American Mineralogist. Volume 76, pages 1662-1673, 1991

Geochemistry and mineralogy of fumarolic deposits, Valley of Ten Thousand Smokes,Alaska: Bulk chemical and mineralogical evolution of dacite-rich protolith

J. J. PlprxoInstitute of Meteoritics, Department of Geology, University of New Mexico, Albuquerque, New Mexico 87131, U.S.A.

T.E.C. KorrHU.S. Geological Survey, MS 910, Menlo Park, California 94025, U.S.A.

M. N. Sprr,onInstitute of Meteoritics, Department of Geology, University of New Mexico, Albuquerque, New Mexico 87131, U.S.A.

K. C. G.ILSREATHCombustion and Environmental Systems Research Institute, Energy and Environmental Research Center,University of North Dakota, Box 8213, University Station, Grand Forks, North Dakota 58202, U.S.A.

C. K. Snn.c.nBnInstitute of Meteoritics, Department of Geology, University of New Mexico, Albuquerque, New Mexico 87131, U.S.A.

J. C. LluLRocky Flats, EG&G, P.O. Box 464, Golden, Colorado 80402, U.S.A.

AssrRAcr

Samples from a fossil fumarole originating in the 1912 ash-flow tuffin the Valley ofTen Thousand Smokes have been analyzed to ascertain chemical changes resulting fromhigh-temperature fumarolic alteration and subsequent cooling and weathering of the pro-tolith. Samples of the underlying, dominantly leached, dacite-rich portion of the ash-flowtuff adjacent to the fumarolic conduit and samples of encrusted fallout from the shallowpart of the fossil fumarole were interpreted using the isocon method of Grant (1986). Theresults show that, relative to unaltered l9l2 dacite, chosen as a standard composition forthe protolith in this fossil fumarole, mass was conserved during the alteration reactionsfor most of the system, but mass gains of l4-2Do/o were determined for three samples inthe leached ash-flow tuff Relative to unaltered dacite protolith, significant enrichmentsoccurred in SOr, LOI (-HrO), Cl, F, Zn, Pb, Cu, Sn, Cr, Ni, As, Sb, Au, Br in variousparts of the fossil fumarole. Some of these were during the high-temperature part of thealteration, and some were during cooling processes when acid alteration becomes promi-nent. The REEs indicate some depletion in highly altered samples relative to dacite pro-tolith and differential mobility of Eu2* relative to trivalent REEs. This is manifested bypositive Eu anomalies in REE patterns normalized against REE in the dacite protolith.

Mineral phases introduced in the alteration assemblages include alunite reflecting highSO, activity, hydrated aluminum hydroxy-fluoride (a ralstonite-like phase) and fluoritereflecting high F activity, smectite, magnetite, hematite, and goethite reflecting oxidationand hydration reactions. Opal and a portion of the a-cristobalite reflect SiO, mobility;however, the abundance of a-cristobalite is formed from pumice leached during high-temperature vapor-phase processes and devitrification of the altered glass.

INrnooucrroN

Spectacular fumaroles developed in the ash-flow sheetemplaced during the 6-8 June l9l2 eruption from No-varupta, causing Robert F. Griggs, who discovered thesteaming valley in 19 16, to name the Valley of Ten Thou-sand Smokes (VTTS) (Griggs, 1922).Early studies of theactive fumaroles of VTTS were reported by Shipley (1920),Allen and Zies (1923), and Zies (1929). Fumarole tem-peratures as high as 645 qC were recorded in l9l9 (AllenandZies, 1923), and these have cooled dramatically since

initial activity. Ramdohr (1962) studied samples of newlydeposited magnetite collected in l9l9 from fumarole 148of Zies (Allen and Zies, 1923) in the mid-valley regionusing reflected light, and he documented a variety of mi-croscopic sulfides included in the magnetite. Lovering( 1957) conducted a geochemical study on a suite of sam-ples from fumarole no. I of Zies (Allen and Zies, 1923),in which he concluded that halogen-acid alteration differsfrom S-acid alteration chiefly in the greater loss of silicarelative to alumina. Keith (1983, 1984, l99l) interpretedthe origin and distribution of fumaroles in the VTTS, in

0003-004xr9 l/09 l0- I 662$02.00 t662

part drawing upon observations of Allen and Zies (1923).Heat to generate and sustain fumaroles was provided bythe ash-flow sheet. Abundant meteoric HrO flowing intoand over the ash-flow sheet helped sustain the fumarolesand also hastened their cooling. The fumarolic fluids alsoincluded exsolved magmatic gases (Allen and Zies, 1923).Most of the fumaroles cooled and died out by the mid-1930s except in the Novarupta vent region, where resid-ual fumaroles remain with temperatures as hot as 90'C(Keirh, 1991).

We have initiated a project to study the geochemistryand mineralogy of the VTTS fossil fumaroles (Papike etal., 1989, 1990a, 1990b, 1990c, l99l). Our major objec-tives are to (l) assess chemical losses and gains in ash-flow and fallout deposit$ that host the fumaroles, (2) de-termine mineralogical changes and reaction paths, and (3)determine chemical and crystal-chemical aspects of phe-nocryst alteration. This paper describes the bulk chemicaland mineralogical evolution of a well-exposed, typicalfossil fumarole (site 212) in the midvalley region (Fig. 1).

Studies of deposits at active fumaroles in other partsof the world (e.9., Stoiber and Rose, 1974; Symonds etal., 1987 , 1990; Quisefit et al., I 989; Kodosky and Kes-kinen, 1990) have greatly increased our understanding ofthe geochemistry and mineralogy of these complex de-posits. In contrast, our studies of VTTS fossil fumarolesdo not provide as simple a picture of the processes op-erative in fumaroles as these referenced studies but in-stead charactetze ejecta that have been subjected not onlyto high-temperature fumarolic activity but also to alter-ation during cooling and weathering (Keith, l99l). Thesamples we are studying represent an integated history ofprocess and time as depicted schematically in Figure 2.

Gnor,ocrc sETTTNG

The 1912 eruption of Novarupta in the Katmai region(Fig. l) is regarded by many as the outstanding igneousevent of this century (Eichelberger and Hildreth, 1986).During June 6-8, 1912, approximately 12 km3 of magmaerupted from the Novarupta vent at the head of the Val-ley of Ten Thousand Smokes (VTTS), Alaska, producing- 17 km3 of fallout, I I + 4 km3 of ash-flow tufl and - Ikm3 of heterogeneous proximal fallout within -60 h (Hil-dreth, I 983, I 987, I 990). Curtis ( I 968) first separated outthe 1912 eruptive stratigraphy and described the falloutdeposits as A-H layers, and Hildreth (1983) presenteddetailed initial petrologic studies of the 1912 deposits.Initial composition of the ash-flow sheet was rhyolite (77o/oSiOr), and increasing amounts of dacite (66-64.50/o SiOr)and andesite (61.5-58.5olo SiOr) were introduced later inthe eruptive sequence (Hildreth, 1983). The exposures ofthe ash-flow sheet presontly frlling the VTTS are rhyoliterich and nonwelded at the distal end -15-20 km fromthe vent; exposures in the middle and upper VTTS aresintered and consist of varying proportions of rhyolite,dacite, and andesite with minor amounts of banded pum-ice and lithic fragments (Hildreth, 1983; Fierstein andHildreth, in preparation). Andesitic scoria is common in

PAPIKE ET AL.: FUMAROLIC DEPOSITS

Thre€ Forks /\

1663

in GdggsA'

Katmai cldeE

Katmai Pass

AObseNation Mn

EXPTANATION

@ l9l2ash.| lowsheet

E Funneltumaroles

- Fi$u@ tumarcles

Fig. l. Location map of Valley of Ten Thousand Smokes(VTTS), Alaska, (insert) and schematic map showing distribu-tion of the l9l2 ash-flow sheet and location of fumarole 212;modified from Keith (1991).

the near-vent, thin ash-flow layers and fallout. Maximumthickness of the ash-flow tuff is unknown but is likely-200 m in the upper River kthe area and almost asthick in the upper Knife Creek area (Curtis, 1968; Kienle,1970).

Phenocrysts of plagioclase, orthopyroxene, titanomag-netite, ilmenite, and trace amounts of apatite and pyr-rhotite are present in all pumice compositions, althoughrhyolitic pumice has only l-2Vo total phenocryst contentand dacite and andesite may have as much as 450lo (Hil-dreth, 1983). Clinopyroxene is common in dacite andandesite; quartz and scarce amphibole occur in rhyolite;andesite contains scarce olivine (Hildreth, 1983). The ini-tial fall unit underlying the ash-flow sheet is 1000/o rhy-olite, fall unit B on top of the ash-flow tuf is mixed up-ward with increasing dacite and andesite relative torhyolite, and fall units C-H on top of the ash flows are>980/o dacite (Fierstein and Hildreth, 1990, and in prep-aration).

Configuration and distribution of fumaroles through-out the VTTS was controlled by the degree of welding ofthe ash-flow sheet, which in turn controlled fracturingand permeability (Keith, l99l). Hotter, longer-lived fu-maroles occurred in the upper VTTS closer to the ventwhere the ash-flow sheet was thicker, more indurated,and more mafic (dacite and andesite) in contrast to thin-ner nonwelded rhyolitic tuff in the distal part of the ash-

ng Mn

A Tdderno 5 h

9 t7 /A 212

\ . "

irt irageikA

PRESENT SAMPLES

Represent an integrated history of process and tim€

FUMAROLES

1 Altered protolith <-Volcanic / meteoric lluid streaming.-

Surface weathering2. Incrustations--.\ Cooler / more oxidizing3. Sublimates -./ portion of system

VOLCANIC FLUIDS

1 Vapor phase exsolulion:HzO, COz, CO, SOz,H2SOa, HCt, HF, HBr

2 Elements €rried: e gNa, K. Si, Fe, Ca, Ti, Mg,Zn, Cu, Pb, Cd, As, Sc,Mo, Sn, Ag, Au

t664 PAPIKE ET AL.: FUMAROLIC DEPOSITS

U

F

zaUrI

z

1912 TO PRESENT

FrE.2. Diagram illustrating the processes involved in fossilfumarole development and thus the mineralogical and geochem-ical complexity of the fossil fumaroles in the VTTS.

flow sheet. The distribution of high-temperature fuma-roles prevalent in the middle and upper VTTS was con-trolled by fractures in the ash-flow sheet.

Keith (1991) determined that high-temperature fossilfumarolic incrustations exposed in the VTTS consistmostly of fall deposits encrusted mainly with either mag-netite or hematite depending on /o, and temperature ofdeposition. Hematite and goethite also formed from al-tered magnetite during cooling processes. Fumarolicallydeposited a-cristobalite coexists with magnetite in high-temperature fossil fumaroles. The outer, cooler parts ofa fossil fumarole typically have a relatively impermeableshell consisting of several of the following minerals: alu-nite, AHF ftydrated aluminum hydroxy-fluoride, a ral-stoniteJike phase described by Desborough and Rostad,1980), fluorite, a- and B-cristobalite, opal, pyrite, S, ka-olinite, illite, goethite, and amorphous hydrous Fe oxide(limonite). A large volume of unstable fumarolic depositswere removed by dissolution while the fumaroles wereactive (Keith, 1991).

The fumarole selected for this study is part of an extincthigh-temperature, fissure fumarole (Keith, 199 1) exposedon the northeast side of River Lethe in the midvalleyregion (Fig. l). The upper 2 m of the ash-flow sheet atthe fumarole site is composed mainly of dacite with sub-

ordinate amounts of andesite, rhyolite, banded pumice,and lithic fragments (Hildreth, 1983; Fierstein and Hil-dreth, unpublished pumice count data, l99l). Below 2 min the ash-flow sheet, the composition decreases in amountof dacite and increases in amounts of andesite and rhy-olite (Hildreth, 1983). The fall deposits overlying the ash-flow sheet at fumarole 212 are of mixed andesite, dacite,and rhyolite at the base and grade quickly upward into>980/o dacite (Curtis, 1968; Hildreth, 1983; Fierstein andHildreth, 1990).

A vertical joint in the ash-flow sheet was the fumarolicconduit for vigorously rising, hot fluids that leached andsintered the adjacent tuff in the deeper parts of the fu-marolic system (Keith, l99l). Upon contact with theloose, permeable fallout overlying the ash-flow sheet, astate ofdisequilibrium was caused by rapid cooling andchanges in gas fugacities that resulted in sublimation ofmineral components carried in the vapor phase, formingincrustations on fallout clasts. Other chemical compo-nents were transported into the atmosphere. Dependingon proximity to the hot fumarolic fluids, various statesofalteration ofthe ash-flow sheet and fall deposits occur.Phenocrysts and pumice clasts in the fall deposits werepartly shielded from fumarolic alteration where fumarol-ic minerals enclosed the clasts and where the clasts werenot in the main path of hoq rising fumarolic fluids. Al-teration during subsequent cooling, particularly thatcaused by the condensation of fumarolic fluids and mix-ing of fumarolic gases with surface HrO at the upper andouter parts of the fumarole, resulted in acid alteration ofboth primary tephra and earlier deposited fumarolic min-erals.

Samples for the present study are bulk samples, con-sisting oflarge fist-sized pieces ofthe altered ash-flow tuffand gollball-sized pieces of encrusted fallout, that werecollected systematically from the leached and deposition-al parts of the fumarolic system (Fig. 3). The ash-flow tufis leached progessively less away from the fumarole con-duit (Fig. 3b); sample J is adjacent to the conduit and isthe most leached, whereas sample O is the farthest fromthe conduit and was expected to be unaltered. Samples Athrough H (Frg. 3a) are from dacite-rich fall deposits en-crusted with fumarolic minerals. Samples A and B arefrom the outer, cooler part of the encrusted fallout and Cthrough H are from the inner part near the conduit wherefumarolic gases reached the surface.

S-lndpI,n pREpARATToN AND ANALyTTCAL METHODS

Powders for X-ray diftaction (XRD), X-ray fluores-cence (XRF), and instrumental neutron activation anal-ysis (INAA) were prepared in a Spex hardened steel ballmill, with XRD powders being further ground by handto approximately l0 pm in a ceramic mortar and pestle.Powders for Fe2*/Fe3* determinations were ground in acorundum mortar. Samples for F analysis were powderedin a Spex hardened steel puck mill.

Whole-rock powders were analyzed for major oxidesand trace elements by energy dispersive X-ray fluores-

cence (EDXRF) and atomic absorption spectroscopy(AAS) at Battelle, Pacific Northwest Laboratories, fuch-land, Washington. Analyses of U.S. Geological Surveystandards AGV-I, BCR-I, and G-2 (Flanagan, 1967,1969,1973) indicate precisions of better than 3olo for Si,Ti, Al, Fe, Mg, Ca, K, andZn;5o/o for Na, V, Y, and Pb;100/o for Mn, Nb, Cu, and Ga. The elements Rb, Ba, Zr,Sr, As, Sb, and REE were analyzed at Battelle, PacificNorthwest Laboratories, by sequential INAA using a high-efficiency 130 cm3 Ge(Li) detector (250lo FWHM 1.8 keVfor 1332 keV of60Co), a 4096 channel analyzer, and co-incidence-noncoincidence Ge(Li) : NaI (Tl) countingsystems (Laul, 1979; Laul et al., 1984). HrO was deter-mined indirectly by loss on ignition (LOI) at Battelle,Pacifi c Northwest Laboratories.

Whole-rock F was measured at South Dakota Schoolof Mines and Technology using a fluoride ion-selectiveelectrode technique (Bodkin, 1977), and Fe2+ was deter-mined using an HF dissolution and potassium dichro-mate titration technique of Goldich (1984). Whole-rockCl was determined at the U.S. Geological Survey, Branchof Geochemistry Laboratory in Menlo Park, Californiaby ion selective electrode (Aruscavage and Campbell,l 983).

XRD procedures and modal determinations

X-ray diffraction, reference intensity method (XRD-RIM) modes were determined by the method describedby Shearer et al. (1988).

Finely ground powders were loaded onto WhatmanGF/C glass-fiber filters using an aerosol suspension tech-nique that significantly decreases preferred orientation(Davis, 1986). Direct beam X-ray transmission analysiswas done on both blank and loaded filters to providesample mass absorption coefficients. The filters were thenmounted in the spinning stage of a Philips diffractometerand scanned from 5 to 60 20 using 0.02'steps and 1.5 scount per step (0.8'/min). Monochromatic CuKa radia-tion at 40 kv and 20 mA and a graphite monochromatorwere used for diffraction and transmission scans.

The reference intensity ratio (RIR) method of Chung(1974a, 1974b) modified by Davis (1984) was used forquantitative mineral analysis of the samples. Samples Athrough H were each mixed with an equal weight of AlrO,(synthetic corundum; a: 4.758 A, s: 12.991 A) as aninternal standard. Integated peak intensities were adjust-ed for the use of automatic divergence slits, and the sam-ple mass absorption coefficient was used in adjusting forsample transparency and matrix efects on the loaded fil-ter. Adjusted intensities (1r) of samples A through H werethen used, along with the RIR (&), in the relation

w,:+ ! (r)' t I " k ,

where W, is the weight fraction of each crystalline com-ponent in the sample; W"and { refer to the added weightfraction and integrated intensity of the corundum stan-dard (Chung, 1974a). The addition of an internal stan-

PAPIKE ET AL.: FUMAROLIC DEPOSITS r665

212 A-H

Venl l=] AFH+labradorite+hematite+|- cristobalitetsmectitel|luorite

212 J-O

0 1 0 2 0

I conouit tr

l-l a-cristouatite >> ptagioctase tr

30 cm

plagioclase > o-cristobalite >>alunite > pyroxene > goethite

plagioclase > a-cristobalite >>alunite > pyroxene > goethile

ffi o-cristobarite=prasiocrase_>> ffi 3jtpf,.ffi;3nT",n0",fi:3 alunite > goethite > pyrcxene H itr"ni"

Fig. 3. (a) Schematic cross section showing locations of sam-ples from fossil fumarole 212 in the upper, encrusted fallout thatoverlies the ash-flow tuffof b. Geologic hammer indicates scale.(b) Schematic map showing location of sarnples from the ash-flow tuf portion of the fossil fumarole. This section underliesthe encrusted fallout deposits of a. Crystalline phase assemblagesand their relative abundances are indicated. Plagioclase, pyrox-ene, ilmenite, and some magnetite are primary and in variousstages ofalteration; a-cristobalite is mostly from devitrificationof pumice during high-temperature vapor-phase leaching; alu-nite and goethite are alteration deposits.

dard allows the independent calculation of the weightfraction of each phase; thereby the total amorphous weightfraction can be found by difference. By subtracting thecrystalline contribution of the mass absorption from thatof the sample, the contribution of one or two amorphousphases may be calculated from the remainder. However,optical studies of the fumarole 212 samples indicate thatas many as three amorphous phases (pumice, opal, andlimonite) can be present. In that case, one phase must bedetermined optically from grain mounts, which is usuallylimonite as it is readily distinguishable from most otherphases.

Optical estimates could be made for the crystalline/amorphous ratio for the leached samples J through O,

b

orange-reo

t666 PAPIKE ET AL.:FUMAROLIC DEPOSITS

TABLE 1. XRD modal analyses of 212 traverse

Crystalline fraction Amorphous fraction

Cristo- Hema-Samole AHF Alunite balite Fluorite tite llmenite

Plagio- Magne- Pyrox- Total Limo- Obsid-clase tite ene Smectite crystal. nite ian

Goe-thite

TotalOpal amorph.

212A 29.09212FJ 12.59212C2120212E 1.48212F212G2't2H212J212K212L212M21 2N2120

0.490.681.491.540.324.301 . 1 80.33

22.9517.8112.7612.6512.8814.70

0.630.871.690.892.29

1 . 1 61.64

1 .O2

0.570.350.55

0.31 0.490.42 0.57

6.3418 .7018.2224.1719.41 1.7115 .9115.3612.7210.65'12.78

17.2117.4818.9321 .56

3.45 40.0033.86

3.85 25.251.19 2.20 30.001.82 30.00't4.79 35.002.81 5.11 25.62

14.6933.60

0.s0 34.001 .15 33.000.96 33.001 .48 34.102.74 40.00

5.60 54.405.15 60.99

61.274.95 65.05

63.403.74 61.26

72.9441.85

0.80 58.1714.39 51 .61

39.5153.1262.6260.00

60.0066.14

13.49 74.7570.00

6.60 70.0065.00

1.44 74.3843.46 85.317.43 66.40

66.0026.49 66.0012.88 66.003.28 65.90

60.00

2.331.311.36

Notej Samples A-H, amorphous phases calculated; J-O amorphous/crystalline ratio estimated optically.

since they contained no isotropic minerals. Therefore thewelght fractions for these samples were calculated usinga slightly different procedure in which no internal stan-dard is added, but the RIR is used in the relation fromChung (1974b):

Further details of both techniques can be found in Davis(1984). All Wj sum to unity, and the relative weight frac-tions of the crystalline components are correct with re-spect to the total crystalline fraction of the sample. Whenone amorphous phase is present, the crystalline/amor-phous ratio can be easily determined from the mass ab-sorption coefficien! however, where two or three amor-phous phases exist, the ratio is found optically, whereasonly the limonite component was optically determined inthe previous calculations for samples A through H. Theweight fractions of pumice and opal are then calculatedby summing the weight fractions of amorphous and crys-talline components times their respective mass absorp-tion coefficients such that the total is equal to the mea-sured sample mass absorption coefficient. Modal analysesusing a combination of both methods are presented inTable l.

RIR factors have been tabulated by Davis et al. (1989)for all of the mineral species present except for AHF

[(A1,6(F,OH)48'12-15 HrO)]. XRD and scanning electronmicroscope studies show that AHF is complexly inter-grown with altered pumice from which it is derived andwith other incrustation material; therefore, it is difficultto separate AHF into the pure rnineral necessary for RIRmeasurements. An impure separate of the compound wasprocessed from sample 2l2A by light crushing andscreening and was mixed with an equal weight of corun-dum. The intensity of the strong XRD peak of AHF wasthen mathematically corrected for pumice, a-cristobalite,and labradorite impurities present to obtain an RIR mea-surement of 0.85. Although plagioclase phenocrysts in

our samples were in various stages of alteration, micro-probe analyses ofthe freshest samples gave labradorite asa reasonable plagioclase composition to use in our mod-els. Hildreth (1983) determined that in the rhyolite pum-ice, plagioclase is oligoclase; in dacite pumice, plagioclaseranges in compostion from Anro to An, but Anru-t. ismost common; and in andesite pumice, plagioclase com-position ranges from An- to An* with Arlor-r, the mostcommon. Because of the mixed composition of the ash-flow tuff protolith, a wide range of compositions wasprobably involved.

Crrnrvrrclr MASS BALANCE

Chemical losses and gains resulting from alterationwithin and adjacent to fossil fumarole 2l2were estimatedusing the isocon method of Grant (1986). This methodis a graphical application of Gresens' (1967) procedure.Measured rock densities are not required, an advantagewhen studying porous pumiceous samples. The chemicaldata for fossil fumarole 212 samples (encrusted falloutA-H and leached ash-flow tutr J-O) and an unaltereddacite ash-flow tuffprotolith sample (VTTS-D) for com-parison are presented in Table 2.

Isocon diagrams for selected samples of fumarole 212are illustrated in Figures 4 and 5. In these diagrams, theconcentrations of specific major and trace elements in theoriginal unaltered rock (CP) are plotted against those inthe altered rock (Ci). If there is no detectable alteration,the data points define a line with a slope of 1, representingconstant mass. In most cases, a linear array is defined bysome combination of the components (e.9., AlrO., TiOr,Ga, Y, Zr, Nb, REE, Hf, Ta, Th) commonly consideredto be relatively immobile during vapor phase/hydrother-mal alteration of volcanic rock (ksher et al., 1986; Stur-chio et al., 1986). This is exemplified in Figures 4 and 5for samples 2l2E from the encrusted fallout and 2l2Lfrom the leached ash-flow tuff. Best-fit isocons to thesedata arrays were calculated using linear regression anal-ysis (Table 3). The slopes of best-fit isocons, MP/M^ (massoriginaVmass altered), were used to assess possible changes

(2),,:(?2,{)'

PAPIKE ET AL.: FUMAROLIC DEPOSITS

TABLE 2. Chemical compositions of the altered and unaltered dacite-rich protolith of fossil fumarole 212

1667

Sample: 212A 2128 212C 212D 212E 212F 2'l2G 212H 212J 212K 212L 212M 212N 2120 VTTS-D

SiO. (wt%)Tio,Al,03FeOFe2OiMnOMgoCaONaroK.oP.OuSO.FclLOIo : c t , Fe

TotalBa (ppm)SrZrThZnRbASCuGaPbBr

Ni

SnNbCrScHfUTaCsSbAu-'CoLaCeNdSmEuGdTbHoTmYbLU

55.4 57.3 56.3 56.0 57.3 52.6 59.0 59.5 79.30.52 0.50 0.70 0.59 0.59 0.55 0.64 0.59 0.61

17 .2 16.2 16.7 17.6 16.7 18.6 16.9 15.9 6.01 .61 1 .95 2.63 3.30 2.10 3.12 3.32 3.10 0.712.51 2.24 2.98 3.73 3.57 7.03 1.91 2.95 1 .080.08 0.07 0.13 0.13 0.11 0.13 0.12 0.12 0.052.2 2.O 3.5 3.1 2.0 2.7 2.2 2.3 1.63.73 5.01 7.11 7.10 6.17 6.72 5.61 5.30 1 .662.78 2.90 3.00 3.50 3.00 3.53 3.30 3.30 1.610.89 1.10 0.61 0.70 1.55 1.10 1.30 1.39 1.00

<0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.10.42 0.07 0.15 <0.05 <0.05 <0.05 <0.05 <0.05 0.174.89 2.59 1 .81 0.66 3.68 1 .04 1.22 1.26 0.020.35 0.38 0.24 0.29 0.31 0.12 0.22 0.15 0.577.4 7.6 4.0 4.3 2.9 2.4 4.1 3.2 5.0

-2.14 -1.18 -0.82 -0.34 -1.62 -O.47 -0.56 -0.56 -0.1497.84 98.73 99.04 100.66 98.36 99.17 99.28 98.50 99.28

450 470 330 360 510 410 500 510 310250 300 360 360 260 320 300 270 80106 109 61.8 72.3 129 71.4 130 123 175

2.10 2.26 1.20 1.45 2.80 1.61 2.70 2.60 3.7566.3 155 590 200 200 730 440 320 28.530.0 29.0 13.3 14.4 31.4 54.3 27.6 27.1 24.7

530 730 630 130 590 410 110 180 11420.6 21.7 29.2 23.9 19.9 405 46.2 190 15.214.9 17 .4 9.2 14.5 11.4 18.9 14.7 15.4 6.048.3 600 990 280 660 240 160 115 9.340.4 41.9 42.7 12.7 48.9 4.3 13.1 4.1 10.233.2 32.3 16.1 18.5 31.6 24.3 24.4 26.5 17.3<3 3.2 5.6 10.9 6.4 10.5 7.4 7.4 4.9170 200 150 210 160 300 110 155 6430 124 106 42 140 52 80 42 163.2 3.3 1.9 2.3 4.1 <1.2 3.7 2.9 3.6

20 20 25 30 20 30 22 20 1619.0 15.5 26.6 22.7 17 .6 23.6 1 8.9 17 .8 1 1 .83.00 3.07 2.00 2.00 3.55 2.30 3.60 3.40 4.801 .1 1 .1 0.64 0.70 1 .3 0.80 1.3 1.3 1.7o.24 0.34 0.30 0.20 0.53 0.16 0.53 0.43 0.351 .77 1 .93 1 .93 1 .60 1 .30 13.8 3.70 3.80 0.90

10.1 15.5 5.6 2.6 3.8 5.3 6.7 3.3 48.0<5.0 <5.0 <5.0 <5.0 3.5 40 7.0 15 10

6.20 5.30 12.9 18.9 10.5 22.O 14.6 16.9 6.2011.7 12.O 6.88 7.90 11.4 9.47 10.7 1 1.3 5.3527.0 25.4 15.1 18.1 26.0 20.7 24.3 25.4 11.119.2 1 8.0 10.0 1 1 .8 17 .2 15.0 16.6 16.8 7 .O5.40 4.86 2.65 3.02 4.57 4.00 3.91 4.20 1.41 .70 1 .50 1 .12 1 .10 1 .18 1 .24 1 .10 1 .06 0 .406.6 6.6 3.6 3.9 5.7 5.5 4.8 5.6 3.01.10 1.02 0.65 0.66 1.03 0.90 0.86 0.84 0.471.5 1 .1 0 .7 1 .0 1 .4 1 .2 1 .2 1 .3 0 .820.53 0.57 0.38 0.34 0.48 0.42 0.42 0.43 0.353.70 3.45 210 2.30 3.25 2.80 2.81 2.85 2.310.54 0.53 0.32 0.34 0.49 0.41 0.43 0.43 0.35

73j 68.40.54 0.46

1 1 .6 15.91.78 2.77

<0.02 <0.020.05 0.071 . 5 1 . 82.41 3.823.00 3.802.12 1 .96

<0.1 <0.10.60 0.340.03 0.040.18 0 .113.1 1 .7

-0.05 -0.0499.96 101 .13

750 720160 220154 130

4.'t0 3.6036.8 43.847.0 43.866 17.517.6 19 .111 .6 12.121.6 19.27.5 3.2

20.2 21.85.4 7.6

40 401 2 < 1 03.7 3.1

15 2014.2 16.64.60 4.',t41 . 9 1 . 90.42 0.511.70 1.43

40.0 2.1<5.0 <5.0

7.90 12.110.2 11.420.0 22.911 .5 12.62.55 2.760.77 0.963.4 3.60.62 0.621 . 1 1 . 00.36 0.422.35 2.820.37 0.44

66.3 66.5 63.s0.47 0.50 0.61

16.0 16 .0 16 .02.61 1.71 n.a.1 .33 1 .10 5.600.10 0.08 0.132.1 2.O 2.14.35 4.O4 4.953.80 3.43 3.851 .90 1-80 1.70

<0.1 <0.1 <0.1<0.05 0.19 <0.05

0.04 0.03 0.040.08 0.02 0.0770.7 2.7 1.6

-0.03 -0.02 -0.0399.75 100.08 100.12

710 640 580220 240 260130 134 150

3.70 3.60 3.5673.5 48.3 3137.8 32j 36.013 18 .5 7 .027.3 14.7 9.315.8 13.6 12.822.0 11 .8 8.02 .9 <1 <1

27j 19.4 28.410.6 8.3 <4

140 170 130< 1 0 < 1 0 < 1 0

3.8 3.2 3.927 17 15.319.1 17.3 20.03.83 3.67 4.131 . 8 1 . 6 1 . 30.62 0.4s 0.33'1 .16 0 .67 1 .501 .3 7.3 0.758.0 10 <3

14.0 25.8 n.a.11.2 9.05 12.724.0 17.3 30.414.0 10.0 20.73.50 2.26 4.400.99 0.84 1.154.4 3.0 5.40.78 0.s0 1.001.3 0.90 1.30.50 0.34 0.513.16 2.37 3.400.48 0.36 0.53

66.10.50

16.51.600.460.071 . 73.853.701.94

<0.10.430.050.072.8

-0.0499.73

690240130

3.21co.o36.524.915 .012.620.5

1 Q

17.713.390

< 1 03.4

1 6't7.44.O21 . 8o.321 . 1 47.1

<5.023.510.119.110.02.300.873.0o.520.880.382.200.35

Note.'n.a. : not analyzed.* Ta value for Hildreth (1983).

" Au in ppb.

in rock mass that ocrcurred during alteration. Isocon slopesfor samples A-H in the encrusted fallout and J, N, andO in the leached tuffdeviate very little from I (Table 3),indicating that rock mass was essentially conserved dur-ing alteration processes. The isocon slopes for samples K,L, and M correspond to mass increases of 20o/o, 160/o, andl4ol0, respectively.

The gains and losses of major oxides, F and LOI(-HrO), calculated for the two chemical profiles docu-mented in Table 2 are reported in Table 4. The massexchange detemined for major oxides approximates themass exchange indicated by the isocons. This is an in-

dependent check on the method because the slope oftheisocon is based on immobile trace elements, not entirelyon major oxides. The values reported in Table 4, along\Mith the results of trace-element mass balance calcula-tions, are normalized to the original unaltered rock con-tent of each component and presented on enrichment-depletion diagrams in Figure 6. These diagrams illustratethat for samples A-H: SO., F, LOI (-HrO), Cl, Zt, As,Cu, Pb, Br, Ni, V, Sn, Cr, Cs, Sb, and Au are enriched,and SiOr, NarO, KrO, Z.r,U,Th, Ba, and Rb are depleted.Total Fe as FerOr, MnO, MgO, CaO, Sr, and Sc generallyshow erratic and insignificant variations. Along the

0 02 Pb ,.\"7.,*..".r%:"

CaO o3

Na2 O

l 668 PAPIKE ET AL.: FUMAROLIC DEPOSITS

Trsle 3. lsocon definitions22

20

9 r oI 1 4

o r o! ^

g 6< 4

20

2220

9 r o

9 1 4

O 1 0

9 6< 4

20

22

2A

i : 18

9 r oQ 1 4

o r o

! eo 6< 4

2

0

22

20

9 r oI 1 4

o r o! ^

9 6< 4

20

222A

121 0I

6

2

0

22

20

1 a

1 2

l 0

I

6

2

0

r 0 2 S i O 2

LOI t0

Sm

1 4 1 6 1 8 2 0 2 2

f i , / . O l Z l

r!'/ . too uno

5 E u

10 Tm

Th

Sam-ple R MolM^ lsocon comDonents'

A

B

0.981 0.99

0.976 0.96

0.995 0.930.992 1.O7

0.990 1.01

0.992 0.990.977 0.800.978 0.84

0.979 0.860.989 0.98

0.983 1.00

Al, Ti, Nb, La, Ce, Nd, Sm, Gd, Tb, Tm, Yb,Lu, Hf, Ta, Th

Al, Ti, Y, Nb, La, Ce, Nd, Sm, Gd, Tb, Tm,Yb, Lu, Hf, Ta, Th

Al, Ti, TaTiAl, Ti, Ga, Y, Nb, La, Ce, Sm, Gd, Tb, Ho,

Tm, Yb, Lu, Hf, Th, UTi, Sm, Gd, Tb, Ho, Tm, YbAl, Ti, Nb, Ce, Sm, Gd, Tb, Ho, Tm, Yb, Lu,

Hf, Th, UAl, Ti, Y, Nb, Ce, Sm, Gd, Tb, Ho, Tm, Yb,

Lu, Hf, Ta, Th, UTi, Cr, Nb, Hf, Ta, ThAl, Ti, Y, La, Ce, Sm, Gd, Tb, Ho, Tm, Yb, LuTi ,Y,Zt , Nb, La, Ce, Sm, Gd, Tb, Ho, Tm,

Yb, LuTi, Zr, Nb, La, Ce, Sm, Gd, Ho, Tm, Yb, LuAl, Ti, Y, Nb, La, Ce, Sm, Gd, Ho, Tm, Yb,

Lu, Hf, ThAl, Ti, Ga, Nb, Ht, Ta, Th

DE

FG

H

JKL

MN

o

0.995 1.07N.A. 0.970.987 1.05

O H

r 5 Tio2

10 LurlT o 01 Ba

"{4,i1

Unaltered Dacite (Ci) Unaltered Dacite (Ci)

Fig. 4. Isocon diagram for sample 212E showing an exampleof constant mass relationships in the encnrsted fallout part oftho fossil fumarole. Isocon diagrams plotted according to themethod of Grant (1986). Major oxides in weight percent; traceelements in ppm except Au in ppb. Numbers above componentsare factor used to bring component concentration in unalteredrhyolite onto range of isocon so that enrichment or depletioncan be determined, e.g., l0 K2O has original concentration of1.55 wto/o (Table 2) and plots at l0 x 1.55 : 15.5 on the y axis.

chemical profile of the leached ash-flow tuff, J-O, SiOr,SO3, LOI (-HrO), Zn, As, Cu, Pb, Br, Ni, Sn, Cr, U, Sb,and Au are enriched and Fe2O3, MnO, CaO, and V aredepleted. AlrOr, CaO, NarO, KrO, F, Ba, Sr, Rb, Ga, Y,Sc, and Cs are significantly depleted in the most intenselyaltered sample, J, which is the sample immediately ad-jacent to the fumarole conduit in the ash-flow tuff andtherefore subject to the greatest interaction between fluidand rock.

0 2 4 6 8 1 0 r 2 1 4 1 6 1 A 2 0 0 12 14 16 18 20 22

Note: R : correlation coefficient; MolM^ : isocon slope; N.A. : notapplicable.

* Elements used to define isocons.

REE srcN,c.TUREs

Eleven REE were determined for 14 samples from fu-marole 212 and for a sample of unaltered dacite pumice(VTTS-D) (Table 2). Hildreth (1983, p. 34) presented REEdata for unaltered rhyolite, dacite, and andesite pumicesamples and high-purity glass separates. The fumaroli-cally encrusted dacite-rich fallout samples (A-H; Fig. 7)show variable REE patterns indicating redistribution asa result of fumarolic alteration. Samples E and H aresimilar to the dacitic protolith; samples F and G are lessso. However, samples A, B, C, and D illustrate some totalREE depletions and, more significantly, positive Euanomalies. Michard (1989) has shown that chondrite-normalized REE concentrations of high-temperatureacidic HrO (pH < 6; T > 230 "C) have high positive Euanomalies. Thus a reasonable interpretation of the posi-tive Eu anomaly is that Eu2* has been mobilized andenriched relative to the trivalent REE. Hot acidic solu-tions transported Eu2* preferentially to the other REE anddeposited it in the more altered samples.

REE patterns in fossil fumaroles in rhyolite-rich pro-tolith (fumaroles 63 and 64) in the lower VTTS (Papikeet al., l99l) indicate that high-temperature fumarolic al-teration that cooled rapidly and was not subjected to sur-ficial acid alteration resulted in a small amount of REEmovement except for Eu (fumarole 64). However, fu-marole 63, which was subjected to extensive acid alter-ation, showed highly variable REE values and especiallyenrichment of Eu in altered samples relative to unalteredrhyolite pumice. Thus REE pattern disturbances mightbe an effective means of evaluating the degree of alter-ation of VTTS fumarolic incrustations, particularly theeffect of acid alteration.

In samples J-O (Fig. 8) from the leached ash-flow sheet,

:o , l yo

+o^ /" ' t rooMno

/'ii,:""'

02star. / / / /

oo1Ba. /a /H t / -

0"""./ot'"./ . o 5 \

212L2 2 -

2 A -

22

20

1 2

1 0

I

6

2

0

/. La'

'4{,.,5 E u

LOtr..? Nh

o 2 4 6 8 1 0 1 2 1 4 1 6 1 5 2 0 2 2 0 2 4 6 I 1 0 1 2 1 4 1 6 1 A 2 0 2 2

Unaltered Dacite (C? ) Unaltered Dacite (C: )

Fig. 5. Isocon diagram for sample 2l2L showing an exampleof mass gain in the leached ash-flow tuff part of the fossil fu-marole. Explanation as for Figure 4.

PAPIKE ET AL.: FUMAROLIC DEPOSITS t669

the most altered sample (J) adjacent to the fumarole con-duit shows Eu depletion that is consistent with the sig-nificant loss of plagioclase by high-temperature fumarolicleaching (Table l, Fig. 9). Sample O, which has the high-est modal plagioclase (2lo/o, Table l) and is the least al-tered, has the geatest positive Eu anomaly. Hildreth (1990,and personal communication) suggests that an additionalfactor that may cause variable REE concentrations and,specifically, Eu2* is mechanical sorting of plagioclase dur-ing tephra deposition.

DrscussroN

Interpretation of fossil fumarole mineralogy and geo-chemistry is complicated by weathering superimposed onthe complex leaching and depositional processes of activefumarolic evolution from high-temperature vapor-phaseprocesses through cooling processes. In addition, thecomplex fine-grained intergowths and thin depositionalcoatings of sublimates and secondary minerals superim-posed upon primary glasses and phenocrysts in variousstages ofalteration (depending upon position in the areaaffected by the fumarole) make the interpretations of thedata difrcult. An additional consideration of the com-plexity of the system is that during fumarolic activity, asubstantial volume of unstable compounds, which con-tain as much as several weight percent of trace metals,are formed (Keith et al., 1981, 1989). These compoundsinclude chlorides and sulfates of all the major constitu-ents except Ti and many minor elements (Shipley, 1920;Keith, I 99 I ). Therefore, substantial amounts of elementsleached by fumarolic gases are removed from the area bydissolution within hours and days of deposition (Keith etal., 1989). As fumaroles cool and die out, late-stage min-erals are deposited in pore spaces where earlier leachingoccurred.

Given that this study is based on bulk samples ratherthan discrete mineral phases, we are of the opinion thatour study provides unique understanding of the bulkchemical leaching and depositional processes in high-temperature fumaroles at the VTTS. Caution must beused in assessing rock masses and associated chemicalgains and losses in these bulk samples because both leach-ing and deposition have occurred in many places becauseof the processes described above. Also the compositionused for original protolith is only an approximation.

To understand the chemical systematics of our results,the mineral assemblages and alteration textures must beconsidered. The major mineral modal analyses along withsome of the chemical data are illustrated in Figure 10.Crystalline phases introduced in the alteration assem-blages by fumarolic alteration and subsequent cooling re-actions include alunite reflecting high SO3 activity; AHFand fluorite reflecting hrgh F activity; smectite, magne-tite, hematite, and goethite reflecting oxidation and hy-dration; and opal and some of the a-cristobalite reflectingSiO, mobility. Although the encrusted fallout (A-H) showsgreat chemical and mineralogical diversity, the leachedash-flow tutr (J-O) shows some systematics. For the

leached part of the system (note Fig. 9), the modal dataand thin sections show that plagioclase, pyroxene, mag-netite, and ilmenite phenocrysts are removed adjacent tothe fumarole conduit and are less altered away from theconduit. The abundance of a-cristobalite increases to-ward the conduit.

Comparing the chemical data shown on Figures 6 and10. substantial enrichments in some elements occur inencrusted fallout samples A-H compared to the leachedash-flow, e.g., Zn, Cu, F, Cl, CaO, Fe,o, as FetOr. AIso,note the enrichments are localized, particularly in sam-ples F and C. The F enrichments clearly correlate withincreases in AHF and fluorite in sample A. Although theassociated mineralogy does not always indicate wherethese enriched elements might be concentrated, referringto Figure 3a, the most enriched samples come from thenear surface part ofthe system and near the conduit wherethese elements would have been sublimated from fuma-rolic fluids. Although primary dacite-rich fall clasts re-main in various stages of alteration in the encrusted fall-out, certain elements are clearly enriched over the contentof the primary clasts, i.e., dacite protolith.

Overall chemical relationships illustrated in Figure 6are determined as follows:

l. The sample adjacent to the conduit, J, is heavilyleached of most elements except LOI (-HrO), Cl, SiO2,SOr,Zr, As, Au, Sb, Cu, Pb, Ni, Cr, Hf, and U.

2.Inached from the ash-flow tuffand deposited in thefallout or deposited directly from the volcanic vapor phaseare FerOr, MgO, CaO, Cl, F, Sr, V, Sc, and Cs.

3. Relatively enriched in the leached ash-flow tuf anddepleted in the fallout are SiOr, KrO, Ba (except adjacentto the conduit were it is strongly depleted), Zr, Th, Rb,Hl U, and Ta.

4. Enriched in both parts of the fossil fumarolic systemare SOr, LOI (-HrO),F,Zn, As, Cr, Pb, Br, Ni, Sn, Sb,and Au. AlrO. and Ga are slightly enriched in both partsofthe system except for strong depletion adjacent to theconduit in the leached ash-flow tuff.

5. Depleted in both parts of the system are NarO, MnO,Y, Nb, and REE.

6. TiO, is close to constant in both parts of the systembut is slightly enriched or depleted in individual sampleswithout any pattern indicating that it is relatively stableunder the various conditions that have affected this fossilfumarole. The variability is most likely a result of vari-able protolith composition.

The enrichments and depletions are similar to thosefor fumarolically altered and incrusted tephra in the dis-tal, rhyolite-rich part ofthe ash-flow sheet (Papike et al.,199 l). Major oxides and trace elements enriched in fossilfumarolic incrustations in both the dacite-rich protolith(fumarole 212) ard the rhyolite-rich protolith (fumaroles63, 64,69; Papike et al., l99l) are FerOr, MgO, CaO,SO. (except 64),LOl (-HrO), Sr,Zn, Cu, Br, and F. Inaddition, enrichments in incrustations from fumarole 212and at least one rhyolite-rich protolith fumarole are Ni,Cr, As, Au, Sn, and Pb. Depletions of NarO occur in all

t670 PAPIKE ET AL.: FUMAROLIC DEPOSITS

TABLE 4. Compositional gains and losses compared to 100 g of unaltered dacite

Encrusted dacite-rich falloul

Average

sio,Tio,Al,o3FerO3MnOMgoCaONa.OK"OSo.FLOIMass change

-8 .1-0.09

1 . 2-1 .30-0.05

0.10-1.22-1.O7-0.81

0.374.855.8

-o.32

-6.2- 0 . 1 1

0.2-1.20-0.06- 0 . 1 0

0.06-0.95-0.06

0.o22.556,00 . 1 5

-7.20.090.70.300.001.402.16

-0.85_1 no

0.101.772.4

-0.22

- I - a

-0.021 . 61 .800.001.002.15

-0.35-1 .00

0.000.622.71 .00

-6.2-0.o2

o.70.30

-o.o2-0 .10

1.22-0.85-0 .15

0.003.641 . 3

-0 .18

- 10 .9-0.06

2.64.900.000.601.77

-0.32-0.60

0.001.000.8

-0.21

-4.50.030.90.00

-0.010 1 00.66

-0.55-0.40

0.001 . 1 82.5

-0.09

-4.O-0.02-0 .1

0.80-0.01

o.200.35

-0.55-0.31

0.001.221 . 6

-o.82

-6.8-0.03

1 . 00.70

-0.020.400.89

-0.69-0.55

0.062.',t02.9

-0.09

four fossil fumaroles. Although the masses of individualcomponents moved in fumaroles in different protolithsvary, the similarities in the chemical species that are en-riched and depleted suggests that the fumarolic gas com-position is the most critical factor in determining whichcomponents are moved, regardless of protolith.

Comparison with Symonds et al. (1987), who reportenrichment in sublimates and incrustation in most ele-ments for high-temperature (500-800 .C) fumaroles atMerapi volcano, shows that the same suite of enrichedcomponents occur in both the leached ash-flow tuffandthe encrusted fallout at fossil fumarole 212 in the VTTS.

A e H : J e O

Fig. 6. Enrichment and depletion diaglam for fumarole 212.Samples A-H are from the encrusted fallout part of the fossilfumarole; locations as in Figure 3a. Samples J-O are from theleached ash-flow tuffas shown in Figure 3b with J, adjacent tothe fumarole conduit, being the most leached.

Thus we conclude that in both parts of the fossil fumarolewe are dealing with the shallow part of the fumarole sys-tem. The most visible component, FerOr, along with MgO,CaO, Cl, F, Sr, V, Sc, and Cs, was leached from the ash-flow sheet and deposited in the fallout (Fig. 6). The com-

' 10-+- 2124

- + - 2 1 2 8-^'.-- 212C

--4- 212Do

(!oO loEo<t>

Io

;l

E

oF

8

o

ooo l

E(!U)

0 ' 11 0

- + - 2 1 2 F

-^- 212G

0 1

1 0

t-1-.-FR==H-F:-E-::=i=---q'-i-i

-d- 212E

- * - 2 1 2 H

La Ce Nd Sm Eu Gd Tb Ho

REE lonic Radii

Tm Yb Lu

Fig. 7. REE patterns for samples from the encrusted falloutpart of fumarole 212 rrormalized against dacite protolith REEconcentrations.

Sm Eu Gd Tb

Sm Eu Gd Tb

Tm Yb Lu

Tm Yb LuLa Ce

o

oo6 1

E(ga

0 1

Fe 203Tod MnO MgO CaO Na2O K2O SO3

0 1

PAPIKE ET AL.: FUMAROLIC DEPOSITS t67 |

Taete A-Continued

Leached dacite-rich ash-flow tuff

M Average

sio,Tio,Alro3FerO"MnOMgoCaONaroGOSo"FLOIMass change

15.80.00

-10.0-3.73- 0.08-0.50-3.29-2.24-0.70

0.12-o.o2

3.4- 1 . 1 8

27.90.06

- 1 . 5-3 .15-0.07-0.22- 1.94- 0 . 1 0

0.950.700.002.3

24.9

17.9-0.06

2.9-2.71-0.05

0.04-0.40

0 6 70.630.350.01o.4

19.7

13.4-0.03

3.2-3.00-0.05-o.12-o.47

0.450.560.450.021 . 7

16.1

2.8-0 .14

0.0-1.37

0.050.00

-0.60-0.05

0.200.000.00

- 0.90.00

3.0-0 .11

0.0-2.60-0.05-0 .10-0.91-0.42

0.100.14

-0.011 . 10 .15

13.5-0.05-0.9-2.76-0.04-0 .15-1.27-0.28

o.290.290.001 . 39.95

ponents enriched in both parts ofthe system were trans-ported during high-temperature fumarolic alteration butalso during cooling processes and became fixed in morestable solids so as to be resistant to further alteration. As,Sb, and Au are a well-known association in shallow ep-ithermal deposits (White and Heropoulos, 1983; Heden-quist and Henley, 1985; Berger, 1985; Romberger, 1988).Pb.Zn. and Cu are also a common association in shallow

212 J-L

- . - 212K

-a-212L

parts of epithermal systems and volcanic exhalative de-posits (Sillitoe and Bonham, 1984; Berger, 1985). Sn andF are a common association in high-silica rhyolites thatyield cassiterite ore deposits (Burt and Sheridan, 1987;Duffield, 1990). SO3 and HrO are added during coolingand acid alteration.

CoNcr,usroNs

This study looked at the upper part of a fossil fumarolicsystem in which ash-flow tuffof mixed rhyolitic, dacitic,and andesitic pumice was leached by highaemperaturefluids that carried chemicals upward, where they weresublimated onto permeable dacite-rich fallout and sub-jected to further alteration, particularly surficial acid al-teration, as the fumarole cooled and died out.

KatmaiTraverse 212 Samples J-O

--O-- 41rni1s--|- cristobalite* Plagioclase---o-- Pyroxene

.. -'L]-l:-= +- - g.;11 -?" " " -

|M

Sample

Fig. 9. Variation of modal proportions of crystalline phasesfrom the conduit (sample J) outward in the lower, leached ash-flow tuff part of the system. Depletion of plagioclase and pyrox-ene from high-temperature vapor-phase and later hydrothermalalteration and crystallization of a-cristobalite from altered pum-ice adjacent to the conduit are clearly demonstrated.

6)o

o6 l

E

U)

0 1

o

oo lCLEu)

30

La Ce

0 1La ce Nd Sm Eu Gd Tb Ho Tm Yb Lu

BEE lonic Radii

Fig. 8. REE patterns for samples from the leached ash-flowtuf part of fumarole 212 normalized against dacite protolith

REE concentrations.

212M-O

\=='A;1:i'l;;-L - 212M- r - 2 1 2 N-o - 2120

20coooo.EEDo)= 1 0

o

E

t672 PAPIKE ET AL.: FUMAROLIC DEPOSITS

A B c D E F e r r i l J

K L M N o

Fig. 10. Schematic modal and selected chemical data for fu-marole 212. Samples A-H are from the encrusted fallout part ofthe fossil fumarole; samples J-O are sequentially outward fromthe leached fumarolic conduit in the ash-flow tuff. Chemical datain parentheses are taken from Table 2. Relative thickness ofbarsrepresents relative abundance ofphase in upper part ofdiagramand abundance of selected maior oxides and trace elements inlower part.

Studies of mineralogy and chemical composition ofwhole-rock samples of fossil fumarole 212 show impor-tant characteristics for understanding and interpretingfossil fumarolic systems. These studies address the entiresequence of alteration processes from high-temperature,vapor-phase fumarolic alteration through cooling pro-cesses during which fumarolic minerals and their chem-ical constituents are subjected to secondary alteration. Fi-nally, they also record weathering processes that includedissolution of unstable and metastable deposits.

This study illustrates the usefulness of the isocon meth-od of Grant (1986) to interpret a complex alteration sys-tem, providing that the geologic constraints are carefullyevaluated.

AcrNowr,nncMENTS

This research was funded by NSF grant EAR-8907712 (J.J.P.), the U.S.Geological Survey (T.E.C.K), and DOE Contract DE-AC06-76RLO-1830(J.C.L.). We gatefully acknowledge this support. We thank D. Tully forcollecting much of the XRD data.

RprnnnNcns crrEDAllen, E.T., and Zies, E.G. (1923) A chemical study of fumaroles of the

Katmai region. National Geographic Society, Contributed TechnicalPapers, Katmai Series, 2, 75-155.

Aruscavage, P.J., and Campbell, E.Y. (1983) An ion-selective electrodemethod for the determination of Cl in geological materials. Talanta,30, 7 45-749.

Berger, B.R. (1985) Geologic-geochemical features of hot-spring precious-

metal deposits. U.S. Geological Survey Bulletin, 1646,47-53.Bodkin, J.B. (1977) Determination of fluorine in silicates by use of an

ion-selective electrode following fusion with lithium metaborate. TheAnalyst, 102,409-413.

Burt, D.M., and Sheridan, M.F. (1987) Types of mineralization relatedto fluorine-rich silicic lava flows and domes. Geological Society ofAmerica Special Paper 212, 1O3- lO9.

Chung, F.H. (1974a) Quantitative interpretation of X-ray diftaction pat-

tems of mixtures. l. Matrix-flushing method for quantitative multi-component analysis. Journal of Applied Crystallography, 7, 5 19 - 525.

-(1974b') Quantitative interpretation of X-ray diftaction patterns

of rnixtures. II. Adiabatic principle of X-ray diftaction analysis ofmixtures. Journal of Applied Cryslallography, 7, 526-53 l.

Cunis, G.H. ( I 968) The stratigraphy of the ejecta from the I 9 I 2 eruptionof Mount Katmai and Novarupta, Alaska. Geological Society of Amer-ica Memoir I 16,153-210.

Davis, B.L. (1984) Reference intensity quantitative analysis using thin-layer aerosol samples. Advances in X-ray Analysis, 27, 339-348.

-(1986) A tubular aerosol suspension chamber for the preparation

of powder samples for X-ray difraction analysis. Powder Diftaction,r .240-243.

Davis, B.L., Smith, D.I(, and Holomany, M.A. (1989) Tables of exper-imental reference intensity ratios, Table no. 2, December, 1989. Pow-der Diffraction. 4. 2Ol-205.

Desborough, G.A., and Rostad, 0. (1980) Hydrated aluminum hydroxy-fluoride, a ralstonite-like mineral at Big Southem Butte, Snake RiverPlain, Idaho. American Mineralogist, 65, I 057- I 05 8.

Duffield, w.A. (1990) Eruptive fountains of silicic magma and their pos-

sible effects on the tin content offountain-fed lavas, Taylor Creek Rhy-olite, New Mexico. Geological Society of America Special Paper 246,251 -26 r .

Eichelberger, J.C., and Hildreth, W. (1986) Research drilling at Katmai,Alaska. Eos. 67. 7 7 8-7 80.

Fierstein, J., and Hildrerh, W. (1990) Contemporaneity of pyroclasticflows and falls; evidence from the eruption at Novarupta (Alaska) in1912. Eos. 71.1690.

Flanagan, F.J. (1967) U.S. Geological Survey rock standards, Geochimicaet cosmochimica Acta. 31. 289-308.

- 0969) U.S. Geoloaical Survey standards-Il. First compilation ofdata for the new U.S.G.S. rocks. Geochimica et Cosmochimica Acta,3 3 , 8 l - 1 2 0 .

-(1973) 1972 values for international geochelnical reference sam-ples. Geochimica et Cosmochimica Acta,37,I I89-I200.

Goldich, S.S. (1984) Determination of fenous iron in silicate rocks'Chemical Geology, 42, 343-347.

Grant, J.A. (1986) The isocon diagram-A simple solution to Gresens'equation for metasomatic alteration. Economic Geology, 81, 1976-1982.

Gresens, R.L. (1 967) Compositional-volume relationships of metasoma-tism. Chemical Geology, 2,47-65.

Griggs, R.F. (1922) The Valley ofTen Thousand Smokes, 340 p. National

Geographic Society, Washinglon, DC.Hedenquist, J.W., and Henley, R.W. (1985) Hydrothermal eruptions in

the Waiotapu geothermal system, New Zaaland:. Their origin, associ-ated breccias, and relation to precious metal mineralization. EconomicGeology, 80, 1640-1668.

( 't47 - 27 3ppn)

(0 12 - 0 38 wr/.) (002-0 twf / . )

(0 02 - 0 05 wtold

(<0 05 - 0.60 wr/.)

(1 66 - 4.35 wto/")CaO (3 73 - 7 11 wty")

(1.87 - 4.23 wF/.)

PAPIKE ET AL.: FUMAROLIC DEPOSITS 1673

Hildreth, W. (1983) The compositionally zoned eruption of 1912 in theValley ofTen Thousand Smokes, Katmai National Park, Alaska. Jour-nal ofVolcanology and Geothermal Research, 18, 1-56.

-(1987) New perspectives on the eruption of l9l2 in the Valley ofTen Thousand Smokes, Katmai National Park, Alaska. Bulletin of Vol-canology, 49,680-693.

-(1990) The Katmai eruption of 1912: Was the magma stored be-neath Novarupta, Trident or Mount Katmai? Petrochemical and tem-poral evidence. Eos, 71, 1691.

Keith, T.E.C. (1983) Mineralogical and chemical changes in fumarolicdeposits with time at surface conditions, p.231-234. Fourth Interna-tional Symposium on Water-Rock Interaction, Proceedings, Misasa,Japan.

-(1984) Preliminary observations on fumarole distribution and al-teration, Valley of 10,000 Smokes, Alaska. U.S. Geological Survey Cir-cular 939, 82-85.

-(1991) Fossil and active fumaroles in the 1912 eruptive deposits,Valley of Ten Thousand Smokes, Alaska. Journal of Volcanology andGeothermal Research, 45, 227 -254.

Keith, T.E.C., Casadevall, T.C., and Johnston, D.A. (1981) Fumaroleencrustations: Occurrence, rnineralogy, and chemistry. U.S. GeologicalSuwey Professional Paper 1250, 239-250.

Keith, T.E.C., Heropoulos, C., and Presser, T.S. (1989) Chemical trans-port during fumarolic alteration, cooling, and weathering, Valley ofTenThousand Smokes, Alaska, p. 377-379. Proceedings, Sixth Interna-tional Symposium on Water-Rock Interaction. Malvern, England.

Kienle, J. (1970) Gravity traverses in the Valley ofTen Thousand Smokes,Katmai National Monument, Alaska. Journal of Geophysical Re-search.75. 6641-6649.

Kodosky, L., and Keskinen, M. (1990) Fumarole distribution, morphol-ogy, and encrustation mineralogy associated with the 1986 eruptivedeposits ofMount St. Augustine, Alaska. Bulletin ofVolcanology, 52,l 75- I 85.

Laul, J.C. (1979) Neutron activation analysis of geological materials.Atomic Energy Review, 17.3, 603-695.

Laul, J.C., Walker, R.J., Shearer, C.K., Papike, J.J., and Simon, S.B.(1984) Chemical migation by contact metamorphism between peg-matite and country rock: Natural analogs for radionuclide migration.Materials Research Society Symposium Proceedings, 26, 9 51-958.

Lesher, C.M., Gibson, H.L., and Campbell, I.H. (1986) Composition-volume changes during hydrothermal alteration ofandesite at Butter-cup Hill, Noranda District, Quebec. Geochimica et Cosmochimica Acta,50,2693-2705.

Lovering, T.S. (1957) Halogen-acid alteration ofash at Fumarole No. 1,Valley of Ten Thousand Smokes, Alaska. Geological Society of Amer-ica Bulletin, 68, 1585-1604.

Michard, A. (1989) Rare earth element systematics in hydrothermal flu-ids. Geochimica et Cosmochimica Acta. 53, 745-750.

Papike, J.J., Spilde, M.N., Galbreath, K.C., Shearer, C.K., Keith, T.E.C.,and Laul, J.C. (1989) Geochemistry and mineralogy of fumarole de-posits, Valley of Ten Thousand Smokes, Alaska: Reference intensitymethod (RIM) XRD modal analysis. Eos,7O, 1412-1413.

-(1990a) Geochemistry and mineralogy of fumarole deposits, Val-ley of Ten Thousand Smokes, Alaska: Alteration of rhyolite ash-flowtuffprotolith. Second V.M. Goldschmidt Conference, Baltimore, Mary-land, Abstracts with Progams, 72.

-(1990b) Geochemistry and mineralogy of furnarole deposits, Yal-ley of Ten Thousand Smokes (VTTS), Alaska: Bulk chemical and min-eralogical evolution ofa dacitic fissure fumarole. Geological Society ofAmerica Abstracts with Programs, A351-A352.

Papike, J.J., Spilde, M.N., Shearer, C.K., and Keith, T.E.C. (1990c) Geo-chemistry and mineralogy of fumarole deposits, Valley of Ten Thou-sand Smokes (VTTS), Alaska: Major element mass exchange and traceelement enrichmenVdepletion systematics. Eos, 7 1, I 690- 1 69 I'

Papike, J.J., Keith, T.E.C., Spilde, M.N., Galbreath, K.C., Shearer, C.K.,and laul, J.C. (1991) Major and trace element mass flux in fumarolicdeposits, Valley of Ten Thousand Smokes, Alaska: Rhyolite-rich pro-

tolith. Geophysical Research Letters, in press.

Quisefit, J.P., Toutain, J P., Bergametti, G., Javoy, M., Cheynet, B., andPerson, A. (1989) Evolution versus cooling of gaseous volcanic emis-sion from Monotombo Volcano, Nicaragua: Thermochemical modeland observations. Geochimica et Cosmochimica Acra, 53,2591-2608.

Ramdohr, P. (1962) Erzmikroskopische unersuchungen an magnetiten derexhalationen im Valley of the I0 000 Smokes. Neues Jahrbuch fiirMineralogie Monatshefte, 3/ 4, 49-59.

Romberger, S.B. (1988) Geochernistry ofgold in hydrothermal deposits.U.S. Geological Survey Bulletin, 1857-A, A9-A25.

Shearer, C.K., Papike, J.J., Simon, S.8., Davis,8.L., and Laul, J.C. (1988)

Mineral reactions in altered sediments from the Califomia State 2-14Well: Variations in the modal mineralogy, mineral chemistry and bulkcomposition of the Salton Sea Scientific Drilling Project Core. Joumalof Geophysical Research, 93, l3lD4-13122.

Shipley, J.W. (1920) Some chemical observations on the volcanic ema-nations and incrustations in the Valley of 10,000 Smokes, Katmai,Alaska. American Journal ofScience, 50, 141-153.

Sillitoe, R.H., and Bonham, H.F., Jr. (1984) Volcanic landforms and oredeposits, Economic Geology, 79, l'286-1298.

Stoiber, R.E., and Rose, W.I., Jr. (197a) Fumarole incrustations at activeCentral American volcanoes. Geochimica et Cosmoschimica Acta, 38,495-516.

Sturchio, N.C., Muehlenbachs, K., and Seitz, M.G. (1986) Element re-distribution during hydrothermal alteration of rhyolite in an activegeothermal system: Yellowstone drill cores Y-7 and Y-8. Geochimicaet Cosmochimica Acta. 50. 1619-1631.

Symonds, R.B., Rose, W.I., Reed, M.H., Lichte, F.E., and Finnegan, D.L.(1987) Volatilization, transport and sublimation ofmetallic and non-metallic elements in high temperature gases at Merapi Volcano, Indo-nesia. Geochimica et Cosmochimica Acta, 51, 2083-2101

Symonds, R.8., Rose, W.L, Gerlach, T.M., Briggs, P.H., and Harmon,R.S. (1990) Evaluation ofgases, condensates, and SO: emissions fromAugustine volcano, Alaska: The degassing ofa Cl-rich volcanic system.Bulletin of Volcanology, 52, 355-37 4.

White, D.E., and Heropoulos, C. (1983) Active and fossil hydrothermal-convection systems ofthe Great Basin. Geothermal Resources CouncilSpecial Report 13, 4l-53.

Zies,E.G. (1929). The Valley of Ten Thousand Smokes: I. The fumarolicincrustations and their bearing on ore deposition. II. The acid gases

contributed to the sea during volcanic activity. National GeographicSociety, Contributed Technical Papers, Katmai Series, 4, 1-79.

Mer.ruscnrp'r REcETvED Jure 18, 1990MeNuscrrsr AccEPTED Mav 20, 1991