Geochemistry, Geophysics, Geosystems

Controls on palagonitization versus pedogenic weathering of basaltic tephra: Evidence from the consolidation and geochemistry of the Keanakako'i Ash Member, Kilauea Volcano



[1] On the summit of Kilauea Volcano, incipient palagonitization and consolidation of vitric tephras from the <500-year-old Keanakako'i Ash Member of the Puna Basalt have occurred only adjacent to caldera-bounding faults. Calcite intergrown with palagonitized glass has δ18O values indicative of a hydrothermal origin. Whereas the δ13C of the calcite suggests that it has sequestered carbon from atmospheric as well as mantle sources, radiocarbon dates from the calcites, in addition to calcites from modern steam vents at Kilauea, indicate that >90% of the carbon is mantle derived. We hypothesize that the palagonitized zones may denote the location of fossil steam vents, similar to active vents exposed along circumferential faults on both the north and the south sides of the Kilauea Caldera. Elsewhere, Keanakako'i tephras are weathering into pedogenic products that closely reflect modern environmental conditions. Under acidic soil conditions (pH <6.0), tephra is undergoing dissolution, with the development of opaline crusts on outcrop faces. Under more neutral soil pH (6.5–7.8), the pedogenic products reflect the degree of desilication of tephra mitigated by leaching: dominantly smectite where mean annual rainfall is <50 cm/yr and dominantly kaolinite, allophane, and/or imogolite where rainfall exceeds 250 cm/yr.

1. Introduction

[2] Palagonitization is a common alteration process affecting glassy, basaltic tephra in subaerial, submarine, lacustrine, subglacial, and, perhaps, extraterrestrial low-temperature environments [Singer and Banin, 1990; Fisher and Schmincke, 1984]. Palagonitization commonly results in a consolidated assemblage of unaltered glass, altered or “palagonitized” glass, and a variety of pore-filling authigenic minerals that include zeolites and calcite. Palagonitized glass, which typically occurs as rinds lining the external and internal surfaces of tephra fragments, is composed of smectite crystallites within an amorphous glassy matrix [Schiffman et al., 2000; Zhou et al., 1992; Eggleton and Keller, 1982].

[3] Fresh basaltic glass is one of the most abundant, yet thermodynamically unstable, materials found on the Earth's surface, such that the process of palagonitization has important consequences for (1) the composition of seawater through weathering of basaltic rocks of the upper oceanic crust [Staudigel and Hart, 1983; Honnorez, 1981; Bonatti, 1965], (2) the sorption/retention properties of glass- or basalt-hosted nuclear-waste repositories and contaminant sites [Jercinovic et al., 1990; Lutze et al., 1985], (3) the composition of meteoric waters in basalt-hosted aquifers [Tilling and Jones, 1996; Hay and Jones, 1972], and (4) the biogeochemical cycling of nutrients in soils developed on basaltic protoliths [Schiffman and Southard, 1996; Staudigel et al., 1995; Berkgaut et al., 1994].

[4] Currently, there are at least three distinct models for the palagonitization of vitric tephras deposited through hydrovolcanism. In the syndepositional model, glassy tephra deposits are thought to be palagonitized by “trapped” pore fluids that condensed from steam entrained within a pyroclastic flow [Wohletz and Sheridan, 1983]. In the hydrothermal model, palagonitization is believed to be affected by fluids (primarily steam) that pass up through unconsolidated tephras at some time after deposition, presumably in response to hypabyssal intrusions within or beneath the tephra pile [Jakobsson, 1978]. In the postdepositional model, meteoric water percolates down through unconsolidated tephra in the unsaturated zone, essentially under ambient weathering conditions and at some time after cessation of magmatic activity [Hay and Iijima, 1968]. This study was initiated to test the feasibility of these three models in accounting for the incipient palagonitization of Holocene tephras deposited by hydrovolcanic eruptions in the region around Kilauea Caldera. These tephra deposits are unusually well suited for testing these models because they are distributed across a region that experiences pronounced variations in (1) rainfall (<50 to >300 cm/yr), (2) soil pH (3.5–7.5), and (3) hydrothermal input (active steam vents at T<100°C). Many Holocene tephras in this region are weathering to soils, which are both texturally and mineralogically distinct from the palagonitized tephras.

2. Geologic and Climatic Setting

[5] Tephra deposits of the Keanakako'i Ash Member of the Puna Basalt [Easton, 1987] are exposed over a wide region across the summit of Kilauea Volcano. The tephras, which are up to 12 m thick in the region immediately west of Keanakako'i Crater (Figure 1a), were deposited from pyroclastic surges and related fall-out, initiated by hydrovolcanic eruptions from within Kilauea Caldera [McPhie et al., 1990; Decker and Christiansen, 1984]. A lower, vitric-rich portion of the Keanakako'i Ash Member is unconformably overlain by an upper, lithic-rich portion [Swanson et al., 1998]. These observations are consistent with a transition from mainly phreatomagmatic to phreatic eruption mechanisms [Mastin, 1997; McPhie et al., 1990; Decker and Christiansen, 1984]. Although deposition of the uppermost, lithic-rich tephras was completed around 1790 A.D., the oldest vitric tephras may have been deposited nearly 300 years earlier (D. A. Swanson et al., unpublished 14C data, 2000). Samples examined for this study were collected from the lowermost vitric tephras (equivalent to layers 1 or 2 of McPhie et al. [1990]) exposed immediately above the lowest reticulite layer of the Keanakako'i Ash. These vitric tephras are mostly <2 mm in grain size and composed of microvesiculated, aphyric basaltic glass.

Figure 1.

(a) Map of the Kilauea summit region showing the circumferential fault system that defines the caldera, isopachs for the vitric layers of the Keanakako'i Ash [after McPhie et al., 1990], annual precipitation contours [from Giambelluca and Schroder, 1998], soil pH, and the location and temperature of active steam vents. (b) Location of samples described in Table 1 and alteration styles of tephra described in text. The double-dashed lines indicate the major roads in the region of Kilauea Caldera.

[6] The Keanakako'i is the youngest in a series of Holocene and Pleistocene tephra deposits that appear to have originated from the vicinity of present-day Kilauea Caldera and the upper east rift zone [Clague et al., 1995; Easton, 1987; Decker and Christiansen, 1984]. Alteration of the older tephras has been described by Hay and Jones [1972], who first recognized that these basaltic ash deposits have weathered in response to a wide range of environmental conditions.

[7] Kilauea Caldera straddles large orographic climatic gradients affecting annual precipitation, soil pH, and vegetation. Rainfall of over 250 cm/yr supports an ‘Ohi'a- and tree fern–dominated rain forest and woodland on the windward side (i.e., northeast flank) of the caldera. Conversely, deposition of sulfuric acid aerosols of volcanic origin on the leeward side of the caldera has dramatically lowered soil pH and created a desert-like environment with virtually no vegetation, even though rainfall is ∼125 cm/yr.

3. Alteration of Vitric Tephras

[8] Mineralogical and geochemical analyses on a suite of Keanakako'i samples reveal that these young, vitric tephras have altered variably in response to the diverse climatic conditions surrounding Kilauea Caldera (Table 1). Although the degree of weathering in some of the samples may be small (as indicated by very low yield of the <2-micron size fraction from the dispersion and centrifugation of bulk tephra samples), the alteration products of these vitric tephra can be unequivocally identified through X-ray diffraction analysis of the clay-size fraction as well as by backscattered electron imaging and wavelength dispersive X-ray microanalysis of epoxide grain mounts.

Table 1. Alteration of Keanakako'i Ashes
SampleDry Tephra ColorLocalityaAnnual Rainfall, cmbSoil pHFloraPresent Thermal ActivityDegree of ConsolidationAlteration Mineralogyc
  • a

    Value in parentheses is the distance in kilometers from the rim of Kilauea Crater.

  • b

    Data from Giambelluca and Schroder [1998] and Taliaffero [1959].

  • c

    Determined by X-ray diffraction and backscattered electron analysis/energy dispersive spectrometry.

116-15YR2/2Sand Wash (0.5)1253.5Very sparse shrubnonesurface crustsopaline crusts
117-57.5YR5/4Arroyo Nanahu (12)507.8Dry grass landnoneweaksmectite, kaolinite, opal
116-35YR6/4Keanakako'i Bluffs (0)1756.5sparsenonemoderatepalagonitized glass, calcite
116-45YR5/6Keanakako'i Bluffs (0)1757.1Ferns near ventsteam ventsstrongpalagonitized glass
918-32.5Y3/2Keanakako'I Bluffs (0)150not recordedSparsenonestrongpalagonitized glass, calcite
919-12.5 Y 5/3East of Keanakako'i Crater (0.5)2007.4‘Ohi'anonenonenone
117-25YR3/2Soil Pit by Thurston Lava Tube (0.5)2257.5Rain forestnonenonenone
117-17.5YR6/6Volcano Village (2.5)2507.1Rain forestnonenonekaolinite
118-17.5YR5/4Wright Road (6)3007.2Rain forestnonenonekaolinite, allophane, imogolite

[9] Nanahu Arroyo, 12 km south of Kilauea Caldera, is the driest region in which the Keanakako'i tephras are exposed, with ∼50 cm of rain/yr [Taliaferro, 1959]. There, an approximately 10- to 15-cm-thick vitric-lithic layer of Keanakako'i Ash is exposed at the top of the cliff face on the Hilina Pali fault scarp [Easton, 1987]. Under slightly alkaline soil conditions (pH = 7.8), glassy pyroclasts are partially altered into a complex, weakly consolidated, pedogenic assemblage of smectite, kaolinite, and opal (Figure 2a). However, at informally named Sand Wash, ∼1 km southwest of Kilauea Caldera, which receives ∼125 cm of rain per year, vitric pyroclasts are undergoing nearly congruent dissolution (Figure 2b) in response to the extremely low soil pH (3.5) created by acidic aerosol fallout. Similar acid dissolution can be found in Keanakako'i tephra on the south side of the caldera, as far eastward as Halema'uma'u, where soil pHs are <6.0. In the regions experiencing this acid dissolution, opaline crusts have formed on the faces of tephra outcrops (presumably because of high rates of evaporation on the exposed surfaces) and have weakly consolidated the tuffs.

Figure 2.

Backscattered electron (BSE) images of Keanakako'i Ash Member tephras depicting dramatic differences in weathering of mainly vitric, ash-sized pyroclasts in the region around Kilauea Crater. In all four images the brightest, largest grains are ash vitroclasts (GL), the black background is the epoxide used to impregnate the tephras, and the intermediate-toned, fine-grained materials coating the ash are the alteration products. (a) Weakly consolidated tephra from Arroyo Nanahu (117-5) in which weathering under relatively low rainfall and slightly alkaline conditions has produced fine-grained mixtures of pedogenic smectite, kaolinite, and opal. (b) Unconsolidated tephra from Sand Wash (116-1) in which weathering at low pH has resulted in dramatic acid dissolution of glass. (c) Consolidated tephra from the Keanakako'i Ash type locality (116-4) that has ∼20-μm palagonitized glass rinds on the exterior and interior surfaces of ash vitroclasts. (d) Unconsolidated tephra from Wright Road (118-1) in which leaching and desilication of glass under high rainfall has resulted in very fine-grained mixtures of pedogenic kaolinite and short-range-order aluminosilicates (i.e., allophane and imogolite).

[10] Alteration of the Keanakako'i Ash tephras is markedly different on the windward side of Kilauea Caldera, where annual precipitation exceeds 250 cm of rain per year [Giambelluca and Schroder, 1998] within a few kilometers northeast of the caldera's rim. Since sulfuric acid aerosols do not greatly affect the upwind region of Kilauea Caldera, soil pH is uniformly close to neutral (7.1–7.5; Table 1). Vitric Keanakako'i tephra is altering into aluminous pedogenic clays: kaolinite at Volcano Village (2.5 km from the rim of Kilauea; 250 cm/yr) and also the short-range-order aluminosilicates allophane and imogolite (Figure 2d) in outcrops along Wright Road (6.0 km from Kilauea; >300 cm/yr). These pedogenically affected tephras are essentially unconsolidated.

[11] Just south of Kilauea Caldera, thick deposits of the Keanakako'i Ash are exposed, locally in large slump blocks, along scarps created by the circumferential faults that bound the summit crater. In this region, which extends from due west of the Keanakako'i Crater to southeast of Halema'uma'u Crater (Figure 1a), the vitric Keanakako'i tephras are moderately to strongly consolidated and exhibit yellow to orange hues indicative of palagonitization.

[12] Petrographic examination of the strongly consolidated vitric-rich sample (116-4) reveals that glassy ash fragments are ubiquitously coated with rinds of isotropic, palagonitized glass, averaging 20 micrometers in thickness (Figure 2c). The palagonitized rinds are found both on the outer surfaces of pyroclasts and on the walls of vesicles. The cementation of tephra has occurred because the palagonitized rinds have “fused” along the contacts between pyroclasts (Figure 2c). Compositionally, the palagonitized rinds are strongly hydrated and markedly depleted in Na and Ca with respect to the fresh glass (Table 2). X-ray diffraction of the <2-μm fraction (mainly the palagonitized rinds) of these samples indicates the presence of smectite with a mean crystallite thickness of <3 nm [Schiffman et al., 2000].

Table 2. Average Compositions of Keanakako'i Glass and Palagonitized Rinda
 Glass (23), wt %Glass, mole %Palagonitized Glass (8), wt %Palagonitized Glass (Range), wt %Palagonitized Glass, mole %Change in Mole %
  • a

    Determined by electron microprobe analyses. The mean of number of individual analyses is indicated in parentheses; H20 is estimated by difference.

H2O0.963.2213.92 36.3633.15

[13] At Kilauea the palagonitization of vitric fragments of the Keanakako'i tephras has generally not been accompanied by the coprecipitation of authigenic minerals as pore-filling cements. However, in some localities (e.g., 116-3 and 918-3), anhedral crystals of low Mg-calcite (MgO = 1.8 wt%) are intergrown with the palagonitized vitric pyroclasts. Texturally, the calcite appears to both predate and postdate the palagonitized rinds (Figure 3). Stable isotopic analyses of calcite (Table 3) separated from samples 116-3 and 918-2 yield δ18O values of 16.5 and 19.6‰ (standard mean ocean water (SMOW)) and δ13C values of 1.8 and 4.3‰ (Peedee belemnite (PDB)), respectively.

Figure 3.

BSE image of vitric ash (GL) with palagonitized rinds (PA) and calcite (CC) cement from sample 918-3. The calcite cements have δ18O values between −9.2 and −3.7‰ (Peedee belemnite (PDB)). If the calcite exchanged with local meteoric water (−5‰ [Scholl et al., 1996]), isotopic equilibrium occurred at temperatures between 36° and 62°C.

Table 3. Stable Isotopic and 14C Analyses of Carbonates
Sampleδ13CVPDBδ18OVSMOW14C Age, years
  • a

    Stable isotope analyses conducted on individual calcite crystals.

  • b

    Stable isotope analyses conducted on ground, homogenized material (carbonates and silicate matrix).

116-3a1.5416.4431,140 ± 210
918-3b4.3419.7219,940 ± 70
Keanakako'I7.9523.9526,300 ± 140
(modern calcite)7.7923.98 

4. Discussion

4.1. Petrogenesis of Palagonitized Glass in the Keanakako'i Tephras

[14] The process of palagonitization is complex, entailing extensive compositional modification of basaltic glass and the coprecipitation of various authigenic minerals as well as amorphous mineraloids. Petrogenetic studies of this process (as reviewed by Singer and Banin [1990], Fisher and Schmincke [1984], and Honnorez [1981]) indicate that the initial stages of palagonitization mainly entail the hydration, oxidation, and dissolution of basaltic glass. The resultant palagonitized glass, an amorphous or nanophase material, is strongly leached of its alkali elements, although not potassium in submarine environments [Staudigel and Hart, 1983]. The Keanakako'i palagonitized glass rinds have lost roughly 90% of their original Na and Ca (Table 2).

[15] Under what conditions can this initial stage of palagonitization occur? Although palagonitized glass has been described from glassy basaltic rocks of many ages and from many settings, only one case of “active” palagonitization has ever been described. Tephras at Surtsey began to palagonitize ∼5 years following their deposition (which ended in 1964) and 2 years after the cessation of all related volcanic activity [Jakobsson, 1978]. Palagonitization was caused by steam passing through the tephra cone, presumably in response to a hydrothermal system driven by mafic dikes intruded within the volcano [Jakobsson and Moore, 1986]. Time series data presented by Jakobsson [1978] demonstrate that tephras exposed to steam at temperatures >50°C had consolidated into dense tuffs through the palagonitization process within 4 years.

[16] Two independent lines of evidence, one structural and the other geochemical, suggest that the Surtsey hydrothermal model best explains the incipient palagonitization of the Keanakako'i tephras. At Kilauea we have thus far identified palagonitized tuffs only in regions adjacent to the circumferential fault systems that define the caldera's walls. These faults also host active steam vents on Kilauea's summit. On the north side of the caldera, numerous individual vents in the Steaming Bluff region have been active for over 100 years [Casadevall and Hazlett, 1983] and have present-day exit temperatures of 72°–79°C (Figure 1a). On the south side of the caldera, two steam vents (informally referred to as “Keanakako'i Vent” and “Keanakako'i Cliff” in Figure 4) that lie midway between Crater Rim Drive and the caldera floor, ∼500 m west of Keanakako'i Crater, have exit temperatures of 55°–60°C (Figure 1a). Since the present water table is ∼500 m beneath the caldera floor [Keller et al., 1979], the steam, which originates from the boiling of local meteoric waters (Figure 4), has presumably reached the surface through these circumferential faults.

Figure 4.

Oxygen and hydrogen isotope analyses of local precipitation and condensed steam from vents in the vicinity of Kilauea Caldera. The location of the two Keanakako'i steam vents is denoted as “KSV” in Figure 1. The two sites at Steaming Bluffs (“SB” in Figure 1) are located north and south of the crater rim road ∼1.5 miles west of the park visitor's center.

[17] Stable isotopic data from the two calcite samples (Table 3) in the consolidated Keanakakako'i tuffs also support a hydrothermal origin for the palagonitized glasses. Oxygen isotope measurements of water condensed from active steam vents along the rim of Kilauea Crater in September 1999 yield δ18OSMOW values ranging between −5.7‰ (at the Steaming Bluff sites on the north rim) and −11.2‰ (at the “Keanakako'i” sites on the south rim). In contrast, local precipitation immediately north of the crater produced δ18OSMOW values of −3.6‰. Oxygen and hydrogen isotope data show that the more positive δ18O values fall on the global meteoric water line (Figure 4) and are consistent with the average local meteoric water value of −5‰ [Scholl et al., 1996]. In contrast, the lower δ18O values suggest kinetic fractionation during boiling at depth. If we assume that the calcite samples precipitated with a median isotopic water value (δ18OSMOW=−7.4‰), application of calcite-water fractionation relationships [O'Neil et al., 1969] yields equilibrium isotopic exchange temperatures between approximately 33° and 49°C. For comparison, if the extreme steam and precipitation δ18O values had been used (−11.2 and −3.6‰) for samples 918-3 and 116-3, the temperature estimates would vary from 8° to 70°C, respectively. Although the colder temperature is probably unrealistic on the basis of an average temperature of 16°C at Hawaii Volcanoes National Park [Giambelluca and Schroder, 1998], the warmer temperature is consistent with steam vent temperatures of 55°–77°C that were measured during our 1999 field campaign.

[18] It is likely that the calcites in the Keanakako'i palagonitized tuffs precipitated in the presence of condensed steam. If we assume that the modern maximum vent temperatures measured in 1999 are representative of steam vent temperatures during the consolidation of the vitric tephras, then the condensate from which calcite precipitation occurred would have had a δ18O value within the range of the measured, modern-day Kilauea waters. This is clearly not the case for all carbonates precipitating in steam vents along the margins of Kilauea Caldera today. Carbonates collected in 1999 at an active steam vent (within the informally named “Keanakako'i” vent field on the south side of the caldera; Table 3) adjacent to site 116-4 (Figure 1) yielded average δ18O values of 24.08‰. If these carbonates precipitated in equilibrium with the vent steam (δ18O=−9.50‰), then the temperature of precipitation would have been 4.4°C. Since the calculated temperature is too cold, the condensate in the microenvironment of precipitation must have been kinetically enriched in 18O by up to 9‰ (presumably by evaporation at the site of calcification) for the calcite to precipitate at the vent temperature of 49.5°C. Because the estimated temperatures of precipitation for the calcites in the Keanakako'i palagonitized tuffs are reasonable, we take a conservative approach and consider the median δ18O value of the modern waters to be a reasonable estimate of the fluid that equilibrated with those fossil calcites. The 13C enrichment observed in the fossil calcite samples (116-3 and 918-3) suggests that the CaCO3 may have precipitated from a mixture of atmospheric and mantle-derived CO2. If we assume that the carbon source for calcite is dissolved HCO3 and that this carbon is in equilibrium with CO2 in the vapor state, we can use appropriate fractionation factors from Mook [1986],

equation image


equation image

to calculate the δ13C value of the CO2(gas) in the precipitation environment. With this assumption, the computed δ13C value for CO2(gas) is approximately −5.7 and −4.1‰ for samples 116-3 and 918-3, respectively (Table 4). These calculated values agree well with δ13C measurements (−5.21 and −4.75‰) made on CO2 collected in 1999 from within two separate vents of the Keanakako'i vent field. If we assume a simple mass balance between magmatic CO2 with a δ13C value of approximately −3.5‰ as reported from Kilauea fumaroles [Rubin et al., 1987] and preanthropogenic atmospheric CO2 with a δ13C value of approximately −6.5‰ [Indermühle et al., 1999], then we obtain a 30–80% contribution of magmatic CO2 carbon to the calcite. If the measured vent CO2δ13C values are more representative of the mantle carbon signature in this region, then the magmatic contribution would be even higher.

Table 4. Isotopic Analysis of Calcites and Calculated Temperatures and Fluids
Sampleδ13CVPDB calcite, ‰δ18CVSMOW calcite, ‰Temperature, °Caδ13C of HCO3 (aqueous)bδ13C of CO2(gas)c% Magmatic CO2d
  • a

    Calculated assuming isotopic equilibrium with a median δ18OSMOW value of −7.4 and the relationship of O'Neil et al. [1969].

  • b

    Calculated using ɛcalciteimage from Mook [1986].

  • c

    Calculated using image −HCO3 from Mook [1986].

  • d

    Assumes two-component mixing model. End-members are magmatic CO2 (δ13C ≈ − 3.5‰ from Kilauea fumaroles by Rubin et al. [1987]) and preindustrial atmospheric CO2 (δ13C ≈ − 6.5‰ [Indermühle et al., 1999]).


[19] To confirm these calculations, we radiocarbon (14C) dated bulk fossil calcite from sites 116-3 and 918-3 as well as bulk modern calcite from the active Keanakako'i vent field. Bulk 14C samples were pretreated in hydrogen peroxide to remove organics, rinsed with deionized water, dried, and placed in individual reaction chambers. Samples were evacuated, heated, and then acidified with orthophosphoric acid at 90°C. The evolved CO2 was purified, trapped, and converted to graphite in the presence of an iron catalyst in individual reactors [e.g., Vogel et al., 1987]. Graphite targets were measured at the Center for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory [Davis et al., 1990]. The 14C results are reported as conventional radiocarbon ages as described by Stuiver and Polach [1977] and include a δ13C correction. Backgrounds were subtracted using a similarly treated (processed) 14C-free calcite.

[20] Because 14C is produced in the upper atmosphere via cosmic ray bombardment of 14N, and mantle-derived CO2 contains no 14C, the apparent age of these calcites should fall on a simple two end-member mixing line. If we assume a modern end-member of 200 years before present (representing an integrated period of calcite formation), then older ages would confirm a mantle carbon contribution to the calcite. We obtained ages of 31,140, 19,940 and 26,300 14C years for the three Kilauea calcite samples (Table 3). These ages correspond to a mantle carbon contribution of 98, 91, and 96%, respectively.

[21] From these geochemical measurements it is clear that magmatic carbon was incorporated into the Keanakako'i calcite because of its close proximity to a steam vent or fumarole. Although mantle-derived δ13C values have been reported for groundwaters in volcanically active regions such as Mount Etna [Allard et al., 1997] and Mount Lassen [Rose et al., 1996], to our knowledge, this is the first time that the isotopic composition of mantle-derived carbon has been recorded in an authigenic mineral from an active volcano.

4.2. Possibility of Palagonitization by Syndepositional and Postdepositional Mechanisms

[22] Although we hypothesize that the initial palagonitization of the Keanakako'i tephras occurred in a hydrothermal environment, can we completely rule out an origin by syndepositional or postdepositional processes? The syndepositional model argues that the controlling factor for palagonitization is the volume of water (i.e., condensed steam) deposited within the tephra during phreatomagmatic eruptions [Wohletz and Sheridan, 1983]. This model predicts that within a given sequence of pyroclastic surge deposits, the “wetter” units (generally thicker and deposited at greater angles of repose) should be preferentially palagonitized with respect to the “dryer” units. However, zones of palagonitization within the Keanakako'i Ash Member are discordant with respect to bedding and are closely correlated with proximity to circumferential faults.

[23] The postdepositional model argues that the driving force for palagonitization is recharge of meteoric fluids through the tephra pile [Hay and Iijima, 1968]. In this model, originally invoked for hydrovolcanically deposited tuffs from Koko Craters, palagonitization is driven by downward percolating fluids whose chemistry has been strongly modified through continuous chemical reaction with the overlying glassy tephras. This model predicts that palagonitization should occur mainly above the existing water table but below a zone of relatively fresh tephra. At Kilauea, although all of the Keanakako'i tephras were deposited hundreds of meters above the water table (which has probably remained more or less in a constant position for the past few hundred years), zones of palagonitization are not overlain by unpalagonitized tephra. Interestingly, as viewed in plane-polarized light (Figure 5), the palagonitized Keanakako'i tephras are texturally similar to those from Koko Craters (albeit pore spaces in the latter tuffs are in-filled with zeolitic cements).

Figure 5.

Basaltic tephras (Figures 5a–5c) and hyaloclastite (Figure 5d) viewed in plane-polarized light. In all of the images the light orange-yellow grains are ash vitroclasts composed of fresh glass and rare (e.g., in Figure 5d) microphenocrysts. The darker, reddish brown coatings on the vitroclasts are palagonitized glass. The white, interclast regions in Figures 5a–5c are void pore spaces. In Figure 5d, which has very little pore space, the white interclast regions are composed of authigenic zeolites. (a) Fresh, unconsolidated, vitric tephra from the Keanakako'i Ash Member (919-1); (b) palagonitized, consolidated, vitric tephra from the Keanakako'i Ash Member (918-3); (c) palagonitized, consolidated vitric tephra from the Koko Craters tuff cone. For comparison, Figure 5d is a submarine hyaloclastite in core recovered from near the base of the Hawaiian Scientific Drilling Project's 3-km-deep hole; in this highly compacted, 500- to 550-ka hyaloclastite [DePaolo et al., 1999], the percentage of palagonitized rinds far exceeds that of the fresh sideromelane (Figure 4d). The scale bar in all four photos is 200 μm.

4.3. Rates of Palagonitization

[24] Regardless of the mechanism by which palagonitization of basaltic glass occurs, this process is generally thought to be both time- and temperature-dependent [Singer and Banin, 1990; Hekinian and Hoffert, 1975; Morgenstein and Riley, 1975; Moore, 1966]. Measured thicknesses of palagonitized rinds developed on glassy (mainly submarine) basalt of known ages suggest a low-temperature, steady state, diffusional (i.e., hydration) mechanism that follows the time-thickness relationship: T=(kt)1/2, where T is the rind thickness (in μm), k is a rate constant (μm2/ka), and t is time (in units of 103 years). Estimates for k vary from as low as 1 μm2/ka (for subglacial basaltic hyaloclastite from Iceland [Le Gal et al., 1999]) to as high as 2000 μm2/ka (for submarine basalt from the rift zones of Mauna Loa and Mauna Kea [Moore, 1966]). Thicknesses of rinds observed on the Keanakako'i vitric clasts (20–40 μm; Figures 2 and 3) might imply a rate constant similar to that of the submarine Hawaiian basalts, provided that palagonitization of the Keanakako'i tephras has been a constant process and one that initiated shortly after deposition.

[25] However, if the Keanakako'i tephras were hydrothermally affected, their palagonitization rates might have been substantially higher. Steam-bathed tephras from Surtsey were palagonitized at rates that effectively doubled with every 12°C increase in temperature [Jakobsson and Moore, 1986]. Sutseyan tephras exposed to temperatures between 40° and 60°C were palagonitized at rates between 0.5 and 1 μm/year. If the Keanakako'i tephras were palagonitized over a temperature range of approximately 30°–50°C, as the δ18O data on the calcite suggest, as few as 40 and as many as 400 years might have been necessary to produce the observed rinds. The latter time estimate implies that if palagonitization of the Keanakako'i tephras was indeed hydrothermally induced, steam vents should still have been active in the nineteenth century.

4.4. Palagonitization Versus Pedogenic Alteration of Keanakako'i Tephras

[26] In this study, we have shown that the texture and mineralogy of palagonitized versus pedogenically weathered Keanakako'i tephras are distinct. The former have undergone consolidation and cementation through the “welding” of palagonitized rinds on glassy fragments (Figures 2, 3, and 5). Moreover, the rinds themselves, although partly noncrystalline, contain crystallites of smectite [Schiffman et al., 2000]. Pedogenically affected tephras may contain smectite as well (e.g., sample 117-5), but only in relatively arid weathering regimes and, more important, as a component of complex pedogenic clay assemblages also containing kaolinite and opal (Table 1). Moreover, all of the pedogenically affected tephras in the study area are nonconsolidated to poorly consolidated and characterized by textures in which the pedogenic minerals, unlike the palagonitized rinds, are physically unattached to the parent glass fragments (compare Figure 1a and 1d with Figure 1c). Applied in unison, texture and mineralogy represent useful criteria for distinguishing these forms of low-temperature alteration of basaltic tephra.

4.5. Basaltic Tephras as Environmental Indicators

[27] Although incipient palagonitization of Keanakako'i tephras is spatially restricted to portions of Kilauea Caldera's circumferential fault system, it may also have been further restricted to regions where present-day soil pH is greater than ∼6.0. Tephras exposed along the major fault scarp west of Keanakako'i Crater have been palagonitized only as far west as roughly the south end of Halema'uma'u Crater (Figure 1b), where present-day soil pHs begin to fall below 6.0 owing to sulfur rain-out. Petrographic examination of these unconsolidated tephras (for example, from sample locality 918-5; Figure 1b) reveals no relict palagonitized rinds, implying that these tephras, presently undergoing acidic dissolution, were apparently never palagonitized. However, our present sampling density in this region is not sufficient to unequivocally demonstrate that soil pH has been a critical factor in controlling palagonitization.

[28] In the unpalagonitized portions of the Keanakako'i Ash Member we have recognized three distinct regimes of weathering, as defined by tephra alteration products that are consistent with the intensities of pedogenic desilication predicted from the precipitation gradient and soil acidity: (1) a relatively dry regime characterized by development of smectite (e.g., locality 117-5); (2) a relatively wet regime characterized by development of kaolinite and, with increasing intensity of desilication, short-range-order aluminosilicates (e.g., locality 118-1); and (3) an acidic regime characterized by marked glass dissolution and the development of opaline pendants on outcrop faces (e.g., locality 116-1). Recognition of these distinct forms of weathering in tephras, which are no greater than 500 years old, implies that vitric basaltic tephras are much more responsive to environmental conditions than previously appreciated. Although annual rainfall, temperature, and soil pH are presumably all critical environmental factors that control tephra alteration around Kilauea Caldera, we still do not know their relative importance and, perhaps more important, how they relate to the mechanisms (e.g., congruent versus incongruent dissolution) that result in palagonitization versus pedogenesis.


[29] This research was supported by grant CS-33-98 from the California Space Institute (P.S.) and a Faculty Research Grant from the University of California, Davis (P.S. and H.J.S.). Radiocarbon analyses were performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory (contract W-7405-Eng-48). We thank Tom Guilderson for facilitating the radiocarbon analyses. We also thank Rob Zierenberg for isotopic insights, Charlene Sailer for invaluable field assistance, and Hubert Staudigel, Bill White, and two anonymous reviewers for their helpful suggestions toward improving this manuscript.