Chemical weathering is seen as an irreversible, nonequilibrium, stochastic process, which implies the necessity of stepwise modeling. At early stages, the process is governed by the dissolution of primary minerals, which remains the limiting step throughout alteration. A pathway of further weathering is governed by a delicate balance between mineral dissolution, precipitation of secondary phases, and oxidation of Fe2+, which are connected by feedbacks. Dissolution of minerals increases pH, which often leads to lower dissolution rates. Increasing concentrations of solutes causes minerals to precipitate, which can stabilize the composition of solution. Precipitation also regulates pH and dissolution rates. Although higher pH (> ∼3.5) increases the rate of Fe2+ oxidation, formation of ferric species releases H+ (e.g., Reaction 7) partially compensating for rising pH. Significant increases in pH related to the formation of zeolites lowers dissolution rates of ferrous minerals and limits the amount of precipitated goethite.
 Despite these compensatory effects, overall decreasing acidity reduces rates of mineral dissolution and makes weathering less efficient in time. A stepwise quasiexponential slowing of weathering can be seen in Figures 1–8. In fact, the major events (formation of amorphous silica, goethite, kaolinite, and smectites) occur early and time intervals between depositions of each new phase grow sequentially. Rapid neutralization and corresponding decrease in dissolution rates work toward preservation of primary minerals. For typical basaltic lava, a complete closed system equilibration at low-temperature aqueous conditions could not be achieved even over geological timescale (Figures 1–3). Short-term episodes of acid weathering do not affect the majority of large rock fragments and the process may end before precipitation of clays, zeolites, and carbonates. This inference is strengthened by typically lower rates of natural weathering compared to laboratory data [e.g., White and Brantley, 2003] used in modeling.
4.1. Fate of Primary Minerals in Weathering Models and Martian Materials
 Different pH dependencies of mineral dissolution rates (Table 2) imply preferential dissolution of specific primary phases at each weathering stage (Figure 1b). It follows that abundances of primary minerals in a weathered material could reflect a stage of alteration. Our results show that basalt weathered below pH ∼3 would be enriched in plagioclases that dissolve slower than olivine, Fe-Mg pyroxenes and magnetite.
 On the Martian surface, variations in abundances of primary basaltic minerals could reflect preferential dissolution of some phases. The interpretation of thermal emission spectra (TES) of low-albedo surface regions reveals broad variations in the pyroxene/plagioclase ratio (0.16 to 1.24) [Bandfield et al., 2000; Bandfield, 2002; Hamilton et al., 2001; Wyatt and McSween, 2002; McSween et al., 2003; Michalski et al., 2006a; Rogers et al., 2007], which may not be entirely related to igneous petrology. Lowest pyroxene/plagioclase ratios are observed in the Northern hemisphere, for example, in Acidalia Planitia (surface type 2 [Bandfield et al., 2000; Bandfield, 2002; Christensen et al., 2001a]). These observations are consistent with faint pyroxene features in near-infrared spectra of northern low-albedo regions [Bibring et al., 2005; Mustard et al., 2005; Mustard and Cooper, 2005]. A kinetic modeling in the pyroxene-plagioclase system open to solution [McAdam et al., 2007a], and our results suggest that the observed deficiency of pyroxene could be related with aqueous weathering below pH 3. In Gusev crater, depletion of rock surfaces in Mg and Fe in the Clovis class rocks may indicate preferential low-pH dissolution of pyroxene and olivine relative to feldspars [Hurowitz et al., 2006]. In heavily altered Wishstone class rocks, the high Na-plagioclase/pyroxene ratio (4−5 [Ruff et al., 2006]) may reveal faster dissolution of pyroxene. A presence of plagioclase in heavily altered jarosite-bearing bedrocks in Meridiani Planum [Glotch et al., 2006] also agrees with its slow low-pH alteration compared to primary Fe-Mg silicates that are not detected in the bedrocks. A high plagioclase/pyroxene ratio inferred from TES spectra of Martian dust [e.g., Bandfield and Smith, 2003] is also consistent with preferable pyroxene alteration (A. C. McAdam et al., Preferential low-pH dissolution of pyroxene in plagioclase-pyroxene mixtures: Implications for Martian surface materials, submitted to Icarus, 2007). Since the spectral data refer to surfaces on mineral grains, the bulk pyroxene/plagioclase ratio in surface materials may not be strongly affected by weathering. It is possible that observed weak pyroxene signatures are related to specifics of surface alteration of pyroxene and/or silica coatings [Rampe et al., 2006] (see section 4.2).
 The presence of magnetite in Martian soil [Morris et al., 2006] and atmospheric dust [Pollack et al., 1979; Goetz et al., 2005] is not consistent with weathering above pH 3–5 when magnetite may dissolve faster than silicates (Figure 1b). Weathering of magnetite could have been slowed because of a presence of Ti [White et al., 1994], a resistance to physical weathering, and secondary Fe3+ oxide coatings, consistent with reddish hue of magnetite-rich dust particles accumulated on MER magnets [Goetz et al., 2005].
4.2. Secondary Silica
 The models show that formation of amorphous silica may designate early stages of basalt weathering by solutions with initial pH below ∼3. Neutralization of these solutions above pH 3 may cause dissolution of silica. Our results agree with the formation of amorphous silica in experiments aimed to model acid weathering of basalts [Tosca et al., 2004; Golden et al., 2005; Hurowitz et al., 2005] and at volcanic locations affected by acid vapors and solutions [Golden et al., 1993; Wolfe et al., 1997; Morris et al., 2000; Schiffman et al., 2006]. An early formation of silica is also inferred from thermodynamic models for acid weathering [Schiffman et al., 2006; McAdam et al., 2007b]. These observations and models are consistent with low solubility and slow dissolution of amorphous silica in acids [Dove and Rimstidt, 1994; Dove, 1995].
 Secondary silica has been reported in the ALH84001 Martian meteorite where it forms veins and coatings probably deposited from low-T aqueous solutions [Westall et al., 1998; McKay et al., 1998]. A likely presence of amorphous silica in surface materials became evident from thermal infrared spectra obtained from the orbit and at MER sites. Indistinguishable “high-silica phases” observed in spectra of some low-albedo regions (surface-type 2) may contain amorphous silica formed through basalt alteration [Wyatt and McSween, 2002; Kraft et al., 2003; Michalski et al., 2005]. Laboratory investigations show that thermal infrared observations could be explained by silica coating on primary mineral grains [Kraft et al., 2003; Rampe et al., 2006]. Abundant (up to 25 vol.%) silica-rich glass and/or impure amorphous silica in layered bedrocks in Meridiani Planum [Glotch et al., 2006] is consistent with acid environments. At the Columbia Hills in Gusev crater, some rock classes (Clovis, Watchtower) may contain high-silica glass [Ruff et al., 2006] that may represent weathering products. The elemental composition of trench soils in Gusev crater indicates a presence of silica [Haskin et al., 2005]. At the Mars Pathfinder landing site, elevated Si content in rock fragments may represent secondary silica on basaltic rock fragments [McLennan, 2003; Hurowitz et al., 2006] that contributes to the observed “andesitic” bulk composition [Foley et al., 2003]. If confirmed, the presence of amorphous SiO2 in Martian surface materials would be consistent with low-pH environments and a lack of thorough neutralization of aqueous solutions in many locations.
 Since the formation of silica coatings on primary minerals decreases their dissolution rates, chemical or mechanical removal of coatings could be needed to facilitate alteration. Silica coatings could be weaken through neutralization of solution, alteration by solution with pH > 3, as well as physical weathering during and between episodes of aqueous weathering. Note that experimental data [Tosca et al., 2004; Golden et al., 2005; Hurowitz et al., 2006] and field observations [Schiffman et al., 2006] do not reveal a significant lingering of acid weathering because of silica coating.
4.3. Fate of Ferrous Iron During Martian Weathering
 Ferrous minerals on the Martian surface have been subjected to irreversible but limited oxidation. Although Martian surface materials contain several ferric phases (nanophase oxides, goethite, hematite, jarosite, and nontronite [Morris et al., 2000, 2006; Klingelhöfer et al., 2004; Poulet et al., 2005; Glotch et al., 2006]), ferrous minerals typically dominate in rock fragments and soils. The presence of magnetite in dust [Pollack et al., 1979; Goetz et al., 2005] and soil [Morris et al., 2006], and olivine and pyroxenes in rock fragments and soils [e.g., Klingelhöfer et al., 2004; Morris et al., 2006; Mustard et al., 2005; Hoefen et al., 2003] are consistent with only partial oxidation of primary minerals. Throughout history, aqueous Fe2+ oxidation at the surface could have been restricted by rare and short appearances of aqueous solutions [Haberle, 1998], and slow oxidation in acidic and O2-deficient environments [Burns, 1993; Hurowitz et al., 2006] (see also our results). Likewise, anhydrous photooxidation of ferrous minerals could have not been efficient [Morris and Lauer, 1980].
 At < ∼10−7 bar, which could represent Martian atmosphere during active volcanism [Catling and Moore, 2003], the role of Fe2+ oxidation is negligible and dissolution of Fe3+ spinels provided the major source for secondary Fe3+ species. The incongruent dissolution of magnetite in acid solutions leads to the formation of metastable maghemite [White et al., 1994],
followed by its conversion to stable goethite or hematite (at low aH2O(l) and in anhydrous conditions in the equatorial region). A paucity of magnetite in primary rocks (as in several Martian meteorites) makes formation of secondary Fe3+ phases entirely dependent on oxidation that becomes efficient only at elevated pH and
 Our results demonstrate that Martian acid weathering supplies Fe2+ in solution faster than it oxidizes by dissolved O2. This may cause precipitation of ferrous phyllosilicates (smectites and chlorites), as well as deposition of ferrous sulfates and chlorides during freezing and/or evaporation. If solution does not freeze or evaporate in O2-deficient conditions < ∼10−4 bar), formation of ferrous phyllosilicates consumes Fe2+ from solution, delays its neutralization, decreases Fe2+ oxidation rate, and leads to smaller amounts of goethite compared to acid weathering at O2-rich Earth-like conditions (Figures 9 and 10). A rapid supply of Fe2+ in solution during weathering of fine particles at low W/R ratios can cause deposition of daphnite and siderite (Figures 4–7). Overall, our models demonstrate that short-time episodes of acid weathering on Mars should not cause significant oxidation, especially at current
 In addition to short duration of O2-rich periods, oxidation could be limited by Fe3+-oxide coatings that restrict dissolution of primary and secondary ferrosilicates, as it happens on Earth. An early deposition of silica and Fe3+ hydroxide during freezing/evaporation of acid solutions would also lead to protective coatings. It is possible that ferric oxide (and silica?) coatings on magnetite dust particles [Goetz et al., 2005] prevented thorough oxidation of the mineral.
 A prolonged single acid weathering episode on Mars could have led to deposition of mostly ferrous phases: sulfates, chlorides, smectites, chlorites, and siderite, while formation of goethite was mainly related to dissolution of magnetite. Subsequent weathering events would first cause dissolution of ferrous salts and make Fe2+ available for oxidation. Dissolution of ferrous phyllosilicates, and primary minerals would supply Fe2+ ions as well. It follows that irreversible Fe2+ oxidation could have occurred through multiple aqueous events in the presence of O2 in the atmosphere > ∼10−7 bar). Aqueous decomposition of H2O2, which could be present in the soil [e.g., Yen et al., 2000], provided an additional O2 source. Oxidation of Fe2+-saponite could have led to formation of nontronite that is tentatively detected in small surface regions [Poulet et al., 2005] and layered rocks at Meridiani Planum [Glotch et al., 2006]. Although oxidation of Fe2+-saponite to poorly crystalline nontronite occurs on Earth [Kohyama and Sudo, 1975], efficiency of Martian oxidation of saponite is limited by low and a necessity of advanced Mg leaching from octahedral sites.
 Formation of significant amounts of Fe3+ compounds, as in hematite-rich regions [Christensen et al., 2001b], requires the involvement of significant masses of O2, which are comparable with the atmospheric abundance. These masses of O2 may not be related to periods of active volcanism [Catling and Moore, 2003] but could have been produced in large impacts (see section 4.5).
4.4. Dust Weathering, Weathering Rinds, and Specifics of Northern Lowlands
 Martian dust contains an abundant semicrystalline silicate component, as seen in the near-infrared [Soderblom, 1992; Murchie et al., 2000; Bibring et al., 2005] and thermal infrared [Michalski et al., 2006a] spectra of high-albedo surface regions. The dust is not rich in primary minerals (except magnetite, and, probably, plagioclase) but sulfates and chlorides are likely to be abundant. The altered semicrystalline nature of Martian dust (< 10 micron) agrees with rapid acid weathering of small particles (Figures 4 and 5b). Although solution could be neutralized within a year in a closed system, subsequent acid attacks may destroy secondary phases, leading to semicrystalline silicates and Si-, Al-, Fe3+-oxides. Inefficient low-temperature crystallization of phyllosilicates is observed on Earth and suggested for Mars [e.g., Michalski et al., 2006b]. High concentrations of S and Cl in Martian fines could reflect an incorporation of these elements during multiple acid attacks.
 Dust particles weather in acid atmospheric aerosols, as a soil component, as well as on surfaces of rocks. In the latter case, weathering would produce a layer of alteration products that contain elements that came from the dust and the rock. Surface layers enriched in Fe3+, S, Cl, and Br on basaltic rocks in Gusev crater may reveal their complex formation [Haskin et al., 2005; Gellert et al., 2006; McSween et al., 2004, 2006; Hurowitz et al., 2006]. As for dust alteration, multiple acid attacks on surface rocks could have prevented formation and accumulation of minerals formed from neutralized solutions. In addition, minerals could be affected by lowering pH during freezing and/or evaporation of solutions (Figures 11 and 12). As a result, silica, Mg-, Fe2+-, and Ca-sulfates and chlorides, amorphous Al-bearing mineraloids (Al-bearing opal [Michalski et al., 2005]), and ferric oxides and sulfates (not modeled here) could compose weathered products on surfaces of basaltic rocks at the end of each weathering episode. The observations of chemically distinct rock surfaces of olivine basalts on plains in Gusev crater [Haskin et al., 2005] agrees with this view. In particular, the lack of evidence for carbonates, smectites, and zeolites on rock surfaces and soils reveals short-term and/or multiple acid attacks.
 Abundant Mg sulfates in Martian surface materials are consistent with freezing and/or evaporation of low-pH (< 7.5) fluids formed through acid weathering of basaltic materials. This notion agrees with inferences in recent works [e.g., Tosca et al., 2005; Clark et al., 2005; Hurowitz et al., 2006]. Despite some differences in mineralogy of freezing and evaporation of acid solutions, Martian soil mineralogy may not reveal the leading process of salt precipitation. Although evaporation results in less hydrated solids than freezing, many salts may then dehydrate to equilibrate with atmospheric H2O, especially at low latitudes. At present T-PH2O conditions in the equatorial belt, ice is not stable near the surface, and precipitated salts could dehydrate to MgSO4 · H2O, FeSO4 · H2O, FeCl2 · 2H2O, MgCl2 · 2H2O, CaSO4, and NaCl [Zolotov, 1989].
 Northern low-albedo regions with strong spectral features of “silica-rich phases” and weak pyroxene signal [e.g., Bandfield et al., 2000; Mustard et al., 2005] could have been affected by more intense, and/or longer acid weathering, as well as lower pH solutions compared to equatorial and southern regions, which do not reveal these features. The spectral observations may reflect lower altitude, higher atmospheric pressure, higher atmospheric humidity [Jakosky and Farmer, 1982], larger amounts of ground ice, and more importantly accumulation of surface ice during periods on high obliquity compared to equatorial and southern highlands. All these conditions may cause slower evaporation/freezing, higher W/R ratios, and longer contacts of solution with rocks and soils in northern lowlands.
 During periods of high obliquity, acid rock weathering could have occurred through accumulation of acid aerosols and raindrops (see section 4.5) on ice followed by the migration of acid brine pockets toward the ice-regolith boundary. An isolation of acid brine pockets from the atmosphere prevented their evaporation. The low temperature of the H2SO4-H2O and HCl-H2O eutectics (−62°C and ∼−87°C, respectively) and low water activities (∼0.8−0.5) of low-temperature acid brines allow temporal solution existence near ice even at present climate. During winters, acid brines pockets could have frozen to form H2SO4 hydrate (see Figure 11). Melting at warmer seasons favored brine migration toward rocks. Accumulated subglacial aqueous solutions could have led to a prolonged acid alteration of rocks compared to rapidly evaporating (and/or freezing) acid precipitates on ice-free surfaces.
 Thermal insulation by thick ice layers could have caused an increase in temperature in lower horizons and also favored downward migration of acid brines and their accumulation at the ice-regolith interface. In addition, this process could have been enhanced by melting at ice-regolith boundaries at low altitudes [Richardson and Mischna, 2005]. This mechanism would cause acid weathering of uppermost regolith layers and is consistent with signs of aqueous migration of silica, and Mg, and Fe3+ species observed in rock coatings and excavated soils in Gusev crater [Haskin et al., 2005; Wang et al., 2006b]. Note that melting at the ice-surface interface has been proposed to account for those observations [Arvidson et al., 2004; Haskin et al., 2005].
4.5. Acid Weathering in Martian History: Role of Impacts and Volcanism
 Volcanic eruptions and impacts could have been responsible for multiple short-term episodes of acid weathering as a dry and cold climate prevailed over history. The effect of volcanism on generation of acid solutions is mostly attributed to degassing of SO2, H2S, and HCl. However, a consumption of photochemically produced atmospheric oxidants (O2, O3, OH, and peroxides [Yung and DeMore, 1999]) in reactions with reduced volcanic gases (H2S, CO, H2, and S2 [Zolotov, 2003]) would limit the generation of sulfuric acid through oxidation of SO2 and H2S. Without photosynthesis, active volcanic degassing could have limited amounts of strong atmospheric oxidants on early Earth [Holland, 1984] and Mars [Catling and Moore, 2003]. Without oxidants, only moderately acid solutions could form because of a limited abundance of volcanic HCl and HF [Symonds et al., 1994; Zolotov, 2003], a mild dissociation of H2S in solutions (H2S H+ + HS−), and a low solubility of SO2 compared to SO3, which forms sulfuric acid. Acid generation through formation of pyrite from dissolved H2S (2H2S + Fe2+ 2H+ + H2 + FeS2) could have been limited by the supply of Fe2+ through dissolution of ferrous silicates. Likewise, formation of low-pH fluids would have been limited in the anoxic Martian subsurface. An alkaline nature of Martian subsurface solutions [Zolotov et al., 2004], high concentrations of cations in brines, and low W/R ratios would have favored rapid neutralization of volcanic acid gases and fluids. Involvement of photochemical O2 into oxidation of SO2 [Settle, 1979] and H2S, as well as Fe sulfides, provided a major contribution into generation of sulfuric acid, but makes acid weathering a near-surface phenomenon that may not be closely related to volcanism. Finally, volcanic eruptions could not have led to acid raining owing to cold and dry climate throughout history [Haberle, 1998].
 Although volcanism was likely to be responsible for the net supply of S and halogens to the Martian surface [e.g., Clark and Baird, 1979; Clark and Van Hart, 1981], the effects of large impacts on aqueous acid weathering and erosion could have been greater than the effects of volcanism. On Earth, the role of impacts in generating acid aerosols and rains has been realized in the context of Chicxulub crater at the Cretaceous-Tertiary boundary [Ivanov et al., 1996; Pope et al., 1997; Pierazzo et al., 1998]. These studies show that the Chicxulub impact into sulfate-bearing sediments released several orders of magnitude more sulfur gases compared to a volcanic eruption. On Mars, the elevated abundance of sulfates in the regolith implies similar effects. Impact generation of SO2 and/or SO3 through decomposition of sulfates [Lyons and Ahrens, 1996; Ivanov et al., 1996; Gupta et al., 2001], as well as other oxidized gases (O2, O3, NO, NO2 [e.g., Prinn and Fegley, 1987; Gerasimov, 2002; Gerasimov et al., 2002]) provided reactants to form acids in cooling impact plumes and clouds. Oxidation of sulfides in target and projectile contributed to the release of S-bearing gases. Note that mass independent fractionation of S isotopes observed in Martian meteorites [Farquhar et al., 2000] does not exclude impact degassing and recycling of sulfur through the atmosphere. On the surface, aqueous interaction of strong oxidants with Fe2+ and sulfides caused oxidation of S and Fe and produced H+. In addition, atmospheric oxidants, which would not be abundant during periods of volcanic activity, would react with impact-generated H2S, SO2, and NO to form SO3 and NO2, as well as oxidize Fe2+ in surface solutions. In H2O-bearing impact environments, hydrogen halides could form through decomposition of soil halides and igneous chlorapatite, which is abundant in Martian meteorites.
 Dissolution of SO3, H2S, NO, NO2, and HCl in water droplets formed through condensation in an impact cloud would have led to an array of acids. Although sulfuric and hydrochloric acids were probably most abundant, cometary impacts could have led to nitrous and nitric acids [Prinn and Fegley, 1987]. Local and global raining from impact clouds, which is expected in the present climate [Segura et al., 2002], generated acid solutions on the surface. Plentiful impact-generated fine-grained material and silicate glasses would be primarily altered by acid rains/aerosols, consistent with an altered nature of Martian dust and air-blown particles. Acid rainfalls also affected upper surface layers and locally penetrated to deeper horizons upon accumulation in surface depressions. Freezing and/or evaporation of acid solutions led to deposition of sulfates and chlorides. We suggest that sulfate-bearing rocks and soils were a cumulative effect of multiple-impact produced acid volatiles, at least initially from volcanic emissions and impacts on sulfide-rich rocks [cf. Burt et al., 2006].
 Penetration of impact-generated acid fluids into surface ice sheets restricted their evaporation and prevented freezing until temperatures of acid-rich eutectics (see Table 3 and Figure 11). During periods of elevated obliquity, subglacial acid weathering by these fluids could have been responsible for the formation of silica-bearing and pyroxene-depleted surfaces in Northern lowlands (e.g., TES surface-type 2 in Acidalia Planitia [Wyatt and McSween, 2002]) and also in middle altitudes in the Southern hemisphere. In addition, these processes could have played a role in the equatorial latitudes, where surface ice could temporary exist at high obliquities.
 Impact-generated acid rains could have been responsible for the vertical migration of Si and Fe3+ suggested for excavated soils in Gusev crater [Haskin et al., 2005; Wang et al., 2006b]. The elevated abundance of Si and Fe3+ observed at depth of ∼10 cm [Wang et al., 2006b] may reflect dissolution of surface minerals in low-pH solutions followed by precipitation of silica and Fe3+ hydroxide in the shallow subsurface. Acid precipitating from impact plumes may also account for the high abundance of amorphous or semicrystalline material and lack of detection of crystalline phyllosilicates, carbonates, and zeolites in heavily altered rocks at Columbia Hills [Ruff et al., 2006]. This reveals a low-temperature acidic alteration occurred over a limited timescale. The diverse and patchy nature of aqueous processes in Columbia Hills [Ming et al., 2006] could be related to mixing from impact event(s), implying transient involvement of aqueous solutions.
 An interaction of heavily altered salt-rich deposits (as in Meridiani Planum [Clark et al., 2005]) with acid rainwater could have not led to rapid neutralization, but increased the amount of salts in deposits. Impact-generated acid rainfalls over proposed impact surge deposits at Meridiani Planum [Knauth et al., 2005] do not contradict with observations and arguments for acidic aqueous conditions [Clark et al., 2005; McLennan et al., 2005] and agree with an addition of S and Cl to basaltic material [McCollom and Hynek, 2005]. A formation of abundant ferric iron phases in layered bedrocks [Klingelhöfer et al., 2004; Glotch et al., 2006] could have been driven by plentiful impact-generated oxidants. A possibility exists that the deposition of impact surge deposits and low-pH aqueous processes were related to a large single impact in Meridiani Planum.
 In addition to our earlier suggestions, a large degree of rock alteration and/or elevated silica content in northern low-albedo regions [Wyatt and McSween, 2002; Kraft et al., 2003; Michalski et al., 2005] could be related to impact degassing of more abundant ground ice, sulfate, and other hydrated salt deposits as compared to equatorial and southern highlands. Finally, a correlation of P with S and Cl observed in MER sites could be explained by multiple acid rainfalls that delivered S and Cl and favored P mobility [Greenwood and Blake, 2006].
 A high impact rate in Noachian implies an efficient acid weathering followed by surface erosion, aqueous transport, and deposition of salts and suspended particles in temporal water reservoirs. In fact, Segura et al.  show that impact-generated rainfalls could have been responsible for the high erosion rate in Noachian time [cf. Craddock and Howard, 2002] and the formation of valley networks. No warm and wet climate is needed for impact-induced exogenic processes. It follows that a majority of aqueous oxidation, formation of silica, and Mg-Fe-Ca sulfates could have occurred in the earliest period of geological history.
 With or without a warm and wet climate, frequent large impacts in Early Noachian caused stronger rainfalls and longer duration of weathering episodes compared to subsequent epochs. Warm/wet climate would have reduced the impact of acid weathering because of dilution of acids in persistent and neutralized surface water reservoirs. The longer duration of weathering episodes favored partial and/or local neutralization in temporal surface water reservoirs. A presence of tentatively detected nontronite in Meridiani Planum [Glotch et al., 2006] may reflect neutralization that facilitated the growth of Fe3+ oxide concretions [Zolotov and Shock, 2005]. The neutralization could have been slowed because of accumulation of H+ through freezing/evaporation (Figures 11 and 12), a conversion of jarosite to goethite, the absence of rapidly dissolving pyroxenes and olivine, and possible silica/Fe3+-oxide coatings. At Columbia Hills, an existence of poorly crystalline [Ruff et al., 2006] kaolinite-type aluminosilicates in some Clovis class rocks indicates mildly acidic conditions [Wang et al., 2006a] (pH > ∼3.5, see Figure 2), which may imply some neutralization of original acid solutions.
 In other locations, more advanced neutralization could have occurred. An evaporitic formation of carbonates in the ALH85001 Martian meteorite [Warren, 1998] about 3.9−4.0 Ga ago implies nonacidic pH, which may also reflect neutralization at some depth. The only local occurrence of smectite deposits [Poulet et al., 2005] in Noachian-age regions indicates spatially limited neutralization. Using terrestrial analogues, the formation of observed layered clay deposits could have occurred through surface weathering followed by aqueous transport in suspension and accumulation in water reservoirs. A massive nature of Noachian clay deposits and low permeability could have limited their weathering by subsequent acid attacks. To conclude, acid weathering in Noachian probably played a larger role compared to subsequent epochs, though achieved larger degrees of neutralization because of longer existence of surface solutions and their temporal accumulation in surface reservoirs, where clays deposited. Although the intensity of impact-induced acid weathering declined in time, low-pH aqueous processes could have also been important during the Hesperian epoch, when volcanism reached its highest levels in history [Tanaka et al., 1992] and large impacts still occurred.