Construction of a composite pressureÐtemperatur e path: revealing...

Construction of a composite pressure–temperature path: revealingthe synorogenic burial and exhumation history of the Sevierhinterland, USAC. R. HARRIS,1 , 2 T . D. HOISCH1 AND M. L. WELLS31Department of Geology, Northern Arizona University, PO Box 4099, Flagstaff, AZ 86011, USA ([email protected])2Isotope Geochemistry and Mineral Resources, ETH Zentrum NW, 8092 Zürich, Switzerland3Department of Geosciences, University of Nevada, Las Vegas, Las Vegas, NV 89154, USA

ABSTRACT Metamorphic pressure–temperature (P–T) paths derived from 16 growth-zoned garnets, nine from thisstudy and seven from a previous study, have been combined to construct a detailed composite path foran area in the hinterland of the Cretaceous to early Tertiary Sevier orogenic belt in southern Idaho andnorth-west Utah. Samples are from two Proterozoic units in the footwall of the Basin-Elba thrust: theschist of Mahogany Peaks in the central Albion Mountains, Idaho, and the schist of Stevens Spring inthe Basin Creek area of the Grouse Creek Mountains, Utah, !40 km to the south. The simulatedportions of the garnets analysed in this study grew from reactions involving the breakdown of chlorite inthe upper greenschist to lower amphibolite facies. Multiple garnets were analysed from three samples.Overlapping segments of P–T paths from different garnets in the same sample correlate with respect toslope and garnet Mn concentration. The composite P–T path documents three episodes of sharplyincreasing pressures separated by two episodes of pressure decrease, all during progressively increasingtemperatures. The path is interpreted to represent alternating episodes of synconvergent thrusting andextensional exhumation in the hinterland of the Sevier orogen. Burial was probably caused by the Basin-Elba fault, the only major thrust exposed in the region. Extensional exhumation may have occurredalong the Mahogany Peaks or Emigrant Spring faults, or by extensional reactivation of the Basin-Elbafault. This method of correlating partial P–T paths to reveal a more complete composite path provides apowerful tool in unveiling orogenic histories in metamorphic terranes, where evidence of majorstructures responsible for burial and exhumation is commonly obscured by later events.

Key words: garnet; Gibbs method; P–T path; Sevier; Utah.

INTRODUCTION

Metamorphic rocks preserved in ancient mountainbelts represent one of the primary and potentiallyenduring records of past orogenesis. The cycle of oro-genic activity experienced by mountain belts commonlyculminates with profound extension and magmatism(Dewey, 1988) which can obscure earlier structures andfabrics formed during shortening. Consequently, thereconstruction of changes in pressure and temperaturefrom metamorphic rocks provides a view of pasttectonism not always apparent from other geologicalobservations (e.g. Selverstone, 1985; Spear et al., 1990).Tectonic events including subduction, continental col-lision, extension, lower lithospheric delamination andarc magmatism impart characteristic changes in pres-sure and temperature to rocks during metamorphism;these changes in pressure and temperature (P–T paths)can be linked to tectonic processes through increasinglysophisticated thermal modelling (Oxburgh & Turcotte,1974; England & Thompson, 1984; Ruppel et al., 1988;Grasemann & Mancktelow, 1993; Platt & England,1994; Ruppel & Hodges, 1994; Ketcham, 1996;

Jamieson et al., 1998, 2002; Hoisch, 2005). Despitetheir great utility, the reconstruction of complete P–Tpaths remains one of the foremost challenges in meta-morphic petrology (Spear, 1993).

The construction of complete P–T paths is hinderedby a number of factors including: (i) lack of devel-opment or preservation of mineral assemblagesappropriate for analysing P–T paths during all seg-ments of the path, (ii) erasure of earlier historiesthrough partial or complete diffusive re-equilibrationif temperatures exceed thresholds for cation diffusion,and (iii) difficulty in correlating P–T paths from rocksrecording different or overlapping segments of pathsfrom the same locality. Although no single methodexists to accurately determine all parts of a P–T pathin a given area, from the earliest burial to the finalexhumation, methods which utilize growth zoning ingarnet, such as the Gibbs method based on Duhem’stheorem (e.g. Spear, 1988), offer the greatest potentialto record the most complete P–T path. Metamorphicgarnet will record the portions of the prograde paththrough which they grew, as long as temperatures donot exceed thresholds for cation volume diffusion

J. metamorphic Geol., 2007 doi:10.1111/j.1525-1314.2007.00733.x

! 2007 Blackwell Publishing Ltd 1

(600–650 "C, depending on cooling rate, e.g. Spear,1989).

Orogenic P–T paths have classically been thoughtto have the form of a simple clockwise loop,reflecting a basic concept of orogenesis starting withprotracted shortening, followed by thermal relaxa-tion, and terminating with exhumation by combinederosion and extension (e.g. England & Thompson,1984). In parallel with the recognition that plateboundary forces, gravitational potential energy, rockrheology and erosion dynamically interact to influ-ence deformation kinematics (Davis et al., 1983;Platt, 1986; Molnar & Lyon-Caen, 1988; England &Houseman, 1989; Royden, 1993; Rey et al., 2001),orogenic histories involving synconvergent extensionand alternations in shortening and extension havebeen documented (Hodges et al., 1992b; Wallis et al.,1993; Wells, 1997; Ferranti & Oldow, 1999; Collins,2002). Despite the theoretical predictions and fieldexamples of alternating shortening and extension,P–T paths confirming such alternations are notwidely reported (e.g. Hoisch et al., 2002; Johnson &Brown, 2004).

In this study we present an unusually completecomposite metamorphic P–T path determined fromexhumed mid-crustal pelitic schists from the Sevierorogenic belt of the western United States. The garnetsanalysed in this study nucleated at different times andthus captured different segments of the path. The P–Tpaths recorded in individual garnets overlap and werecorrelated to construct a composite P–T path. Thecomposite path is much more complex than the simpleP–T loops commonly assumed for tectonic burial anduplift, and is interpreted to represent alternating epi-sodes of synconvergent thrusting and extensionalexhumation.

REGIONAL TECTONIC HISTORY

The metamorphic rocks in this study are from thehinterland of the Late Mesozoic to Early CenozoicSevier orogen of the western United States. The Sevierorogen represents a classic retroarc orogenic wedge,broadly similar in gross architecture to the modernAndean orogen (Allmendinger, 1992; Burchfiel et al.,1992; Camilleri et al., 1997; DeCelles, 2004). TheSevier orogen is subdivided into an eastern forelandfold-thrust belt and a western hinterland region thatincludes the isolated exposures of greenschist andamphibolite facies Barrovian metamorphic rocks ofthe Cordilleran metamorphic core complexes (Crit-tenden et al., 1980). Initial shortening in the Sevierorogen progressed eastward from hinterland to fore-land from Middle Jurassic to Early Eocene time(Armstrong & Oriel, 1965; Wiltschko & Dorr, 1983;Elison, 1991; Allmendinger, 1992; Taylor et al., 2000;DeCelles, 2004). The progressive development of theforeland fold-thrust belt from Early Cretaceous toEarly Eocene is well understood, with an eastward

migrating thrust front punctuated by local minor out-of-sequence deformation internal to the orogenicwedge (DeCelles, 1994; Lawton et al., 1997; Yonkeeet al., 1997; DeCelles, 2004).The kinematic history of the hinterland is less clear

and includes a sparse record of Mid-Late Jurassicshortening (Allmendinger et al., 1984; Elison, 1991;Miller & Allmendinger, 1991; Hudec, 1992) as well asLate Cretaceous shortening and extension (Milleret al., 1988; Hodges & Walker, 1992; Camilleri &Chamberlain, 1997; Wells, 1997; Wells et al., 1998;McGrew et al., 2000). Original stratigraphic juxtapo-sitions and deformation fabrics recording early short-ening are commonly obscured by reactivation andoverprinting by normal faults and extensional flowfabrics (Camilleri & Chamberlain, 1997; Wells, 1997;Lewis et al., 1999). The metamorphism provides thestrongest evidence for large magnitude shortening inthe hinterland of the orogen as: (i) systematic lateralgradients in metamorphic pressure in stratigraphicallyequivalent and contiguous metamorphic rocks provideevidence for lateral gradients in thrust loading (Hoisch& Simpson, 1993; Camilleri & Chamberlain, 1997;Lewis et al., 1999); (ii) Barrovian metamorphism ofmiogeoclinal metasedimentary rocks to pressures ashigh as 8–10 kbar indicates structural burial of twoto three times that of stratigraphic depth (e.g. Hod-ges et al., 1992a; McGrew et al., 2000; Hoisch et al.,2002).

GEOLOGY OF THE RAFT RIVER-ALBION-GROUSECREEK MOUNTAINS

In the Raft River, Albion and Grouse Creek moun-tains of north-west Utah and southern Idaho (Fig. 1),a succession of tectonically-thinned greenschist toupper amphibolite facies Mississippian to Proterozoicmetasedimentary rocks, termed the Raft RiverMountains sequence by Miller (1983), unconformablyoverlies Archean basement (Armstrong, 1968; Comp-ton et al., 1977; Wells, 1997). This succession occursover an area greater than 4000 km2 in all parts of theregion except in the hangingwall of the Basin-Elbathrust fault in the northern Albion Mountains (Miller,1980, 1983), which is composed of a different sequenceof rocks referred to by Miller (1983) as the !quartziteassemblage", a >3500 m thick succession of over-turned Late Cambrian and Neoproterozoic quartziteand schist (Mt. Harrison sequence) overlaintectonically by a sequence of upright quartzite andschist of inferred Late Cambrian and Neoproterozoicage (Robinson Creek sequence). The Basin-Elba faultis apparently the only preserved thrust fault in theregion as it places older strata over younger (Figs 1& 2).Exhumation of the metamorphic rocks took place

during episodic Cenozoic extension along both westand east vergent detachment fault systems (Comptonet al., 1977; Malavieille, 1987; Wells et al., 2000b;

2 C . R . HARR I S ET AL .

! 2007 Blackwell Publishing Ltd

Fig. 1. Generalized geological map of the Raft River-Albion-Grouse Creek Mountains, showing location of the Basin-Elba fault,Basin Creek area, and approximate sample locations. Bottom right inset shows location of metamorphic core complexes (solid areas) inthe hinterland of the Sevier thrust belt and in the southern Basin and Range. The thick barbed line represents the leading edge of thethrust belt. RCS, Robinson Creek sequence; MHS, Mount Harrison sequence (Miller, 1983). Modified from Hoisch et al. (2002).

P–T PATH FROM THE SEV I ER H INTERLAND 3


Egger et al., 2003); synconvergent Mesozoic extensionwas also important (e.g. Wells et al., 1990, 1998;Hodges & Walker, 1992). The oldest and most widelydistributed extensional fabric (Table 1) is a flat-lyingfoliation and generally N-trending stretching lineationinterpreted to have developed during orogen-parallelextensional flow at 105 ± 6 Ma (revised from Wellset al., 2000a). This fabric affects Archean to Permianrocks and accomplished variable but significant thin-ning of rock units. Additionally, low-angle, generallybedding-parallel, top-to-west faults, including theMahogany Peaks and the Emigrant Spring faults(Fig. 2c), accomplished significant attenuation of therock units and are interpreted as extensional faults(Table 1) (Wells, 1997; Wells et al., 1998). TheMahogany Peaks fault juxtaposes Ordovician marbleover the Proterozoic schist of Mahogany Peaks with anestimated excision of 4–5 km, and its age is bracketedby 40Ar/39Ar muscovite cooling ages of c. 90 and60 Ma from the hangingwall and footwall respectively(Wells et al., 1998). The Emigrant Spring fault (Wells,1997) places Pennsylvanian(?) calcitic marble overOrdovician and Silurian(?) dolomitic marble and re-moves !5 km of stratigraphic section. Motion on theEmigrant Spring fault has been previously interpretedas prior to or synchronous with cooling recorded by an40Ar/39Ar muscovite age of c. 89–88 Ma from agreenschist facies calcitic marble mylonite (Wells et al.,1990; Wells, 1997); however, the cooling age may pre-date deformation as the temperatures of deformationmay have been less than argon closure in muscovite.Kilometre-scale recumbent folds deform the EmigrantSpring and Mahogany Peaks faults, leading to aninterpretation of renewed contraction followingextension (Wells, 1997; D3 in Table 1). These recum-bent folds are overprinted by an Eocene–Oligoceneextensional shear zone and detachment fault (D4 inTable 1).

The P–T paths generated in this study are from twogarnet-bearing units of Proterozoic schist in the foot-wall of the Basin-Elba fault. The schist of MahoganyPeaks was sampled close to the Basin-Elba fault in the

Albion Mountains (Figs 1 & 2a), and the !upperhorizon" of the schist of Stevens Spring was sampledfrom the Basin Creek area of the Grouse CreekMountains, !40 km to the south (Figs 1 & 2b). Asthese two localities lie within a contiguous exposure ofthe Raft River Mountains sequence and no beddingparallel faults are recognized between them, the twounits are considered to have shared a similar P–Thistory.Hoisch et al. (2002) reported P–T paths from the

Basin Creek area from two different horizons of theschist of Stevens Spring, a !lower horizon" and an!upper horizon". The upper horizon garnets grew at!475 "C and records one episode of approximatelyisothermal pressure increase, interpreted to representrapid burial associated with thrusting. Garnet in thelower horizon grew at higher temperatures (600–630 "C) from a different reaction, and record a secondepisode of burial following an episode of majordecompression (Hoisch et al., 2002). In this study, fivenew P–T paths are presented from two localities of theschist of Mahogany Peaks, and four new P–T pathsare presented from the upper horizon of the schist ofStevens Spring. These are integrated with the sevenpaths previously determined from the schist of StevensSpring (five from the upper horizon and two from thelower horizon) (Hoisch et al., 2002) (Fig. 2).

ANALYTICAL METHODS

Mineral compositions were determined using theCameca MBX electron microprobe at NorthernArizona University. A spot size of 1 lm and a beamcurrent of 25 nA was used for garnet traverses. A 5 lmspot size and a beam current of 10 nA was used toacquire matrix mineral data. The accelerating voltagein both cases was 15 kV. Line traverses across garnetwere run with a point spacing of 10–20 lm (Fig. 3).The garnet traverse data were processed to exclude anyanalysis that failed to produce a good stoichiometry,within certain tolerances. Generally, excluded analysesrepresented inclusions or points that overlapped

Table 1. Sequence of Mesozoic to early Cenozoic deformation proposed in this and previous studies (see text) for the footwall to theBasin-Elba thrust fault, western Raft River, northern Grouse Creek and Albion Mountains.

Tectonic event Structure and kinematics Interpretation Timing Metamorphism

M1 First motion on Basin-Elba fault Contraction Pre-105 Ma,

Jurassic or early

Cretaceous

Metamorphism of Raft River assemblage;

isothermal P increasea,b

D1 Top-to-north shearingd Orogen-parallel extensionc 105 ± 12 Mac

D2 Emigrant Spring and Mahogany Peaks Fault Extensione,f c. 90–60 Maf Inferred decompression P–T pathb

D3 Recumbent folding and second motion along Basin-Elba fault Contractione c. 60–45 Mag P increaseb

D4A Top-to-WNW shearing along amphibolite facies MMSZ Extension Late Eoceneh

aThis study.bHoisch et al. (2002).cWells et al. (in press).dMalavieille (1987).eWells (1997).fWells et al. (1998).gRevised from Hoisch & Wells (2004).hWells et al. (2000b).



(a)

(b)

Emigrant Spring Fault

Middle Detachment Fault

(c)

Chainman Shale - DiamondPeak Formations

Mcd

41°42'30"

41°55'

113°47'30"

113°45'UH

Mount Harrisonsequence

113°45' 113°37'30"113°42'30" 113°40'

42°15'

42°17'30"

42°17'30"

Fig. 2. Geological maps of the sampled areas(see Fig. 1) and regional stratigraphy. (a) Map ofthe northern Albion Mountains (based on Miller,1983 and R.L. Armstrong, unpublished map-ping) and locations for sample sites THAL4 andTHAL6. (b) Map of the Basin Creek area of thenorthern Grouse Creek Mountains (modifiedfrom Hoisch et al., 2002) and locations for sam-ple sites UH and LH. All map units are keyed tothe tectonostratigraphic column (c), with theexception of Oligocene granite (black fill).



inclusions. Multiple analyses of each matrix phasewere performed in each sample and averaged(Table S1). Garnet element maps for Ca, Mg, Fe andMn were generated by counting with wave dispersivespectrometers on a JEOL-733 electron microprobeat Rensselaer Polytechnic Institute and by countingwith an energy dispersive system on a JEOL JSM-6480LV scanning electron microscope at NorthernArizona University (Fig. 4). Images were processedwith NIH IMAGENIH IMAGE or IMAGE JIMAGE J to produce maximumcontrast.

NUMERICAL SIMULATION OF GARNET GROWTHUSING THE GIBBS METHOD WITH DUHEM’STHEOREM

Pressure–temperature paths were determined fromgarnet growth zoning by numerically simulating garnetgrowth based on the Gibbs method with Duhem’stheorem (Spear, 1988, 1993). Garnet in this study dis-play bell-shaped Mn profiles, consistent with Rayleighfractionation during equilibrium growth (e.g. Hollister,1966). Zoning is generally concentric except for post-growth modifications (Figs 3 & 4). The zoning profilesrange from highly symmetric to slightly asymmetric.

Garnet growth is accompanied by changes in inten-sive variables (composition of solid solution phases,pressure and temperature) and extensive variables(amounts of phases). According to Duhem’s theorem,chemical systems possess two degrees of freedom. Byspecifying the changes in any two variables (the !mon-itors parameters"), the changes in all the others can besolved. The program GIBBSGIBBS (Spear et al., 1991), versionFeb. 12, 1999, was used to perform the calculations.Items required for setting up the calculation includespecifying the nucleation density (number of garnetnuclei per 100 cm3 of rock), the phases in equilibriumwith garnet during its growth, the initial compositionsof the solid solution phases including garnet, the modalabundances of the phases at the time garnet beginsgrowing, the temperature and pressure correspondingto initiation of garnet growth, the choice of monitorsand the values of the monitors. Results obtained by thismethod are statistically robust (Kohn, 1993).

For all garnet growth simulations in this study, amodel system appropriate to pelitic schist wasassumed: K2O-Na2O-CaO-Al2O3-SiO2-MgO-FeO-MnO-H2O. The monitors were selected to be thegrossular content of garnet (Xgr) and the amount(mole) of garnet (MGar) in 100 cm

3 of rock. Thegrossular content of garnet was selected because it hasbeen shown to be the least susceptible of the eight-fold cations to diffusion (e.g. Spear, 1993). Thechanges in grossular content (DXgr) during garnetgrowth may be calculated from data contained in thezoning profile, and the changes in the amount ofgarnet is determined through trial and error, that is,when the simulation has produced garnet of thecorrect radius, then the correct value of DMGar has

been entered. The calculation solves for all remainingextensive and intensive parameters, including pressureand temperature, from which the P–T paths areconstructed. For all simulations in this study, theideal one-site mixing activity models pre-specified inthe GIBBSGIBBS program were used for staurolite (Fe, Mg,Mn), garnet (Fe, Mg, Ca, Mn), chlorite (Fe, Mg,Mn), paragonite (Ca-K-Na) and plagioclase (Ca-Na),and pure end-members with activities of 1.0 were forassumed for muscovite and quartz. The thermody-namic data set used is contained in the Thermo.Datfile included with the GIBBSGIBBS program, which isdescribed in Spear & Cheney (1989, their table 2) forthe minerals of interest.

Initial compositions and modes

Of all the initial conditions required to be specified,only the initial composition of the garnet, corre-sponding to the core, is directly measurable. In mostsimulations, the actual (measured) compositions ofthe other minerals were used as initial values(Table S2) because the simulations predicted onlyslight changes in composition and mode during gar-net growth. In all cases in this study, garnet wasdetermined to have grown through chlorite-break-down reactions (discussed later on for each case), andin all cases, chlorite was eliminated from the actualrock following garnet growth. Consequently, both theinitial mode and initial composition of chlorite had tobe estimated. In general, the initial mode of chloritein the simulation was assumed to be !1.5 times thatof the actual garnet mode. The chlorite initial Fe/Mgratio was estimated by back calculating usingthe garnet-chlorite thermometer of Grambling(1990), the assumed initial temperature for garnetgrowth, and the measured garnet core composition(Table S1). A chlorite stoichiometry typical of peliticschist (R2+, Al, Si in the proportions 4.5:3:2.5) wasassumed, which is a pre-specified option in the GIBBSGIBBSprogram.

Initial temperature and pressure

The initial temperature and pressure assumed in themodels were determined through a variety of methods,as described later on for each sample. Mineral assem-blages in relation to known stability constraints andthermobarometry were useful in estimating initialconditions, but in all cases in this study, the determi-nations carry large uncertainties.

Nucleation density

The values assigned to nucleation density act upon thecalculation to divide the garnet grown into a specifiednumber of nuclei. Increasing the nucleation densityresults in smaller garnet, all other things being equal.This gives the user a second way to adjust the garnet



radius in the simulations, the other being altering thevalue of DMGar. However, the two methods are notequivalent, as altering the value of DMGar will modifythe extent of reaction, whereas changing only thenucleation density will not.

Mineral assemblages

In all cases in this study, the mineral assemblagepresent at the time of garnet growth was determined

to be different than the mineral assemblage of theactual rock. Thus, it was necessary to determine thefull reaction history of each sample so that the correctminerals could be associated with garnet growth inthe simulations. Reaction histories are discussed foreach sample in the subsequent sections. All garnetssimulated in this study are interpreted to havegrown from dehydration reactions, so it was neces-sary to include H2O fluid in the list of phases in allsimulations.

Ato

m fr

actio

nA

tom

frac

tion

T (

°C)

Ato

m fr

actio

nT

(°C

)

Fig. 3. Microprobe garnet traverses (shaded lines) and superimposed models (black lines). Segment boundaries utilized in the simu-lations are marked with vertical black lines. For the UH samples, calculated garnet-biotite thermometry temperatures (Holdaway,2000) are plotted assuming 20% of total Fe in biotite is Fe3+ (Hoisch et al., 2002). Prograde re-equilibration features, interpreted fromthe thermometry, are labelled (c, crack; i, inclusion margin; r, rim).



Fig.4.

Ca,

Fe,Mgan

dMnelem

entmap

sforga

rnets.Lightershad

escorrespondto

higher

elem

entconcentrations.Lines

correspondto

thelocationsofmicroprobelinetrav

erses.

!L"istheleft

endofthetrav

erse

and!R"istherigh

tend,correspondingto

Fig.3.



Determination of segment boundaries

To simulate garnet zoning profiles, each garnet wasdivided up into segments. Segment boundaries werelocated at changes in slope along the Ca profile. Pointsthat display prograde reequilibration, as judged fromgarnet-biotite thermometry calculations (Fig. 3), wereavoided. Each segment is simulated in sequence fromthe core outward. To perform the calculations, valuesof the monitor parameters DXgr and DMGar weredetermined for each interval of growth (Table S3).

Obtaining a good fit

The fit of the simulated profile to the analysed profilewas judged by visual comparison. Only two parame-ters, Xgr and the garnet radius (controlled by DMGar),are guaranteed to fit. In some cases, good fits to Xalm,Xpy and Xsp were obtained through the trial and erroradjustment of three parameters: the nucleation density,the initial value of XMn-Chl and DMGar (adjusted foreach segment). The initial value of XMn-Chl makes littledifference to the calculated P–T path, but is importantbecause it controls the size of the Mn reservoir, andtherefore the rate at which Rayleigh fractionationcauses the garnet Mn concentration to drop towardsthe rim. In the simulations, only slight changes in thecompositions accompanied garnet growth for mostminerals (Tables S2 & S4). Adjustments were madeto the initial compositions to account for changes inthose cases where the changes appeared significant(Table S2). Adjustments to modes were also made insome cases (Table S5).

An important complication to consider when mak-ing adjustments to initial modes and initial composi-tions is that all garnets in this study are partiallyresorbed because of prograde reactions (and in theTHAL samples, also because of minor chlorite alter-ation). The resorption was accompanied by changes tothe modes and compositions of the matrix phases, andin some cases, changes to the mineral assemblage.Because only the early portion of the history corre-sponding to the growth of the garnet, from the core tothe resorbed rim, could be simulated, it is not validto judge the quality of fit by detailed comparisonsbetween the final simulated mineral compositions andmodes to the actual values from the rock. Our expe-rience is that simulations are very sensitive to changesin the monitor parameters, but not very sensitive tosmall changes in the values input for initial matrixcompositions or modes, so it is only necessary to knowthose values approximately.

Simulations using measured values as initial valuessometimes failed because of a phase or phase compo-nent being completely consumed prior to the comple-tion of garnet growth. This was the result of aninsufficient reservoir of a component needed for garnetgrowth. When this occurred and how it was dealt withis discussed below for the individual cases.

Thal4e

Petrography and reactions

This sample is schist of Mahogany Peaks from theAlbion Range and contains garnet, staurolite, musco-vite, paragonite, quartz and kyanite. Small grains(5–10 lm) of graphite occur in the matrix, as sparseinclusions in garnet, and as abundant inclusions instaurolite. Ilmenite occurs in the matrix and as inclu-sions in garnet and staurolite. Accessory rutile andtourmaline occur sparsely throughout. Chlorite occursonly along the rims of garnet as a secondary alterationproduct. The garnet is porphyroblastic, up to 8 mm indiameter and corroded. Embayments are commonlyfilled by staurolite (Fig. 5), suggesting a reaction inwhich staurolite grew at the expense of garnet. Thestaurolite is porphyroblastic (2–7 mm) and corroded.Small (20–50 lm) kyanite grains are localized alongthe margins of corroded staurolite, suggesting a reac-tion that consumed staurolite and produced kyanite.Muscovite, quartz and paragonite make up the fine-grained (

additional staurolite and garnet (Fig. 7). At this stage,staurolite should have been idioblastic, as no reactionto this point has consumed it. With further prograda-tion, the kyanite + garnet + staurolite triangle sweptto the right to encompass the bulk composition,causing kyanite and garnet to grow at the expense ofstaurolite (Fig. 7). The reaction is confirmed by

regression of measured mineral compositions in therock (Table S1, using the rim composition for G3):

0:23pgþ 0:53stþ 1:61qtz¼ 0:64grtþ 0:08msþ 4:26kyþ 0:07ilmþH2O:

The presence of kyanite grains localized along themargins of corroded staurolite grains is consistent withthis interpretation.The above discussion predicts that garnet growth

was followed by partial consumption, then by twofurther episodes of growth. A hiatus in the garnetzoning profiles is plainly seen in all three of the garnetsanalysed from this sample. Petrographically, the hiatusis represented by concentrations of graphite inclusionsthat align with discontinuities in the compositionalprofiles (Fig. 6). A portion of the rim of one garnet wasmapped for Y, and displays an annulus consisting of aspike in the Y content just beyond the discontinuity.Yttrium-rich annuli in garnet have been interpreted toresult from consumption followed by regrowth whenconsumption reaction involves the breakdown ofminerals that host significant reservoirs of Y (Pyle &Spear, 1999). In this case, the garnet-consumingreaction also consumed muscovite, which is a majorhost for Y in pelitic schist (e.g. Grauch, 1989). Whenmuscovite was consumed, Y was released into the areaalong the consuming garnet rim, and then incorpo-rated into the garnet when the garnet regrew. Thegarnet-consuming reaction releases H2O, and it may bespeculated that the fluid dissolved small amounts ofgraphite from within the matrix and precipitated itinto the area along the garnet rim. When the garnetregrew, the graphite that had been deposited alongthe corroded rim was occluded, forming the graphiteconcentrations that mark the discontinuity in theprofile.

Thermobarometry

Well-calibrated thermometers such as garnet-biotiteand garnet-chlorite could not be applied to this sampledue to the lack of appropriate mineral assemblages.For the same reason, well-calibrated barometers, suchas GASP or MBPG, could not be applied. The onlyreliable constraint on pressure and temperature is thepresence of kyanite and the inferred discontinuousreaction garnet + chlorite ¼ staurolite + biotite(Spear & Cheney, 1989), which was crossed duringprogradation. The garnet growth segments to bemodelled grew at temperatures below the discontinu-ous reaction. Consistent with these constraints,conditions of 500 "C and 5 kbar were assumed for theinitiation of garnet growth.

Simulation of garnet growth zoning

Three garnet fromTHAL4Ewere simulated (designatedG1, G2 and G3), all from the same polished section. Of

THAL4E

St

StSt

St

St

Grt

St

Ky

Ky

Ky

Chl

Grt

St St

THAL6B

Bt

Chl

UH-3

Grt

Bt

Fig. 5. Plane light photomicrographs showing textural charac-teristics of the different bulk compositions. Scale bars are 1 mm.Mineral abbreviations defined in Fig. 7. See text for descriptionand interpretation of textures.

10 C . R . HARR IS ET AL .


the three stages of garnet growth, only the first stageof growth, which took place within the assemblagechlorite + staurolite + muscovite + paragonite +quartz, could be simulated numerically to determineP–T paths. Although the beginning of the second stageof growth, following the hiatus, could theoretically besimulated, there is no feature that would allow one tointerpret where the change from the second stage to thethird stage occurs. Because the mineral assemblagechanged between the second and third stages of garnetgrowth, and because it is essential to associate the cor-rect portion of the garnet profile with the correctassemblage in the simulations, neither could be simu-lated. Thus, garnet from this sample was numericallysimulated from the cores to the hiatuses. G3 contains afaint irregularly shaped cluster of graphite inclusionsnear the centre, which corresponds to a distinct area ofincreased Ca concentration (Figs 3 & 4). This feature isinterpreted as a resorbed older garnet, or possibly arelict of an overgrown phase (e.g. Hirsch et al., 2003),and so was avoided in the modelling.

The profile for G1 was well simulated in a singlesegment of growth, G2 was well simulated in two

segments, and G3 was well simulated in seven seg-ments. All three garnet display a high degree of sym-metry. G3 provided the most complete path of anysingle garnet modelled in this study.

In this rock, the reservoir for Ca needed to make thegrossular component of garnet is contained entirely inparagonite as the margarite component. When themodel was run using measured values as initial values,the simulations ran out of margarite component beforegarnet growth was completed. In order to providesufficient Ca for garnet growth, the initial values ofboth the paragonite composition and modal abun-dance were adjusted to higher values than determinedfor the actual rock (Tables S2 & S5).

Thal6b and Thal6c

Petrography and reactions

These two samples are schist of Mahogany Peaks fromthe Albion Range (Fig. 2a). Both contain garnet,staurolite, muscovite, biotite, paragonite, plagioclaseand quartz. Small grains (

Grt

+ C

hl

St +

Bt

Fe Mg

Ky

Grt Chl

St

Bt

Fe Mg

Grt Chl

Bt

St

Fe Mg

Ky

Grt Chl

St

Bt

Fe Mg

Ky

Grt Chl

St

Bt

Fe Mg

Ky

Grt Chl

St

Bt

Fe Mg

Ky

Grt Chl

St

Bt

Ky

102-3

102-5

THAL4E

THAL6CTHAL4B

UH-3UH-5

St

Ky

Gar

BtBt

6B

6C

4E

6C6B

BULK MINERAL

TIE LINES

COMPOSITIONS

Fe Mg

Fe-G

rt +

Ky

Fe-S

t

(a) (b) (c)

(d)(e) (f)

+Qz+Ms+H2O±Pg±Pl

(g)

Chloritoid

Fig. 7. AFM projections (Thompson, 1957) showing reactions responsible for garnet growth. Bulk compositions are shown as stars.(g) Labelled bulk compositions, composition field for normal pelitic chloritoid (grey shaded area) and precisely plotted mineralcompositions. Arrows show migration of three-phase triangles in direction of net dehydration. (a–f) Prograde sequence; (a–c)migration of three-phase triangles responsible for initial garnet growth. In samples THAL4E, THAL6B and THAL6C, progradegarnet rim consumption, growth of clear rims on staurolite, and the complete consumption of chlorite resulted from discontinuousreaction grt + chl ¼ st + bt across (c) and (d). Further progradation resulted in consumption of biotite and the corrosion ofstaurolite in THAL4E as the bt + grt + st triangle migrated to the right (e, f). In THAL4E, texturally late kyanite and re-growth ofgarnet occurred due to prograde migration of ky + grt + st triangle (e, f). Biotite analyses plotted on triangles in (g) are shown todemonstrate slight offset from assumed biotite compositions. Assumed biotite compositions compensate for low alkali totals thatresulted from using small spot size during analysis, necessary due to small grain size. Bt, biotite; Chl, chlorite; Grt, garnet; Ms,muscovite; Ky, kyanite; Pg, paragonite; Pl, plagioclase; Qtz, quartz; St, staurolite.



the matrix, as sparse inclusions in garnet and asabundant inclusions in staurolite. Ilmenite occurs inthe matrix and as inclusions in garnet and staurolite.Chlorite occurs only along the rims of garnets as asecondary alteration product. The garnets is porphy-roblastic (up to 8 mm in diameter) and embayed.Embayments are commonly filled by staurolite (Figs 5& 6), suggesting a reaction in which staurolite grew atthe expense of garnet. The staurolite is porphyroblasticand subidioblastic. Fine-grained (

inclusion trails defined by quartz and ilmenite and arecorroded along rims (Fig. 5). The corroded rimstruncate the interior zoning patterns, except for fea-tures related to post-consumption prograde diffusivereequilibration, which caused Fe-enrichment alongrims, cracks and some inclusion margins (Fig. 4). Theconsumed portions of the garnet rims are notablydevoid of muscovite, suggesting a muscovite-consum-ing reaction. Unlike the THAL samples, these containno paragonite and the matrix is dominated by plagio-clase. There is also less late shear deformation, eventhough the rocks occur within a well-developedTertiary shear zone. The resistance to shearing maybe attributable to the abundance of plagioclase inthe sample, which may have provided greater resis-tance to deformation overprints relative to more micaand quartz-rich adjacent lithologies (e.g. Passchier &Trouw, 2005).

The estimated bulk composition plots below thegarnet-chlorite join on an AFM plot (Table S1), indi-cating a low-alumina (common) variety of peliticschist. Using the estimated bulk composition andinterpreting from the AFM plots of Spear & Cheney(1989), garnet growth is predicted to have occurred viaessentially the same reaction as the THAL samples,that is, by the breakdown of chlorite. Garnet corrosionis interpreted to have occurred via a fluid-absentreaction resulting from a major drop in pressure duringprogradation.

Thermobarometry

These samples contain mineral assemblages appropri-ate for the application of well-calibrated thermo-barometry methods, specifically, garnet-biotitethermometry and MBPG barometry. Unlike theTHAL samples, garnet is interpreted to have grown inthe presence of biotite. For these samples, garnet-bio-tite thermometry (Hodges & Spear, 1982; Holdaway,2000) and MBPG barometry (Hoisch, 1991) were cal-culated using garnet compositions as described belowand average matrix compositions for biotite and pla-gioclase, in an effort to determine the P–T conditionsof initiation of garnet growth (Fig. 8).

The results of the calculations must be interpretedbearing in mind a number of complicating factors.Garnet consumption causes the Fe-content of matrixbiotite to increase, causing the calculated temperatureto artificially increase (Kohn & Spear, 2000). Garnetconsumption also releases Ca from the consumedgarnet rim, which becomes incorporated intoplagioclase, thus increasing the anorthite contentbeyond what it was at the end of garnet growth.Because matrix compositions change during garnetgrowth, combining matrix compositions with garnetcore compositions introduces error in thermobarom-etry calculations. For rocks from this outcrop,Hoisch et al. (2002) estimated that changes in biotitecomposition accompanying garnet growth could

result in temperatures being underestimated by nomore than 18 "C. However, associating garnet coreswith matrix plagioclases in barometry calculationsdid not produce good results in this study, sobarometry calculations were made using garnet rims(Fig. 8).Considering these complicating factors, the results

should be considered to possess large uncertainties, butgarnet rims yielded generally consistent temperatures of450–460 "C and consistent pressures for the intersec-tions of the barometry equilibria of 4.5–5.5 kbar(Fig. 8). The determined conditions are appropriate forthe growth of garnet via chlorite breakdown in peliticrocks (Spear & Cheney, 1989). For the simulations, theinitial pressure was assumed to be 5 kbar and the initialtemperature of garnet growth was taken to be the cal-culated temperatures using the garnet cores at 5 kbarand average matrix biotite assuming 20% of Fe to beFe3+ (Hoisch et al., 2002).

Simulation of garnet growth zoning

Garnet growth in these samples was simulated withinthe assemblage quartz + muscovite + biotite +chlorite + plagioclase. Two garnets from each of thetwo samples were analysed and simulated. Garnet-biotite thermometry, plotted for each point alongeach traverse in Fig. 3, was used as a guide forevaluating whether analysed points were acceptableoptions for segment boundaries. Points near rims,along cracks, and along inclusion margins displayprograde reequilibration, and so could not be used assegment boundaries. Two of the garnets possessprofiles that are slightly asymmetric (UH-3-G1 andUH-3-G2), so one half of each garnet profile waschosen for simulation. The profiles of the other twogarnets (UH-5-G1 & UH-5-G2) are highly symmetric,and the half with the most complete record waschosen for simulation. UH-5-G2 and UH-3-G1 wereeach simulated in two segments of growth, whileUH-3-G2 required three segments and UH-5-G1required four segments.

CREATING THE COMPOSITE P–T PATH

In this study, the simulated portions of all garnets grewfrom very similar reactions involving the breakdown ofchlorite. However, because of differences in the bulkchemistry of the rocks and possibly other factors, eachgarnet initiated growth at a different point along theP–T path. Because the initial P–T conditions assumedfor the initiation of the growth of each garnet couldonly be approximated, the generated paths are unlikelyto be in their correct relative positions in P–T space.However, the determined P–T paths may be shiftedslightly in P–T space without introducing significanterrors because the path calculations are not sensitive tosmall changes in the assumed initial pressure andtemperature.



Several approaches were used to determine the rel-ative positions of the P–T paths in P–T space. Garnetsanalysed within the same polished section grew in closeproximity to each other, and therefore drew from thesame Mn reservoir during growth. This allows Mnconcentration along garnets profiles to be used totemporally correlate P–T path segments from differentgarnets (e.g. Spear & Daniel, 2003). The correlation ofthe paths of the three garnets analysed in THAL4E isshown in Fig. 9. The correlation is consistent withgarnets growth initiating at different times along thesame P–T path, and with growth of the garnets over-lapping to varying degrees. The approach appears towork with the exception of the first segment of the pathgenerated for G2, which is slightly out of place withrespect to Mn composition compared with the simi-larly-sloping final segment of G3. This could be due tothe development of slight heterogeneities in the Mn

reservoir during garnets growth, such as might occur ifsubtle Mn depletion halos developed. Similarly, pathsmay be correlated based on Mn content for the twogarnets analysed from UH-3 and the two garnetsanalysed from UH-5. Each pair of garnets substan-tially overlap in their Mn contents, and therefore, soshould their P–T paths.

A second approach involves correlating path seg-ments with similar slopes. The repositioning of thethree paths from THAL4E based on Mn content de-scribed above places into alignment similarly slopingsegments (Fig. 9). The same is true of the two garnetsfrom UH-3 and the two garnets from UH-5 (Fig. 10).In addition, the slope of the first segment of the P–Tpath from THAL4E-G3 (schist of Mahogany peaksfrom the Albion range) is in good agreement withsimilar segments of P–T paths from the upper horizonof the schist of Stevens Spring from the northern

2

3

4

5

6

7

8

9

10

P (

kbar

)

400 500 600 700T (°C)

UH-3 G1

20% Fe3+10% Fe3+H&S

not involving Tschermak components

involving Tschermak components

MBPG geobarometry:

Garnet-biotite geothermometry:

2

3

4

5

6

7

8

9

10

T (°C)

UH-3 G2

P (

kbar

)

T (°C) T (°C)

UH-5 G2

2

3

4

5

6

7

8

9

10UH-5 G1

2

3

4

5

6

7

8

9

10

400 500 600 700

400 500 600 700 400 500 600 700

Fig. 8. Geothermobarometry calculated from analysed mineral compositions. MBPG geobarometry (Hoisch, 1991) and garnet-biotitegeothermometry were calculated using garnet rims as follows (see Fig. 3): UH-3-G1 and G2 used point 3, UH-5-G1 used point 4,UH-5-G2 used point 2. Geothermometry (Holdaway, 2000) shown for calculations assuming 10% and 20% of total Fe in biotite isFe3+. Geothermometry labelled H&S (Hodges & Spear, 1982) assumes all Fe in biotite is Fe2+.



Grouse Creek Mountains (Hoisch et al., 2002; and thetwo UH-5 garnets in this study). The paths of the twogarnets from UH-3 were found to be similar to thosefrom the same outcrop reported in Hoisch et al. (2002),in that they possess a similar characteristic inflectionwhere temperatures change from increasing todecreasing. To align them with those in Hoisch et al.(2002), the paths in the current study were shifted asindicated in Table S6. When the starting pressure forthe P–T paths for UH-5-G1 and G2 is shifted downfrom 5 to 4 kbar, the paths align with the slope andlocation of the first path segments of the UH-3 paths,and thus the UH-5 paths appear to represent a cooler,earlier portion of the P–T history.

A third approach involves considerations of mineralequilibria. Qualitatively, the interpretation of reactionsbased on AFM plots indicates that garnet growth inTHAL6B and THAL6C should have begun afterTHAL4E (Fig. 7). In addition, the repositioning of thefour UH paths, described above, is supported by thecalculated rim temperatures for all four garnets, as allthe four paths appear to have ended at similar tem-peratures (Fig. 8). Conceivably, one could use theGibbs method of Spear & Selverstone (1983) todetermine the relative displacement of paths fromrocks that possess the same mineral sub-assemblagerepresenting a thermodynamic variance of three orless, which would allow the calculation to be run onthe basis of the differences in garnet core compositions.However, this was not case for the samples in thisstudy.

Using all these approaches, the individual pathswere combined into a single composite path in Fig. 10by shifting each calculated path as specified inTable S6. The distance between samples sites in thenorthern Grouse Creek Mountains (UH samples,schist of Stevens Spring) and sample sites in the AlbionMountains (THAL samples, schist of MahoganyPeaks) is !40 km (Fig. 1). The distance between theselocalities and the uncertainties in the determinedpressures of metamorphism allows for significant dif-ferences of structural level. Although similar P–T pathsegments for the two rock units most probably recordthe same tectonic events, the absolute pressures may besomewhat offset from the depiction in Fig. 10. How-ever, major structures that would vertically separaterocks between the two areas have not been identified.Synorogenic low-angle normal faults have been iden-tified in the region (Wells et al., 1998), but the areas ofinterest lie in the footwalls of these faults and so couldnot be offset by them.

TECTONIC IMPLICATIONS OF THE COMPOSITEP–T PATH

The composite P–T path overlaps and extends thepreviously published P–T paths from Basin Creekreported in Hoisch et al. (2002). Some general state-ments regarding tentative correlations between pathsegments and geological structures and the tectonicsignificance of these paths can be made, althoughcorrelations with mapped geological structures and

0.068

XSp =

T (°C)450 500 550 600

4

5

6

7

8

9P

(kb

ar)

0.051

G1

G3 G2

0.012 (G3)0.015 (G1)0.021 (G2)

0.008

0.074

G3 G1 G2

step XSp XSpXSpstep step

core 0.0739

1 0.07192 0.0680

3 0.0619

4 0.0508core 0.0391

core 0.0374

1 0.0210

5 0.0190

1 0.0151

6 0.01232 0.0080

T, P

T, P

T, P

T, P

Inferredtime

Most recent

Least recent

Dec

reas

ing

XS

p

T, P

(b)(a)

Fig. 9. Correlation of XSp values among garnet from THAL4E. (a) Pressure–temperature plot of P–T paths with Xsp values labelledfor selected segment boundaries. (b) Xsp values at segment boundaries v inferred time.



deformation fabrics in the region (e.g. Wells, 1997)and with dated events in the fold-thrust belt to the eastremain speculative.

Emplacement of the hangingwall rocks of the Basin-Elba fault most probably caused the burial eventsrecorded in the P–T paths as well as widespread uppergreenschist to amphibolite facies metamorphism of thefootwall across the exposure of the Raft River Moun-tains sequence (M1 in Table 1). The relationship be-tween the thrust at the base of the Robinson Creeksequence and the Basin-Elba fault may indicate that theBasin-Elba fault represents the floor thrust to a duplexfault zone of tectonically thickened Late Cambrian andNeoproterozoic rocks, or that the Robinson Creeksequence represents a faulted upright limb to a foldnappe, with the Mount Harrison sequence representingthe lower overturned limb (Miller, 1980, 1983). TheBasin-Elba fault accomplished significant crustalthickening as the hangingwall sheet had an estimatedstratigraphic thickness greater than 12 km.

By inference, initial motion on the Basin-Elba faultis interpreted to be older than the 105 (±12) Maorogen-parallel extensional fabric, as this fabric isprograde and the metamorphic conditions requiresignificant tectonic burial (Table 1). The Willard thrustsheet, and other !megathrust sheets" that carry thickProterozoic–Lower Cambrian clastic successions attheir bases, including the Canyon Range and SheepRock, were initially translated in Early Cretaceoustime (144–110 Ma) (DeCelles, 2004). As the Basin-Elba fault lies in a more western position within theorogenic wedge than the Willard thrust (Camilleriet al., 1997), and because it is unlikely that the Basin-Elba fault fed slip into the Willard thrust (see below),we infer that its age of initial motion is older thaninitial motion on the Willard thrust.

It is unlikely that the Willard thrust, and otherthrusts of the foreland fold-thrust belt, received slipfrom the Basin-Elba fault. The metamorphic rocks inthe footwall of the Basin-Elba fault include upper

greenschist-lower amphibolite facies Ordovician strata.The Ordovician rocks lie in the hangingwall of theMahogany Peaks fault, a west-directed low-anglenormal fault that post-dates initial motion on theBasin-Elba fault and juxtaposes a hangingwall flat inOrdovican rocks against a footwall flat in Neoprote-rozoic rocks. The simplest interpretation of theserelationships is that the Ordovician metasedimentaryrocks were derived from an up-dip segment of thewest-tilted footwall of the Basin-Elba fault and dis-placed down-dip, minimally cutting out the Cambriansection, to rest on the Neoproterozoic schist ofMahogany Peaks. The foreland thrusts sole into abasal decollement in Late Proterozoic siliciclastic rocksto the west, and are interpreted to eventually cut downwestward into crystalline basement (Allmendinger,1992; Yonkee, 1992; Camilleri et al., 1997). Thus, it isunlikely that the basal decollement to the forelandfold-thrust belt buried rocks as young as Ordovician inthe study area, requiring a western thrust (e.g. theBasin-Elba fault), now internal to the orogenic wedge,to have buried the metamorphic rocks of the RaftRiver sequence. The Basin-Elba fault may have fed slipinto the Manning Canyon detachment, a hinterlandthrust, thought to be Late Jurassic, which forms abedding-parallel fault over a broad region of north-western Utah and places Pennsylvanian over Missis-sippian strata (Allmendinger & Jordan, 1981).

Field geological relationships provide evidence for atleast one episode of renewed thrust motion on theBasin-Elba fault following an episode of extension.The extensional Mahogany Peaks fault is folded intoan east-vergent, tight to isoclinal, overturned synclinein the immediate footwall of the Basin-Elba fault(Miller, 1980, 1983), suggesting a kinematic linkbetween east-vergent folding and motion on the fault(D3 in Table 1). We tentatively correlate the youngestsegment of the composite P–T path, a pressureincrease, with this renewed thrusting event. The ageconstraints of renewed thrusting, c. 60–45 Ma, deter-

650

UH-3 G2

UH-3 G1

UH-5 G1

from Hoisch et al. (2002)

T (°C)

P (

kbar

)

3

4

5

6

7

8

9

Ky

AndSil

T (°C)450 500 550 600

Schist of Stevens Spring, Basin Creek Schist of Mahogany Peaks, Albion Range All

Ky

AndSil

Ky

AndSil

upper horizon

lowerhorizon

cores

rims

cores

rims

450 500 550 600400 450 500 550 600T (°C)

cores

rims

THAL4E G3

UH-5 G2

THAL4E G1

THAL4E G2

THAL6C G1

THAL6B G1

exhumation - D

2

D3

M1

(a) (b) (c)

Fig. 10. P–T paths generated from garnet growth simulations and shifted as indicated in Table S6. Al2SiO5 stability fields fromPattison (1992); And, andalusite; Ky, kyanite; Sil, sillimanite). (a) Paths from schist of Stevens Spring samples in Basin Creek. (b) Pathsfrom schist of Mahogany Peaks samples, Albion Mountains. (c) Paths from all samples. Tentative correlations of path segments totectonic events (Table 1) are shown.



mined by Th-Pb dating of monazite inclusions ingarnet (revised from Hoisch and Wells 2004), areconsistent with garnet growth following motion on theMahogany Peaks fault that is bracketed between c. 90and 60 Ma (Table 1).

The exhumation events recorded in the compositeP–T path may have occurred by a combination ofextensional reactivation of the Basin-Elba fault (Hod-ges & Walker, 1992), motion along the extensionalMahogany Peaks and Emigrant Spring faults (Wells,1997; Wells et al., 1998), extensional flow associatedwith mid-Cretaceous orogen-parallel extension oralong structures not yet recognized. We tentativelycorrelate the two major decompressional segments ofthe path with motion along the Mahogany Peaks andEmigrant Spring extensional faults, because the Pro-terozoic pelitic schists lie in the footwall of both andthis is permitted by the available geochronology. Thetemperature increase of !50 "C that accompanied thefirst decompression is attributable to thermal relaxa-tion associated with thrusting along the Basin-Elbafault (Hoisch et al., 2002); however, heating followingthe second decompression may require an additionalheat source, such as might result from lower litho-spheric delamination.

ACKNOWLEDGEMENTS

Reviews by A. Martin, J. Cheney and an anonymousreviewer greatly improved the paper. We thankJ. Wittke for assistance with microprobe analysis atNAU, K. Becker at RPI for generating some of thegarnet element maps and E. Kelly for drafting thegeological map in Fig. 1. This work was supported byNSF grants EAR-9805076 and EAR-0061048 to TDH,EAR-9805007 and EAR-00610089 to M.L.W., andgrants from the Colorado Scientific Society andGeological Society of America to C.R.H. NSF grantEAR-0421452 supported the acquisition of the JEOLJSM-6480LV used to produce some of the elementmaps in this study.

REFERENCES

Allmendinger, R. W., 1992. Fold and thrust tectonics of theWestern United States exclusive of the accreted terranes. In:The Cordilleran Orogen; Conterminous U.S., G-3 (eds Bur-chfiel, B.C. Lipman, P.W. & Zoback, M.L.), pp. 583–607.Geological Society of America, Boulder, CO.

Allmendinger, R. W., & Jordan, T. E., 1981. Mesozoicevolution, hinterland of the Sevier orogenic belt. Geology, 9,308–313.

Allmendinger, R. W., Miller, D. M. & Jordan, T. E., 1984.Known and inferred Mesozoic deformation in the hinterlandof the Sevier belt, northwestern Utah. In: Geology of North-western Utah, Southern Idaho and Northeast Nevada, (edsKerns, G.J. & Kerns, R.L.), pp. 21–34. Utah GeologicalAssociation, Salt Lake City, Utah.

Armstrong, R. L., 1968. Mantled gneiss domes in the AlbionRange, southern Idaho. Geological Society of America Bulle-tin, 79, 1295–1314.

Armstrong, F. C. & Oriel, S. S., 1965. Tectonic development ofIdaho-Wyoming thrust belt. American Association of Petro-leum Geologists Bulletin, 49, 1847–1866.

Burchfiel, B. C., Cowan, D. S. & Davis, G. A., 1992. Tectonicoverview of the Cordilleran orogen in the western UnitedStates. In: The Cordilleran Orogen; Conterminous U.S., G-3(eds Burchfiel, B.C. Lipman, P.W. & Zoback, M.L.), pp. 407–479. Geological Society of America, Boulder, CO.

Camilleri, P. A. & Chamberlain, K. R., 1997. Mesozoic tectonicsand metamorphism in the Pequop Mountains and Wood Hillsregion, northeast Nevada: implications for the architectureand evolution of the Sevier orogen. Geological Society ofAmerica Bulletin, 109, 74–94.

Camilleri, P., Yonkee, W. A., Coogan, J. C., DeCelles, P. G.,McGrew, A. & Wells, M., 1997. Hinterland to forelandtransect through the Sevier orogen, NE Nevada to SWWyoming: structural style, metamorphism, and kinematichistory of a large contractional orogenic wedge. In: Protero-zoic to Recent Stratigraphy, Tectonics, and Volcanology, Utah,Nevada, Southern Idaho and Central Mexico, Vol. 42 (edsLink, P.K. & Kowallis, B.J.), pp. 297–309. Brigham YoungUniversity Geology Studies, Provo, Utah.

Collins, W. J., 2002. Hot orogens, tectonic switching, and crea-tion of continental crust. Geology, 30, 535–538.

Compton, R. R., Todd, V. R., Zartman, R. E. & Naeser, C. W.,1977. Oligocene and Miocene metamorphism, folding, andlow-angle faulting in northwestern Utah. Geological Society ofAmerica Bulletin, 88, 1237–1250.

Crittenden, M. J., Coney, P. J., & Davis, G. H. (eds), 1980.Cordilleran Metamorphic Core Complexes, Geological Societyof America Memoir 153. Geological Society of America,Boulder, CO.

Davis, D., Suppe, J. & Dahlen, F. A., 1983. Mechanics of fol-d-and-thrust belts and accretionary wedges. Journal of Geo-physical Research, 88, 1153–1172.

DeCelles, P. G., 1994. Late Cretaceous-Paleocene synorogenicsedimentation and kinematic history of the Sevier thrust belt,Northeast Utah and Southwest Wyoming. Geological Societyof America Bulletin, 106, 32–56.

DeCelles, P. G., 2004. Late Jurassic to Eocene evolution of theCordilleran thrust belt and foreland basin system, westernUSA. American Journal of Science, 304, 105–168.

Dewey, J. F., 1988. Extensional collapse of orogens. Tectonics, 7,1123–1139.

Egger, A. E., Dumitru, T. A., Miller, E. L., Savage, C. F. I. &Wooden, J. L., 2003. Timing and nature of Tertiary plutonismand extension in the Grouse Creek Mountains, Utah. Inter-national Geology Review, 45, 497–532.

Elison, M. W., 1991. Intracontinental contraction in westernNorth America: continuity and episodicity. Geological Societyof America Bulletin, 103, 1226–1238.

England, P. C. & Houseman, G., 1989. Extension during conti-nental convergence, with application to the Tibetan Plateau.Journal of Geophysical Research, 94, 17,561–17,579.

England, P. C. & Thompson, A. B., 1984. Pressure-temperature-time paths of regional metamorphism; I, Heat transfer duringthe evolution of regions of thickened continental crust. Journalof Petrology, 25, 894–928.

Ferranti, L. & Oldow, J. S., 1999. History and tectonic impli-cations of low-angle detachment faults and orogen parallelextension, Picentini Mountains, southern Apennines fold andthrust belt, Italy. Tectonics, 18, 498–526.

Grambling, J. A., 1990. Internally consistent geothermometryand H2O barometry in metamorphic rocks: the example ofgarnet-chlorite-quartz. Contributions to Mineralogy andPetrology, 105, 617–628.

Grasemann, B. & Mancktelow, N. S., 1993. Two-dimensionalthermal modeling of normal faulting; the Simplon fault zone,Central Alps, Switzerland. Tectonophysics, 225, 155–165.

Grauch, R. I., 1989. Rare earth elements in metamorphic rocks.Reviews in Mineralogy 21, 147–167.



Hirsch, D. M., Prior, D. J. & Carlson, W. D., 2003. An over-growth model to explain multiple, dispersed high-Mn regionsin the cores of garnet porphyroblasts. American Mineralogist,88, 131–141.

Hodges, K. V. & Spear, F. S., 1982. Geothermometry, geoba-rometry, and the Al2SiO5 triple point at Mt. Moosilauke, NewHampshire. American Mineralogist, 67, 1118–1134.

Hodges, K. V. & Walker, J. D., 1992. Extension in the Creta-ceous Sevier orogen, North American Cordillera. GeologicalSociety of American Bulletin, 104, 560–569.

Hodges, K. V., Snoke, A. W. & Hurlow, H. A., 1992a. Thermalevolution of a portion of the Sevier hinterland: the northernRuby Mountains-East Humboldt Range and Wood Hills,northeastern Nevada. Tectonics, 11, 154–164.

Hodges, K. V., Parrish, R. R., Housh, T. B. et al., 1992b.Simultaneous Miocene extension and shortening in theHimalayan Orogen. Science, 258, 1466–1470.

Hoisch, T. D., 1991. Equilibria within the mineral assemblagequartz + muscovite + biotite + garnet + plagioclase, andimplications for the mixing properties of octahedrally-coor-dinated cations in muscovite and biotite. Contributions toMineralogy and Petrology, 108, 43–54.

Hoisch, T. D., 2005. Thermod7: a general two-dimensionalnumerical modeling program for heat conduction and advec-tion, with special application to faults. Computers & Geo-sciences, 31, 698–703.

Hoisch, T. D. & Simpson, C., 1993. Rise and tilt of metamorphicrocks in the lower plate of a detachment fault in the FuneralMountains, Death Valley, California. Journal of GeophysicalResearch, 98, 6805–6827.

Hoisch, T. D. & Wells, M. L., 2004. Correlation between Th-Pbion microprobe ages of monazite inclusions in garnet andgarnet growth zoning from upper amphibolite facies schist –implications for garnet dating. Geological Society of AmericaAbstracts with Programs, 36, 134.

Hoisch, T. D., Wells, M. L. & Hanson, L. M., 2002. Pressure-temperature paths from garnet-zoning: evidence for multipleepisodes of thrust burial in the hinterland of the Sevier oro-genic belt. American Mineralogist, 87, 115–131.

Holdaway, M. J., 2000. Application of new experimental andgarnet Margules data to the garnet-biotite geothermometer.American Mineralogist, 85, 881–892.

Hollister, L. S., 1966. Garnet zoning: an interpretation basedon the Rayleigh fractionation model. Science, 154, 1647–1651.

Hudec, M. R., 1992. Mesozoic structural and metamorphichistory of the central Ruby Mountains metamorphic corecomplex, Nevada. Geological Society of America Bulletin, 104,1086–1100.

Jamieson, R. A., Beaumont, C., Fullsack, P. & Lee, B., 1998.Barrovian regional metamorphism: where’s the heat? In: Whatdrives metamorphism and metamorphic reactions? SpecialPublication (eds Treloar, P.J. & O’Brien, P.J.), pp. 23–51.Geological Society, London.

Jamieson, R. A., Beaumont, C., Nguyen, M. H. & Lee, B., 2002.Interaction of metamorphism, deformation and exhumation inlarge convergent orogens. Journal of Metamorphic Geology,20, 9–24.

Johnson, T. & Brown, M., 2004. Quantitative constraints onmetamorphism in the variscides of Southern Brittany – acomplementary pseudosection approach. Journal of Petrology,45, 1237–1259.

Ketcham, R. A., 1996. Thermal models of core-complex evolu-tion in Arizona and New Guinea: implications for ancientcooling paths and present-day heat flow. Tectonics, 15, 933–951.

Kohn, M. J., 1993. Uncertainties in differential thermodynamic(Gibbs" method) PT paths. Contributions to Mineralogy andPetrology, 113, 24–39.

Kohn, M. J. & Spear, F. S., 2000. Retrograde net transferreaction insurance for pressure–temperature estimates. Geol-ogy, 28, 1127–1130.

Lawton, T. F., Sprinkel, D., DeCelles, P. G., Mitra, G. &Sussman, A. J., 1997. Thrusting and synorogenic sedimenta-tion in the central Utah Sevier thrust belt and foreland basin.Brigham Young University Geology Studies, 42, 33–67.

Lewis, C., Wernicke, B. P., Selverstone, J. & Bartley, J. M., 1999.Deep burial of the footwall of the northern Snake Range de-collement, Nevada. Geological Society of America Bulletin,111, 39–51.

Mahar, F. M., Baker, J. M., Rowell, R., Holland, T. J. B. &Howell, N., 1998. The effect of Mn on mineral stability inmetapelites. Journal of Metamorphic Geology, 15, 223–238.

Malavieille, J., 1987. Kinematics of compressional and exten-sional ductile shearing deformation in a metamorphic corecomplex of the northeastern Basin and Range. Journal ofStructural Geology, 9, 541–554.

McGrew, A. J., Peters, M. T. & Wright, J. E., 2000. Thermo-barometric constraints on the tectonothermal evolution of theEast Humboldt Range metamorphic core complex, Nevada.Geological Society of America Bulletin, 112, 45–60.

Miller, D. M., 1980. Structural geology of the northern AlbionMountains, south-central Idaho. In: Cordilleran MetamorphicCore Complexes, Geological Society of America Memoir 153(eds Crittenden, M.D., Coney, P.J. & Davis, G.H.), pp. 399–423. Geological Society of America, Boulder, CO.

Miller, D. M., 1983. Allochthonous quartzite sequence in theAlbion Mountains, Idaho, and proposed Proterozoic Z andCambrian correlatives in the Pilot Range, Utah and Nevada.In: Tectonic and Stratigraphic Studies in the Eastern GreatBasin, Geological Society of America Memoir 157 (eds Miller,D.M., Todd, V.R. & Howard, K.A.), pp. 191–213. GeologicalSociety of America, Boulder, CO.

Miller, D. M. & Allmendinger, R. W., 1991. Jurassic normal andstrike-slip faults at Crater Island, northwestern Utah. Geo-logical Society of America Bulletin, 103, 1239–1251.

Miller, E. L., Gans, P. B., Wright, J. E. & Sutter, J. F., 1988.Metamorphic history of east-central Basin and Range Prov-ince. In: Metamorphism and Crustal Evolution of the WesternUnited States, Vol. 7 (ed. Ernst, W.G.), pp. 649–682. PrenticeHall, Englewood Cliffs, NJ.

Molnar, P. & Lyon-Caen, H., 1988. Some simple physical as-pects of the support, structure and evolution of mountainbelts. In: Processes in Continental and Lithospheric Deforma-tion, Geological Society of America Special Paper 218 (ed.Clark, S.P.J.), pp. 197–207. Geological Society of America,Boulder, CO.

Oxburgh, E. R. & Turcotte D. L., 1974. Thermal gradients andregional metamorphism in overthrust terrains with specialreference to the eastern Alps. Schweizerische Mineralogischeund Petrographische Mitteilungen, 54, 641–662.

Passchier, C. W. & Trouw, R. A. J., 2005. Microtectonics.Springer-Verlag, New York, 366 pp.

Pattison, D. M. R., 1992. Stability of andalusite and sillimaniteand the Al2SiO5 triple point: constraints from the Ballachulishaureole, Scotland. Journal of Geology, 100, 423–446.

Platt, J. P., 1986. Dynamics of orogenic wedges and uplift ofhigh-pressure metamorphic rocks. Geological Society ofAmerica Bulletin, 97, 1037–1053.

Platt, J. P., & England, P. C., 1994. Convective removal oflithosphere beneath mountain belts: thermal and mechanicalconsequences. American Journal of Science, 293, 307–336.

Pyle, J. M. & Spear, F. S., 1999. Yttrium zoning in garnet:coupling of major and accessory phases during metamorphicreactions. Geological Materials Research, 1, 1–49.

Rey, P., Vanderhaeghe, O. & Teyssier, C., 2001. Gravitationalcollapse of the continental crust; definition, regimes andmodes. Tectonophysics, 342, 435–449.

Royden, L. H., 1993. The tectonic expression of slab pull atcontinental convergent boundaries. Tectonics, 12, 303–325.

Ruppel, C. & Hodges, K. V., 1994. Pressure-temperature-time paths from two-dimensional thermal models; prograde,retrograde, and inverted metamorphism. Tectonics, 13, 17–44.



Ruppel, C., Royden, L. & Hodges, K. V., 1988. Thermal mod-eling of extensional tectonics; application to pressure–tem-perature-time histories of metamorphic rocks. Tectonics, 7,947–957.

Selverstone, J., 1985. Petrologic constraints on imbrication,metamorphism, and uplift in the SW Tauern window, easternAlps. Tectonics, 4, 687–704.

Spear, F. S., 1988. The Gibbs and Duhem’s theorem: thequantitative relationships among chemical potential, phasecomposition and reaction progress in igneous and metamor-phic systems. Contributions to Mineralogy and Petrology, 99,249–256.

Spear, F. S., 1989. Petrologic determination of metamorphicpressure–temperature-time paths. In: Metamorphic Pressure–Temperature–Time Paths, Short Course in Geology Vol. 7 (edsSpear, F.S. & Peacock, S.M.), pp. 1–55. American Geophys-ical Union, Washington, DC.

Spear, F. S., 1993. Metamorphic Phase Equilibria and Pressure–Temperature–Time Paths. Monograph 1. Mineralogical Soci-ety of America, Washington DC, 799 pp.

Spear, F. S. & Cheney, J. T., 1989. A petrogenetic grid for peliticschists in the system SiO2-Al2O3-FeO-MgO-K2O-H2O. Con-tributions to Mineralogy and Petrology, 101, 149–164.

Spear, F. S. & Daniel, C. G., 2003. Three-dimensional imagingof garnet porphyroblast sizes and chemical zoning; nucleationand growth history in the garnet zone. American Mineralogist,88, 245–289.

Spear, F. S. & Selverstone, J., 1983. Quantitative PT paths fromzoned minerals: theory and tectonic applications. Contribu-tions to Mineralogy and Petrology, 83, 348–357.

Spear, F. S., Hickmott, D. D. & Selverstone, J., 1990. Meta-morphic consequences of thrust emplacement, Fall Mountain,New Hampshire. Geological Society of America Bulletin, 102,1344–1360.

Spear, F. S., Peacock, S. M., Kohn, M. J., Florence, F. P. &Menard, T., 1991. Computer programs for petrologic PT-tpath calculations. American Mineralogist, 76, 2009–2012.

Taylor, W. J., Bartley, J. M., Martin, M. W. et al., 2000. Rela-tions between hinterland and foreland shortening: Sevierorogeny, central North American Cordillera. Tectonics, 19,1124–1143.

Thompson, J. B., 1957. The graphical analysis of mineralassemblages in pelitic schists. American Mineralogist, 42, 842–858.

Wallis, S. R., Platt, J. P. & Knott, S. D., 1993. Recognition ofsyn-convergence extension in accretionary wedges withexamples from the Calabrian Arc and the eastern Alps.American Journal of Science, 293, 463–495.

Wang, P. & Spear, F. S., 1991. A field and theoretical analysis ofgarnet + chlorite + chloritoid + biotite assemblages fromthe tri-state (MA, CT, NY) area, USA. Contributions toMineralogy and Petrology, 106, 217–235.

Wells, M. L., 1997. Alternating contraction and extension in thehinterlands of orogenic belts; an example from the Raft RiverMountains, Utah. Geological Society of America Bulletin, 109,107–126.

Wells, M. L., Dallmeyer, R. D. & Allmendinger, R. W., 1990.Late Cretaceous extension in the hinterland of the Sevierthrust belt, northwestern Utah and southern Idaho. Geology,18, 929–933.

Wells, M. L., Hoisch, T. D., Peters, M. T., Miller, D. M., Wolff,E. D. & Hanson, L. M., 1998. The Mahogany Peaks fault, aLate Cretaceous-Paleocene(?) normal fault in the hinterland ofthe Sevier orogen. Journal of Geology, 106, 623–634.

Wells, M. L., Spell, T. L. & Zanetti, K., 2000a. Laserprobe argondating of strain fringes: Albian orogen-parallel extension inthe hinterland of the Sevier orogen. Geological Society ofAmerica Abstracts with Programs, 32, A-106.

Wells, M. L., Spells, T. L., Arriola, T. & Hoisch, T. D., in press.A mid-cretaceous age for synorogenic orogen-parallel exten-sion in the interior of the sevier orogen constrained bylaserprobe 40Ar ⁄ 39Ar dating of strain fringes. GeologicalSociety of America Abstracts with Programs.

Wells, M. L., Snee, L. W. & Blythe, A. E., 2000b. Dating ofmajor normal fault systems using thermochronology: anexample from the Raft River detachment, Basin and Range,western United States. Journal of Geophysical Research, 105,16,303–16,327.

Wiltschko, D. V. & Dorr, J. A., 1983. Timing of deformation inoverthrust belt and foreland of Idaho, Wyoming, and Utah.American Association of Petroleum Geologists Bulletin, 67,1304–1322.

Yonkee, W. A., 1992. Basement-cover relations, Sevier orogenicbelt, northern Utah. Geological Society of America Bulletin,104, 280–302.

Yonkee, W. A., DeCelles, P. G. & Coogan, J., 1997. Kinematicsand synorogenic sedimentation of the eastern frontal part ofthe Sevier orogenic wedge, northern Utah. Brigham YoungUniversity Geological Studies, 42, 355–380.

Received 23 August 2005; revision accepted 28 July 2007.

SUPPLEMENTARY MATERIAL

The following supplementary material is available forthis article online http://www.blackwell-synergy.com:Table S1. Mineral and calculated bulk composi-

tions.Table S2. Parameters used in garnet growth simu-

lations.Table S3. Monitor parameters, DXgr and DMgar,

used in garnet growth simulations.Table S4. Measured v simulated mineral composi-

tions.Table S5. Initial and final mineral modes from gar-

net growth simulations, compared with visually esti-mated modes.Table S6. P–T path displacements.



Construction of a composite pressureÐtemperatur e path: revealing...

Documents

Transcript of Construction of a composite pressureÐtemperatur e path: revealing...