Mineralogy and diagenesis of Mars-analog paleosols from eastern Oregon, 1 USA 2

Ancient (4.1-3.7-billion-year-old) layered sedimentary rocks on Mars are rich in clay minerals which formed from aqueous alteration of the Martian surface. Many of these sedimentary rocks appear to be composed of vertical sequences of Fe/Mg clay minerals overlain by Al clay minerals 33 that resemble paleosols (ancient, buried soils) from Earth. The types and properties of minerals in paleosols can be used to constrain the environmental conditions during formation to better 35 understand weathering and diagenesis on Mars. This work examines the mineralogy and diagenetic 36 alteration of volcaniclastic paleosols from the Eocene-Oligocene (43-28 Ma) Clarno and John Day 37 Formations in eastern Oregon as a Mars-analog site. Here, paleosols rich in Al phyllosilicates and 38 amorphous colloids overlie paleosols with Fe/Mg smectites that altogether span a sequence of ~500 39 individual profiles across hundreds of meters of vertical stratigraphy. Samples collected from three 40 of these paleosol profiles were analyzed with visible/near-infrared (VNIR) spectroscopy, X-ray 41 diffraction (XRD), and evolved gas analysis (EGA) configured to operate like the SAM-EGA 42 instrument onboard Curiosity Mars Rover. (oxy)hydroxides, respectively. These and other forms of diagenesis are common in terrestrial paleosols that formed from weathering of volcaniclastic sediments 18 798 and may also explain occurrences of zeolites 92 and hematite 51 detected from orbit in putative weathering 799 sequences on Mars. 800 This study provides a protocol for constraining climate and habitability from the geochemistry of 801 weathering profiles on Mars. Molecular weathering ratios that have been well studied in terrestrial 802 paleosols (e.g., 31,93 ) could be useful for interpreting climate and habitability of weathering profiles on 803 Mars. By using a suite of molecular weathering ratios and geochemical climofunctions (Figure 4), a 804 reconstruction of the climate and nature of weathering can be inferred from weathering profiles on Mars. 805 However, differences in the nature of weathering and diagenesis between Earth and Mars present 806 challenges for making direct comparisons. Such differences include a presumably anoxic early Mars 807 atmosphere that perhaps led to Fe 2+ mobility during subaerial weathering 24 , and the apparent absence of 808 plate tectonics which has implications for the nature and severity of diagenesis of weathering profiles on 809 Mars 3,94 . One additional consideration is application of the chemical index of alteration to weathering 810 profiles on Mars that were subject to weathering by acidic and sulfur-rich fluids. Weathering indices such 811 as CIA may not accurately reflect acid sulfate weathering of mafic Fe/Mg rich sediments because 812 weathering rates of mafic materials such as olivine proceeds more efficiently than feldspars, especially 813 under acidic conditions 95 . In addition, acidic conditions also affect the mobility of alkaline elements which may further confuse interpretations of weathering intensity by examining CIA 51 . Martian weathering profiles that were alterated by fluids with circumneutral pH are better candidates for application of molecular weathering ratios, weathering indices, and geochemical climofunctions to terrestrial paleosols. from the Clarno and lower John 897 Day Formations. Pedogenic features observed in this work include dioctahedral smectite mineralogy, a clay mineral doublet feature observed with VNIR spectroscopy possibly resulting from isomorphous 900 substitutions during pedogenic weathering, destruction of sedimentary bedding, sub-meter scale 901 differences in composition and color, and illuvial accumulation of clay minerals into subsurface horizons, all of which resulted from precipitation-driven pedogenic weathering of andesitic to rhyodacitic volcanic 903 ash and tuff. Results from this work can help distinguish paleosols and weathering profiles from other 904 types of sedimentary rocks in the geological record of Mars.


Introduction 58
Today the surface of Mars is frigid, wind-deflated and barren, but there is extensive geological 59 evidence for transient warm and wet habitable surface conditions in the Noachian (4.1-3.7 Ga) period of 60 early Mars [1][2][3][4] . Orbital sensing of the Martian surface has revealed clay mineral deposits in thousands of 61 locations, wherever Noachian-age terrains are not obscured by dust, sand, or overlying strata 5-8 . These 62 ancient deposits are rich in smectite clay minerals and other hydrated phases, suggesting formation in 63 subsurface and surface environments from the weathering of mafic rocks and sediments with liquid water. 64 While some phyllosilicates on Mars are associated with lacustrine deposition in surface 65 environments (e.g., 9 ) many phyllosilicate detections occur in regionally widespread deposits, inconsistent 66 with deposition in basin settings. Two hypotheses to explain the formation, occurrence and distribution of 67 these regionally widespread phyllosilicates on Mars are: 1) subsurface hydrothermal activity, diagenesis 68 and/or authigenesis 10-14 and, 2) surface pedogenic alteration (e.g., subaerial chemical weathering) 2,6,14,15 . 69 In some locations, trioctahedral smectites exhibiting lateral variations in Al and Fe/Mg smectites 70 intermixed with chlorite, serpentinite, talc and zeolite are consistent with formation in hydrothermal 71 subsurface environments, diagenesis, and/ or authigenesis during sediment emplacement 13,16,17 . However, 72 dioctahedral smectites often outcrop as extensive vertical profiles of Fe/Mg smectites overlain by Al 73 smectites, suggesting subaerial formation in surface environments, consistent with pedogenesis (soil 74 formation) or large-scale leaching of sediments 14 . 75 Paleosols are ancient, buried soils that are lithified into sedimentary rocks. On Earth, paleosols 76 are a geological record of the atmospheric composition, climate, topography and organisms present before 77 soil burial 18 . Paleosols are created by removal from their soil-forming factors, sometimes because of 78 change in those factors, but most often by rapid burial. The deposition of volcanic ash, flood basalts, 79 sedimentation from flooding, landslides, and tsunamites all rapidly bury surface environments. Sequences 80 of paleosols form when soils are periodically buried, for example by repeated volcanic eruptions which 81 emplace tephra or lava onto soil surfaces, followed by successive pedogenic weathering of emplaced 82 tephra or lava, and then subsequent burial by another eruption. Flooding by rivers also buries paleosols 83 within alluvial sequences, and dune migration buries paleosols within eolian sequences. On Earth, 84 sequences of paleosols formed by periodic burial can record weathering, paleoclimate and diagenetic 85 alteration over 10 7 year timescales 19,20 . 86 Paleosols can be useful tools for interpreting ancient climates of Earth and Mars 21-23 . The types 87 and properties of minerals in Mars-analog paleosols can be used to help understand the nature of 88 weathering and diagenesis on Mars 24,25 but they remain relatively understudied as Mars-analog sites. The 89 objective of this study was to examine the mineralogy and diagenetic alterations of Mars-analog paleosols 90 from eastern Oregon, USA using analytical techniques similar to those onboard current and future 91 missions to Mars.

128
The stratigraphic level of paleosols analyzed in this work is indicated (black arrow In this study, we examined early Oligocene (33 Ma) paleosol profiles from the Mars-analog paleosol 229 sequence in eastern Oregon (Figure 2). Orbital and in-situ visible-near infrared (VNIR) spectroscopic 230 techniques are used by current and future missions to Mars and are a useful tool for identifying minerals 231 and diagenetic features across the global Martian surface 5,52,53 . Here, in-situ visible-near infrared 232 spectroscopy was performed to identify clay minerals, zeolites and oxides in bulk paleosol samples. The 233 Curiosity Mars rover employs in-situ X-ray diffraction (XRD) for identification and quantification of 234 crystalline minerals in rocks, sediments and soils on Mars 54,55 . Qualitative X-ray diffraction was used in 235 this study to identify crystalline minerals in bulk paleosol samples from the three consecutive paleosol 236 profiles. Lastly, samples were analyzed with an instrument calibrated to use analytical conditions similar 237 to the Sample Analysis at Mars evolved gas analysis (SAM-EGA) instrument onboard the Curiosity rover 238 56 . The purpose of this analysis was to constrain the mineralogy of hydrated phases in samples, 239 specifically by examining evolutions of H2O and SO2 from bulk paleosol samples during heating. Since 240 thermal techniques such as evolved gas analysis will fly onboard future missions to Mars (e.g. ExoMars 241 2022 Rosalind Franklin rover 57 ), a detailed characterization of terrestrial paleosol mineralogy via evolved 242 gas analysis can help constrain the formation mechanism(s) of dioctahedral clay-bearing sedimentary 243 rocks on Mars. 244 (e.g., badland toeslope) that may be accessible for in-situ analysis on Mars by future landed missions 255 including rovers and /or astronauts. Furthermore, previous 40 Ar/ 39 Ar dating of volcanic tuffs above and 256 below the sampling location allowed for a constrained age of 33.0 +/-0.10 to 32.7 +/-0.03 Ma (Biotite 257

Methods
Tuff and Overlook Tuff, respectively) 28 . 258 Samples were collected by first trenching to ~30 cm into the outcrop to remove the modern 259 weathering zone and to expose the underlying claystone paleosols. This was followed by sampling with a 260 rock hammer down a vertical transect (parallel with the hillslope) at approximately 10 cm intervals, 261 similar to sampling the horizons of a modern soil profile. Large, ~0.2 kg lithified blocks were removed 262 from the brick-like paleosol surface for mineralogical analyses. The morphology, Munsell color and 263 qualitative calcareousness of samples were also described during collection (Figure 3). 264 The same paleosol profiles were previously analyzed for bulk geochemistry by Retallack et al. 265 (2000) and those values were used to calculate chemical index of alteration and molar weathering ratios to 266 augment our mineralogical assessment of the same samples (see Results microns. Spectra were not gathered from the thinnest and least developed paleosol (Entisol), but all three 298 paleosol types were subject to X-ray diffraction and evolved gas analysis, discussed below. 299

X-ray diffraction of paleosol samples 300
Paleosols were powdered and homogenized to < 45 μm grain size, then unoriented samples were 301 mounted on aluminum holders and measured using a PANalytical X'pert Pro MPD XRD at NASA 302 Johnson Space Center. XRD patterns were collected using an X'celerator detector and Co Kα X-ray 303 source, with a Fe filter to reduce Kβ peak intensities. Samples were analyzed under the following 304 conditions: 45 kV, 40 mA, ½° antiscatter slit, ¼° fixed divergence slit, and a beam knife to reduce low-305 angle scattering. Samples were measured from 4° to 80° 2θ with a 0.02° step size at 100s/step. Mineral 306 identification was accomplished using HighScore and Jade MDI software by comparing XRD patterns to 307 International Center for Diffraction Data (ICDD) database patterns, and with Crystallography Open 308 Database (COD) patterns. 309 Semi-quantitative XRD (no internal standard) was used to provide a relative estimate of phyllosilicate 310 abundances. Rietveld refinement 59 was carried out using MDI Jade Software with initial structure 311 parameters for crystalline phases from the RRUFF database (http://rruff.info/). Background patterns were 312 fit by a polynomial and peaks were modeled by a pseudo-Voigt profile function. Pattern overlays of 313 standard phyllosilicates with known relative intensity ratio (RIR) and full width at half maximum 314 (FWHM) from the Clay Minerals Society 60 were used in Rietveld refinements to estimate abundances of 315 phyllosilicates in bulk samples. Since no internal standard was used, semi-quantitative XRD was used 316 only for estimating the relative abundances of phyllosilicates in each sample (Table S4). 317

Thermal and evolved gas analysis (EGA) of paleosol samples 318
A Setaram Labsys Evo differential scanning calorimeter (DSC) / thermal gravimeter (TG)  319 connected to a Pfeiffer Omnistar quadrupole mass spectrometer (QMS) was configured to operate 320 similarly to the SAM evolved gas analyzer. The Sample Analysis at Mars (SAM) onboard Curiosity Mars 321 Rover does not have TG/DSC capabilities, but these components permit a better understanding of phase 322 transitions and chemical reactions in laboratory experiments. Approximately 50 mg ± 3 mg of ground 323 paleosol sample were placed in an Al2O3 sample crucible which was previously ashed at 500° C to 324 remove organic contaminants before use. The sample crucible and an identical empty reference crucible 325 were placed in the furnace and then the system was purged twice with helium gas and then set to a 326 pressure of 30 mbar. Helium was chosen as a carrier gas because it is inert and because it used as a carrier 327 gas in the SAM instrument. The crucibles were heated from approximately 35 °C to 1000 °C at a heating 328 rate of 35°C/min and at a flow rate of 10 sccm. Volatiles ranging from mass/charge (m/z) 1 -100 were 329 measured. All analyses were performed in duplicate. 330 Evolved water abundances were determined using a Netzsch TG/DSC coupled to a Pfeiffer QMS. An 331 Al2O3 sample crucible and an identical reference crucible were placed in the furnace. The instrument was 332 purged twice with ultra-high purity O2 and set to a pressure of 1000 mbar prior to sample analyses to 333 remove any contamination in the system. Leitz Orthoplan Pol research microscope. Accuracy of point counting was determined to be ± 2 wt. % for 353 common constituents. Bulk density was also previously determined by the clod method 18 first by 354 determining raw weight, then weight of clods coated in paraffin of known density (0.86g cm -3 ) in and out 355 of chilled (6°C) water (1.00 g cm -3 ). Major element chemistry of paleosols was determined by X-ray 356 fluorescence, atomic absorbance, and X-ray diffraction at Washington State University, Pullman (Table  357 S1). These previously published data were used to calculate chemical index of alteration and molar 358 weathering ratios of each paleosol profile examined in this work. Major element chemistry from Retallack 359 et al. (2000) and calculated CIA and molar weathering ratios are both included as supplementary data 360 (Table S1). 361 showed slight gleyization (FeO/Fe2O3 > 0) most likely from the onset of chemically reducing 387 conditions shortly after burial. This has been attributed to a diagenetic phenomenon known as burial 388 gleization (discussed in detail in Section 3.5) that typically affects the organic matter-rich surface 389

Results and Discussion
horizons of rapidly buried paleosols. 390 The chemical index of alteration (CIA) ranged from 72.1 to 80.8 and generally decreased with 391 depth across the Alfisol and Inceptisol (Figure 4). The highest values (80.8) were in the subsurface 392 clay (Bt) horizon of the Alfisol and lowest (72.1) in the C-horizon of the lowermost Inceptisol. The 393 CIA in clayey paleosols is generally highest in subsurface horizons due to illuvial accumulation of 394 clay minerals during top-down hydrolytic weathering and therefore these horizons are thought to be 395 the most reliable for paleoclimate estimations 31 . In contrast, the unweathered, lowermost C and R 396 horizons of paleosols (e.g., saprolite) reflect the characteristics of the soil parent material rather than 397 alteration from weathering, and thus CIA is lower in these horizons. The thinnest and least developed 398 paleosol (Entisol) preserves a parent material of redeposited tuffaceous clayey siltstone that was 399 minimally altered by soil formation, inferred from relict bedding in the C-horizon. The high clay 400 content of this paleosol (~75 wt. %) was most likely inherited from preexisting soils by sheet erosion 401 or flooding. It is unlikely that the Entisol was developed for long enough to develop characteristics 402 indicative of paleoclimate, and therefore estimations of CIA are unreliable and not shown. 403 404 405

Visible/ near infrared spectroscopy of paleosols 411
The mineralogy of paleosols observed with VNIR spectroscopy was dominated by dioctahedral 412 phyllosilicates and occasionally zeolites and hematite. All paleosol samples had strong spectral signatures 413 of 2:1 Al/Fe dioctahedral smectites ( Figure 5)

456
The spectral "doublet" feature with bands near 2.21 and 2.23-2.25 µm was observed in all 457 samples and was attributed to OH stretching and bending combination vibrations in phyllosilicates 458 ( Figure 6).

491
Beginning with the stratigraphically highest sample (Alfisol 4 cm), major phases identified from 492 patterns were montmorillonite and nontronite while minor phases identified were andesine, Opal-CT and 493 cristobalite, with lesser abundances of clinoptilolite, quartz, gypsum, jarosite, and anatase (Figure 7). 494 Stratigraphically below this sample, the Alfisol A-horizon at 14 cm showed a similar mineral assemblage 495 with the addition of albite as a minor phase (Table 1). A deeper sample of the Alfisol at 46 cm (Bt 496 horizon) had montmorillonite and nontronite as major phases and albite, cristobalite, opal-CT, 497 clinoptilolite, hematite, quartz and anatase as minor phases. Directly below, the pattern from Entisol A-498 horizon at 7 cm was consistent with montmorillonite and nontronite as major phases and cristobalite, 499 anatase, clinoptilolite, and orthoclase as minor phases (Figure 7). Stratigraphically below this sample, the 500 pattern for the near-surface (A) horizon of the Inceptisol (3 cm) showed montmorillonite, nontronite and 501 Opal-CT as major phases and clinoptilolite, cristobalite, andesine, orthoclase, quartz, and gypsum as 502 minor phases. The pattern from the stratigraphically lowest sample, the Inceptisol Bw-horizon at 21 cm, 503 showed the same major phases as the A-horizon (3 cm) sample, but with the additions of hematite, 504 ilmenite (FeTiO3) and anatase along with clinoptilolite, cristobalite, quartz and andesine as minor phases. 505 unweathered volcanic glass and/or poorly ordered phases) which presumably was diagenetically altered 516 via zeolitization to clinoptilolite. There were no XRD detections of illite which suggests minimal or 517 absent potash metasomatism despite burial by an estimated ~2 km of overburden 69 . 518 519 Evolved water abundances ranged from 3.24 ± 0.47 wt. % to 5.03 ± 0.12 wt. % H2O across all samples (n 527 = 20) and trends in water abundances were apparent across the three paleosols. The Alfisol evolved the 528 lowest amount of water observed with 3.24 ± 0.47 in the subsurface (Bt) horizon whereas the Entisol 529 averaged 5.03 ± 12 wt. % which was the highest amount of evolved water observed in the experiment. 530

Evolved Gas Analysis
The Inceptisol ranged from 4.32 ± 0.03 wt. % to 4.81 ± 0.13 wt. % H2O and showed a trend of decreasing 531 abundance with depth. Despite significant differences in evolved water abundances between profiles, each 532 profile generally showed a decrease in evolved water abundance with depth. Though many factors control 533 the abundance and persistence of hydrated phases in paleosols, trends in evolved water abundances could 534 have resulted from lateral and vertical diversity in mineralogy within each of the paleosol profiles.

Alfisol (4 cm) Alfisol (14 cm) Alfisol (46 cm) Entisol (3 cm) Entisol (7 cm) Inceptisol (3 cm) Inceptisol (7 cm)
Differences in evolved H2O peaks < 450 °C across the three paleosols examined here may have 544 resulted from differences in abundance and composition of x-ray amorphous components. Though 545 amorphous phase composition was not examined in this work, a previous study identified basaltic glass 546 (5.6 wt. %), allophanes (7.9 wt. %) and ferrihydrite (0.6 wt. %) as the dominant amorphous phases in 547 paleosols from the Oligocene (~28 Ma) Turtle Cove member of the John Day Formation (Smith et al.,548 2018) which are stratigraphically higher than paleosols examined here (Figure 1) there were no sharp water release peaks at ~290 °C (Figure 8), possibly because samples are composed 558 primarily of strongly crystalline clay minerals rather than amorphous colloids, especially the Alfisol with 559 up to 95 wt. % clay minerals ( Figure 3). However, a minor 300 °C endotherm in all samples is consistent 560 with small amounts of amorphous phases, which were also detected with XRD as minor phases in the 561 Entisol and Inceptisol (Table 1). In contrast to the moderately weathered Alfisol, the Entisol and 562 Inceptisol were only minimally weathered before burial, inferred from morphological features such as 563 absence of clay illuviation and persistence of relict bedding in subsurface horizons ( Figure 3). As such, 564 differences in the duration of weathering before burial can explain the persistence of amorphous phases in 565 the Entisol and Inceptisol and absence of amorphous phases in the Alfisol. 566 Overall

581
Evolutions of H2O above 450° C are consistent with the dehydroxylation of the octahedral layer 582 of a 2:1 phyllosilicate 71, 75 , and there was a ~200° C difference in H2O peak release temperature between 583 the three paleosols. The Alfisol with ~95 wt. % smectite evolved H2O with peaks centered at ~500° C, 584 while the Entisol (~75 wt. % smectite) and the Inceptisol (~78 wt. % smectite) evolved H2O with peaks 585 centered at ~ 700 °C. Differences in clay mineralogy between the soils may be responsible for shifting the 586 peaks and shoulders of the high-temperature (> 450° C) evolutions. The considerable difference in peak 587 H2O release temperature between the red Alfisol (~500° C) and the other soils (~700° C, Figure 8) could 588 have resulted from differences in the occupation of the octahedral sheet of a 2:1 phyllosilicate 70 which 589 leads to differences in the high temperature H2O peak release temperature during SAM-EGA 76,77 . Clay 590 minerals with Fe in the octahedral layer (e.g., nontronite or Fe-montmorillonite) dehydroxylate at a lower 591 temperature (~500° C) relative to smectite with Al in the octahedral layer (e.g., Al-montmorillonite, 700° 592 C ) 77 , though mixed illite-smectite mineralogy also show peak H2O release at 698° C. The Alfisol (~500° 593 C peak H2O release) exhibited strong VNIR signatures of an Al-smectite with minor Fe-smectite and 594 hematite. The brown Inceptisol (~700° C peak H2O release) had strong VNIR signatures of an Al/Fe 595 smectite and lacked hematite ( Figure 5). Differences in clay mineralogy across the samples likely caused 596 the large difference (~200° C) of peak water release temperatures from smectite dehydroxylation. 597 Together these results show that EGA in conjunction with XRD and VNIR spectroscopy are suitable 598 techniques to constrain smectite mineralogy in paleosols. 599 600 3.4.2 SO2 evolutions 601 All samples evolved minor amounts of SO2 primarily above 450° C (Figure 9). Minor SO2 peaks 602 below 400° C were observed in the Bt-horizon of the Alfisol and the A-horizon of the Entisol (Figure 9) 603 which most likely resulted from instrument background sources. A distinct SO2 peak at 400° C in the A-604 horizon of the Entisol is consistent with the presence of minor amounts of sulfides such as pyrite and/or 605 pyrrhotite which thermally decompose at temperatures above 400° C under SAM-EGA analog conditions 606 78 . Oxidative sulfite decomposition directly to SO2 could have resulted from trace amounts of oxygen in 607 the instrument furnace even after successive purges with helium and are the likely source of the broad 608 400° C SO2 peak noted in the Entisol. 609

614
Evolutions of SO2 above 500° C are consistent with the thermal decomposition of Ca and Fe 615 sulfates ranging from crystalline (gypsum and jarosite) to amorphous and/or adsorbed sulfate 73,78 . Also 616 possible are contributions from organo-sulfur compounds and/or S phase inclusions in volcanic glass. 617 Crystalline sulfate species including jarosite and gypsum have peak SO2 release temperatures near 900° C 618 and 1200° C, respectively 74 . Thus, jarosite and gypsum most likely account for the evolutions of SO2 619 above 600° C and both were confirmed as minor phases with XRD (Table 1). Samples from the Bt and C 620 horizons of the Alfisol had a broad SO2 release with a peak at ~790° C that was absent in the Entisol and 621 Inceptisol. Trace amounts of Mg sulfates in the Alfisol could account for minor SO2 releases > 700° C 622 including the ~790° C SO2 peaks 71,79 At higher temperatures, all soils showed a major release of SO2 623 beginning at 900° C which co-occurred with an endotherm, both of which are consistent with the thermal 624 decomposition of crystalline sulfates 75,79 . Since the samples were only heated to ~1000° C for this work, 625 the maximum peak height of this release cannot be ascertained. 626 The presence of sulfate minerals is uncommon in smectite-rich soils such as those examined here. 627 Sulfate minerals such as gypsum and jarosite tend to form at low pH and low water:rock ratios whereas 628 pedogenic smectites such as montmorillonite typically form at circumneutral pH and increased water:rock 629 ratios. Both gypsum and jarosite are unlikely to be original minerals in the paleosols, but more likely 630 formed in the current weathering zone. One possibility for the origin of sulfate minerals in these paleosols 631 is leaching from the modern soils forming atop paleosols (e.g., the weathered paleosol surface). These 632 modern soils that mantle paleosol outcrops have visibly accumulated pedogenic gypsum into a thin (~1 633 cm) subsurface gypsic (By) horizon, allowing for their classification as Gypsids (gypsum-rich desert 634 soils) in US Soil Taxonomy  an early diagenetic alteration because water-rock interactions during pedogenesis alters the physical and 644 chemical properties of sediments. After burial, however, soils are subject to additional early and late 645 diagenetic alterations ranging from minor (e.g., decomposition of organic matter) to severe (e.g., 646 metamorphic alteration). Four types of alteration after burial that have affected paleosols examined in this 647 work are 1) Drab olive-green surface horizons attributed to burial gleization; 2) brick-red color from 648 burial-induced dehydration of ferric hydroxide minerals; 3) zeolitization of volcanic glass and/or poorly 649 crystalline phases; and 4) significant mechanical compaction. 650 Burial gleization, also known as gley overprinting 80 , has been envisaged as the chemical 651 reduction of iron oxides and hydroxides by anaerobic bacteria in the near-surface horizons of paleosols 652 and is thought to occur shortly after soil burial 81,82 . Burial gleization is an early diagenetic process in 653 paleosols which involves the reduction of Fe 3+ to Fe 2+ in clays, oxides and other minerals after rapid 654 burial, and promotes anaerobic decay of organic matter 18 , even in soils that originally formed under 655 oxidizing conditions before burial. Typical burial gleization is closed system alteration, without depletion 656 of total iron, and is usually limited to the surface (A) horizons where organic matter is most concentrated. 657 The surface (A) horizon of the uppermost paleosol (Alfisol) examined in this work showed classic 658 evidence of burial gleization with drab-colored mottles and tubular features predominantly in the A-659 horizon with minor radiation downward into the subsurface (Bt) horizon ( Figure 3) as well as 660 accumulations of Fe 2+ exclusively in the A-horizon (Table S1). Spectroscopic techniques such as VNIR 661 employed here can readily identify burial gleization by observing absorbance features attributed to Fe 2+ in 662 the surface horizons of paleosols. This was presumably the cause of the broad 1.1-micron Fe 2+ band in the 663 surface of the Alfisol, and the shoulders at 2.23-2.25 microns noted in the gleyed surface horizon of the 664 Alfisol (Figures 5 and 6). 665 Burial gleization can be distinguished from other original redoximorphic features such as 666 groundwater alteration because it is limited to the surface (A) horizon unlike the gleyed subsurface (Bg)  667 horizons of seasonally or perennially waterlogged soils, and because the mineralogy and morphology of 668 this soil provide evidence of well-drained and oxidizing conditions during soil formation. Morphological 669 evidence of burial gleization in paleosols is a drab greenish-gray color exclusively in near-surface 670 horizons directly below the burial contact 82 . Geochemical evidence of burial gleization can include 671 depletions of Fe 3+ coupled with increases in Fe 2+ and the formation of siderite and pyrite. Most soils 672 accumulate organic matter in near-surface horizons, which after burial and the onset of anoxic conditions 673 is generally the horizon most affected by burial gleyization. Though the timing of burial gleization 674 remains poorly constrained, reduction haloes around buried organic matter such as roots can form in tens 675 to hundreds of years after burial 83 . In the Alfisol, burial gleyization can be distinguished from 676 groundwater alteration because it is limited to the surface horizon directly below the white biotite-bearing 677 tuff layer which buried the uppermost soil (Figure 3). The original color of the surface horizon of the 678 Alfisol was most likely brown and darkened by accumulation of organic matter. Burial gleization thus 679 indicates the surface horizon of the Alfisol may be enriched in organic matter relative to unaffected 680 subsurface horizons (Bt and C horizons), but burial degradation of organic carbon over geological time 681 scales has likely reduced the original organic carbon content by up to two orders of magnitude 43 . 682 Directly below the drab surface horizon, the remainder of the Alfisol profile was brick-red in 683 color (Figure 3, Munsell 10R 3/1) which most likely resulted from dehydration of iron oxyhydroxides 84 . 684 This phenomenon, also known as "burial reddening", is one of the most common types of diagenetic 685 alteration in the fossil record of soils Other factors which cause reddening of paleosols include heating by lava flows 85 which can also cause 696 iron oxyhydroxides to dehydrate to hematite and maghemite, but the effects of reddening are typically 697 limited to the near-surface horizons due to the thermal insulation properties of soils 86 . Another possibility 698 is that the protolith of the soil was red in color, but relict bedding and volcanic shards in the subsurface 699 (C) horizon ( Figure 3) imply a tuffaceous parent material unlikely to be red in color. 700 The presence of clinoptilolite detected with XRD in all samples (Table 1)  First, the aqueous flushing model by Hay (1963) 88 proposed that volcanic glass in the lower John Day 706 Formation was diagenetically altered to zeolites by open-system interactions with groundwater that added 707 Ca and H2O while removing Si, Na and K. Zeolitization may have occurred during early Miocene deep 708 burial at depths of 380-1200 meters and temperatures of 27-55° C during deposition of the overlying 709 Mascall and Rattlesnake Formations. Though the timing of zeolitization is well constrained, the 710 mechanism of alteration by aqueous flushing is not likely because many silty tuff beds in the John Day 711 Formation escaped zeolitization while the clay mineral-rich paleosols did not 29 . These permeable silty 712 tuff beds would have presumably been channels for groundwater, but they include both zeolitized and 713 non-zeolitized segments along strike, suggesting a more localized or patchy distribution of zeolitized 714 facies. These observations are consistent with the hypothesis that some tuffs and paleosols were subject to 715 zeolitization, while others were not. 716 Rather than zeolites forming from groundwater alteration, diagenesis may have been more localized 717 and the formation of clinoptilolite may have instead resulted from burial-induced recrystallization of 718 amorphous colloids and/or volcanic glass 89,90 . The "burial ripening" model of Retallack (2000)  which perhaps makes groups of individual paleosol profiles appear as "units" with similar mineralogy. 774 Observations of the Oregon paleosol sequence suggest that changes in climate (e.g. Eocene-Oligocene 775 cooling and drying) led to stratigraphic changes in mineralogy rather than the intense and uninterrupted 776 leaching that is characteristic of a deep weathering profile. Instead, the Oregon paleosol sequence reflects 777 a cycle of continuous soil formation and burial that occurred during ~15 million years of climate change. 778 These changes are reflected in the tens of individual paleosol profiles composing the basal Fe/Mg 779 smectite and oxide unit (Clarno Formation, Figure 1)  This study provides a protocol for constraining climate and habitability from the geochemistry of 801 weathering profiles on Mars. Molecular weathering ratios that have been well studied in terrestrial 802 paleosols (e.g., 31,93 ) could be useful for interpreting climate and habitability of weathering profiles on 803 Mars. By using a suite of molecular weathering ratios and geochemical climofunctions (Figure 4), a 804 reconstruction of the climate and nature of weathering can be inferred from weathering profiles on Mars. 805 However, differences in the nature of weathering and diagenesis between Earth and Mars present 806 challenges for making direct comparisons. Such differences include a presumably anoxic early Mars 807 atmosphere that perhaps led to Fe 2+ mobility during subaerial weathering 24 , and the apparent absence of 808 plate tectonics which has implications for the nature and severity of diagenesis of weathering profiles on 809 Mars 3,94 . One additional consideration is application of the chemical index of alteration to weathering 810 profiles on Mars that were subject to weathering by acidic and sulfur-rich fluids. Weathering indices such 811 as CIA may not accurately reflect acid sulfate weathering of mafic Fe/Mg rich sediments because 812 weathering rates of mafic materials such as olivine proceeds more efficiently than feldspars, especially 813 under acidic conditions 95 . In addition, acidic conditions also affect the mobility of alkaline elements 814 which may further confuse interpretations of weathering intensity by examining CIA 51 . Martian 815 weathering profiles that were alterated by fluids with circumneutral pH are better candidates for 816 application of molecular weathering ratios, weathering indices, and geochemical climofunctions 817 commonly applied to terrestrial paleosols. 818 The mineralogy and diagenetic alteration of paleosols also has implications for biosignature 819 preservation in Martian weathering profiles. Biosignatures in paleosols can include biomarkers, 820 biominerals, macro and microstructures and textures, chemistry, and isotopes 26 , and the rapid burial that 821 characteristically entombs paleosols often creates favorable taphonomic environments for the preservation 822 of biosignatures. However, the preservation of chemical and isotopic biosignatures often relies on the 823 bulk abundance of organic matter preserved in a sample. Many factors contribute to the preservation and 824 degradation of organic matter in terrestrial paleosols including redox state prior to burial, clay 825 mineralogy, amorphous phase compositions and abundance, diagenetic alterations, and interactions with 826 sulfur (e.g., sulfurization) 15,43 . Redox state provides a first-order control on the preservation of organic 827 carbon in rapidly buried soils 43 ; for example, soils forming under reducing conditions (e.g., wetlands) 828 generally preserve higher abundances of organics relative to those forming in oxidized, well-drained 829 conditions. Oxidized and Al-smectite rich paleosols such as those examined here are associated with 830 longer organic carbon residence time relative to kaolin group clays 96 , but well-drained, oxidizing 831 conditions before burial are associated with severe losses of organic C after burial 18,43 . In addition, most 832 types of diagenetic alterations commonly observed in terrestrial paleosols are associated with the 833 degradation of organic matter. Illitization, zeolitization and celadonization may facilitate desorption of 834 organic carbon held on mineral surfaces, interlayer spaces, and crystal edges, thus possibly contributing to 835 burial-induced degradation of organic matter 97,98 , though several authors have reported that early 836 diagenetic smectite-illite transformation may be facilitated by microbes 98-100 which has implications for 837 biosignature preservation in illite-rich soils. In this work, illitization and celadonization of smectite was 838 not observed, but zeolitization of amorphous colloids and/or poorly crystalline smectite may have 839 liberated adsorbed or chemisorbed organic carbon 101 and likely contributed to the degradation of the bulk 840 organic fraction. One the other hand, diagenetic features of paleosols such as burial gleization observed in 841 this work (Table S1 and Figure 3) may indicate organic carbon enrichment and the preservation of 842 chemical biosignatures in the surface horizons of paleosols. The drab green surface layer of the uppermost 843 paleosol examined in this work showed an accumulation of Fe 2+ attributed to diagenetic burial gleization 844 via anaerobic microbial decay of organic matter. Previous investigations showed that this gleyed layer 845 was enriched in organic carbon relative to deeper layers in the paleosol 43 , and thus burial gleization 846 features most likely constitute a chemical biosignature in paleosols. If features resembling burial 847 gleization are detected in upper layers of weathering profiles on Mars, they should be considered a high-848 priority location for in-situ biosignature investigation. 849

850
Conclusions 851 The objective of this study was to analyze the mineralogy and diagenetic alterations of paleosols 852 from eastern Oregon, USA using techniques similar to those utilized by current and future missions to 853 Mars. Samples were gathered from three successive paleosol profiles in the early Oligocene (33 Ma) 854 middle Big Basin Member of the John Day Formation that formed from pedogenic weathering of volcanic 855 ash and tuff. Visible/near infrared spectroscopy, X-ray diffraction and evolved gas analysis confirmed 856 dioctahedral smectite was the major phase in all samples, with most samples primarily containing a 857 mixture of montmorillonite (Al smectite) and nontronite (Fe smectite). Minor phases detected with x-ray 858 diffraction included Opal-CT, cristobalite, andesine and gypsum. All samples contained minor amounts of 859 the zeolite mineral clinoptilolite which most likely formed from the diagenetic Ostwald ripening of 860 amorphous or nanocrystalline phases such as allophane, imogolite and/ or poorly crystalline smectite. 861 Across all samples only minor (< 5 wt. %) abundances of amorphous phases were observed; instead, most 862 samples contained between 70-95 wt. % crystalline clay minerals. 863 The mineralogy and morphology of paleosols examined here is consistent with formation under 864 well-drained, oxidizing conditions with moderate weathering rates. Geochemical climofunctions based on 865 molecular weathering ratios applied to these paleosols indicate soil formation under mean annual 866 precipitation of ~600 mm and mean annual temperature of approximately 10° C. Pedogenic weathering of 867 volcanic ash under these climatic conditions was sufficient to transform volcanic glass and 868 amorphous/nanocrystalline phases into strongly crystalline dioctahedral clay minerals. 869 Four types of alteration after burial that have affected paleosols examined in this work are 1) 870 Drab green surface horizons due to burial gleization of organic matter; 2) brick-red color from burial-871 induced dehydration of ferric oxides and hydroxides; 3) zeolitization of volcanic glass and/or poorly 872 crystalline phases; and 4) significant mechanical compaction. Burial gleization, limited to the surface 873 horizons of paleosols, was most likely an early diagenetic alteration that resulted from the chemical 874 reduction of iron hydroxides and oxides by anaerobic bacteria consuming buried organic matter at or 875 below the water table. The timing of burial dehydration of (oxy)hydroxides remains poorly constrained, 876 but late diagenetic alterations such as zeolitization may have occurred during early Miocene burial at 877