Terry E.C. Keith

Minerals produced during cooling and hydrothermal alteration of ash flow tuff from Yellowstone drill hole Y-5

Abstract A rhyolitic ash-flow tuff in a hydrothermally active area within the Yellowstone caldera was drilled in 1967, and cores were studied to determine the nature and distribution of primary and secondary mineral phases. The rocks have undergone a complex history of crystallization and hydrothermal alteration since their emplacement 600,000 years ago. During cooling from magmatic temperatures, the glassy groundmass underwent either devitrification to alkali feldspar + α-cristobalite ± tridymite or granophyric crystallization to alkali feldspar + quartz. Associated with the zones of granophyric crystallization are prismatic quartz crystals in cavities similar to those termed miarolitic in plutonic rocks. Vapor-phase alkali feldspar, tridymite, magnetite, and sporadic α-cristobalite were deposited in cavities and in void spaces of pumice fragments. Subsequently, some of the vapor-phase alkali feldspar crystals were replaced by microcrystalline quartz, and the vapor-phase minerals were frosted by a coating of saccharoidal quartz. Hydrothermal minerals occur primarily as linings and fillings of cavities and fractures and as altered mafic phenocrysts. Chalcedony is the dominant mineral related to the present hydrothermal regime and occurs as microcrystalline material mixed with various amounts of hematite and goethite. The chalcedony displays intricate layering and was apparently deposited as opal from silica-rich water. Hematite and goethite also replace both mafic phenocrysts and vapor-phase magnetite. Other conspicuous hydrothermal minerals include montmorillonite, pyrite, mordenite, calcite, and fluorite. Clinoptilolite, erionite, illite, kaolinite, and manganese oxides are sporadic. The hydrothermal minerals show little correlation with temperature, but bladed calcite is restricted to a zone of boiling in the tuff and clearly was deposited when CO2 was lost during boiling. Fractures and breccias filled with chalcedony are common throughout Y-5 and may have been produced by rapid disruption of rock caused by sudden decrease of fluid pressure in fractures, most likely a result of fracturing during resurgent doming in this part of the Yellowstone caldera. The chalcedony probably was deposited as opal or β-cristobalite from a pre-existing silica floc that moved rapidly into the fractures and breccias immediately after the sudden pressure drop.

Fossil and active fumaroles in the 1912 eruptive deposits, Valley of ten thousand smokes, Alaska

Fumaroles in the ash-flow sheet emplaced during the 1912 eruption of Novarupta were intensely active throughout the Valley of Ten Thousand Smokes (VTTS) when first studied in 1917. Fumarole temperatures recorded in 1919 were as hot as 645°C. Influx of surface waters into the hot ash-flow sheet provided the fluid flow to sustain the fumaroles but also enhanced cooling so that by the mid-1930s vigorous activity survived only in the vent region. Configuration and distribution of high-temperature fissure fumaroles tens of meters long, that are prevalent in the middle and upper VTTS, were controlled largely by sintering and degree of welding, which in turn controlled fracturing and permeability of the ash-flow tuff. One fracture type developed parallel to the enclosing valley walls during compaction of the ash-flow sheet. Another type extends across the VTTS nearly perpendicular to the flow direction. A third type of randomly oriented fractures developed as cooling contraction cracks during vapor-phase devitrification. In distal parts of the ash-flow sheet where the tuff is nonwelded, prominent fumaroles have irregular funnel-shaped morphologies. Fumarole distribution in the nonwelded part of the ash-flow sheet is concentrated above pre-emplacement river channels. The hottest, longest-lived fumaroles occurred in the upper VTTS near the 1912 vent where the ash-flow sheet is thicker, more indurated, and on average more mafic (richer in dacite and andesite) in contrast to the thinner, nonwelded rhyolitic tuff in the distal part of the sheet. Fumarolic activity was less intense in the distal part of the tuff because of lower emplacement temperatures, more diffuse fumarole conduits in the nonwelded tuff, and the thinness of the ash-flow sheet. Chemical leaching of ash-flow tuff by hot rising fluids took place adjacent to fumarolic conduits in deep parts of the fumaroles. Deposition of incrustation minerals, the components of which were carried upward by fumarolic gases, took place in the upper part of the ejecta, mostly in the fallout layers. The permeability difference between the ash-flow tuff and the overlying coarse dacite fallout was a critical factor in promoting the abrupt gradients in temperature, pressure, and fO2 that resulted in deposition of minerals from the fumarolic gases. The permeability difference between nonwelded ash-flow tuff and overlying fine-grained fall layers in the lower VTTS is less pronounced. The total mass of fumarolically deposited minerals appears large at first glance owing to the conspicuous coloration by Fe minerals; the mass is appreciably less than is apparent, however, because most incrustations are composed largely of ejecta coated or cemented by fine-grained fumarolic minerals. A large mass of unstable incrustation minerals, mainly chlorides and sulfates, reported during the 1917–1919 studies have since been removed by dissolution and weathering. In the vent region, argillic alteration that followed high-temperature degassing is localized along arcuate subsidence fractures in fallback ejecta. At widely scattered residual orifices, fumarolic gases presently are near-neutral steam, and temperatures are as hot as 90°C.

Oxygen and carbon isotope ratios of hydrothermal minerals from Yellowstone drill cores

Abstract Oxygen and carbon isotope ratios were measured for hydrothermal minerals (silica, clay and calcite) from fractures and vugs in altered rhyolite, located between 28 and 129 m below surface (in situ temperatures ranging from 81 to 199°C) in Yellowstone drill holes. The purpose of this study was to investigate the mechanism of formation of these minerals. The δ18O values of the thirty-two analyzed silica samples (quartz, chalcedony, α-cristobalite, and β-cristobalite) range from −7.5 to +2.8‰ . About one third of the silica 7samples have δ18O values that are consistent with isotopic equilibrium with present thermal waters; most of the other silica samples appear to have precipitated from water enriched in 18O (up to 4.7‰) relative to present thermal water, assuming precipitation at present in situ temperatures. Available data on fluid-inclusion homogenization temperatures in hydrothermal quartz indicate that silica precipitation occurred mostly at temperatures above those measured during drilling and imply that 15O enrichments in water during silica precipitation were generally larger than those estimated from present conditions. Similarly, clay minerals (celadonite and smectite) have δ18O values higher (by 3.5 to 7.9‰) than equilibrium values under present conditions. In contrast, all eight analyzed calcite samples are close to isotopic equilibrium with present thermal waters. The frequent incidence of apparent 18O enrichment in thermal water from which the hydrothermal minerals precipitated may indicate that a higher proportion of strongly 18O-enriched deep hydrothermal fluid once circulated through shallow portions of the Yellowstone system, or that a recurring transient 18O-enrichment effect occurs at shallow depths and is caused either by sudden decompressional boiling or by isotopic exchange at low water/rock ratios in new fractures. The mineralogy and apparent 18O enrichments of hydrothermal fracture-filling minerals are consistent with deposition during transient boiling or rock-water exchange (fracturing) events.

Selective concentration of cesium in analcime during hydrothermal alteration, Yellowstone National Park, Wyoming

Abstract Chemical and mineralogical studies of fresh and hydrothermally altered rhyolitic material in Upper and Lower Geyser Basins, Yellowstone National Park, show that all the altered rocks are enriched in Cs and that Cs is selectively concentrated in analcime. The Cs content of unaltered rhyolite lava flows, including those from which the altered sediments are derived, ranges from 2.5 to 7.6 ppm. The Cs content of analcime-bearing altered sedimentary rocks is as high as 3000 ppm, and in clinoptilolite-bearing altered sedimentary rocks Cs content is as high as 180 ppm. Altered rhyolite lava flows which were initially vitrophyres, now contain up to 250 ppm Cs, and those which were crystallized prior to hydrothermal alteration contain up to 14 ppm. Mineral concentrates of analcime contain as much as 4700 ppm Cs. The Cs must have been incorporated into the analcime structure during crystallization, rather than by later cation substitution, because analcime does not readily exchange Cs. The Cs Cl of the fluids circulating through the hydrothermal system varies, suggesting that Cs is not always a conservative ion and that Cs is lost from upflowing thermal waters due to water-rock interaction resulting in crystallization of Cs-bearing analcime. The source of Cs for Cs enrichment of the altered rocks is from leaching of rhyolitic rocks underlying the geyser basins, and from the top of the silicic magma chamber that underlies the area. Analcime is an important natural Cs sink, and the high Cs concentrations reported here may prove to be an important indicator of the environment of analcime crystallization.

Geochemistry of waters in the Valley of Ten Thousand Smokes region, Alaska

Abstract Meteoric waters from cold springs and streams outside of the 1912 eruptive deposits filling the Valley of Ten Thousand Smokes (VTTS) and in the upper parts of the two major rivers draining the 1912 deposits have similar chemical trends. Thermal springs issue in the mid-valley area along a 300-m lateral section of ash-flow tuff, and range in temperature from 21 to 29.8°C in early summer and from 15 to 17°C in mid-summer. Concentrations of major and minor chemical constituents in the thermal waters are nearly identical regardless of temperature. Waters in the downvalley parts of the rivers draining the 1912 deposits are mainly mixtures of cold meteoric waters and thermal waters of which the mid-valley thermal spring waters are representative. The weathering reactions of cold waters with the 1912 deposits appear to have stabilized and add only subordinate amounts of chemical constituents to the rivers relative to those contributed by the thermal waters. Isotopic data indicate that the mid-valley thermal spring waters are meteoric, but data is inconclusive regarding the heat source. The thermal waters could be either from a shallow part of a hydrothermal system beneath the 1912 vent region or from an incompletely cooled, welded tuff lens deep in the 1912 ash-flow sheet of the upper River Lethe area. Bicarbonate-sulfate waters resulting from interaction of near-surface waters and the cooling 1953–1968 southwest Trident plug issue from thermal springs south of Katmai Pass and near Mageik Creek, although the Mageik Creek spring waters are from a well-established, more deeply circulating hydrothermal system. Katmai caldera lake waters are a result of acid gases from vigorous drowned fumaroles dissolving in lake waters composed of snowmelt and precipitation.

Fluid-inclusion evidence for past temperature fluctuations in the Kilauea East Rift Zone geothermal area, Hawaii

Heating and freezing data were obtained for fluid inclusions in hydrothermal quartz, calcite, and anhydrite from several depths in three scientific observation holes drilled along the lower East Rift Zone of Kilauea volcano, Hawaii. Compositions of the inclusion fluids range from dilute meteoric water to highly modified sea water concentrated by boiling. Comparison of measured drill-hole temperatures with fluid-inclusion homogenization-temperature (Th) data indicates that only about 15% of the fluid inclusions could have formed under the present thermal conditions. The majority of fluid inclusions studied must have formed during one or more times in the past when temperatures fluctuated in response to the emplacement of nearby dikes and their subsequent cooling. The fluid-inclusion data indicate that past temperatures in SOH-4 well were as much as 64°C hotter than present temperatures between 1000 and 1500 m depth and they were a maximum of 68°C cooler than present temperatures below 1500 m depth. Similarly, the data show that past temperatures near the bottoms of SOH-1 and SOH-2 wells were up to 45 and 59°C, respectively, cooler than the present thermal conditions; however, the remainder of fluid-inclusion Th values for these two drill holes suggest that the temperatures of the trapped waters were nearly the same as the present temperatures at these slightly shallower depths. Several hydrothermal minerals (erionite, mordenite, truscottite, smectite, chlorite-smectite, chalcedony, anhydrite, and hematite), occurring in the drill holes at higher temperatures than they are found in geothermal drill holes of Iceland or other geothermal areas, provide additional evidence for a recent heating trend.

Factors controlling the geochemical evolution of fumarolic encrustations, Valley of Ten Thousand Smokes, Alaska

Abstract Factor and canonical correlation analysis of geochemical data from eight fossil fumaroles suggest that six major factors controlled the formation and evolution of fumarolic encrustations on the 1912 ash-flow sheet in the Valley of Ten Thousand Smokes (VTTS). The six-factor solution model explains a large proportion (low of 74% for Ni to high of 99% for Si) of the individual element data variance. Although the primary fumarolic deposits have been degraded by secondary alteration reactions and up to 75 years of weathering, the relict encrustations still preserve a signature of vapor-phase element transport. This vapor-phase transport probably occurred as halide or oxyhalide species and was significant for As, Sb and Br. At least three, and possibly four, varied temperature leaching events affected the fumarolic deposits. High-temperature gases/liquids heavily altered the ejecta glass and mineral phases adjacent to the fumarolic conduit. As the fumaroles cooled. Fe-rich acidic condensate leached the ejecta and primary fumarolic deposits and resulted in the subsequent precipitation of Fe-hydroxides and/or Fe-oxides. Low- to ambient-temperature leaching and hydration reactions generated abundant hydrated amorphous phases. Up to 87% of the individual element data variance is apparently controlled by the chemistry of the ejecta on which the relict encrustations are found. This matrix chemistry factor illustrates that the primary fumarolic minerals surrounding the active VTTS vents observed by earlier workers have been effectively removed by the dissolution reactions. Element enrichment factors calculated for the VTTS relict encrustations support the statistical factor interpretations. On the average, the relict encrustations are enriched, relative to visibly unaltered matrix protolith, in As, Br, Cr, Sb, Cu, Ni, Pb, Fe, and LOI (an indirect measure of sample H 2 O content).

Chemical and thermal constraints on models of Thermal Springs Valley of Ten Thousand Smokes, Alaska

Thermal (15–30°C) springs issuing from 1912 ash-flow tuff in the mid-valley region of the Valley of Ten Thousand Smokes consist of meteoric water that may have been heated by an incompletely cooled lens of welded tuff upvalley from the springs. Conductive cooling of the thermal waters along the flow path is the likely cause for the difference between the observed spring temperatures and the source temperature inferred from chemical geothermometry. Conductive cooling alone can not easily account for the seasonal fluctuations in spring temperatures, however. Mixing of the thermal waters with cold meteoric waters would seem a likely possibility; but thermal water chemistry is constant, indicating that mixing does not occur after the water leaves the zone of heating. Even if mixing occurred, simple mixing models do not account totally for the observed temperature differences of the springs from late spring to summer. The geochemical and thermal data argue for a complex hydrological and thermal regime.

Isotope geochemistry of minerals and fluids from Newberry volcano, Oregon

Isotopic compositions were determined for hydrothermal quartz, calcite, and siderite from core samples of the Newberry 2 drill hole, Oregon. The δ15O values for these minerals decrease with increasing temperatures. The values indicate that these hydrothermal minerals precipitated in isotopic equilibrium with water currently present in the reservoirs. The δ18O values of quartz and calcite from the andesite and basalt flows (700–932 m) have isotopic values which require that the equilibrated water δ18O values increase slightly (− 11.3 to −9.2‰) with increasing measured temperatures (150–265°C). The lithic tuffs and brecciated lava flows (300–700 m) contain widespread siderite. Calculated oxygen isotopic compositions of waters in equilibrium with siderite generally increase with increasing temperatures (76–100°C). The δ18O values of siderite probably result from precipitation in water produced by mixing various amounts of the deep hydrothermal water (− 10.5 ‰) with meteoric water (− 15.5 ‰) recharged within the caldera. The δ13C values of calcite and siderite decrease with increasing temperatures and show that these minerals precipitated in isotopic equilibrium with CO2 of about −8 ‰. The δ18O values of weakly altered (<5% alteration of plagioclase) whole-rock samples decrease with increasing temperatures above 100°C, indicating that exchange between water and rock is kinetically controlled. The water/rock mass ratios decrease with decreasing temperatures. The δ18O values of rocks from the bottom of Newberry 2 show about 40% isotopic exchange with the reservoir water. The calculated δ18O and δD values of bottom hole water determined from the fluid produced during the 20 hour flow test are −10.2 and −109‰, respectively. The δD value of the hydrothermal water indicates recharge from outside the caldera.

Further insights into the geochemical evolution of fumarolic alteration, Valley of Ten Thousand Smokes, Alaska

Abstract Factor analysis of geochemical data from nine fossil fumaroles and four warm ground argillic alteration sites indicates that the fossil and extant fumarolic alteration of the 1912 ash-flow sheet in the Valley of Ten Thousand Smokes (VTTS) integrate a complex overprinting of primary and secondary alteration events. The five-factor solution model explains 77% of the complete data set variance and a large proportion (60–92%) of the individual element data variance. These data support halide-species (e.g., chlorides and fluorides) vapor-phase element transport having occurred during the cooling of the fumaroles and that this process was significant for As, Sb, B and Bi, and of lesser importance for Cu and Zn. Similarities in geochemical element covariance between the extant argillic alteration and fossil fumarolic encrustations suggest that many of the fossil fumaroles experienced low-temperature argillic alteration events. The combination of the results of this work with those of Kodosky and Keith (1993) enables a general history of the VTTS fumaroles to be reconstructed. During the higher-temperature stages of fumarolic activity, substantial halide-species vapor-phase element transport likely occurred. The high-temperature gases heavily altered the ejecta glass and mineral phases adjacent to the fumarolic conduit. As the fumaroles cooled, Fe-rich acidic condensate leached the ejecta and primary fumarolic deposits (Kodosky and Keith, 1993); aqueous chloride complexes were the primary mechanism of element transport. With further declines in temperature, many of the fumaroles developed moderate- to low-temperature argillic alteration. Low- to ambient-temperature leaching and ongoing hydration reactions subsequently produced abundant hydrated amorphous phases (Kodosky and Keith, 1993). Although relict mineralogical evidence of argillization remains at some of the fossil fumaroles, at many of the sites the secondary alteration reactions have removed or obscured the record of prior argillization events.