TTG, which constitute the Earth’s early continental crust) and intraplate volcanism (representative of nonsubduction-related volcanism on Earth). We first selected samples from our Earth data set in the range

of the felsic compositions from Gale crater (SiO2 >50 wt % and Na2O+K2O >3 wt %, Figure 2e). In an

UDRY ET AL. 1534 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Figure 6. Comparison of major element geochemical data (wt % oxides) for igneous rocks from Earth and Mars. Note the similarities between data from Earth’s intraplate volcanoes and in situ rover data collected on Mars and Martian meteorites. Mid-ocean ridge basalts are not included, as spreading center magmas are too low in alkalis compared to the in situ rover data collected on Mars. Data sources as in Figure 2. SNC = shergottites, nakhlites, and chassignites; TTG = tonalite-trondhjemite-granodiorite.

attempt to limit the effects of fractionation and to offer the means of comparison with spectroscopically produced ratios, we produced major element ratios normalized to Si and compared them using a t-test at a 95% confidence level (2σ). Figure 5 presents the results from this statistical evaluation. For all major element ratios, the Gale evolved compositions (Sautter et al., 2015) are more consistent with Earth’s intraplate volcanoes than with TTG, the only exception being Mg/Si that is also statistically equivalent with Iceland data. There is a clear bimodality in K/Si of the felsic samples from Gale crater; thus, we compared sam- ples with low K/Si (Aangmat and Sparkle) separately. As previously mentioned, Payré et al. (2017) showed that Group 3 is more enriched in K (and Ba). These samples were described as having a closer affinity to Earth’s average continental crust (Sautter et al., 2015). Nevertheless, Figure 5 clearly shows that they are statistically more similar to felsic Iceland magmas than to TTG.

UDRY ET AL. 1535 Journal of Geophysical Research: Planets 10.1029/2018JE005602

We also compare Gale crater evolved rocks and terrestrial rock major element compositions (Figure 6) to understand further the formation of the Martian evolved compositions. We included examples of intra- plate volcanoes developed on oceanic plates and away from ancient continental terranes (Canary Islands, Cape Verde, Hawaii, and Iceland) known to produce alkaline and felsic rocks. We contrasted these data with terrestrial felsic continental crust represented by TTG suites (Figure 6). We excluded mid-ocean ridge basalts and derivative melts from this comparison, as they are too low in alkali contents compared to the in situ rover data collected on Mars. From these comparisons, in terms of major element compositions, the Gale crater felsic and alka- line compositions are more similar to Earth’s intraplate volcanoes than to terrestrial continental crust as represented by TTG suites (Figure 6). In fact, the evolutionary trends for Martian magmas are remarkably similar to intraplate volcanoes (especially Hawaii and Iceland, Figures 6a–6c), Figure 7. Comparisons of the liquid lines of descent of Iceland volcanic rocks with the exceptions of CaO, which is generally lower in the Martian with in situ rover data collected on Mars, tonalite-trondhjemite-granodiorite fi (TTG), and rhyolite-MELTS fractional crystallization models for seven Martian magmas (Figure 6d), and FeO, which is signi cantly more abundant parental magmas. Note that the felsic rocks from Gale crater (Sautter et al., 2015) in Martian magmas (Figure 6e). TiO2 is also notably different between are more consistent with fractional crystallization processes than with early the Gale felsic and the terrestrial suites (see Stolper et al., 2013). One continental crust formation on Earth as represented by TTG compositions possible explanation is early fractionation of a Fe-Ti oxide. The differ- (Moyen, 2011). The rhyolite-MELTS models were calculated at 4 kbars with an ences in FeO can simply be explained by Mars’ mantle being more FMQ + 1 buffer and with an initial H O abundance of 0.5 wt %. 2 enriched in Fe relative to our ’s mantle (Taylor, 2013). Evolved lithologies from Gale crater also show different mineralogical compositions from TTG samples. Continental crust on Earth, as repre- sented by TTG suites, contains quartz, sodic plagioclase, and K-feldspar (Rudnick & Gao, 2003). Minor amounts of quartz and cristobalite were detected by CheMin (Bish et al., 2013; Rampe et al., 2017; Vaniman et al., 2013), and of igneous origin was detected by the Curiosity rover (Morris et al., 2016). However, in our results, only the Sparkle, Aangmat, and Sledgers samples from Gale crater are quartz-normative (Table 2). In addition, evidence from orbital thermal infrared spectra from Mars’ surface indicates that quartz is insignificant at a global scale, if it occurs at all (Christensen et al., 2005). Moreover, orbi- tal visible/near-infrared spectral observations have identified feldspathic rocks, but these are not felsic in the sense of bearing quartz (Rogers & Nekvasil, 2015). Moreover, rocks from terrestrial TTG suites are character- ized by containing water-bearing phases (amphibole and mica; Moyen, 2011), which have not been observed in the Gale felsic rocks. The nondetection of these phases might be explained breakdown of the hydrous phases to anhydrous ones (e.g., and magnetite) during storage or transport to the surface as magma degas at greater depth in Mars due to the lower gravity. To further evaluate the presence of continental crust on Mars, we compare the evolution of Mg# (molar MgO/[FeO + MgO]×100) versus silica content of melts from our six Martian starting compositions with Iceland , where abundant felsic rocks have been formed by fractional crystallization and/or assimilation-fractional crystallization processes (MacDonald et al., 1987; Willbold et al., 2009), and with TTG (Figure 7). Iceland magmas evolved from basaltic melts in equilibrium with the terrestrial mantle (Mg# = ~70–72) into felsic magmas, with trends that can be reproduced by fractional crystallization processes that agree with the rhyolite-MELTS models produced in this study. Note that the highly evolved Icelandic magmas (rhyolitic compositions) formed through fractionation of plagioclase-clinopyroxene-magnetite assemblage from ferrobasaltic parental melts (MacDonald et al., 1987). This fractionation phases are similar

to our models, except the presence of magnetite, which shows high H2O content in terrestrial magmas. The TTG compositions also start at both high Mg# and silica contents (62–77 wt %) suggesting the necessity

of equilibrium with the mantle for their primary magmas (Figure 6) but at much higher SiO2 contents than magmas produced by simple decompression of the upper mantle. This typical characteristic of continental crust on Earth has been explained by reaction between melts from an eclogitic/amphibolitic source and man- tle peridotite (Kelemen et al., 1998), and is necessary for the genesis of continental crust in the Archean and modern analogs in arc settings (see additional discussion in Gazel et al., 2015). Therefore, by comparing these two contrasting data sets (Martian and terrestrial) and processes that best explain their origin, we conclude

UDRY ET AL. 1536 Journal of Geophysical Research: Planets 10.1029/2018JE005602

the Martian felsic rocks are more consistent with fractional crystallization of basaltic compositions (Figure 7) rather than continental crust development as it happened on Earth. The current geochemical evidence collectively suggests that, on early Mars, basaltic magmas fractionated to produce limited amounts of felsic rocks, comparable to magmas produced in intraplate volcanoes on Earth. The dominant composition of more recently produced Martian crust appears to have been tholeiitic basalt. In other words, felsic magmas on Mars are not representative of a true continental crust. The presence of evolved rocks may have been favored on early Mars due to the fact that older magmas may have been more water rich than the latter magmas, because early melting dried out the mantle sources (McCubbin et al., 2010). Although hydrous minerals confirming wetter magmas were not analyzed in Gale crater, these minerals might have been in small quantity. Until more robust evidence is found on Mars and on other ter- restrial planets, we conclude that continental crust, defined by large amounts of felsic rocks, may be a uniquely terrestrial attribute. The apparent lack of large amounts of felsic rocks on Mars and the similarity in major element compositions between Martian rocks and samples from terrestrial intraplate volcanoes suggest that the most likely process for magmatism on Mars was adiabatic melting of the mantle, possibly in localized thermal-chemical upwellings like mantle plumes, followed by fractional crystallization and possibly assimilation at crustal levels.

5. Conclusions We have modeled fractional crystallization of various Martian starting compositions with MELTS in order to understand the formation of the felsic igneous compositions analyzed at Gale crater. We conclude that 1. pressure differences have a strong effect on liquid lines of descent, with almost all basaltic starting com- positions evolving toward the trachyte and rhyolite fields, except for Clast 59, which is depleted in alkali elements compared to the other starting compositions; 2. most Gale crater felsic compositions could have formed from fractional crystallization at different degrees of a Gusev crater Backstay-like composition; and

3. our models yield compositions lower in Al2O3 and higher in CaO, suggesting that the Gale crater igneous rocks may not represent true liquids but underwent plagioclase fractionation. In addition, our models show enrichment in plagioclase and depletion in orthoclase compared to the felsic Gale crater rocks. A rather exhaustive comparison indicates that Gale felsic rocks are compositionally and mineralogically more comparable to fractionated terrestrial intraplate volcanic rocks than to the Archean continental crust, as represented by TTG. Based on this comparison and the modeling results, we conclude that felsic rocks in Gale crater do not represent an early Martian continental crust but instead are localized products of the frac- tionation of mildly alkaline to alkaline basaltic magmas that were more common in early Mars history.

Acknowledgments References This study was supported in part by NASA Mars Data Analysis Program Agee, C. B., Wilson, N. V., McCubbin, F. M., Ziegler, K., Polyak, V. J., Sharp, Z. D., et al. (2013). Unique meteorite from early Amazonian Mars: – 80NSSC17K0477 to A. U. and E. G., NSF Water-rich basaltic breccia Northwest Africa 7034. Science, 339(6121), 780 785. https://doi.org/10.1126/science.1228858 grant EAR-12019033 to E. G., and NASA Balta, J. B., & McSween, H. Y. (2013). Application of the MELTS algorithm to Martian compositions and implications for magma crystallization. – Cosmochemistry grant NNX13AH86G to Journal of Geophysical Research: Planets, 118, 2502 2519. https://doi.org/10.1002/2013JE004461 H. Y. M. The authors thank John Bridges, Bellucci, J. J., Nemchin, A. A., Whitehouse, M. J., Humayun, M., Hewins, R., & Zanda, B. (2015). Pb-isotopic evidence for an early, enriched crust – Mariek Schmidt, and two anonymous on Mars. Earth and Planetary Science Letters, 410,34 41. https://doi.org/10.1016/j.epsl.2014.11.018 reviewers for their helpful reviews. We Bish, D. L., Blake, D. F., Vaniman, D. T., Chipera, S. J., Morris, R. V., Ming, D. W., et al. (2013). X-ray diffraction results from Mars science also thank Justin Filiberto and David laboratory: Mineralogy of at Gale crater. Science, 341(6153). https://doi.org/10.1126/science.1238932 Baratoux for the editorship. Original Boynton, W. V., Taylor, G. J., Evans, L. G., Reedy, R. C., Starr, R., Janes, D. M., et al. (2007). Concentration of H, Si, Cl, K, Fe, and Th in the low- and data sources are described in section 2 mid- regions of Mars. Journal of Geophysical Research, 112, E12S99. https://doi.org/10.1029/2007JE002887 and the supporting information. Carter, J., & Poulet, F. (2013). Ancient plutonic processes on Mars inferred fromthe detection of possible anorthositic terrains. Nature Geoscience, 6(12), 1008–1012. https://doi.org/10.1038/ngeo1995 Cartwright, J. A., Ott, U., Herrmann, S., & Agee, C. B. (2014). Modern atmospheric signatures in 4.4 Ga Martian meteorite NWA 7034. Earth and Planetary Science Letters, 400,77–87. https://doi.org/10.1016/j.epsl.2014.05.008 Christensen, P. R., McSween, H. Y., Bandfield, J. L., Ruff, S. W., Rogers, A. D., Hamilton, V. E., et al. (2005). Evidence for magmatic evolution and diversity on Mars from infrared observations. Nature, 436(7050), 504–509. https://doi.org/10.1038/nature03639 Cousin, A., Sautter, V., Payré, V., Forni, O., Mangold, N., Gasnault, O., et al. (2017). Classification of igneous rocks analyzed by ChemCam at Gale crater, Mars. Icarus, 288, 265–283. https://doi.org/10.1016/j.icarus.2017.01.014 Dann, J. C., Holzheid, A. H., Grove, T. L., & McSween, H. Y. Jr. (2001). Phase equilibria of the Shergotty meteorite: Constraints on preeruptive H2O contents of Martian magmas and fractional crystallization under hydrous conditions. Meteoritics and Planetary Science, 36(6), 793–806. https://doi.org/10.1111/j.1945-5100.2001.tb01917.x Dreibus, G., & Wänke, H. (1985). Mars, a volatile-rich planet. Meteoritics, 20, 367–381.

UDRY ET AL. 1537 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Dreibus, G., & Wänke, H. (1987). Volatiles on Earth and Mars—A comparison. Icarus, 71(2), 225–240. https://doi.org/10.1016/ 0019-1035(87)90148-5 Edwards, P. H., Bridges, J. C., Wiens, R., Anderson, R., Dyar, D., Fisk, M., et al. (2017). Basalt–trachybasalt samples in Gale crater, Mars. Meteoritics & Planetary Science, 52(11), 2931–2410. Filiberto, J. (2008). Experimental constraints on the parental liquid of the Chassigny meteorite: A possible between the Chassigny meteorite and a Gusev basalt. Geochimica et Cosmochimica Acta, 72(2), 690–701. https://doi.org/10.1016/j.gca.2007.11.007 Filiberto, J. (2017). Geochemistry of Martian basalts with constraints on magma genesis. Chemical Geology, 466,1–14. https://doi.org/ 10.1016/j.chemgeo.2017.06.009 Filiberto, J., Dasgupta, R., Kiefer, W. S., & Treiman, A. H. (2010). High pressure, near-liquidus phase equilibria of the Home Plate basalt Fastball and melting in the Martian mantle. Geophysical Research Letters, 37, L13201. https://doi.org/10.1029/2010GL043999 Filiberto, J., Gross, J., & McCubbin, F. M. (2016). Constraints on the water, , and fluorine content of the Martian mantle. Meteoritics & Planetary Science, 51(11), 2023–2035. https://doi.org/10.1111/maps.12624 Filiberto, J., Gross, J., Trela, J., & Ferre, E. C. (2014). Gabbroic Shergottite Northwest Africa 6963: An intrusive sample of Mars. American Mineralogist, 99(4), 601–606. https://doi.org/10.2138/am.2014.4638 Filiberto, J., Gross, J., Udry, A., Trela, J., Wittmann, A., Cannon, K. M., et al. (2018). Shergottite Northwest Africa 6963: A Pyroxene-Cumulate Martian Gabbro. Journal of Geophysical Research: Planets, 120. https://doi.org/10.1029/2018JE005635 Gattacceca, J., Rochette, P., Scorzelli, R. B., Munayco, P., Agee, C., Quesnel, Y., et al. (2014). Martian meteorites and Martian magnetic anomalies: A new perspective from NWA 7034. Geophysical Research Letters, 41, 4859–4864. https://doi.org/10.1002/2014GL060464 Gazel, E., Hayes, J. L., Hoernle, K., Kelemen, P., Everson, E., Holbrook, W. S., et al. (2015). Continental crust generated in oceanic arcs. Nature Geoscience, 8(4), 321–327. https://doi.org/10.1038/ngeo2392 Gellert, R., Rieder, R., Brückner, J., , B. C., Dreibus, G., Klingelhöfer, G., et al. (2006). Alpha Particle X-Ray Spectrometer (APXS): Results from Gusev crater and calibration report. Journal of Geophysical Research, 111, E02S05. https://doi.org/10.1029/2005JE002555 Ghiorso, M. S., & Sack, R. O. (1995). Chemical mass transfer in magmatic processes IV: A revised and internally consistent thermodynamic model for the interpolation and extrapolation of liquid-solid equilibria in magmatic systems at elevated temperatures and pressures. Contributions to Mineralogy and Petrology, 119(2-3), 197–212. https://doi.org/10.1007/BF00307281 Giesting, P. A., Schwenzer, S. P., Filiberto, J., Starkey, N. A., Franchi, I. A., Treiman, A. H., et al. (2015). Igneous and shock processes affecting chassignite amphibole evaluated using chlorine/water partitioning and isotopes. Meteoritics & Planetary Science, 50(3), 433–460. https://doi.org/10.1111/maps.12430 Green, D. H. (1970). A review of experimental evidence on the origin of basaltic and nephelinitic magmas. Physics of the Earth and Planetary Interiors, 3, 221–235. https://doi.org/10.1016/0031-9201(70)90060-9 Gross, J., & Filiberto, J. (2014). Granitic composition in gabroic Martian meteorite NWA 6963: Evidence for extreme fractional crsytallization of a hydrous magma, Workshop on Volatiles in the Martian Interior. Grotzinger, J. P., Crisp, J. A., Vasavada, A. R., & Science Team, M. S. L. (2015). Curiosity’s mission of exploration at Gale crater, Mars. Elements, 11(1), 19–26. https://doi.org/10.2113/gselements.11.1.19 Gualda, G. A. R., Ghiorso, M. S., Lemons, R. V., & Carley, T. L. (2012). Rhyolite-MELTS: A modified calibration of MELTS optimized for silica-rich, fluid-bearing magmatic systems. Journal of Petrology, 53(5), 875–890. https://doi.org/10.1093/petrology/egr080 Humayun, M., Nemchin, A., Zanda, B., Hewins, R. H., Grange, M., Kennedy, A., et al. (2013). Origin and age of the earliest Martian crust from meteorite NWA 7533. Nature, 503(7477), 513–516. https://doi.org/10.1038/nature12764 Irvine, T. N. J., & Baragar, W. R. A. F. (1971). A guide to the chemical classification of the common volcanic rocks. Canadian Journal of Earth Sciences, 8(5), 523–548. https://doi.org/10.1139/e71-055 Kelemen, P. B., Hart, S. R., & Bernstein, S. (1998). Silica enrichment in the continental upper mantle via melt/rock reaction. Earth and Planetary Science Letters, 164(1-2), 387–406. https://doi.org/10.1016/S0012-821X(98)00233-7 Le Deit, L., Hauber, E., Fueten, F., Pondrelli, M., Rossi, A. P., & Jaumann, R. (2013). Sequence of infilling events in Gale crater, Mars: Results from morphology, stratigraphy, and mineralogy. Journal of Geophysical Research: Planets, 118, 2439–2473. https://doi.org/10.1002/2012JE004322 Le Deit, L., Mangold, N., Forni, O., Cousin, A., Lasue, J., Schröder, S., et al. (2016). The potassic sedimentary rocks in Gale crater, Mars, as seen by ChemCam on board Curiosity. Journal of Geophysical Research: Planets, 121, 784–804. https://doi.org/10.1002/2015JE004987 MacDonald, R., Sparks, R. S. J., & Sigurdsson, H. (1987). The 1875 eruption of Askja volcano, Iceland: Combined fractional crystallization and selective contamination in the generation of rhyolitic magma. Mineralogical Magazine, 51(360), 183–202. https://doi.org/10.1180/ minmag.1987.051.360.01 Mangold, N., Forni, O., Dromart, G., Stack, K., Wiens, R. C., Gasnault, O., et al. (2015). Chemical variations in Yellowknife Bay formation sedimentary rocks analyzed by ChemCam on board the Curiosity rover on Mars. Journal of Geophysical Research: Planets, 120, 452–482. https://doi.org/10.1002/2014JE004681 Mangold, N., Schmidt, M. E., Fisk, M. R., Forni, O., McLennan, S. M., Ming, D. W., et al. (2017). Classification scheme for sedimentary and igneous rocks in Gale crater, Mars. Icarus, 284,1–17. https://doi.org/10.1016/j.icarus.2016.11.005 Maurice, S., Clegg, S., Wiens, R. C., Gasnault, O., Rapin, W., Forni, O., et al. (2016). ChemCam activities and discoveries during the nominal mission of the Mars Science Laboratory in Gale crater. Mars. JAAS, 31(4), 863–889. https://doi.org/10.1039/c5ja00417a McCubbin, F. M., Boyce, J. W., Novák-Szabó, T., Santos, A. R., Tartèse, R., Muttik, N., et al. (2016). Geologic history of martian regolith breccia Northwest Africa 7034: Evidence for hydrothermal activity and lithologic diversity in the martian crust. Journal of Geophysical Research: Planets, 121, 2120–2149. https://doi.org/10.1002/2016JE005143 McCubbin, F. M., Boyce, J. W., Srinivasan, P., Santos, A. R., Elardo, S. M., Filiberto, J., et al. (2016). Heterogeneous distribution of H2O in the Martian interior: Implications for the abundance of H2O in depleted and enriched mantle sources. Meteoritics & Planetary Science, 51, 2036–2060. https://doi.org/10.1111/maps.12639 McCubbin, F. M., Hauri, E. H., Elardo, S. M., Vander Kaaden, K. E., Wang, J., & Shearer, C. K. Jr. (2012). Hydrous melting of the Martian mantle produced both depleted and enriched shergottites. Geology, 40(8), 683–686. https://doi.org/10.1130/G33242.1 McCubbin, F. M., Nekvasil, H., Harrington, A. D., Elardo, S. M., & Lindsley, D. H. (2008). Compositional diversity and stratification of the Martian crust: Inferences from crystallization experiments on the picrobasalt Humphrey from Gusev crater, Mars. Journal of Geophysical Research, 113, E11013. https://doi.org/10.1029/2008JE003165 McCubbin, F. M., Smirnov, A., Nekvasil, H., Wang, J., Hauri, E., & Lindsley, D. H. (2010). Hydrous magmatism on Mars: A source of water for the surface and subsurface during the Amazonian. Earth and Planetary Science Letters, 292(1–2), 132–138. https://doi.org/10.1016/ j.epsl.2010.01.028 McSween, H. Y., Ruff, S. W., Morris, R. V., Bell, J. F. III, Herkenhoff, K., Gellert, R., et al. (2006). Alkaline volcanic rocks from the Columbia Hills, Gusev crater, Mars. Journal of Geophysical Research, 111, E09S91. https://doi.org/10.1029/2006JE002698

UDRY ET AL. 1538 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Médard, E., & Grove, T. L. (2006). Early hydrous melting and degassing of the Martian interior. Journal of Geophysical Research, 111, E11003. https://doi.org/10.1029/2006JE002742 Ming, D. W., Gellert, R., Morris, R. V., Arvidson, R. E., Brückner, J., Clark, B. C., et al. (2008). Geochemical properties of rocks and soils in Gusev crater, Mars: Results of the Alpha Particle X-Ray Spectrometer from Cumberland Ridge to Home Plate. Journal of Geophysical Research, 113, E12S39. https://doi.org/10.1029/2008JE003195 Monders, A. G., Medard, E., & Grove, T. L. (2007). Phase equilibrium investigations of the class basalts from the Gusev plains, Gusev crater, Mars. Meteoritics & Planetary Science, 42(1), 131–148. https://doi.org/10.1111/j.1945-5100.2007.tb00222.x Morris, R. V., Vaniman, D. T., Blake, D. F., Gellert, R., Chipera, S. J., Rampe, E. B., et al. (2016). Silicic evidenced by tridymite in high-SiO2 sedimentary rock at Gale crater. Proceedings of the National Academy of Sciences, 113(26), 7071–7076. https://doi.org/10.1073/ pnas.1607098113 Moyen, J. F. (2011). The composite Archaean grey gneisses: Petrological significance, and evidence for a non-unique tectonic setting for Archaean crustal growth. Lithos, 123(1–4), 21–36. https://doi.org/10.1016/j.lithos.2010.09.015 Nekvasil, H., Filiberto, J., McCubbin, F. M., & Lindsley, D. H. (2007). Alkalic parental magmas for Chassignites? Meteoritics and Planetary Science, 42(6), 979–992. https://doi.org/10.1111/j.1945-5100.2007.tb01145.x Nekvasil, H., McCubbin, F. M., Harrington, A., Elardo, S., & Lindsley, D. H. (2009). Linking the Chassigny meteorite and the Martian surface rock Backstay: Insights into igneous crustal differentiation processes on Mars. Meteoritics and Planetary Science, 44(6), 853–869. https://doi.org/ 10.1111/j.1945-5100.2009.tb00773.x Nemchin, A. A., Humayun, M., Whitehouse, M. J., Hewins, R. H., Lorand, J. P., Kennedy, A., et al. (2014). Record of the ancient martian hydrosphere and atmosphere preserved in zircon from a martian meteorite. Nature Geoscience, 7(9), 638. Nyquist, L. E., Shih, C. Y., McCubbin, F. M., Santos, A. R., Shearer, C. K., Peng, Z. X., et al. (2016). Rb-Sr and Sm-Nd isotopic and REE studies of igneous components in the bulk matrix domain of Martian breccia Northwest Africa 7034. Meteoritics & Planetary Science, 51(3), 483–498. https://doi.org/10.1111/maps.12606 Ollila, A. M., Newsom, H. E., Clark, B. III, Wiens, R. C., Cousin, A., Blank, J. G., et al. (2014). Trace element geochemistry (Li, Ba, Sr, and Rb) using Curiosity’s ChemCam: Early results for Gale crater from site to Rocknest. Journal of Geophysical Research: Planets, 119, 255–285. https://doi.org/10.1002/2013JE004517 Payré, V., Fabre, C., Cousin, A., Sautter, V., Wiens, R. C., Forni, O., et al. (2017). Alkali trace elements in Gale crater, Mars, with ChemCam: Calibration update and geological implications. Journal of Geophysical Research: Planets, 122, 650–679. https://doi.org/10.1002/2016JE005201 Peslier, A. H., Hnatyshin, D., Herd, C. D. K., Walton, E. L., Brandon, A. D., Lapen, T. J., & Shafer, J. T. (2010). Crystallization, melt inclusion, and redox history of a Martian meteorite: Olivine-phyric shergottite Larkman Nunatak 06319. Geochimica et Cosmochimica Acta, 74(15), 4543–4576. https://doi.org/10.1016/j.gca.2010.05.002 Rampe, E. B., Ming, D. W., Blake, D. F., Bristow, T. F., Chipera, S. J., Grotzinger, J. P., et al. (2017). Mineralogy of an ancient lacustrine succession from the , Gale crater, Mars. Earth and Planetary Science Letters, 471, 172–185. https://doi.org/10.1016/ j.epsl.2017.04.021 Rogers, A. D., & Nekvasil, H. (2015). Feldspathic rocks on Mars: Compositional constraints from infrared spectroscopy and possible formation mechanisms. Geophysical Research Letters, 42, 2619–2626. https://doi.org/10.1002/2015GL063501 Rudnick, R. L. (1995). Making continental crust. Nature, 378(6557), 571–578. https://doi.org/10.1038/378571a0 Rudnick, R. L., & Gao, S. (2003). The composition of the continental crust. Treatise On Geochemistry, 3,1–64. Santos, A. R., Agee, C. B., McCubbin, F. M., Shearer, C. K., Burger, P. V., Tartèse, R., & Anand, M. (2015). Petrology of igneous clasts in Northwest Africa 7034: Implications for the petrologic diversity of the martian crust. Geochimica et Cosmochimica Acta, 157,56–85. https://doi.org/ 10.1016/j.gca.2015.02.023 Sarbadhikari, A. B., Goodrich, C. A., Liu, Y., Day, J. M. D., & Taylor, L. A. (2011). Evidence for heterogeneous enriched shergottite mantle sources in Mars from olivine-hosted melt inclusions in Larkman Nunatak 06319. Geochimica et Cosmochimica Acta, 75(22), 6803–6820. https://doi. org/10.1016/j.gca.2011.09.001 Sautter, V., Fabre, C., Forni, O., Toplis, M. J., Cousin, A., Ollila, A. M., et al. (2014). Igneous mineralogy at Bradbury Rise: The first ChemCam campaign at Gale crater. Journal of Geophysical Research: Planets, 119,30–46. https://doi.org/10.1002/2013JE004472 Sautter, V., Toplis, M. J., Beck, P., Mangold, N., Wiens, R., Pinet, P., et al. (2016). Magmatic complexity on early Mars as seen through a combination of orbital, in-situ and meteorite data. Lithos, 254-255,36–52. https://doi.org/10.1016/j.lithos.2016.02.023 Sautter, V., Toplis, M. J., Wiens, R. C., Cousin, A., Fabre, C., Gasnault, O., et al. (2015). In situ evidence for continental crust on early Mars. Nature Geoscience, 8(8), 605–609. https://doi.org/10.1038/ngeo2474 Schmidt, M. E., Campbell, J. L., Gellert, R., Perrett, G. M., Treiman, A. H., Blaney, D. L., et al. (2014). Geochemical diversity in first rocks examined by the Curiosity Rover in Gale crater: Evidence for and significance of an alkali and volatile-rich igneous source. Journal of Geophysical Research: Planets, 119,64–81. https://doi.org/10.1002/2013JE004481 Schmidt, M. E., Schrader, C. M., & McCoy, T. J. (2013). The primary fO2 of basalts examined by the Spirit rover in Gusev crater, Mars: Evidence for multiple redox states in the Martian interior. Earth and Planetary Science Letters, 384, 198–208. https://doi.org/10.1016/ j.epsl.2013.10.005 Siebach, K. L., Baker, M. B., Grotzinger, J. P., McLennan, S. M., Gellert, R., Thompson, L. M., & Hurowitz, J. A. (2017). Sorting out compositional trends in sedimentary rocks of the Bradbury group (), Gale crater, Mars. Journal of Geophysical Research: Planets, 122, 295–328. https://doi.org/10.1002/2016JE005195 Squyres, S. W., Aharonson, O., Clark, B. C., Cohen, B. A., Crumpler, L., De Souza, P., et al. (2007). Pyroclastic activity at home plate in Gusev crater, Mars. Science, 316(5825), 738–742. https://doi.org/10.1126/science.1139045 Stolper, E. M., Baker, M. B., Newcombe, M. E., Schmidt, M. E., Treiman, A. H., Cousin, A., et al. (2013). The petrochemistry of Jake_M: A martian . Science, 341(6153). https://doi.org/10.1126/science.1239463 Symes, S. J. K., Borg, L. E., Shearer, C. K., & Irving, A. J. (2008). The age of the martian meteorite Northwest Africa 1195 and the differentiation history of the shergottites. Geochimica et Cosmochimica Acta, 72(6), 1696–1710. https://doi.org/10.1016/j.gca.2007.12.022 Taylor, G. J. (2013). The bulk composition of Mars. Chemie der Erde - Geochemistry, 73(4), 401–420. https://doi.org/10.1016/ j.chemer.2013.09.006 Taylor, S. R., & McLennan, S. (2009). Planetary crusts: Their composition, origin and evolution (Vol. 10). Cambridge University Press. Taylor, L. A., Nazarov, M. A., Shearer, C. K., McSween, H. Y. Jr., Cahill, J., Neal, C. R., et al. (2002). Martian meteorite Dhofar 019: A new shergottite. Meteoritics & Planetary Science, 37(8), 1107–1128. https://doi.org/10.1111/j.1945-5100.2002.tb00881.x Thomson, B. J., Bridges, N. T., Milliken, R., Baldridge, A., Hook, S. J., Crowley, J. K., et al. (2011). Constraints on the origin and evolution of the layered mound in Gale crater, Mars using Mars Reconnaissance Orbiter data. Icarus, 214(2), 413–432. https://doi.org/10.1016/ j.icarus.2011.05.002

UDRY ET AL. 1539 Journal of Geophysical Research: Planets 10.1029/2018JE005602

Treiman, A. H., Bish, D. L., Vaniman, D. T., Chipera, S. J., Blake, D. F., Ming, D. W., et al. (2016). Mineralogy, , and diagenesis of a potassic basaltic on Mars: CheMin X-ray diffraction of the Windjana sample (Kimberley area, Gale crater). Journal of Geophysical Research: Planets, 121,75–106. https://doi.org/10.1002/2015JE004932 Tuff, J., Wade, J., & Wood, B. J. (2013). Volcanism on Mars controlled by early oxidation of the upper mantle. Nature, 498(7454), 342–345. https://doi.org/10.1038/nature12225 Udry, A., Balta, J. B., & McSween, H. Y. (2014). Exploring fractionation models for Martian magmas. Journal of Geophysical Research: Planets, 119,1–18. https://doi.org/10.1002/2013JE004445 Udry, A., Lunning, N. G., McSween, H. Y. Jr., & Bodnar, R. J. (2014). Petrogenesis of an impact melt clast in the martian meteorite breccia NWA 7034. Geochimica et Cosmochimica Acta, 141, 281–293. https://doi.org/10.1016/j.gca.2014.06.026 Usui, T., McSween, H. Y., & Clark, B. C. (2008). Petrogenesis of high-phosphorous Wishstone Class rocks in Gusev Crater, Mars. Journal of Geophysical Research, 113, E00B17. https://doi.org/10.1029/2008JE003225 Usui, T., O’D Alexander, C. M., Wang, J., Simon, J. I., & , J. H. (2012). Origin of water and mantle-crust interactions on Mars inferred from hydrogen isotopes and volatile element abundances of olivine hosted melt inclusions of primitive shergottite. Earth and Planetary Science Letters, 357–358, 119–129. Vaniman, D. T., Bish, D. L., Ming, D. W., Bristow, T. F., Morris, R. V., Blake, D. F., et al. (2013). Mineralogy of a mudstone at Yellowknife Bay, Gale crater, Mars. Science, 1243480. https://doi.org/10.1126/science.1243480 Vasavada, A. R., Grotzinger, J. P., Arvidson, R. E., Calef, F. J., Crisp, J. A., Gupta, S., et al. (2014). Overview of the Mars Science Laboratory mission: Bradbury Landing to Yellowknife Bay and beyond. Journal of Geophysical Research: Planets, 119, 1134–1161. https://doi.org/10.1002/ 2014JE004622 Wheater, C. P., & Cook, P. A. (2006). Using statistics to understand the environment (245 p.). Routledge, New York. Whitaker, M. L., Nekvasil, H., and Lindsey, D. H. (2005). Potential magmatic diversity on Mars, Lunar Planetary Sci. [CD-ROM], XXXVI, Houston, Abstract 1440. Willbold, M., Hegner, E., Stracke, A., & Rocholl, A. (2009). Continental geochemical signatures in dacites from Iceland and implications for models of early Archaean crust formation. Earth and Planetary Science Letters, 279(1–2), 44–52. https://doi.org/10.1016/j.epsl.2008.12.029 Wray, J. J., Hansen, S. T., Dufek, J., Swayze, G. A., Murchie, S. L., Seelos, F. P., et al. (2013). Prolonged magmatic activity on Mars inferred from the detection of felsic rocks. Nature Geoscience, 6, 1013–1016. https://doi.org/10.1038/ngeo1994 Yin, Q.-Z., McCubbin, F. M., Zhou, Q., Santos, A. R., Tartese, R., Li, X., et al. (2014). An Earth-like beginning for ancient Mars indicated by alkali- rich volcanism at 4.4 Ga. Lunar Planet. Sci. [CD-ROM], XLV, Houston, Abstract 1320. Yingst, R. A., Kah, L. C., Palucis, M., Williams, R. M. E., Garvin, J., Bridges, J. C., et al. (2013). Characteristics of pebble-and cobble-sized clasts along the curiosity rover traverse from Bradbury Landing to Rocknest. Journal of Geophysical Research,: Planets, 118, 2361–2380. https://doi.org/10.1002/2013JE004435 Zipfel, J., Schroder, C., Jolliff, B. L., Gellert, R., Herkenhoff, K. E., Rieder, R., et al. (2011). -a shergottite-like basalt encountered at , Mars. Meteoritics & Planetary Science, 46(1), 1–20.

UDRY ET AL. 1540