American Mineralogist, Volume 87, pages 115Ð131, 2002

Pressure-temperature paths from garnet-zoning: Evidence for multiple episodes of thrust burial in the hinterland of the Sevier orogenic belt

THOMAS D. HOISCH,1,* MICHAEL L. WELLS,2 AND LORI M. HANSON1,†

1Department of Geology, Northern Arizona University, Flagstaff, Arizona 86011, U.S.A. 2Department of Geosciences, University of Nevada, Las Vegas, Nevada 89154, U.S.A.

ABSTRACT Garnets are present in two horizons of the schist of Stevens Spring from the Basin Creek area of the Grouse Creek Mountains in northwest Utah. The two horizons possess different bulk composi- tions, which resulted in garnet growth by different reactions along the same pressure-temperature (P-T) path. Garnet in the upper horizon grew from the breakdown of chlorite at upper greenschist- facies conditions and garnet in the lower horizon grew from the breakdown of staurolite at upper amphibolite-facies conditions. From the upper horizon, five garnets from three samples were ana- lyzed. All display growth zoning, ragged morphologies associated with secondary rim consumption, and yield garnet-biotite geothermometry temperatures of ~460Ð490 °C. From the lower horizon, one garnet from each of two samples was analyzed. These also display growth zoning, but differ from garnet in the upper horizon in that they are dominantly idioblastic and yield garnet-biotite geothermometry temperatures of ~635 °C. Garnet-biotite geothermometry calculated for every point along detailed compositional traverses across the garnets revealed localized reequilibration along rims, cracks and inclusions in both generations of garnet. Garnets from the upper horizon display prograde reequilibration and the garnets from the lower horizon display retrograde reequilibration. Numerical simulations of garnet growth using the Gibbs method with Duhem’s theorem were car- ried out to determine P-T paths. The P-T path defined by the modeling of five garnets from the upper horizon is an isothermal pressure increase of ~1.7 kbar. The P-T path defined by the modeling of two garnets from the lower horizon has a steep P-T trajectory (dP/dT = 32 bars/°C) and a total pressure change of ~0.9 kbar. Both paths are indicative of thrust burial; however, the two paths cannot be reconciled as products of a single thrust episode. These data are interpreted to indicate two different episodes of thrust burial during the Sevier Orogeny, separated by ~150 °C of heating and partial exhumation.

INTRODUCTION 1983; Selverstone and Spear, 1985; Spear et al. 1990). Addi- Since the early efforts in thermal modeling of tectonic pro- tionally, arrays of P-T determinations have been interpreted as cesses (e.g., Oxburgh and Turcotte 1974; England and Thomp- recording P-T paths (e.g., Hodges and Royden 1984; McGrew son 1984), construction of pressure-temperature (P-T) paths et al. 2000). from metamorphic rocks has proven to be a powerful tool in The samples analyzed in this study are schist of Stevens studies of ancient orogenic belts. Events including thrust burial, Spring from the Basin Creek area in the northern Grouse Creek intrusive heating, and extensional exhumation have now been Mountains, Utah (Figs. 1 and 2). Garnet growth was simulated widely interpreted from the study of P-T paths (e.g., Florence using the Gibbs method with Duhem’s theorem (Spear 1989, and Spear 1993; Lang 1996; Selverstone and Spear 1985; Spear 1993) to determine P-T paths. Garnets grew in two horizons of et al. 1990; Whitney et al. 1999). differing bulk composition; each records growth at different The general approaches used in the construction of P-T paths temperatures and along different segments of the P-T path, are varied and include geothermobarometry of mineral inclu- making it possible to determine the P-T path more completely sion suites in porphyroblasts that may record progressive equi- than would have been possible with either one alone. Garnet in librium conditions during porphyroblast growth (St-Onge the upper horizon grew during upper greenschist-facies meta- 1987), inferences made from reaction textures that show the morphism along the prograde path, and garnet in the lower direction of reactions in P-T space (e.g., Thost et al. 1991), and horizon grew later during upper amphibolite-facies metamor- studies of chemical zoning in minerals, specifically garnet (e.g., phism (discussed in detail below). Florence and Spear, 1993; Lang 1996; Spear and Selverstone The two sampled horizons of the schist of Stevens Spring are the uppermost and lowermost horizons, each about 5Ð10 meters thick. The distance orthogonal to foliation separating * E-mail: [email protected] the two sampled horizons is about 300 m (Fig. 2). Between the † Present address: Department of Natural Resources/Air Qual- two horizons is a package of interlayered amphibolite, quartz- ity Bureau, 7900 Hickman Road, Suite 1, Urbandale, IA 50322. biotite gneiss, and garnet-free pelitic schist. There are no high-

0003-004X/02/0001Ð115$05.00 115 116 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT

FIGURE 1. Simplified geologic and location map for parts of the Raft River, Grouse Creek, and Albion Mountains. Abbreviations are as follows: MM, Middle Mountain; CR, City of Rocks; VM, Vipont Mountain. Modified from Armstrong (unpublished mapping), Compton (1972, 1975), Miller (1980), Todd (1980), Wells (1996), and new mapping by M.L. Wells (this study). Top inset, tectonic map of the western U.S., shows location of metamorphic core complexes (black fill), leading edge of the Sevier fold and thrust belt (thick line with sawteeth), and outline of the Cenozoic Basin and Range Province. Bottom inset is a generalized map illustrating location of northern Great Basin metamorphic core complexes (diagonal wavy pattern) and study area in the hinterland of the Sevier orogenic belt. Thrusts of the foreland fold and thrust belt are shown by barbed lines; hatchured lines are major normal faults of the Wasatch fault system.

angle faults between the sample sites, and there is no evidence (Holdaway 2000). Pressure constraints were obtained in rocks for foliation-parallel faulting within the schist of Stevens Spring, from the upper horizon using MBPG (muscovite-biotite-pla- and thus the two sampled horizons are considered to have un- gioclase-garnet) geobarometry. Although rocks from the lower dergone the same P-T history. Sample stations UH1 and UH2 horizon possess a mineral assemblage appropriate for the ap- are from the upper horizon on the south side of Basin Creek, plication of MBPG geobarometry and geobarometry based on and LH1 is from the lower horizon on the north side of Basin the anorthite breakdown reaction (GASP geobarometry), they Creek (Fig. 2). were not used because the low grossular content of the garnets Throughout the schist of Stevens Spring, a schistosity con- and the low anorthite content of the plagioclases create prob- sisting of granoblastic quartz and feldspar and planar aligned lems with the propagation of uncertainties (Todd 1998). Crude micas is overprinted by a greenschist-facies ductile fabric that pressure constraints for rocks in the lower horizon were ob- is expressed as shear bands, strain shadows around tained through considerations of staurolite and sillimanite sta- porphyroblasts, and dynamically recrystallized quartz. Samples bilities. were examined petrographically to determine which were best suited for detailed microprobe analysis. Samples showing the GEOLOGIC SETTING least visible alteration, the least overprint of the younger fab- The metamorphic rocks studied in the Grouse Creek Moun- ric, and a mineral assemblage appropriate for thermobarometry tains are part of a discontinuous belt of isolated occurrences of and for determining P-T paths based on garnet growth zoning amphibolite-facies Barrovian metamorphic rocks in the meta- (at least quartz + muscovite + biotite + plagioclase + garnet) morphic core complexes of the western Cordillera (Fig. 1). The were judged to be suitable. The temperatures of garnet growth regional metamorphism resulted from crustal thickening within were determined using garnet-biotite geothermometry the hinterland of the Sevier orogenic belt, mostly in Cretaceous HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 117

FIGURE 2. Geologic map (a), geologic cross section (b), and tectonostratigraphic column (c) of the Basin Creek area, northern Grouse Creek Mountains. Sample localities for studied pelitic schist are shown with labeled circles. Locations of kyanite- bearing metamorphic rocks are indicated with triangles on a, in close proximity to studied samples. All map units are keyed to the tectonostratigraphic column c, with the exception of Oligocene granite (black fill).

time, with local enhancement of deformation fabrics and con- graphically elevated hinterland crust experienced gravitation- tact-regional metamorphic effects around Cretaceous plutons ally driven horizontal extension synchronous with protracted (e.g., Snake Range, Miller and Gans 1989). The principal ar- shortening within the foreland fold and thrust belt. guments for the thrust-burial genesis for the metamorphism Throughout the Raft River, Grouse Creek, and Albion Moun- are metamorphic pressures locally recording burial of two to tains, a remarkably consistent yet greatly attenuated sequence three times that of stratigraphic burial (e.g., Hodges et al. 1992) of Proterozoic to Permian greenschist- to amphibolite-facies and systematic lateral metamorphic pressure gradients within metasedimentary rocks overlie the Archaean Green Creek com- stratigraphically equivalent metamorphic rocks (Hoisch and plex (Fig. 1) (see Compton et al. 1977; Miller et al. 1983; and Simpson 1993; Camilleri and Chamberlain 1997; Lewis et al. Wells 1997 for reviews of the stratigraphy, deformation his- 1999). Studies of metamorphism within these core complexes tory, and metamorphism). The Archaean Green Creek complex have determined conditions of peak prograde metamorphism (Armstrong and Hills 1967; Armstrong 1968), dominantly ~2.5 (Hodges et al. 1992; Camilleri and Chamberlain 1997; Lewis Ga adamellite gneiss (Compton 1972, 1975), is nonconformably et al. 1999), decompressional P-T paths recording exhumation overlain by the Elba Quartzite (Fig. 2), which forms the basal following peak burial (Hodges and Walker 1990; Hodges et al. unit of a sequence of alternating quartzite and psammatic, 1992; McGrew et al. 2000), and retrograde conditions of meta- pelitic, and amphibolitic schists. This sequence includes in as- morphism during extensional exhumation (Hurlow et al. 1991). cending order: the schist of Upper Narrows, quartzite of Yost, Following Mesozoic metamorphism and crustal shortening, schist of Stevens Springs, quartzite of Clarks Basin and schist strata within the Raft River, Albion, and Grouse Creek Moun- of Mahogany Peaks. Although previously interpreted to be of tains were attenuated by intralayer plastic flow and by Cambrian age based on map relationships (Armstrong 1968; excisement along low-angle faults and shear zones (see below) Compton 1972, 1975; Compton et al. 1977), a Neoproterozoic during regional extension in the Cretaceous (Wells 1997; Wells or Paleoproterozoic age for the quartzite of Clarks Basin is et al. 1998). A similar record of Cretaceous extension within suggested from carbon isotope studies of marble interbeds the Sevier orogenic belt has emerged from a number of regions (Wells et al. 1998), which also suggests a Proterozoic age for struc- (e.g., Hodges and Walker 1992; Camilleri and Chamberlain turally lower quartzite and schist units within this succession. 1997), and has led to the notion that the thickened and topo- Ordovician marbles, quartzites, and phyllites, correlative to 118 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT the Pogonip Group, Eureka Quartzite, and Fishhaven Dolomite, one from sample UH1, were analyzed with line traverses (Figs. everywhere overlie the Proterozoic metasedimentary rocks 3aÐe); element maps (Fe, Mn, Ca, and Mg) were generated on along a contact recently interpreted as a syn-orogenic normal three of these (Fig. 4). In addition, line traverses were done on fault that excised 4Ð5 km of structural section, named the Ma- two garnets from the lower horizon, one each from samples hogany Peaks fault (Fig. 2c) (Wells et al. 1998). Metamorphic LH1A and LH1B (Figs. 3fÐg). Muscovite, biotite, and plagio- conditions of hanging wall Ordovician and footwall Protero- clase in all samples were analyzed with spot analyses. Element zoic strata are upper greenschist and middle amphibolite fa- maps of garnets were generated at Rensselaer Polytechnic In- cies, respectively, and quantitative geothermometry indicates stitute using a JEOL-733 instrument. Spot analyses were col- a temperature discontinuity of about 100 °C. Muscovite 40Ar/ lected on two different machines, both of which were set to a 39Ar ages from Ordovician marbles document cooling through spot size of 10 µm and an accelerating voltage of 15 kV. Spot the closure temperature of ~350 °C at ~90 Ma (Wells et al. analyses were done on samples LH1A, UH1, UH2A, and UH2B 1990), which constrains the metamorphism to be older than 90 using the JEOL 8600 Superprobe at Arizona State University Ma and provides a maximum age for faulting. with a beam current of 10 nA. Characteristic X-rays were ac- A second syn-orogenic normal fault, the Emigrant Spring cumulated for 30 seconds or until the counting statistics fault (Fig. 2c), separates Pennsylvanian marble from Ordovi- achieved a standard deviation of 1%. Sample LH1B was ana- cian strata. Both the Emigrant Spring and Mahogany Peaks lyzed using the ETEC Autoprobe at Northern Arizona Univer- faults are folded about recumbent folds. This structural sequence sity. Characteristic X-rays were accumulated for 30 seconds or has been interpreted to record renewed crustal shortening, fol- until the counting statistics achieved a standard deviation of lowing extensional exhumation (Wells 1997). 0.5%. A beam current of 25 nA was used to analyze garnet; 10 The late Mesozoic to early Cenozoic ductile structures and nA was used for all other minerals. Different domains on indi- metamorphic rocks, developed at mid-crustal levels within the vidual polished sections are referred by F0, F1, or F2 after the hinterland of the Sevier orogenic belt, were exhumed along sample name. Garnet traverses were run with equal spacing two oppositely rooted Cenozoic extensional fault and shear zone between points. Poor analyses and analyses of inclusions were systems. The first is the Middle Mountain shear zone (Fig. 1), excluded using computer codes that tested the quality of the a west-rooted middle Eocene to early Oligocene shear zone analyses based on preset tolerances. exposed in the Grouse Creek, western Raft River, and Albion Mountains that was reactivated in the Miocene following Oli- PETROGRAPHY AND GARNET CHARACTERISTICS gocene intrusion (Compton 1983; Saltzer and Hodges 1988; Wells et al. 1997b). The second is the Miocene east-rooted Raft Upper horizon River detachment and shear zone (Fig. 1) of the central and All three samples from the upper horizon (UH1, UH2A, eastern Raft River Mountains (Malavieille 1987; Wells et al. and UH2B) contain the mineral assemblage quartz + musco- 2000). The Basin Creek area is located entirely within the vite + garnet + biotite + plagioclase. Ilmenite, tourmaline and Middle Mountain shear zone. monazite occur as accessory phases in all samples, and In the Basin Creek area, Archaean to Pennsylvanian rocks clinozoisite is an accessory phase in UH1 and UH2A. All are deformed by top-to-the-WNW shearing within the Middle samples lack graphite. Traces of chlorite altering from biotite Mountain shear zone. Studies in the Basin Creek and Vipont in sample UH2A constitute the only visible alteration in all Mountain areas (Fig. 1) (Wells et al. 1997a, 1997b), Middle three samples. All garnets are strongly and unevenly corroded Mountain and the City of Rocks (Saltzer and Hodges 1988; along the margins, having been reduced from initial diameters Miller and Bedford 1999), and in the southern Grouse Creek of ~3.5 mm to <2.5 mm, as judged from textural evidence and Mountains (Compton et al. 1977; Compton 1983), collectively from the longest preserved dimensions. All garnets contain in- suggest two periods of ductile shearing along the zone, the first clusions of ilmenite and quartz. Garnets in UH1 and UH2A in middle Eocene to early Oligocene time, and the second in also contain inclusions of clinozoisite. Monazite and tourma- the early to late Miocene. In the Basin Creek area, sparse cen- line occur as rare inclusions in garnets in all samples. timeter- to meter-thick sills of peraluminous leucogranite ema- The garnet element maps (Fig. 4) show that the corroded nating from the Vipont pluton intrude Archaean to Proterozoic rims are discordant to interior growth zoning. Secondary rocks. Further north and at an equivalent structural level, the reequilibration along the rims penetrated up to 300 µm, and is abundance of these intrusive rocks increases to the degree of associated with enrichments of Mg and Mn and a depletion of pervasively intruding the metasedimentary country rock (Fig. Fe (Figs. 3aÐe). The tendency for Mn to become enriched along 2a). U-Pb studies of monazite from the Vipont pluton indicate the margins of corroded garnets is well documented (e.g., Tracy a crystallization age of 27 ± 2 Ma (Wells et al. 1997a), consis- 1982) and results from Mn diffusing into the garnet from the tent with determined ages of 29 Ma for two correlative intru- consumed rim. Secondary reequilibration is also found along sive bodies to the north, the Middle Mountain injection complex cracks and around quartz and ilmenite inclusions, where Ca is and the Almo pluton (Armstrong 1968; Miller et al. 1983; depleted and Mg and Mn are enriched. These features resulted Forrest and Miller 1994). from cracks and inclusion margins acting as conduits for cat- ion migration. Whitney et al. (1996) described similar features MICROPROBE ANALYSIS and attributed them to metasomatism. Five garnets from the upper horizon of the schist of Stevens No garnet-consuming reaction, either prograde or retrograde, Spring, two from sample UH2A, two from sample UH2B, and can be written among the phases present in these rocks to ex- HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 119

FIGURE 3. Garnet zoning profiles (left) compared to numerical simulations of garnet growth (right). Thin dashed lines on right diagrams duplicate the portions of the zoning profiles on corresponding left diagrams used to produce the models. Garnet-biotite geothermometry, calculated using the calibration of Holdaway (2000) and assuming that 20% of Fe in biotite is Fe3+, is plotted above each profile (right). Anomalies in the profiles are explained as follows: points near a crack (C), points near an embayment (E), points near a quartz inclusion (Q). For three of the garnets (UH1, UH2A-F0, UH2A-F1) the locations of the traverses are shown on element maps in Figure 4. The endpoints of individual segments within the profiles are marked with Roman numerals. The values of all parameters used in the simulations are given in Tables 1 and 3. Abbreviations: Al = almandine; Gr = grossular; Py = pyrope; Sp = spessartine. 120 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT

9 Mn Ca Fe Mg

10 UH1

Mn Ca Fe Mg

8 9

UH2A-F0

Mn Ca Fe Mg

10

11

UH2A-F1

FIGURE 4. Element maps (Fe, Mn, Mg, and Ca) for garnets UH1, UH2A-F0, and UH2A-F1, from the upper horizon. Data was collected in a square grid pattern with a point spacing of 8Ð10 µm, a spot size equal to the point spacing, and a count time of 0.05 second at each point. The maps are grey-level representations of raw counts, with lighter shades representing higher counts. Lines represent microprobe traverses; endpoints are labeled on the Mn maps to correspond with profiles in Figure 3. The images of the different garnets are not proportionally to scale. For scales, see composition profiles in Figure 3. plain the consumption of garnet. Numerical simulations of gar- rolite grains occur as inclusions in plagioclase. Garnets also net growth (discussed later) document that for every volume contain abundant inclusions of ilmenite and quartz. The tex- percent of chlorite consumed, approximately one volume per- ture indicates a sequence of reactions in which staurolite growth cent of garnet is produced. If the garnet had become corroded was followed by the simultaneous growth of sillimanite and by way of this reaction operating in the retrograde direction, plagioclase, which was followed by the simultaneous break- then major amounts of chlorite should have been produced. down of staurolite and growth of garnet. The ductile fabric of Because chlorite is virtually absent from these rocks, this ex- the Middle Mountain shear zone weakly overprints the schis- planation is untenable. One plausible explanation, possibly re- tosity and there is no chloritization or any other visible alter- lated to the secondary reequilibration along cracks and inclusion ation. The garnets are dominantly idioblastic, but have margins, is fluid dissolution. undergone slight modification by dissolution along bounding shear surfaces. The outer surfaces of garnets as seen in outcrop Lower horizon are striated parallel to the northwest transport direction associ- Both samples (LH1A and LH1B) from the lower horizon ated with the ductile deformation, documenting that the defor- contain the mineral assemblage quartz + muscovite + garnet + mation post-dated the growth of garnet. Garnets from the lower biotite + plagioclase + staurolite + sillimanite. Ilmenite, tour- horizon are much larger than in the upper horizon, averaging maline, and monazite occur as accessory phases. The silliman- about 1 cm in diameter with some being as large as 2 cm. Some ite is partly fibrolitic and locally interwoven with biotite. garnets contain linear to slightly wavy inclusion trails of quartz Corroded staurolite grains occur in the matrix and as inclu- and ilmenite, and others are largely free of inclusions. Both sions within garnet, and both sillimanite and idioblastic stau- staurolite and plagioclase preserve inclusion trails of quartz HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 121

and ilmenite that range from linear to tight microfolds. plagioclase = garnet + muscovite + H2O. This reaction is typi- Line traverses of both analyzed garnets (Figs. 3fÐg) display cal of the upper greenschist facies. The KFMASH petrogenetic simple zoning profiles with no enrichment of Mn or Ca along grid of Spear and Cheney (1989) shows a narrow stability field rims, as seen in garnets from the upper horizon. The profiles for the mineral assemblage quartz + muscovite + chlorite + are interpreted as growth-related based on the decrease of Mn biotite + garnet, but with additional components (Ca and Mn) from core to rim. Iron increases smoothly from core to rim, in garnet, the field expands. Because the analyzed garnets con- and Ca and Mg are nearly flat. tain up to 32% extra components (18% spessartine and 14% grossular), the Spear and Cheney (1989) grid does not offer a CONDITIONS OF METAMORPHISM meaningful constraint on the conditions of metamorphism. Lower horizon. Garnet growth in samples collected from Mineral reactions the lower horizon is explained by a reaction involving the other The two horizons of the schist of Stevens Spring sampled minerals present: quartz + staurolite + muscovite = sillimanite in this study represent two different bulk compositions of pelitic + biotite + garnet + H2O. In the KFMASH system and on AFM schist. The lower horizon represents a more Al-rich and Ca- petrogenetic grids (e.g., Spear and Cheney 1989), this is a dis- poor composition than the upper horizon. Garnet in the upper continuous reaction; however, the presence of extra compo- horizon grew in the upper greenschist facies from the break- nents, notably Mn and Ca in garnet, make it continuous and down of chlorite, whereas garnet in the lower horizon grew in thus the assemblage is stable over a range of temperatures. The the upper amphibolite facies from the breakdown of staurolite. extra components in garnet cause the stability of the assem- Garnet-biotite geothermometry supports this interpretation (dis- blage to be displaced toward lower temperatures than in the cussed below), as do the following arguments concerning min- ideal system. Because the staurolite in the sample was not com- eral reactions. pletely consumed, the reaction must have progressed up to the Upper horizon. In the three samples from the upper hori- peak temperature, and ceased only when temperatures began zon (UH1, UH2A, and UH2B), no net-transfer reaction result- to decline. Using the Gibbs method, it was calculated that the ing in garnet growth is possible from the other minerals present temperature displacement from the end-member reaction due (quartz, biotite, muscovite, and plagioclase). Therefore, the to Mn and Ca in the garnet is 22.8 °C (see Appendix 1 for garnet growth reaction must have involved a mineral that was details of the calculation). The calculated stability field for the completely consumed by the reaction. In pelitic schist of com- breakdown of staurolite and growth of garnet for sample LH1A mon composition, this mineral would ordinarily be chlorite. If is shown on Figure 5. chlorite is added to the mineral assemblage, then the following The late crystallization of plagioclase in samples from the reaction explains garnet growth: quartz + biotite + chlorite + lower horizon is attributable to the breakdown of paragonite.

FIGURE 5. Calculated constraints on the 10 peak conditions of metamorphism in sample Calculated for KFMASH system LH1A, from the lower horizon of the schist of Stevens Springs. The conditions are Calculated using garnet rim from LH1A 9 constrained by the mineral assemblage and Calculated using garnet core from LH1A textural relations to be within the sillimanite stability field at higher temperatures than the 8 Qz + Mus + St paragonite-breakdown reaction, at lower temperatures than the muscovite-breakdown Gar + Als + Bio + H2O 7 reaction, within the field of the staurolite- Garnet-biotite geothermometry breakdown reaction (stripe pattern), and along for LH1A using garnet rim the staurolite-breakdown line calculated for 6 the garnet rim. Garnet-biotite geothermometry

(model “5AV” of Holdaway 2000) is shown (kbar) P calculated using the garnet rim composition 5 assuming that 20% of the total Fe in biotite is s u 2O Fe3+. The shaded area represents a ±25 °C M H Ky g 2O + + 4 P H z ls uncertainty in both the garnet-biotite + + Q A Sil z .5 + geothermometry and the staurolite-breakdown Q n4 sp A K line calculated for the garnet rim. All reactions And + 3 s were calculated using the thermodynamic data Al and method of Berman (1988), except for the staurolite-breakdown reactions (see Appendix 2 1). The paragonite-breakdown reaction was 500 550 600 650 700 750 calculated using the plagioclase activity model T (°C) of Fuhrman and Lindsley (1988) and the plagioclase composition shown for LH1A in

Tables 1 and 2. Abbreviations: Als = Al2SiO5; An4.5 = plagioclase of composition An4.5; And = andalusite; Bio = biotite; Gar = garnet; Ksp = Sanidine; Ky = kyanite; Mus = muscovite; Qz = quartz; Sil = sillimanite; St = staurolite. 122 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT

This interpretation is supported by the highly sodic composi- tion may be evaluated. When temperatures are calculated for tion (An4.5 to An7.2) of the plagioclase (Tables 1 and 2). The the garnet cores using biotite compositions that are adjusted to paragonite-breakdown reaction shown on Figure 5 was calcu- account for the change in biotite composition during garnet lated from the thermodynamic data of Berman (1988), the al- growth (core to rim I as shown on Fig. 3), the values increase bite activity model of Fuhrman and Lindsley (1988), and the by 8.3, 9.6, 17.5, 2.8, and 1.8 °C for UH1, UH2A-F0, UH2A- measured plagioclase composition for sample LH1A. The tem- F1, UH2B-F0, and UH2B-F2, respectively, and decrease by perature of the reaction was calculated to be slightly lower than 10.1 and 8.0 °C for LH1A and LH1B, respectively. Thus, use that of the first staurolite breakdown for pressures less than ~6 of the average matrix composition does not introduce large er- kbar (Fig. 5), which is consistent with petrographic observa- rors for the applications in this study. tions of idioblastic staurolite inclusions in plagioclase and Many reequilibrated points along the garnet traverses are heavily corroded staurolite in the matrix, indicating that pla- not immediately apparent from a visual examination of the gioclase grew before staurolite began to break down. garnet zoning plots, but become apparent after calculating gar-

Although sillimanite is the only Al2SiO5 polymorph found net-biotite geothermometry. Positive (prograde) temperature in the Basin Creek area, kyanite has been found in several deviations along rims, cracks, inclusion margins, and near nearby locations in Proterozoic metasedimentary rocks and in embayments as large as ~100 °C are easily distinguished and the Archaean basement (Wells et al. 1998; Hanson 1997). The document the effects of secondary reequilibration (Fig. 3). To kyanite occurrence closest to Basin Creek is 2 km to the north, evaluate whether the deviations could arise from counting sta- where boulders eroded from the Archaean basement contain tistics, a Monte Carlo simulation (Box and Muller 1958) of the kyanite blades up to 10 cm in length (Fig. 2a). At that locality, Holdaway (2000) “5AV” geothermometer was performed. The the kyanite is partially replaced by sillimanite. A kyanite pre- simulation was applied directly to the oxide wt% values for cursor for sillimanite in rocks from the lower horizon of the biotite and garnet prior to formula normalization using micro- schist of Stevens Spring in the Basin Creek area is, however, probe counting statistics as measures of one sigma variability inconsistent with the morphology of the sillimanite bundles. for each oxide wt% value. Based on one hundred determina- tions, using the rim I garnet composition and the matrix biotite Garnet-biotite geothermometry composition for sample domain UH2A-F1 (Table 2), a mean This study makes use of the Holdaway (2000) formulation and standard deviation of 455.2 ± 2.66 °C was determined (Fig. of the garnet-biotite geothermometer, which was derived for 7). Thus, the observed order-of-magnitude larger deviations rocks that are buffered to low values of oxygen fugacity by the along the garnet traverses cannot arise from counting statis- presence of graphite. Based on the Mössbauer data contained tics. Holdaway (2000) estimated absolute uncertainties in ap- within the study of Guidotti and Dyar (1991), Holdaway (2000) plications of his garnet-biotite geothermometer to be ±25 °C. incorporated the assumption that 11.6% of total Fe in biotite is Upper horizon. For samples UH1, UH2A, and UH2B, tem- Fe3+. The rocks in the present study lack graphite but contain perature profiles in the garnet interiors are flat (Figs. 3aÐe) and ilmenite, which is permissive of higher values of oxygen fugac- range from ~460 to ~490 °C. Thus, the temperature associated ity and greater Fe3+ contents in biotite. The composition of il- with growth of the garnet core is interpreted to be ~475 °C. menite may be interpreted as an indication of redox conditions Prograde reequilibration is displayed along rims and within the (e.g., Rumble 1973). Microprobe analyses of ilmenite grains interiors along cracks and/or inclusion margins on all garnets in rocks from both horizons show no hematite component within analyzed. analytical uncertainties, consistent with reducing conditions; Lower horizon. For sample LH1A, the non-reequilibrated however, higher values of Fe3+ in biotite ranging from 20.3 to garnet rim composition, represented by the analysis labeled “I” 33.0% were found using wet chemistry methods applied to four on Figure 3f, yielded a garnet-biotite geothermometry value of samples from the upper horizon (Mahlen et al. 2000). For the 633 °C at an assumed pressure of 5 kbar. Retrograde purposes of calculating the temperatures shown on Figure 3, it reequilibration along the rims and along a crack passing through was assumed that 20% of the total iron is Fe3+ in biotite, based the core is clearly displayed. For sample LH1B, the highest on the lowest value obtained from wet chemistry, although it is temperature calculated along the traverse is 640 °C (Fig. 3g); possible that all of the rocks have been affected by secondary many points along this traverse appear to have undergone vary- oxidation and that the primary values were less than 20%. In- ing degrees of retrograde reequilibration. From these results, corporating an assumed value of 10% into the geothermometry garnets in the lower horizon are interpreted to have ended calculation results in temperatures ~10 °C higher than assum- growth at ~635 °C. ing 20% for upper-horizon samples (Fig. 6) and about ~25 °C Interpretation. The temperatures of garnet growth and the higher for lower-horizon samples. In addition, it was assumed nature of the reequilibrations, prograde in the upper-horizon that 3% of total Fe in garnet is octahedrally coordinated Fe3+ rocks and retrograde in the lower-horizon rocks, along the gar- (Holdaway et al. 1997; Holdaway 2000). net traverses are consistent with a P-T path in which garnets in Garnet-biotite geothermometry (Holdaway 2000, model the upper horizon grew before those in the lower horizon. Gar- “5AV”) was applied to every spot analysis along the garnet nets from rocks from the upper horizon began growth at tem- traverses using the average matrix biotite composition (Fig. peratures of ~475 °C, then underwent partial dissolution and 3). Using the changes in biotite composition predicted in the prograde reequilibration along rims, cracks and inclusion mar- numerical simulations of garnet growth (Table 1), the errors gins, as the rocks heated to temperatures over 600 °C, required introduced by the assumption of a constant biotite composi- to produce garnet in the lower horizon. Garnets from rocks in HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 123

10 10

9 9

8 8

7 7

6

(kbar) 6

P 5 5

4 4

3 3 UH2A-F0 UH2A-F1 2 2 350 400 450 500 550 600 650 700 750 350 400 450 500 550 600 650 700 750 10 10

9 9

8 8

7 7

6

(kbar) 6

P 5 5

4 4

3 3 UH2B-F0 UH2B-F2 2 2 350 400 450 500 550 600 650 700 750 350 400 450 500 550 600 650 700 750 10 T (°C)

9

8 Garnet-Biotite Geothermometry 7 10% Fe3+ 20% Fe3+ 6 H&S (kbar) P MBPG Geobarometry 5 involving Tschermak components not involving Tschermak components 4

3 UH1 2 350 400 450 500 550 600 650 700 750 T (°C)

FIGURE 6. MBPG geobarometry (Hoisch 1991) and garnet-biotite geothermometry (model “5AV” of Holdaway 2000) for upper horizon samples. Geothermometry was calculated using Holdaway (2000) assuming 10% of total Fe is Fe3+, labeled 10% Fe3+, and assuming 20%, labeled 20% Fe3+, and using Hodges and Spear (1982), labeled H&S (assumes all Fe as Fe2+). MBPG equilibria involving Tschermak components of micas are distinguished from those that do not by different line patterns (see legend on figure). The calculations used average matrix compositions for biotite, muscovite and plagioclase, and garnet rim compositions labeled “I” on Figure 3 (Table 2). 124 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT

TABLE 1. Values used in the numerical simulations of garnet growth and calculated P-T paths UH2B-F2 UH2B-F0 UH1 UH2A-F1 UH2A-F0 LH1B LH1A Initial Values

Xann* 0.6629 0.6629 0.6090 0.5835 0.5835 0.6358 0.6446 XMn-Bio* 0.0048 0.0048 0.0070 0.0053 0.0053 0.0015 0.0013 Xab* 0.7872 0.7872 0.7614 0.6876 0.6876 0.9276 0.9547 Xan* 0.2128 0.2128 0.2386 0.3124 0.3124 0.0724 0.0453 Xclin† 0.2695 0.2695 0.2478 0.2602 0.2602 n/a n/a Xdaph† 0.7205 0.7205 0.7422 0.7298 0.7298 n/a n/a XMn-Chte† 0.0100 0.0100 0.0100 0.0100 0.0100 n/a n/a XMg-St‡ n/a n/a n/a n/a n/a 0.1294 0.1294 XFe-St‡ n/a n/a n/a n/a n/a 0.8674 0.8674 XMn-St‡ n/a n/a n/a n/a n/a 0.0032 0.0032 Xpy*,¤ 0.0283 0.0296 0.0271 0.0310 0.0302 0.0954 0.0944 Xal*,¤ 0.6542 0.6658 0.6737 0.6956 0.6958 0.7433 0.7595 XSp*,¤ 0.1799 0.1401 0.1397 0.1329 0.1180 0.1125 0.1284 XGr*,¤ 0.1377 0.1646 0.1595 0.1405 0.1561 0.0480 0.0177 T (ûC) 475 475 475 475 475 600 600 P (bars) 5000 5000 5000 5000 5000 3700 3700 garnets/100 cm3 || 65 400 800 1800 1200 2.5 3

Monitor parameters (#) and calculated values of ∆T and ∆P Rim I-Core Rim I-Core Rim V-Core Rim IV-Core Rim II-Core Rim I-Core Rim I-Core

∆Xal* 0.1206# 0.1141# 0.0466# 0.0673# 0.0754# 0.0794# 0.0646# ∆XSp* n/a n/a n/a n/a n/a Ð0.0496# Ð0.0735# ∆XGr* Ð0.0234# Ð0.0343# Ð0.0248# Ð0.0157# Ð0.0110# n/a n/a ∆T (°C) 22.4 21.9 9.7 12.7 13.9 34.3 29.2 ∆P (bars) 375.4 269.0 Ð77.6 173.0 388.5 972.5 861.1 Rim IV-Rim V Rim III-Rim IV Rim I-Rim II

∆Xal* ÐÐ0.0145# 0.0020# n/a Ð Ð ∆XGr* ÐÐ0.0257# 0.0252# 0.0828# Ð Ð ∆MGar (mols) Ð Ð n/a n/a 0.000012# Ð Ð ∆T (°C) Ð Ð 0.6 Ð1.7 Ð10.9 Ð Ð ∆P (bars) Ð Ð 566.4 496.6 1021.9 Ð Ð Rim IIIÐRim IV Rim II-Rim III

∆Xal* ÐÐ0.0439# n/a Ð Ð Ð ∆XGr* ÐÐÐ0.0215# 0.0453# Ð Ð Ð ∆MGar (mols) Ð Ð n/a 0.000008# Ð Ð Ð ∆T (°C) Ð Ð 9.1 Ð2.5 Ð Ð Ð ∆P (bars) Ð Ð Ð12.7 797.9 Ð ÐÐ Rim II-Rim III Rim I-Rim II

∆Xal* ÐÐÐ0.0419# n/a Ð Ð Ð ∆XGr* ÐÐ0.0662# Ð0.0059# Ð Ð Ð ∆MGar (mols) Ð Ð n/a 0.00001# Ð Ð Ð ∆T (°C) ÐÐÐ11.3 5.3 Ð Ð Ð ∆P (bars) Ð Ð 759.4 315.3 Ð Ð Ð Rim I-Rim II

∆Xal* ÐÐ0.0168# ÐÐÐÐ ∆XGr* ÐÐÐ0.0060# ÐÐÐÐ ∆T (°C) ÐÐ3.8ÐÐÐÐ ∆P (bars) Ð Ð 167.4 ÐÐÐÐ

∆Xphl** 0.0071 0.0107 0.0337 0.0725 0.0387 Ð0.0235 Ð0.0193 Note: Abbreviations are Al = Almandine; Ann = Annite; Ab = Albite; An = Anorthite; Bio = Biotite; Ch = Chlorite; Clin = Clinochlore; Daph = Daphnite; Gar, garnet; Gr = Grossular; M = moles of garnet; n/a = not applicable; Phl = Phlogopite; Py = Pyrope; Sp = Spessartine; St = Staurolite; X = fraction. * Calculated from microprobe data. † Chlorite initial values were estimated based on 500 °C and 5 kbar and the garnet-chlorite Fe-Mg exchange equilibrium of Grambling (1990). See text for further explanation. ‡ Compositions for LH1B staurolite were taken from measured compositions for LH1A. ¤ Garnet core value. || Nucleation density. # Monitor parameter value.

** Cumulative change in Xphl value. the lower horizon ended growth at temperatures of ~635 °C, with the biotite except narrowly along rims, cracks, and inclusion then underwent partial retrograde reequilibration along rims margins, which are volumetrically insignificant. and cracks as the rocks cooled, with the process ending when the temperatures dropped below the closure temperature for cation Geobarometry diffusion in garnet, ~600 °C for most geologically realistic cool- The 43 geobarometers of Hoisch (1991) were applied to ing rates (Spear 1989). The preservation of comparatively low samples from the upper horizon using mineral compositions temperatures in the rocks from the upper horizon is made possible from Table 2 (muscovite, biotite, plagioclase, and rim I com- by a lack of net-transfer reactions through the additional ~150 °C positions for garnet). The geobarometers are based on equilib- of heating that followed garnet growth, and by the fact the Fe and ria among quartz, muscovite, biotite, plagioclase, and garnet Mg contained in the garnets were largely unavailable for exchange (MBPG geobarometry). The mineral assemblage in samples HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 125

TABLE 2A. Microprobe analyses of garnets at rim I TABLE 2C. Microprobe analyses of plagioclase wt% UH2A-F1 UH2A-F0 UH2B-F0 UH2B-F2 UH1 LH1A LH1B wt% UH2A UH2B UH1 LH1B LH1A

SiO2 37.860 38.320 38.050 37.550 38.190 37.232 37.590 SiO2 61.128 63.267 62.987 65.960 66.696 Al2O3 21.625 21.500 21.300 21.020 21.610 20.61 20.890 Al2O3 25.815 24.015 24.555 22.597 22.075 Fe2O3 0.000 0.000 0.000 0.000 0.000 0.000 0.134 FeO* 0.020 0.020 0.075 0.057 0.022 FeO 33.180 31.260 35.130 33.675 33.890 36.177 36.909 CaO 6.455 4.460 5.037 1.600 0.968

MnO 1.145 1.460 2.360 3.025 0.680 2.386 2.870 Na2O 7.970 9.107 8.875 10.593 11.282 MgO 0.880 0.970 0.930 0.985 0.880 2.727 2.367 K2O 0.125 0.160 0.110 0.103 0.100 CaO 6.550 7.950 4.580 3.880 6.910 0.36 0.497 Total 101.512 101.030 101.640 100.910 101.143 Total 101.240 101.460 102.350 100.135 102.160 99.492 101.257 Formula proportions based on 8 oxygen atoms Formula proportions based on 8 cations and 12 oxygen atoms Si 2.678 2.770 2.746 2.868 2.892 Si 3.009 3.025 3.015 3.032 3.011 3.026 3.016 Al 1.333 1.239 1.262 1.158 1.128 Al 2.025 2.001 1.989 2.000 2.008 1.975 1.973 Fe2+ 0.001 0.001 0.003 0.002 0.001 Fe3+,* 0.000 0.000 0.000 0.000 0.000 0.000 0.008 Ca 0.303 0.209 0.235 0.075 0.045 Fe2+ 2.205 2.064 2.328 2.274 2.234 2.459 2.476 Na 0.677 0.773 0.750 0.893 0.949 Mn 0.077 0.098 0.158 0.207 0.045 0.164 0.195 K 0.007 0.009 0.006 0.006 0.006 Mg 0.104 0.114 0.110 0.119 0.103 0.330 0.283 * All Fe assumed to be Fe2+. Ca 0.558 0.672 0.389 0.336 0.584 0.031 0.043 * For garnet-biotite geothermometry calculations, 3% of the total Fe was assigned to octahedrally-coordinated Fe3+, consistent with Holdaway et al. (1997) and Holdaway (2000).

TABLE 2B. Microprobe analyses of micas wt% UH2A UH2A UH2B UH2B LH1B LH1B UH1 UH1 LH1A LH1A Biotite Musovite Biotite Musovite Biotite Musovite Biotite Musovite Biotite Musovite

SiO2 35.450 47.197 34.390 45.980 35.087 47.440 34.999 46.355 33.782 44.368 Al2O3 20.210 35.070 19.257 34.896 19.875 36.520 19.020 35.058 20.884 37.749 TiO2 2.085 0.425 2.197 0.438 1.965 0.000 2.203 0.555 1.630 0.370 FeO* 21.065 0.998 23.295 1.190 22.560 2.140 21.494 1.143 23.299 1.476 MnO 0.180 0.005 0.160 0.016 0.053 0.000 0.254 0.005 0.048 0.006 MgO 8.325 0.755 6.550 0.500 7.223 0.730 7.599 0.675 7.183 0.454 CaO 0.013 0.015 0.025 0.020 0.047 0.000 0.083 0.003 0.106 0.003

Na2O 0.233 0.860 0.240 0.882 0.188 0.590 0.143 0.768 0.085 1.733 K2O 8.475 8.838 8.238 9.046 8.850 8.550 8.064 9.133 8.072 8.035 F 0.000 0.000 0.000 0.000 n/a n/a 0.029 0.000 n/a n/a Total 96.035 94.162 94.352 92.968 95.847 95.970 93.875 93.693 95.179 94.199

Formula proportions based on 11 oxygen atoms Si 2.678 3.137 2.682 3.108 2.684 3.100 2.715 3.108 2.611 2.962 Al 1.800 2.747 1.770 2.780 1.792 2.812 1.739 2.770 1.897 2.971 Ti 0.118 0.021 0.129 0.022 0.113 0.000 0.128 0.028 0.094 0.019 Fe2+ 1.331 0.055 1.520 0.067 1.443 0.117 1.394 0.064 1.502 0.082 Mn 0.012 0.000 0.011 0.001 0.003 0.000 0.017 0.000 0.003 0.000 Mg 0.938 0.075 0.762 0.050 0.823 0.071 0.878 0.067 0.825 0.045 Ca 0.001 0.001 0.002 0.001 0.004 0.000 0.007 0.000 0.009 0.001 Na 0.034 0.111 0.036 0.116 0.028 0.075 0.021 0.100 0.013 0.224 K 0.817 0.749 0.820 0.780 0.863 0.713 0.798 0.781 0.794 0.684 F 0.000 0.000 0.000 0.000 n/a n/a 0.007 0.000 n/a n/a Note: n/a = not analyzed. * All Fe assumed to be Fe2+.

from the lower horizon also allowed for the application of equilibrium and leaves room for a broad interpretation of pres- MBPG geobarometry, and geobarometry based on the anorth- sure. Disequilibrium involving the mica Tschermak components ite-breakdown reaction involving sillimanite (GASP barom- may have developed during the heating that took place subse- etry), but neither were applied owing to problems with the quent to the growth of the garnets. Of the 43 geobarometers propagation of uncertainties. Large uncertainties are to be ex- calculated, 41 of them involve Tschermak components of mus- pected in the application of both GASP and MBPG barometers covite and/or biotite (Fig. 6). for these samples because of the Ca-poor compositions of both Second, the geobarometers of Hoisch (1991) contain large the garnet (Xgr = 0.0143) and plagioclase (XAn = 0.04Ð0.07). uncertainties that derive from the garnet activity model used in Todd (1998) estimated that uncertainties in the GASP the calibration (Hodges and Spear 1982) and the empirical na- geobarometer under these circumstances exceed ±3 kbar. The ture of the calibration. More recent versions of garnet activity MBPG geobarometers of Hoisch (1991) behave similarly with models (Berman 1990; Berman and Aranovich 1996; respect to error propagation. Mukhopadhyay et al. 1997; Holdaway 2000) incorporate newer Upper horizon. The results of the MBPG geobarometry sources of data and calculate substantially different activity equilibria (Fig. 6) are difficult to interpret for two reasons. First, coefficients for the pyrope, almandine, and grossular compo- the scatter among the MBPG equilibria is suggestive of dis- nents than the Hodges and Spear (1982) model. These newer 126 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT

30 simulating the growth of each segment independently. The in- µ ± s dividual segments are defined as shown on the garnet profiles 455.2 ± 2.66 °C in Figure 3. Points labeled I, II, III, etc., along the traverses mark the endpoints of the segments and specify the points used in the calculations. The endpoints of segments were located at breaks in the slope of one or more elements along the profile, 20 and chosen so that they avoided reequilibrated points near rims, cracks, inclusion margins, and embayments. For three of the garnets (UH2A-F1, UH2B-F2, and LH1B), analogous points within both left and right halves of the profiles were averaged.

Frequency The profiles of the other four garnets (LH1A, UH1, UH2A-F0, 10 and UH2B-F0) allowed use of only half of the profile because of having to avoid reequilibrated points. The usable portion of a garnet profile displayed a decrease in manganese from the core outward to the point labeled “I” (Fig. 3). Increases in Mn toward the rims are known to be products of secondary reequilibration during rim consumption (e.g., Tracy 1982), 0 which is confirmed in this study by a strong correlation with 445 450 455 460 465 perturbations in the garnet-biotite geothermometry (Fig. 3). T (°C) Each step in the simulations assumed a duration of 0.3 m.y., during which time diffusion of 8-coordinated cations took place FIGURE 7. Histogram of one hundred temperatures determined in a within the garnet. The amount of diffusion that took place re- Monte Carlo simulation of garnet-biotite geothermometry for UH2A-F1. sulted in imperceptible changes in the zoning profiles. The cat- ion diffusion coefficients of Chackraborty and Ganguly (1990) were used. In the simulations, water released from the net-trans- models may be more accurate, but cannot be used in conjunc- fer reaction left the system. Garnet interiors were fractionated tion with the MBPG geobarometers of Hoisch (1991) because but did not leave the system; matrix phases were in equilib- the Hodges and Spear (1982) model is embedded in the cali- rium only with the immediate garnet rim. Ideal activity models bration of the geobarometers. We interpret the pressure associ- were assumed for chlorite, staurolite, garnet, biotite, and pla- ated with initial garnet growth to have been ~5 kbar, which is gioclase. Muscovite, quartz, and the H2O-fluid were assumed broadly consistent with the intersections of the Hodges and to be pure end-members with activities of one. The simulation Spear (1982) garnet-biotite geothermometry lines with the of each segment of garnet growth was subdivided into 50 equal geobarometry equilibria not involving Tschermak components intervals to reduce descretization error. (Fig. 6), and with a moderate steady-state geothermal gradient According to Duhem’s theorem, there are only two degrees unperturbed by plutons at temperatures of 460Ð490 °C (there of freedom in closed heterogeneous chemical systems when are no Mesozoic plutons in the region). mass balance is taken into consideration (e.g., Spear 1993). Lower horizon. Pressures are not well constrained in rocks Any two variables that describe changes in the system during a from the lower horizon. The presence of the staurolite break- segment of garnet growth may be selected as monitors of the down assemblage, including sillimanite, offers loose con- system. These include T, P, mole fractions of the end members straints. For sample LH1A, the staurolite breakdown reaction in solid solution, and the modal proportion of any phase. The calculated for the garnet rim crosses the garnet-biotite exchange changes in garnet composition are recorded in the profile, as equilibrium at 4.7 kbar (Fig. 5). Combining a ±25 °C uncer- are changes in the mode of garnet. Consequently, two garnet tainty in the garnet-biotite geothermometer (Holdaway 2000) mole fractions, or one garnet mole fraction and the garnet mode, with a ±25 °C assumed uncertainty in the staurolite-breakdown were selected as monitors. The simulations are strongly depen- reaction yields a pressure uncertainty of +2.8/Ð2.3 kbar, which dent on the values of the changes specified for the monitors. represents essentially the same range of pressures at which sil- The method also requires that the initial compositions of all limanite is stable (Fig. 5). Spear and Cheney (1989) offered no solid solution phases be specified. Only the garnet initial com- assessment of uncertainties in their calculation of the KFMASH positions, represented by the cores, were actually measurable. staurolite breakdown reaction, upon which the calculated dis- The initial compositions of the matrix phases can never be placements shown on Figure 5 are based. known because the present-day matrix compositions were in- herited after the garnets had grown, and the garnets do not con- NUMERICAL SIMULATIONS OF GARNET GROWTH tain inclusions of these phases. However, the simulations are Growth of the seven garnets analyzed in this study was simu- largely insensitive to the choice of compositions of the matrix lated numerically using the Gibbs method with Duhem’s theo- phases and the models predict only small changes during gar- rem, as calculated by the program GIBBS4.7 (distribution of net growth. Thus, it was satisfactory to use average measured 5/7/97) or GIBBS (version Feb6_99) (Spear et al. 1991 de- compositions for the matrix phases. Values used in the simula- scribed the earliest version). The simulations were carried out tions are listed in Tables 1 and 3. Table 1 lists initial mineral by first subdividing the garnet traverses into segments, then compositions, initial pressures and temperatures, assumed HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 127 nucleation densities, choices and values for monitor param- garnet to be simulated, leaving much of the chlorite in the simu- eters, and the calculated changes in P and T. Table 3 lists the lations unreacted. initial mineral modes assumed for the model, the mineral modes The simulations of both garnets from sample UH2A uti- reported by the model at the end of garnet growth to rim I, and lized the modal abundance of garnet as one of the two monitor the visually estimated mineral mode. parameters for simulating the outer segments of the zoning pro- files (Table 1). The failure of using only compositional data to Upper horizon simulate outer segments can be attributed to the effects of rim The initial mineral assemblage for modeling samples from reequilibration with respect to Mn. Although the Mn content the upper horizon is quartz + muscovite + biotite + plagioclase declines toward the rims through these intervals, reequilibration + chlorite. Because there is no primary chlorite left in any of lessened the steepness of the decline concomitant with a de- the three samples in which the garnets are inferred to have crease in Fe. Calcium appears to have been unaffected through grown from chlorite (UH1, UH2A, and UH2B), the composi- these intervals; hence the grossular content was used as the tions had to be estimated. The chlorite was assumed to have a second monitor parameter. stoichiometry typical for pelitic schist: (Fe,Mg,Mn,Al)6.0 Although very similar in their profiles and yielding similar VI VI (Al1.5Si2.5)O10(OH)8, where Al is set to 1.5 and Mn is set to P-T paths, the two garnets from UH2B were simulated with 0.01. The Fe/Mg ratio was calculated using the garnet-chlorite very different nucleation densities: 400 garnets per 100 cm3 of Fe-Mg exchange geothermometer of Grambling (1990), assum- rock for domain F0, and 65 garnets per 100 cm3 of rock for ing conditions of 500 °C and 5 kbar, and using the garnet core domain F2. This result suggests highly localized domains of compositions (Table 1). The simulations assumed initial con- equilibrium and sluggish rates of intergranular diffusion, with ditions corresponding to the growth of garnet cores of 475 °C effective nucleation densities varying almost an order of mag- and 5 kbar, as discussed earlier. nitude on the scale of a thin section. The simulations of garnet growth from the upper horizon consumed only a small fraction of the chlorite. In the real rocks, Lower horizon the chlorite was completely consumed, resulting in the growth The initial mineral assemblage for modeling samples from the of much larger garnets than are presently preserved in the rocks. lower horizon is quartz + muscovite + biotite + staurolite + pla- Subsequent dissolution and reequilibration along the garnet rims gioclase + sillimanite. For samples LH1A and LH1B, conditions permitted only a small fraction of the original volume of any of 600 °C and 3.7 kbar were chosen to represent the garnet cores.

TABLE 3. Mineral modes in volume percents for rocks and numerical simulations of garnet growth Quartz Muscovite Biotite Plagioclase Chlorite Garnet Staurolite Sillimanite UH2B-F2 Model initial values 10.00 8.00 13.00 59.00 10.00 0.00 n/a n/a After growth to rim I 9.87 8.11 12.89 58.78 9.38 0.68 n/a n/a Visual estimate 10.00 10.00 16.00 60.00 0.00 4.00 n/a n/a

UH1-F2 Model initial values 7.00 8.00 14.00 57.00 14.00 0.00 n/a n/a After growth to rim I 6.84 8.56 13.28 55.52 11.52 3.04 n/a n/a Visual estimate 5.00 10.00 18.00 58.00 0.00 9.00 n/a n/a

UH2B-F0 Model initial values 10.00 8.00 13.00 59.00 10.00 0.00 n/a n/a After growth to rim I 9.80 8.15 12.80 58.65 9.15 1.00 n/a n/a Visual estimate 10.00 10.00 16.00 60.00 0.00 4.00 n/a n/a

UH2A-F0 Model initial values 10.00 8.00 13.00 59.00 10.00 0.00 n/a n/a After growth to rim I 9.52 8.48 12.40 57.76 7.92 2.56 n/a n/a Visual estimate 10.00 10.00 16.00 60.00 0.00 4.00 n/a n/a

UH2A-F1 Model initial values 10.00 8.00 13.00 59.00 10.00 0.00 n/a n/a After growth to rim I 9.24 8.88 12.00 57.00 6.60 4.20 n/a n/a Visual estimate 10.00 10.00 16.00 60.00 0.00 4.00 n/a n/a

LH1B Model initial values 8.50 25.00 35.00 0.50 n/a 0.00 25.00 6.00 After growth to rim I 6.66 20.23 40.20 0.47 n/a 0.36 12.69 19.83 Visual estimate 10.00 40.00 21.00 10.00 n/a 7.00 2.00 10.00

LH1A Model initial values 25.00 31.00 9.00 5.00 n/a 0.00 30.00 0.00 After growth to rim I 24.46 29.93 10.17 4.98 n/a 0.28 26.68 3.59 Visual estimate 23.00 30.00 10.00 5.00 n/a 10.00 20.00 2.00 Note: n/a = not applicable. 128 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT

This choice resulted in the simulations arriving at ~630Ð635 °C Spear 1989). Temperatures along the P-T path reached ~635 and ~4.6Ð4.7 kbar at the completion of garnet growth, which con- °C, suggesting that some diffusion should have taken place. forms to the garnet core-to-rim staurolite-breakdown field shown Cation diffusion was associated with prograde reequilibration on Figure 5, and is consistent with garnet rim geothermometry for along garnets rims, cracks and inclusion margins in the upper- LH1A. Both garnet crystals have simple profiles, which justified horizon rocks, and retrograde reequilibration along garnet rims only a single step in the growth simulations. and cracks in lower-horizon rocks. Presently, there are no in- The simulation of LH1B resulted in significant mismatches dependent data on the ages of garnet growth or the length of with respect to the modal proportions of muscovite and silli- time the rocks in the upper horizon experienced the higher tem- manite (Table 3). The actual rock contains less sillimanite and peratures, however, it can be inferred that the duration was not more muscovite than is predicted in the simulation and it was long enough or the temperatures not hot enough to cause ho- not possible to rectify this with different choices for initial mogenization of garnets in either horizon. modes. This discrepancy might be a result of the polished section The P-T paths derived from the garnet growth simulations possessing a non-representative sampling of modal abundance, document two separate episodes of thrust burial (Fig. 8). Gar- which is consistent with the rocks being strongly heterogeneous net crystals from the upper horizon record the lower tempera- on the scale of a few centimeters. Similar problems arose in the ture episode, which involved a nearly isothermal pressure modeling of LH1A for possibly similar reasons, where the simu- increase of ~1.7 kbar at ~475 °C. All five garnet crystals ap- lated abundance of garnet is much less than actual abundance. In pear to be defining the same path, with two preserving only order to obtain a good fit for LH1A, the model required the use of small portions of it. Because two of the three samples analyzed initial modes that resulted in final modes much different from those from the upper horizon contain trace amounts of clinozoisite observed in the sample (Table 3). (UH1 and UH2A), it is possible that the garnet growth reaction involved clinozoisite, either at an early stage or throughout the DISCUSSION growth history. Furthermore, the unevenness of the Ca-enriched Several lines of evidence support the interpretation that gar- zone at the margins of two of the garnets (UH1 and UH2A-F1, net grew in the upper and lower horizons at different grades of Fig. 4) suggests a complex growth history involving one or metamorphism by different reactions along the same P-T path. more episodes of consumption. The fact that all garnet crystals In the upper horizon, the growth of garnet by chlorite break- from upper-horizon samples appear to be defining the same P- down in the upper greenschist facies is consistent with garnet- T path suggests that clinozoisite was unimportant to the garnet biotite geothermometry, which yielded temperatures of growth reaction and that there are no hiatuses in the zoning ~460Ð490 °C. The mineral assemblage that resulted following profiles, but neither possibility can be ruled out conclusively. chlorite breakdown, quartz + plagioclase + biotite + muscovite The higher-temperature episode, defined by garnet in two + garnet, is stable throughout the ~150 °C higher temperature samples from the lower horizon, involved a steep P-T trajec- range experienced along the P-T path. Exposure to higher tem- tory (dP/dT = 32 bars/°C) and a total pressure change of ~0.9 peratures resulted in narrow (<300 µm) zones of prograde kbar ending at ~635 °C. reequilibration along garnet rims, cracks, and inclusion (quartz Steep or isothermal P-T paths are a logical consequence of and ilmenite) margins, owing to diffusional exchange with the thrusting; rocks in the footwalls of thrusts spontaneously in- matrix biotite. The biotite composition remained largely un- crease in pressure as sheets of rock are piled on top (e.g., En- changed during the higher temperatures because the reservoir gland and Thompson 1984; Spear et al. 1984). The P-T of Fe and Mg available for exchange in the garnet was extremely trajectories rule out the possibility that heat from the 27 Ma small in comparison to the biotite, resulting in the preservation Vipont intrusion, which is exposed mainly several kilometers of primary garnet-biotite geothermometry temperatures. to the north of the sampled area, could have been responsible Garnet in the lower horizon grew in rocks of a more Al-rich for the metamorphism. Contact metamorphism should produce and Ca-poor bulk composition than the upper horizon. Garnet- isobaric P-T paths (e.g., Spear et al. 1984). Although the actual biotite geothermometry temperatures of ~635 °C are consis- placement of the paths with respect to pressure carries large tent with garnet growth as a result of the staurolite breakdown uncertainties, the paths themselves are quite robust (Kohn 1993) reaction, and with textural relations that suggest that garnet and the temperatures are well determined by garnet-biotite growth took place after paragonite broke down. Garnet zoning geothermometry. Because of the ~150 °C difference in the tem- profiles are smooth and simple, suggesting only a single stage perature of garnet growth between horizons, it is not possible of growth. Although garnet may have been present at an earlier to reconcile both paths to a single thrusting episode, which lower-grade stage in the history of the rock, there is no evi- would require the paths to connect in a single line without in- dence that any of the garnet currently in the lower-horizon flection, retracement, or discontinuity. samples survives from an earlier metamorphism. Such evidence Steady-state geotherms will not pass through the silliman- might include compositional discontinuities between rims and ite stability field given the normal range of physical constants cores, or two populations of garnet compositions or morpholo- for the continental crust (e.g., England and Thompson 1984). gies, but none of these are observed. Along the cooling path, In order for rocks to enter into the sillimanite stability field, garnets underwent retrograde diffusive reequilibration along heat must either be advectively transferred into the area by melts rims and cracks. or the area must be rapidly exhumed. Because there are no Rates of diffusion in garnet for Mn, Mg, and Fe become Mesozoic plutons in the region, the simplest interpretation of geologically important at temperatures above ~600 °C (e.g., the two paths is that of two thrusting episodes separated by a HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 129

a b 7.0 7.0 Upper Horizon rims 6.5 UH2B-F2 6.5 UH2B-F0 UH2A-F1 6.0 UH2A-F0 6.0 UH1

5.5 5.5 Exhumation

Thrusting 5.0

(kbar) 5.0

P cores rims Lower Horizon 4.5 T uncertainty LH1B 4.5 for upper horizon LH1A models 4.0 Ky 4.0 Ky Thrusting cores Sil Sil 3.5 And 3.5 And P-T uncertainty for LH1A garnet rim from Figure 7 3.0 3.0 450475 500 525 550 575 600 625 650 450475 500 525 550 575 600 625 650 T (°C) T (°C) FIGURE 8. P-T paths. The kyanite (Ky)-sillimanite (Sil)-andalusite (And) reactions were calculated using data from Berman (1988). (a) Paths determined from simulations of garnet growth zoning using the Gibbs method based on Duhem’s theorem (e.g., Spear 1993). The values of all parameters used in the simulations are given in Tables 1 and 3. Shaded fields denote the uncertainties as indicated. Upper horizon paths possess a minimum uncertainty of ±25 °C inherited from garnet-biotite geothermometry and are constrained to the kyanite field by the presence of kyanite in rocks 2 km to the north. The uncertainty in the lower horizon path is from Figure 5. (b) Inferred P-T path and tectonic interpretation. period of exhumation (Fig. 8). After an episode of thrusting, REFERENCES CITED rocks at depth should undergo heating associated with thermal Armstrong, R.L. (1968) Mantled gneiss domes in the Albion Range, southern Idaho. relaxation (e.g., England and Thompson 1984). Consequently, Geological Society of America Bulletin, 79, 1295Ð1314. Armstrong, R.L. and Hills, F.A. (1967) Rb-Sr and K-Ar geochronologic studies of heating is inferred to follow the isothermal pressure increase mantled gneiss domes, Albion Range, southern Idaho, U.S.A. Earth and Plan- associated with the first thrusting episode. The inferred pres- etary Science Letters, 3, 114Ð124. sure drop between the two episodes of thrusting was sufficiently Berman, R.G. (1988) Internally consistent thermodynamic data for minerals in the system Na2O-K2O-CaO-MgO-FeO-Fe2O3-Al2O3-SiO2-TiO2-H2O-CO2. Journal rapid to bring the rocks into the sillimanite stability field (Fig. of Petrology, 29, 445Ð522. 8) suggesting a high rate of exhumation consistent with tec- Berman, R.G. and Aranovich, L. Ya. (1996) Optimized standard state and solution tonic denudation. properties of minerals. I. Model calibration for olivine, orthopyroxene, cordier- ite, garnet, and ilmenite in the system FeO-MgO-CaO-Al2O3-TiO2-SiO2. Con- The deformation history of the Raft River Mountains dur- tributions to Mineralogy and Petrology, 126, 1Ð24. ing Sevier orogenesis has been previously interpreted to record Box, G.E.P. and Muller, M.E. (1958) A note on the generation of random number deviates. Annals of Mathematical Statistics, 29, 610Ð611. crustal thickening by thrusting and folding and exhumation by Camilleri, P.A. and Chamberlain, K.R. (1997) Mesozoic tectonics and metamor- plastic flow and normal faulting, in which tectonic contraction phism in the Pequop Mountains and Wood Hills region, northeast Nevada: Im- and extension alternated in late Mesozoic to early Cenozoic plications for the architecture and evolution of the Sevier orogen. Geological Society of America Bulletin, 109, 74Ð94. time (Wells 1997). The P-T paths presented here provide inde- Chackraborty, S. and Ganguly, J. (1990) Compositional zoning and cation diffusion pendent support of this interpretation, which was based on in garnets. In J. Ganguly, Ed., Diffusion, Atomic Ordering, and Mass Trans- mapped geologic structures and rock fabrics from the eastern port—Selected Problems in Geochemistry, p. 120Ð175. Springer, New York. Compton, R.R. (1972) Geologic map of Yost quadrangle, Box Elder County, Utah, part of the Raft River Mountains. The exhumation event docu- and Cassia County, Idaho. U.S. Geological Survey Miscellaneous Investiga- mented here cannot be correlated to mapped geologic struc- tions Map I-672. ———(1975) Geologic map of Park Valley quadrangle, Box Elder County, Utah, tures without dating garnet growth. However, the Emigrant and Cassia County, Idaho, U.S. Geological Survey Miscellaneous Investiga- Spring and Mahogany Peaks faults (Fig. 2b) are two likely can- tions Map I-873. didates for structures that may have caused the inferred ———(1983) Displaced Miocene rocks on the west flank of the Raft River-Grouse Creek core complex, Utah. In D.M. Miller, V.R. Todd, and K.A. Howard, Eds., synorogenic decompression. Tectonic and Stratigraphic Studies in the Eastern Great Basin, p. 271Ð279. Geo- logical Society of America Memoir 157. ACKNOWLEDGMENTS Compton, R.R. and Todd, V.R. (1979) Oligocene and Miocene metamorphism, fold- ing, and low-angle faulting in northwestern Utah, Reply. Geological Society of We thank James Wittke (NAU), James Clark (ASU), and David Wark (RPI) America Bulletin, 90, 307Ð309. for assistance with various aspects of microprobe data collection. We thank Compton, R.R., Todd, V.R., Zartman, R.E., and Naeser, C.W. (1977) Oligocene and Michael Holdaway, Helen Lang, and Robert Tracy for helpful and thoughtful Miocene metamorphism, folding, and low-angle faulting in northwestern Utah. reviews. This study was funded by the National Science Foundation through Geological Society of America Bulletin, 88, 1237Ð1250. grants EAR-9317044 and EAR-9805076 to T.D.H. and grants EAR-9317387 England, P.E. and Thompson, A.B. (1984) Pressure-temperature-time paths of re- and EAR-9805007 to M.L.W. gional metamorphism, I. Heat transfer during the evolution of regions of thick- 130 HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT

ened crust. Journal of Petrology, 97, 894Ð928. phism in overthrust terrains with special reference to the eastern Alps. Florence, F.P. and Spear, F.S. (1993) Influences of reaction history and chemical Schweizerische Mineralogische und Petrographische Mitteilungen, 54, 641Ð diffusion on P-T calculations for staurolite schists from the Littleton Forma- 662. tion, northwestern New Hampshire. American Mineralogist, 78, 345Ð359. Rumble, D. III, (1973) Fe-Ti oxide minerals from regionally metamorphosed quartz- Forrest, S.D. and Miller, E.L. (1994) Oligocene plutonism and associated crustal ites of eastern New Hampshire: Contributions to Mineralogy and Petrology, 42, thinning in the southern Albion Mountains, Idaho. Geological Society of America 181Ð195. Abstracts with Programs, 26, A192. Saltzer, S.D. and Hodges, K.V. (1988) The Middle Mountain shear zone, southern Fuhrman, M.L. and Lindsley, D.H. (1988) Ternary-feldspar modeling and thermom- Idaho: Kinematic analysis of an early Tertiary high-temperature detachment. etry. American Mineralogist, 73, 201Ð215. Geological Society of America Bulletin, 100, 96Ð103.

Grambling, J.A. (1990) Internally consistent geothermometry and H2O barometry Selverstone, J. and Spear, F.S. (1985) Metamorphic P-T paths from pelitic schists in metamorphic rocks: The example of garnet-chlorite-quartz. Contributions to and greenstones in the southwest Tauern Window, Eastern Alps. Journal of Mineralogy and Petrology, 105, 617Ð628. Metamorphic Geology, 3, 439Ð465. Guidotti, C.V. and Dyar, M.D. (1991) Ferric iron in metamorphic biotite and its Spear, F.S. (1989) Petrologic determination of metamorphic pressure-temperature- petrologic and crystallographic implications, 76, 161Ð175. time paths. In Spear, F.S. and Peacock, S.M., Eds., Metamorphic Pressure-Tem- Hanson, M.L. (1997) Metamorphic petrology of pelitic schist from the Raft River perature-Time Paths, Short Course in Geology: volume 7, 102 p., American and Grouse Creek Mountains, northwest Utah. M.S. thesis, Northern Arizona Geophysical Union, Washington, D.C. University, Flagstaff. ———(1993) Metamorphic phase equilibria and pressure-temperature-time paths. Hodges, K.V. and Royden, L. (1984) Geologic thermobarometry of retrograded Monograph 1, 799 p., Mineralogical Society of America, Washington D.C. metamorphic rocks: an indication of the uplift trajectory of a portion of the Spear, F.S. and Cheney, J.T. (1989) A petrogenetic grid for pelitic schists in the

northern Scandinavian Caledonides. Journal of Geophysical Research, 89, 7077Ð system SiO2-Al2O3-FeO-MgO-K2O-H2O. Contributions to Mineralogy and Pe- 7090. trology, 101, 149Ð164.

Hodges, K.V. and Spear, F.S. (1982) Geothermometry, geobarometry, and the Al2SiO5 Spear, F.S. and Selverstone, J. (1983) Quantitative P-T paths from zoned minerals: triple point at Mt. Moosilauke, New Hampshire. American Mineralogist, 67, theory and tectonic applications. Contributions to Mineralogy and Petrology, 1118Ð1134. 83, 348Ð357. Hodges, K.V. and Walker, J.D. (1992) Extension in the Cretaceous Sevier orogen, Spear, F.S., Selverstone, J., Hickmott, D., Crowley, P., and Hodges, K.V. (1984) P- North American Cordillera. Geological Society of America Bulletin, 104, 560Ð T paths from garnet zoning—A new technique for deciphering tectonic pro- 569. cesses in crystalline terrains. Geology, 12, 87Ð90. Hodges, K.V., Parrish, RR., Housh, T.B., Lux, D.R., Burchfiel, B.C., Royden, L.H., Spear, F.S., Hickmott, D.D., and Selverstone, J. (1990) Metamorphic consequences and Chen, Z. (1992) Simultaneous Miocene extension and shortening in the of thrust emplacement, Fall Mountain, New Hampshire. Geological Society of Himalayan Orogen. Science, 258, 1466Ð1470. America Bulletin, 102, 1344Ð1360. Hoisch, T.D. (1991) Equilibria within the mineral assemblage quartz + muscovite + Spear, F.S., Peacock, S.M., Kohn, M.J., Florence, F.P., and Menard, T. (1991) Com- biotite + garnet + plagioclase. Contributions to Mineralogy and Petrology, 108, puter programs for petrologic P-T-t path calculations. American Mineralogist, 43Ð54. 76, 2009Ð2012. Hoisch, T.D. and Simpson, C. (1993) Rise and tilt of metamorphic rocks in the St-Onge, M.R. (1987) Zoned poikiloblastic garnets: P-T paths and syn-metamor- lower plate of a detachment fault in the Funeral Mountains, Death Valley, Cali- phic uplift through 30 km of structural depth, Wopmay Orogen, Canada. Jour- fornia. Journal of Geophysical Research, 98, 6805Ð6827. nal of Petrology, 28, 1Ð22. Holdaway, M.J. (2000) Application of new experimental and garnet Margules data Thost, D.E., Motoyoshi, Y., and Henson, B.J. (1991) Two-stage decompression in to the garnet-biotite geothermometer. American Mineralogist, 85: 881Ð892. garnet-bearing mafic granulites from Sostrene Island, Prydz Bay, East Antarc- Holdaway, M.J., Mukhopadhyay, B., Dyar, M.D., Guidotti, C.V., and Dutrow, B.L. tica. Journal of Metamorphic Geology, 9, 245Ð256. (1997) Garnet-biotite geothermometry revised: New Margules parameters and Todd, C.S. (1998) Limits on the precision of geobarometry at low grossular and a natural specimen data set from Maine. American Mineralogist, 82, 582Ð595. anorthite content. American Mineralogist, 83, 1161Ð1167. Hurlow, H.A., Snoke, A.W., and Hodges, K.V. (1991) Temperature and pressure of Todd, V.R. (1980) Structure and petrology of a Tertiary gneiss complex in north- mylonitization in a Tertiary extensional shear zone, Ruby Mountains-East western Utah. In Crittenden, M.D. Jr., Coney, P.J. and Davis, G.H., Eds., Cor- Humboldt Range, Nevada: Tectonic implications. Geology, 19, 82Ð86. dilleran Metamorphic Core Complexes, p. 349Ð383, Geological Society of Kohn, M.J. (1993) Uncertainties in differential thermodynamic (Gibbs’ method) P- America Memoir 153. T paths. Contributions to Mineralogy and Petrology, 113, 24Ð39. Tracy, R.J. (1982) Compositional zoning and inclusions in metamorphic minerals. Lang, H.M. (1996) Pressure-temperature-reaction history of metapelitic rocks from In J.M. Ferry, Ed., Characterization of metamorphism through mineral equilib- the Maryland Piedmont on the basis of correlated garnet zoning and plagio- ria, 10, 355Ð397. Reviews in Mineralogy, Mineralogical Society of America, clase-inclusion composition. American Mineralogist, 18, 1460Ð1475. Washington, D.C. Lewis, C.J., Wernicke, B.P., Selverstone, J., and Bartley, J.M. (1999) Deep burial of Wells, M.L. (1997) Alternating contraction and extension in the hinterlands of oro- the footwall of the northern Snake Range decollement, Nevada. Geological genic belts: An example from the Raft River Mountains, Utah. Geological So- Society of America Bulletin, 111, 39Ð51. ciety of America Bulletin, 109, 107Ð126. Mahlen, N.J., Hoisch, T.D., and Dufrane, S.A. (2000) Use of volumetric titration to Wells, M.L., Dallmeyer, R.D., and Allmendinger, R.W. (1990) Late Cretaceous ex- determine ferrous iron content in biotite from pelitic schist from the Grouse tension in the hinterland of the Sevier thrust belt, northwestern Utah and south- Creek Mountains, Utah: Geological Society of America Abstracts with Pro- ern Idaho. Geology, 18, 929Ð933. grams, 32, A365. Wells, M.L., Hoisch, T.D. Hanson, L.M., Struthers, J., and Wolfe, E.D. (1997a) Malavieille, J. (1987) Extensional shearing and kilometer-scale “a” type folds in a Large magnitude crustal thickening and repeated extensional exhumation in Cordilleran metamorphic core complex (Raft River Mountains, northwestern the Raft River, Grouse Creek, and Albion Mountains. Brigham Young Univer- Utah). Tectonics, 6, 423Ð448. sity Geological Studies, 42, pt. 1, 325Ð340. McGrew, A.J., Peters, M.T., and Wright, J.E. (2000) Thermobarometric constraints Wells, M.L., Struthers, J.S., Snee, L.W., Walker, J.D., Blythe, A.E., and Miller, D.M. on the tectonothermal evolution of the East Humbolt Range metamorphic core (1997b) Miocene extensional reactivation of an Eocene extensional shear zone, complex, Nevada. Geological Society of America Bulletin, 112, 45Ð60. Grouse Creek Mountains, Utah. Geological Society of America Abstracts with Miller, D.M. (1980) Structural geology of the northern Albion Mountains, south- Programs, 29, A162. central Idaho. In M.D. Crittenden Jr., P.J. Coney, and G.H. Davis, Eds., Cordil- Wells, M.L., Hoisch, T.D., Peters, M.T., Miller, D.M., Wolff, E.D., and Hanson, leran Metamorphic Core Complexes, p. 399Ð423. Geological Society of America L.M. (1998) The Mahogany Peaks fault, a Late Cretaceous-Paleocene normal Memoir 153. fault in the hinterland of the Sevier Orogen. Journal of Geology, 106, 623Ð634. Miller, D.M. and Bedford, D.R. (1999) Pluton intrusion styles, roof subsidence and Wells, M.L., Snee, L.W., and Blythe, A.E. (2000) Dating of major normal fault stoping, and timing of extensional shear zones in the City of Rocks National systems using thermochronology: An example from the Raft River detachment, Reserve, Albion Mountains, southern Idaho. In L.E. Spangler and C.J. Allen, Basin and Range, western United States. Journal of Geophysical Research, 105, Eds., Geology of Northern Utah and Vicinity, p. 11Ð26. Utah Geological Asso- 16,303Ð16,327. ciation Publication 27. Whitney, D.L., Mechum, T.A., Dilek, Y., and Kuehner, S.M. (1996) Modification of Miller, E.L. and Gans, P.B. (1989) Cretaceous structure and metamorphism in the garnet by fluid infiltration during regional metamorphism in garnet through hinterland of the Sevier thrust belt, western U.S. Cordillera. Geology, 17, 59Ð62. sillimanite-zone rocks, Dutchess County, New York. American Mineralogist, Miller, D.M., Armstrong, R.L., Compton, R.R., and Todd, V.R. (1983) Geology of 81, 696Ð705. the Albion-Raft River-Grouse Creek Mountains area, northwestern Utah and Whitney, D.L., Miller, R.B., and Paterson, S.R. (1999) P-T-t evidence for mecha- southern Idaho. Utah Geological and Mineral Survey Special Studies, 59, 1Ð nisms of vertical tectonic motion in a contractional orogen: north-western US 59. and Canadian Cordillera. Journal of Metamorphic Geology, 17, 75Ð90. Mukhopadhyay, B., Holdaway, M.J., and Koziol, A.M. (1997) A statistical model of thermodynamic mixing properties of Ca-Mg-Fe2+ garnets. American Mineralo- MANUSCRIPT RECEIVED AUGUST 23, 2000 gist, 82, 165Ð181. MANUSCRIPT ACCEPTED SEPTEMBER 18, 2001 Oxburgh, E.R. and Turcotte D.L. (1974) Thermal gradients and regional metamor- MANUSCRIPT HANDLED BY ROBERT J. TRACY HOISCH ET AL.: P-T PATHS FROM GARNETS, SEVIER OROGENIC BELT 131

APPENDIX 1. METHOD OF CALCULATING THE STAUROLITE-BREAKDOWN FIELD SHOWN ON FIGURE 5 The first step in calculating the staurolite-breakdown reaction field was to determine the offset due to the Mn and Ca compo- nents of garnet in sample LH1A. Using the mineral compositions in Table 1 (LH1A), starting conditions of 600 °C at 5300 bars, and ideal activity models for all solid solutions, the Gibbs method (program Gibbs4.7; Spear et al. 1991 described an earlier version) without mass-balance constraints calculated an offset of 22.8 °C from the Mn-Ca-free system. To calculate the reaction involving the garnet core composition for LH1A, the starting conditions were determined by reading one point from the KFMASH reaction as shown by Spear and Cheney (1989), 657 °C at 5300 bars, then subtracting the temperature displacement to give 634.2 °C at 5300 bars. The full reaction involving the garnet core composition (left edge of the staurolite-breakdown reaction field on Fig. 6a) was then calculated using the Gibbs method without mass balance constraints; the system has a phase rule variance of three, Xgr and Xsp were held constant while specifying values for ∆P. The reaction involving the garnet rim was calculated in an analogous way, starting with the garnet core reaction at conditions of 634.2 °C at 5300 bars. The displacement due to the change in garnet composition from rim to core (∆Xsp = Ð0.0735, ∆Xgr = 0.0073) was calculated at constant pressure, then the full reaction was calculated by holding Xgr and Xsp constant while specifying values for ∆P as before. To determine the KFMASH (Mn-Ca- free) reaction (right edge of the reaction field on Fig. 6a), compositions from Table 1 (initial values) for LH1A were used together with starting conditions of 634.2 °C at 5300 bars. The Gibbs method was used to displace the reaction at constant pressure associated with a shift in garnet composition from that of the garnet core to a Mn-Ca-free composition (Xsp and Xgr both equal to 0.0001). Then the Gibbs method was used to calculate the full KFMASH reaction as before. The results duplicate the reaction shown by Spear and Cheney (1989).