ORIGINAL PAPER

Kurt J. Ste!en Æ Jane Selverstone

Retrieval of P–T information from shear zones: thermobarometricconsequences of changes in plagioclase deformation mechanisms

Received: 28 April 2005 / Accepted: 31 January 2006 / Published online: 21 March 2006! Springer-Verlag 2006

Abstract Changes in deformation mechanism coupledwith spatial and temporal variations in reaction ratescan result in preservation of disequilibrium mineralcompositions in rocks a!ected by synmetamorphicshearing. Thermobarometric calculations on such rocksmay thus yield meaningless results. We use Garbens-chiefer samples from a shear zone in the Eastern Alps tostudy the e!ects of di!erent deformational processes oncalculated pressures and temperatures in samples thatexperienced the same overall PTt history. We focus onplagioclase, which accommodates strain by a variety ofdeformation mechanisms and is a key mineral in manythermobarometers. Plagioclase that deformed largely viadislocation creep mechanisms shows concentric chemicalzoning, whereas plagioclase that experienced dissolu-tion-precipitation creep preserves complex zoning. Rimcompositions in the latter domains are not necessarilythe youngest compositions, nor did they typicallyequilibrate with other phases in the assemblage. Thetiming of hornblende breakdown reactions relative tochanges in plagioclase deformation mechanism also af-fected chemical zoning. Samples that escaped shearstrain while near the thermal maximum yield internallyconsistent thermobarometric results, whereas samplesthat experienced shearing near the thermal maximum

yield scattered results. Some of the variability in theresults likely represents real di!erences in the P–Tconditions at which equilibration occurred duringdeformation. However, much of the variability repre-sents spurious results obtained by pairing mineralcompositions that were never in equilibrium with oneanother. Extraction of useful P–T information fromsamples that experienced synmetamorphic deformationrequires careful documentation of the relationships be-tween deformation mechanisms and chemical zoning inorder to select appropriate mineral compositions forthermobarometric calculations.

Introduction

In rocks containing zoned minerals, the selection ofappropriate compositions to use in thermobarometriccalculations typically contributes the greatest uncer-tainty to calculated P–T conditions (Kohn and Spear1991). The mechanisms responsible for chemical zoningmust be understood in order to properly select compo-sitions for P–T calculations. The chemical analyses usedin the calculations must represent compositions thatexisted in equilibrium with one another (‘‘local equilib-rium assumption’’), and if possible, the processes thatled to equilibration should also be identified. In somesamples, preserved mineral compositions never equili-brated with one another, and these samples will alwayslead to spurious thermobarometric results.

Development of chemical zoning in minerals solelydue to metamorphic processes can result from changes inexternal conditions (e.g., Tracy 1976; Spear and Selver-stone 1983; Frost and Tracy 1991) or changes in thee!ective bulk composition of the chemical system (Stuwe1997) during mineral growth. Changes in the e!ectivebulk composition during mineral growth may relate tometasomatism (Wintsch and Knipe 1983; Yardley et al.1991) or to the partitioning of components into the rela-tively non-reactive cores of mineral phases, thus changing

Electronic Supplementary Material Supplementary material isavailable for this article at http://dx.doi.org/10.1007/s00410-006-0073-8 and is accessible for authorized users.

Communicated by T. L. Grove

K. J. Ste!en Æ J. SelverstoneDepartment of Earth and Planetary Sciences,University of New Mexico, Albuquerque, NM 87131, USA

K. J. Ste!en (&)ExxonMobil Upstream Research Company,P.O. Box 2189, Houston, TX 77252, USAE-mail: kurt.j.ste!en@exxonmobil.comTel.: +1-713-4316541Fax: +1-713-4317279

Contrib Mineral Petrol (2006) 151: 600–614DOI 10.1007/s00410-006-0073-8

the composition of the system available for chemicalreaction (Tracy 1982; Cherno! and Carlson 1997).

Many studies show that deformation can also a!ectthe rates and mechanisms of chemical equilibration(Brodie 1981; Brodie and Rutter 1985; Koons et al.1987; Tullis and Yund 1991; Yund and Tullis 1991;Tullis et al. 1996; Farver and Yund 1999; Stunitz andTullis 2001; Baxter and DePaolo 2004). For example,changes in plagioclase composition aided by grain-boundary migration recrystallization have been docu-mented experimentally by Tullis and Yund (1991) and innatural samples by Stunitz (1998). Because most meta-morphic rocks exist in a state of partial chemical dis-equilibrium (Carlson 2002), enhanced chemical kineticsrelated to deformational processes could exert a strongcontrol on chemical communication between mineralsand, by extension, on pressures and temperatures cal-culated from thermobarometry.

Processes that increase rates of chemical equilibra-tion during deformation include (Yund and Tullis1991; Tullis et al. 1996): (1) grain-size reduction, whichincreases mineral surface area and decreases meandistances for volume di!usion; (2) motion of high-angle grain boundaries during recrystallization-accommodated dislocation creep; and (3) changes ingrain-boundary fluid distribution, which can enhancestrain accommodation via dissolution and reprecipita-tion. After deformation ceases, due either to changesin the far-field stress field or to rheological changesthat partition strain into other domains, deformation-enhanced rates of chemical equilibration will slow.Local variations in the timing of deformation mayresult in significant di!erences in the P–T conditionsat which equilibration occurs. Thermobarometric cal-culations may thus yield accurate but widely varyingpressures and temperatures for samples that followedthe same PTt path. Strain heterogeneities and changesin deformation mechanism can also lead to variationsin reaction kinetics that allow non-equilibrium mineralcompositions to be preserved. Pressures and tempera-tures calculated from these rocks may be spurious.

The Greiner shear zone in the Tauern Window,Eastern Alps, provides an excellent natural laboratoryfor studying the e!ects of coupled deformation andmetamorphism on calculated P–T conditions. The re-gional PTt history is well constrained by previouswork and provides a reference frame for interpretationof the data. Large local variations in strain-accom-modation mechanisms and preserved textures allowcalculated P–T conditions to be evaluated in samplesthat experienced di!erent deformational processes, butsimilar overall PTt history. This study documents theroles that structural history and plagioclase deforma-tion mechanisms play in controlling mineral equili-bration within the shear zone. Knowledge of thesecontrols is essential to the selection of appropriatemineral compositions for use in thermobarometriccalculations and for interpreting the results of thosecalculations.

Regional geology and previous work

The Greiner shear zone is a subvertical, ENE-strikingzone of concentrated pure and simple shear located inthe western Tauern Window along the Austrian-Italianborder (Fig. 1). The dominant metamorphism anddeformation in the area relate to the Cretaceous-Ter-tiary closure of the Neo-Tethys ocean basin and col-lision of the Eurasian plate with the Adriatic plate.This collision resulted in west- to north-directedemplacement of Adriatic nappes over oceanic andthinned continental margin rocks now exposed in theTauern Window. Subsequent transpressional motion ofthe Adriatic plate led to east–west extension of over-thickened crust and exhumation of the Tauern Win-dow (Selverstone 1988; Ratschbacher et al. 1991;Lammerer and Weger 1998).

There are three major lithotectonic packages presentin the Tauern Window (Morteani 1974; Selverstone et al.1984). The structurally lowest, the Hercynian-age Zen-tralgneis (ZG), consists of granodiorite and tonalite thatdisplay textures ranging from igneous to gneissic.Structurally above the Zentralgneis is the Lower Schi-eferhulle (LSH), which consists of a lower, autochtho-nous, Paleozoic sequence (PLSH) and an upper,parautochthonous sequence that includes rocks ofMesozoic and Tertiary age (MLSH). The PLSH pre-dates emplacement of the Zentralgneis and is domi-nantly composed of interleaved aluminous amphibolite(Garbenschiefer), biotite schist, and graphitic schist. TheMLSH consists of metaconglomerate, quartzite, marble,and pelitic and calcareous schists that have correlativesthroughout the Alps. The structurally highest package,the Upper Schieferhulle (USH), is an allochthonous se-quence of greenstone, marble, calcareous and peliticschists derived from the Neo-Tethys basin.

The Greiner shear zone is located in synformal po-sition between two basement (ZG)-cored antiforms andis defined by an ENE-striking subvertical foliationformed by extreme transposition of earlier fabrics. TheGreiner shear zone cuts all structural levels, indicatingthat it was active after Alpine nappe emplacement, al-though some earlier motion cannot be ruled out.Metamorphic grade within the synform increases fromwest (upper greenschist) to east (amphibolite facies).

This study primarily focuses on coarse-grained,hornblende Garbenschiefer horizons within the PLSH.The hornblende Garbenschiefer is composed of large (upto 20 cm in length), radiating hornblende porphyroblaststhat typically cross both the lineation and the foliation,although in some horizons hornblende lies within folia-tion planes and has grown or been rotated to define alineation. Typical mineral assemblages include horn-blende + plagioclase + quartz + epidote + ilmenite ±garnet ± biotite ± white mica ± chlorite ± stauro-lite ± kyanite ± rutile (see Selverstone et al. 1984, fordetails). Garbenschiefer is most abundant in the PLSH,but also occurs in some areas of the MLSH and highly

601

strained domains in the ZG. In the PLSH,Garbenschieferis interlayered with foliated biotite schist at the scale ofcentimeter to meter. Small amounts of fine-grainedamphibolite that lack radiating bundles of hornblendealso are present in the PLSH; these typically also lackthe aluminous minerals (garnet, staurolite, kyanite)that characterize many Garbenschiefer horizons.Fine-grained amphibolite is most abundant along thenorthern margin of the PLSH portion of the Greinershear zone where a large strain gradient exists between thePLSH and the relatively undeformed ZG north of theshear zone.

Hornblende-consuming reactions occurred locallywithin some portions of the Greiner synform and shearzone, as evidenced by the presence of biotite-richpseudomorphs after hornblende. The pseudomorphsrecord reaction between hornblende and white mica toproduce biotite + plagioclase + H2O ± garnet ±staurolite ± chlorite in rocks with relatively low aH2O

(Selverstone and Munoz 1987). In some areas, strainlocalization during or after pseudomorph developmentresulted in pervasive transformation of Garbenschieferinto biotite schist (Selverstone 1993; Ste!en et al.2001).

Quantitative P–T paths from PLSH samples in thewestern and central parts of the study area show nearlyisothermal decompression from 10 to !4 kbar at!550"C (Selverstone et al. 1984; Selverstone 1985).

These paths were based in part on data from zonedminerals in kyanite- and staurolite-bearing hornblendeGarbenschiefer from the Greiner zone. Maximum tem-peratures increase by about 50"C towards the eastern-most sample localities shown in Fig. 1.

The Selverstone et al. (1984) study focused on Gar-benschiefer with no obvious pseudomorphs after horn-blende. The samples in that study were selected in partfor their lack of obvious shear fabrics, although sub-sequent work (Ste!en et al. 2001) has shown thatdevelopment of the Garbenschiefer texture is favored bysynmetamorphic shearing. Rim thermobarometry andGibbs method results obtained by Selverstone et al.(1984) showed a high degree of consistency amongsamples. Owing to the presence of aluminous mineralssuch as staurolite and kyanite, the samples used bySelverstone et al. (1984) had unusually low thermody-namic variance for amphibolites. Arnold et al. (2000)demonstrated that coexistence of these aluminous pha-ses with hornblende relates both to compositional fac-tors and to the region of P–T space traversed by therocks. In general, the kyanite stability field for ‘‘alumi-nous’’ amphibolites expands to a wider range of bulkcompositions at higher pressures. However, it is impor-tant to note that the model of Arnold et al. (2000) doesnot include potassium, which precludes the consider-ation of the biotite and white mica equilibria that are ofcritical importance in the Greiner shear zone.

Schlegeis 2500

2000

2500

2500

3000

2500

2500

2500

2500

3000

2000

2000

2000 Austria

Italy

Ochsner

Berliner

Schwarz SeeArea

7020

00

7050

00

7080

00

7110

00

7140

00209000

212000

206000

Hochsteller99-5-1

RotbachlspitzePfitscherJoch

FHSamples

Furtschaglkopf

BH Samples

Alpenrose

GrosserGreinerPJ-105-98

KleinerGreiner

FurtschaglspitzeFurtschaglhaus

Schonbichler Horn

Innsbruck

Sterzing

Northern Calcareous Alps

Mayrhofen

47 N

12 E 13 E

ItalyAustria

Southern Alps

(a)

(b)

Austroalpine

Regional Key

Lower SchieferhulleUpper Schieferhulle

ZentralgneisN25 km

NLocality Key

1 km

Int BoundaryRoadContourWaterTrailUTM Coord

HutPeak

Greiner SZ

Fig. 1 a Map showing location of the Tauern tectonic window in the eastern Alps. b Detail of study area showing extent of Greiner shearzone and sample localities

602

For this study, we sampled a much broader array oftextural types than those considered by Selverstone et al.(1984), including stable hornblende Garbenschiefer,samples with varying degrees of pseudomorph forma-tion, and biotite schists in which most traces of the priorGarbenschiefer assemblage and texture were destroyed.Our intention was to obtain a wide range of deforma-tional features and mineral assemblages in order toevaluate the influence of di!erent deformational andmetamorphic processes on calculated P–T conditions.Most samples have a higher thermodynamic variancethan the Selverstone et al. (1984) suite, and the samplesalso exhibit a wider variety of deformation mechanismsand chemical zoning features.

Deformation history and rheological heterogeneityin the Greiner shear zone

Di!erent textures and lithologies are preserved in dis-continuous lenses that are centimeters thick and extendalong strike for centimeters to tens of meters within thePLSH portions of the shear zone (Fig. 2a). The pene-trative foliation generally conceals evidence of earlyfolding in the PLSH, but rare folded quartz veins(Fig. 2b) are locally apparent (e.g., FH sample area,Fig. 1b). These quartz veins are characterized by

upright, open to tight folds with axes that plunge shal-lowly to the west, subparallel to the dominant Greinershear zone lineation; the axial plane of the folds iscoincident with the shear foliation. Folded PLSH,including Garbenschiefer (Fig. 2c), is locally exposed inlow-strain areas of the Greiner shear zone in the FHsample area and also to the ESE of Pfitscher Joch(Fig. 1b). Fold orientations in these localities are similarto those observed in the quartz veins; these folds arecorrelative with regional F3 folding identified by Lam-merer (1988).

Two di!erent groups of folds occur in the eastern-most PLSH exposures in the Greiner synform (SchwarzSee area, Fig. 1b). The dominant group is upright, withaxial planes that strike ENE and fold axes that plungeshallowly to the west (Fig. 2d). The similarity in orien-tation and style of these folds to those described aboveindicates that these are F3 folds. Tight, recumbent F2

folds with N- or S-plunging axes are only locally presentand are refolded by the F3 folds.

Initial development of the Garbenschiefer assemblageand texture occurred in some domains during D2 and inother domains during formation of the later, highlytransposed D3 deformation (Fig. 2a, c). Once the Gar-benschiefer texture developed in a given horizon, sub-sequent strain was partitioned into weaker areas, such asbiotite-rich domains and regions of pseudomorph

Fig. 2 Field photos ofdiachronous Garbenschieferdevelopment. a Elongate bandsof hornblende Garbenschieferinterspersed with biotite schisthorizons in S3. This is thedominant texture withinGreiner shear zone. b Quartzveins showing upright F3 folds,in contrast to transposedfoliation in surroundingGarbenschiefer and biotiteschist. c Garbenschieferpreserved in foliation that wasfolded by F3. d Garbenschieferoverprinting hinge of uprightD3 folds. Black circle is 10 cmin diameter

603

formation (Ste!en et al. 2001). Episodes of chemicalreaction and equilibration during deformation thus donot necessarily represent event markers throughout theentire shear zone: rocks with similar textures that are inclose proximity to one another could have equilibratedunder di!erent external conditions. This potential fordiachronous equilibration as a function of strain local-ization greatly complicates the interpretation of ther-mobarometric results from shear zones that were activeunder changing P–T conditions.

Plagioclase deformation mechanisms and developmentof chemical zoning

Plagioclase types

In the Greiner PLSH samples, plagioclase plays a keyrole in recording interactions between deformationaland metamorphic processes, and we focus on it for thefollowing reasons: (1) Plagioclase is abundant in all of

Fig. 3 Photomicrographs, BSE images and quantitative microprobe traverses for plagioclase. a Type A plagioclase. b Type B plagioclase.c Type C plagioclase

604

the relevant rock types (Garbenschiefer, pseudomorph-rich rocks, and interlayered biotite schists). (2)Plagioclase constitutes an important phase in manythermobarometers. (3) Plagioclase compositions andzoning patterns provide data on the metamorphicconditions under which plagioclase grew and/or equili-brated. (4) Much is known about the mechanisms bywhich plagioclase deforms under both experimental andnatural conditions, thus allowing us to tie the meta-morphic history to the rheologic evolution of theGreiner zone.

Based on textural and compositional characteristics,plagioclase from the Greiner zone was separated intothree di!erent types:

Type A plagioclase is coarse-grained (‡100 lm) andshows radially concentric zoning from core composi-tions of An5–20 towards rim compositions of An25–33(Fig. 3a). Optically visible subgrains are evident in somegrains. In most samples, Type A plagioclase exhibitsa prominent shape-preferred orientation and a moder-ately well-developed lattice-preferred orientation (asobserved with a gypsum plate). Type A is the dominantplagioclase type in some hornblende Garbenschiefersamples and is also locally present in pseudomorph-bearing samples and in biotite schist domains.

Type B plagioclase is typified by complexly zonedgrains (Fig. 3b). Grain size varies widely (20–200 lm),but is generally less than 50 lm. Compositions rangefrom An10 to An40, but zoning is non-concentric withsharp truncations across which compositions can changeby 10–15 mol% over distances of 5–10 lm within indi-vidual grains. Compositional heterogeneity is alsoinvariably present between neighboring grains. Thisplagioclase type shows little to no shape-preferred ori-entation and no lattice-preferred orientation, and hasvery low dislocation densities (Ste!en et al. 2001). TypeB plagioclase is common in Garbenschiefer horizons andlocally in biotite schists within the Greiner shear zone.

Type C plagioclase shows a wide range in grain sizes(20–300 lm) and is generally more An rich (An25–40)than the other types (Fig. 3c). Many type C grains arenearly homogeneous, but others show roughly concen-tric zoning. Type C plagioclase shows a moderatelywell-developed lattice-preferred orientation, but littleshape-preferred orientation. This variety of plagioclaseoccurs only in pseudomorph-bearing Garbenschieferand in areas where hornblende Garbenschiefer hasundergone nearly wholesale conversion to biotite schist.

Textural and chemical evolution of plagioclase

We hypothesize that the three plagioclase types found inthe Greiner shear zone formed by di!erent combinationsof deformational and metamorphic processes, as illus-trated in Fig. 4. The formation of type A plagioclaselikely was dominated by prograde metamorphic reac-tions. In intermediate to mafic bulk compositions, pla-gioclase with concentric zoning and increasing anorthite

content from core to rim is consistent with plagioclasegrowth during heating and/or decompression into theamphibolite facies (Thompson et al. 1982). Althoughevidence of climb-accommodated dislocation creep(Hirth and Tullis 1992) exists in type A plagioclase,deformation apparently played a relatively minor role inthe initial development of this texture and its associatedchemical zoning (Fig. 4a).

In contrast, formation of type B plagioclase wasdominated by dissolution-precipitation creep (DPC)following grain-size reduction (Ste!en et al. 2001). TheDPC regime is favored by small grain size, the presenceof a grain-edge fluid phase, and relatively low di!erentialstress (Kerrich et al. 1977; Tullis and Yund 1991; Tullis

overgrowth fromnet-transfer rxn

asymmetricdissolution

dissolutionand uniformovergrowth

concentric growth zoning

cataclasis and/ordynamic recrystallization

Regime 3 dislocation creepor incipient GBDC

high and/or lowdislocation density

repeated dissolutionand reprecipitation

Deformation dominated:dissolution-precipitation creep

Reaction dominated

= measured "rim" compositions

Type BType C

concentric growth zoning

Regime 2 dislocation creepcore-and-mantlestructure concentric zoning

preserved

hi Nalo Na

subgrains Type A

Type A

hi Nalo Na

(a)

(b)

Fig. 4 Model of plagioclase evolution in the Greiner zone. a TypeA plagioclase develops via growth zoning. Concentric zoning ispreserved through modest degrees of climb-accommodated dislo-cation creep, and rim compositions are homogeneous. b Grain-sizereduction of type A plagioclase occurs via either cataclasis ordynamic recrystallization and exposes old interior compositions onrims of new grains. Fine grain size can result in strain accommo-dation by dissolution-precipitation creep. Ongoing DPC results in arange of rim compositions, and youngest composition cannot easilybe identified. Net-transfer reactions can accompany DPC andproduce both new type C grains and homogeneous overgrowths oftype C plagioclase on older cores. Black dots represent rimcompositions that might be measured for P–T calculations; largecompositional range will produce a similarly large range incalculated P and T

605

et al. 1996), and results in dissolution of material inhigh-stress areas and reprecipitation in low-stress areas.Nearly dislocation-free grains with complex chemicalzoning (Fig. 4b) and no lattice-preferred orientation areproduced by this process (Tullis and Yund 1991). Dis-solution and reprecipitation typically lead to develop-ment of complex chemical zoning patterns involvingtruncations and localized overgrowths. Experimentalstudies (Tullis and Yund 1991; Yund and Tullis 1991) ofDPC show precipitation of plagioclase of variable andintermediate compositions in an originally bi-composi-tional system, and data from three natural shear zonesshow similar compositional e!ects of DPC (Ste!en et al.2001; Wintsch and Yi 2001; Baxter and DePaolo 2004).

During active DPC, changes in plagioclase composi-tion depend on a combination of metamorphic anddeformational factors (Tullis and Yund 1991; Tullis et al.1996; Wintsch and Yi 2001; Baxter and DePaolo 2004).In rocks with concentrically zoned original plagioclasegrains (e.g., type A), grain-size reduction via any mech-anism (cataclasis, grain-boundary bulging, or grain-boundary migration recrystallization) exposes interiorcompositions that were previously chemically isolated byslow intragranular di!usion rates. Dissolution and rep-recipitation thus can produce asymmetric chemical zon-ing patterns composed of original core and rimcompositions, as well as the full range of intermediatecompositions. Many of these compositions will be out ofequilibrium with all other phases in the rock (Fig. 5), andhave the potential to provide chemical energy to drivereaction and neocrystallization. DPC may thus increase

the kinetics of metamorphic reactions and lead to localproduction of new plagioclase of intermediate composi-tion (Yund and Tullis 1991). Because this new plagio-clase composition decreases the local chemical energy ofthe system (Fig. 5), its precipitation represents a form ofchemically driven GBM. Thus, DPC can change pla-gioclase composition via a combination of processesdriven by both strain energy and chemical energy.

The formation of type C plagioclase is dominantly dueto metamorphic reaction accompanying chemically dri-ven GBM (Fig. 4b). Type C plagioclase occurs only inareas where hornblende was partially or totally consumedby reactions that converted Al- andK-rich portions of theGarbenschiefer into garnet-biotite schist (hbl + whitemica fi gar + bio + plag ± epi ± chl ±st ± H2O;Selverstone and Munoz 1987). The stoichiometry andP–T conditions of this reaction in any sample are func-tions of bulk rock composition, which varies over shortdistances in the Greiner shear zone. In all cases, however,plagioclase is a reaction product. The weak lattice-pre-ferred orientation of plagioclase suggests nucleation onpreexisting grains (Fig. 4b), and the absence of a shape-preferred orientation is consistent with field and petro-graphic observations that strain was localized into thenew biotite-rich horizons during reaction rather than intoplagioclase-rich horizons.

The primary method of material transport and strainaccommodation during chemically driven GBM andDPC is di!usion along and across grain boundaries.Grain-boundary di!usion is a rate-limiting factor forchemical reaction in most metamorphic rocks (Kretz1966, 1973; Carlson 2002). Within low-strain domains orbetween deformation episodes, slow reaction kineticsmay prevent the chemical system from fully equilibrating.Stored chemical energy (a"nity or DG) may build due toP–T changes during these periods. At appropriate dif-ferential stresses, increased grain boundary di!usionrates will allow strain accommodation by DPC andchemical reaction due to chemically driven GBM. Be-cause of the chemical a"nity accumulated in the system,initial reaction rates will be relatively rapid (reaction rateis proportional to a"nity for the same kinetic pathway)but will decrease as the system equilibrates. Althoughstrain initially would be required to provide increaseddi!usion rates and reaction kinetics, GBM may bedominantly driven by chemical a"nity stored in the sys-tem (Fig. 5). As this stored chemical a"nity is consumed,GBM likely would be increasingly driven by DPC.

Retrieval of pressure–temperature data

Methods

Mineral compositions were analyzed with the JEOL 733electron microprobe in the Department of Earth andPlanetary Sciences, University of New Mexico, usingoperating conditions of 20 keV and 20 nA. A finely fo-cused (!1 lm diameter) beam was used for all analyses,

chem

ical

ene

rgy

composition

originalrim

phase 1

phase 2

heterogeneousequilibrium

original core -now exposedvia grain-size

reduction

new compositionof phase 1

during GBDC

spurious“equilibrium”

Fig. 5 Schematic chemical energy vs. composition diagram show-ing the e!ect of grain-size reduction on local equilibrium in asample. Rim compositions of phases 1 and 2 are initially inequilibrium with one another. Grain-size reduction exposesdisequilibrium core compositions of phase 1. During DPC,dissolution and reprecipitation of phase 1 produces a new, lowerenergy composition. This composition will be out of equilibriumwith phase 2, a mineral that does not undergo DPC. Pairing of thenew composition of phase 1 with phase 2 will thus result in spuriouscalculated P–T conditions

606

except for plagioclase in which a defocused (!10 lmdiameter) beam was used to minimize Na di!usion un-der the beam. Natural and synthetic oxides were used asstandards and sample compositions were corrected formatrix e!ects using ZAF procedures. Mineral compo-sitions used in the calculations below are available aselectronic supplementary material.

Due to the complexity of plagioclase zoning patternsinitially observed using optical microscopy, quantitativeelectron microprobe traverses were conducted for pla-gioclase in all samples. The data from these traversescombined with qualitative information derived fromBSE images allowed individual plagioclase grains to becategorized into the types described above.

Bulk-rock chemistry (major elements only) wasdetermined using a Rigaku RIX 2100 XRF at theDepartment of Earth and Planetary Sciences, Universityof New Mexico. Representative samples were crushed inagate jars in a mixer mill and a mullite mortar andpestle, sieved, and fused into glass disks.

Thermobarometric calculations were performed on 9sample regions containing varying mineral textures,plagioclase types, and assemblages (Table 1). Thepurpose of this study is neither to evaluate the accuracyof the thermobarometric calibrations nor to place exactconstraints on the P–T history of the Greiner zone;rather, we seek to determine how variations in calculatedP–T conditions reflect the interplay between deforma-tional and metamorphic processes in deep-seated shearzones. In order to explore this variability, we used fourdi!erent calibrations/approaches: (1) garnet-hornblendethermometry (Graham and Powell 1984), (2) garnet-hornblende-plagioclase-quartz barometry (Kohn andSpear 1989), (3) garnet-biotite-muscovite-plagioclase-quartz barometry (Ghent and Stout 1981), and (4) theAX/Average P–T method incorporated in THERMOCALC

(Powell and Holland 1994; Powell et al. 1998). Thecalibrations were applied to all sample regions thatcontained appropriate assemblages.

In samples that experienced dissolution-precipitationcreep of plagioclase (type B), it is generally not possible todetermine which composition, if any, equilibrated withother phases in the sample. The selection of a plagioclasecomposition (Fig. 4b) thus introduces a significant errorinto the thermobarometric calculations. In order to

determine the maximum magnitude of the ‘‘plagioclasee!ect’’ on retrieved pressure–temperature values, allthermobarometric calculations were performed usingfour di!erent plagioclase compositions for each samplearea: (1) themost albitic composition observed (AbMax),(2) the most albitic composition observed on a rim (AbRim), (3) the most anorthitic composition observed on arim (An Rim), and (4) the most anorthitic compositionobserved (AnMax). These plagioclase compositions wereselected for analysis because they span the total range ofcompositions present in the samples and also representthe analyses typically used in thermobarometry (rimcompositions).With the exception of garnet, other phasespresent in the samples show little chemical zoning. Garnetcompositions used in the calculations were collected fromnear-rim locations to avoid the e!ects of retrogression.

Thermobarometric results as a function of plagioclasetype and rock texture

Mineral compositions from three groups of samples wereused in the thermobarometric calculations in order toassess the influence of textural development and plagio-clase type on apparent P–T conditions: (1) Garbens-chiefer with type A plagioclase (Fig. 6a–c); (2)Garbenschiefer with type B plagioclase (Fig. 6d–f); and(3) Garbenschiefer containing pseudomorphs afterhornblende and associated type C plagioclase (Fig. 6g–i).The first group is most similar to the samples used bySelverstone et al. (1984) in that hornblende is stable andplagioclase is concentrically zoned. However, samplesused here lack epidote, kyanite, and staurolite, givingthem a higher thermodynamic variance than the samplesused by Selverstone et al. (1984).

The relative thermobarometric results for the fourdi!erent plagioclase compositions vary by plagioclasetype (Fig. 6). In samples containing concentricallyzoned, type A plagioclase, the Ab rim, An Rim, and AnMax results cluster together because these compositionstend to be similar; the Ab max composition in thesegrains typically corresponds solely to mineral cores, andhence yields very di!erent P–T results. Calculationsusing type B plagioclase yield widely scattered results(e.g., Fig. 6g) because dissolution and reprecipitationduring DPC resulted in a large range in plagioclase rim

Table 1 Assemblage, texture and plagioclase type

Sample Plag type Texture Hb Gar Mu Pa Bio Q Pl Chl Ep St

BH-00-86-1H2 A–B GS X X X X X XBH-00-86-1K2 A GS X X X X X XFH-00-89-1F C PM X X X X X X X XFH-00-810-1A A–B GS X X X X X X X XFH-00-810-1H C PM X X X X X XFH-00-810-1K2 A GS X X X X X XFH-00-810-1L1 C PM X X X X X X X XPJ-105-98 A GS X X X X X X99-5-1 B GS X X X X X

GS Hornblende Garbenschiefer, PM pseudomorphs after hornblende Garbenschiefer texture

607

compositions, even in adjacent grains. Type C plagio-clase also produces some scatter in the calculated P–Tconditions (Fig. 6g, i); although most type C grains arerelatively homogeneous, some include irregular over-growths of pseudomorph-related plagioclase on preex-isting grains, leading to a larger compositional range.

For calculations using a single thermometer orbarometer, the selection of plagioclase compositionswithin a single sample a!ects the position of the resultingKD lines by as much as 2 kbar and 50"C. Plagioclasecomposition has a somewhat smaller e!ect on conditionscalculated from the Average P–T multi-equilibriummethod (Powell and Holland 1994). The additional con-

straints provided by equilibria that do not involve pla-gioclase in the latter method tend to decrease the e!ects ofthe extremes in plagioclase composition. Although thevariation in calculated P–T conditions is significant foreach sample, the total variation in calculated P–T condi-tions for the Greiner shear zone is much larger. For singleequilibrium calibrations, calculated conditions span arange of 6 kbar and 175"C. For multi-equilibrium calcu-lations, the variation is approximately 4 kbar and 100"C.

Separation of the thermobarometric data by locality,plagioclase type, or assemblage does not immediatelyresolve the wide scatter in P–T results into obviousgroups. Limited and di!erential equilibration of

23

4

5

6

7

8

9

10

11

12

Temperature°C

1

3

Pre

ssur

e (k

bar)

Pre

ssur

e (k

bar)

Pre

ssur

e (k

bar)

450 500 550 600 650 700 750 800400

13

3

1

3

2

(a) PJ-105-98

2

3

4

5

6

7

8

9

10

11

12

3

4

5

6

7

8

9

10

11

12

3

1(d) 99-5-1

3

1

3

1

(e) FH-00-810-1A

Garbenschiefer Plag type B

(b) FH-00-810-1K2

(f) BH-00-86-1H2

3

(c) BH-00-86-1K2

Garbenschiefer Plag type A

400 450 500 550 600 650 700 750 800

Temperature°C

3

1

1

3

3

1

(h) FH-00-810-1L

Pseudomorphs Plag type C

(g) FH-00-89-1F

(i) FH-00-810-1H

400 450 500 550 600 650 700 750 800Temperature°C

Ab MaxAb RimAn RimAn Max

Fig. 6 Thermobarometric results obtained from Garbenschiefersamples containing type A (a–c) and type B plagioclase (d–f), andpartially pseudomorphed Garbenschiefer containing type C pla-gioclase (g–i). a and d are from westernmost sample localities, and cand f are from easternmost localities. Shaded rectangles in a and cshow final equilibration conditions determined by Selverstone et al.(1984) for western localities. Thermobarometers: 1 hornblende-plagioclase (Holland and Blundy 1994), 2 garnet-biotite-muscovite-

plagioclase-quartz (Ghent and Stout 1981), 3 hornblende-garnet-plagioclase-quartz (Kohn and Spear 1989). Ellipses = AveragePT/AX method (Powell and Holland 1998). Ab Max, AnMax = the most albitic and most anorthitic plagioclase compo-sitions observed anywhere in a sample. Ab Rim, An Rim = themost albitic and most anorthitic rim compositions observed in asample

608

individual samples at di!erent stages in the deformationand P–T history of the Greiner zone certainly a!ectedboth mineral chemistry and calculated thermobaromet-ric results, but not in a clear-cut fashion. Below, we usethermodynamic modeling to predict average changes inmineral assemblage, modal abundance, and mineralcompositions, and then use this information to identifythe features that enhance or inhibit our ability to retrievemeaningful P–T data from sheared samples.

Thermodynamic modeling

Predicted equilibrium assemblages for a range of PLSHcompositions over the previously determined PTt path(Selverstone et al. 1984) were determined using apseudosection approach (Ste!en 2004). Bulk composi-

tions for two Garbenschiefer samples (one mafic, oneintermediate to felsic), a sample with abundant pseud-omorphs after hornblende, and a biotite schist were used.Endmember thermodynamic data were taken from Hol-land andPowell (1998) and theCORKmodelwas used forH2O (Holland and Powell 1991). Activity-compositionmodels and bulk compositions are listed in the Appendix.

Figure 7a is a pseudosection for the pseudomorph-bearing sample composition, which generally lies be-tween the compositional extremes. The gray lines showthe P–T path determined by Selverstone et al. (1984)from Gibbs method calculations, inclusion and mineralrim thermobarometry, and fluid-inclusion data. TheP–T path used in our thermodynamic model wasshifted up-temperature by !35"C relative to theSelverstone et al. (1984) path in order to obtain abetter match between observed assemblages and the

0 0.1 0.2 0.3 0.4 0.54

5

6

7

8

9

10

P(k

bar)

Mode Plagioclase

4

5

6

7

8

9

10

P(k

bar)

0 0.05 0.1 0.15 0.2 0.25 0.3 0.350

20

40

60

80

100

120

Num

ber

of A

naly

sis

Anorthite Content

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Anorthite Content

450 500 550 600 650 7002

3

4

5

6

7

8

9

10

11

12Pa Q Pl Gl Mu Hb

Q Pl G M

u Hb

Q Pl G Mu Hb

Pa Q Pl Chl Mu Hb

Q Pl G Bio Mu

Q Pl G BioQ Pl Chl Mu Hb

Q Ep Pl Chl Mu Hb

Q Pl Bio Chl Mu

Q Pl Bio

ChlQ Pl Crd G Bio

Q Pl G BioChl

Q Pl G Bio

Chl Mu

Q Pl G Bio Chl

Mu Hb

Pa Q GPl Chl Mu Hb

Pa Q PlG Mu Hb

Q Pl Crd

G Bio Chl

0 0.1 0.2 0.3 0.4 0.5 0.64

5

6

7

8

9

10

Mode Hbl

0 0.05 0.01 0.154

5

6

7

8

9

10

Mode Garnet

P(k

bar)

P(k

bar)

P(k

bar)

99-4-1 (GS)99-5-1 (GS)PJ-107H (PS)

99-2-1 (Bio)

99-4-1 (GS)99-5-1 (GS)PJ-107H (PS)

99-2-1 (Bio)

99-4-1 (GS)99-5-1 (GS)PJ-107H (PS)

99-2-1 (Bio)

99-4-1 (GS)99-5-1 (GS)PJ-107H (PS)

99-2-1 (Bio)

T ˚C

55 Ma 34 Ma

30 Ma

Q Pl Bio Chl M

u Hb

(a) (b)

(d)

(f)

(c)

(e)

Fig. 7 Thermodynamicmodeling results.a Pseudosection produced usingbulk composition frompseudomorphed Garbenschiefersample PJ107-H. b Heavy blackline is P–T path used for theassemblages calculated in b–e.Heavy gray line is P–T pathdetermined by Selverstone et al.(1984); dates are fromChristensen et al. (1994).b–d Calculated modes of garnet,hornblende, and plagioclase.e Calculated plagioclasecomposition for all fourrepresentative bulkcompositions. f Histogram ofplagioclase compositionscollected from 22 sampletraverses (697 total analyses)

609

thermodynamic results (see below). The model predictsgarnet growth over a pressure range (10–6 kbar;Fig. 7b) similar to that previously determined byGibbs method modeling in the PLSH (10-7 kbar,Selverstone et al. 1984). Hornblende mode decreasesalong this P–T path (Fig. 7c) for all samples. Themode of hornblende decreases to less than 3% for thebiotite schist and pseudomorph assemblages; continueddecompression would decrease the mode of hornblendeto zero for these two samples. Because the hornblende-consuming reactions are dehydration reactions, theirposition in P–T space is sensitive to aH2O; decreasingaH2O (fixed at unity for these calculations) would shiftthese reactions to higher pressures and lower temper-atures and would decrease the mode of hornblende forall samples.

The equilibrium plagioclase compositions (Fig. 7e)predicted by the model become increasingly anorthiticwith decreasing pressure. The predicted increase in pla-gioclase mode along this P–T path would favor thedevelopment of growth zoning, with individual grainsgenerally becoming zoned from sodic to more calciccompositions during decompression due to hornblendeconsumption. These predicted compositions can becompared with all plagioclase data from the naturalsamples in Fig. 7f (697 spot analyses collected alongrim-core-rim plagioclase traverses in 22 di!erent sampleregions; note that although there are modes in the his-togram, there is also continuous variation in plagioclasecomposition and no evidence in the data for a solvus).For type A and type C plagioclase, core compositionsare always more sodic than the rims; truncations in typeB plagioclase grains make it di"cult to interpret therelative ages of di!erent compositions. Nonetheless,there is good agreement between the actual plagioclasecompositions and the trends predicted by Fig. 7d and efor each of the di!erent modeled bulk compositions. Thewide range in plagioclase compositions within eachsample can thus be largely ascribed to di!erentialequilibration and development of growth zoning duringdecompression.

In summary, the thermodynamic modeling indicatesthat hornblende abundance should decrease, plagioclasemode should increase, and plagioclase should becomeincreasingly calcic along a P–T path similar to thatdetermined by Selverstone et al. (1984). Plagioclaseproduced by reaction should thus become more anor-thitic with time. However, DPC will tend to expose coresof previously zoned grains, and hence will lower theanorthite content to a value that is less than the pre-dicted equilibrium composition. These results are con-sistent with plagioclase zoning patterns observed in theGreiner shear zone. Because all relevant chemical reac-tions produce increasingly anorthitic plagioclase, theappropriate composition to use in thermobarometriccalculations in these rocks is the most anorthitic one,even if that composition is not a rim composition. Forsamples that were not significantly a!ected by DPC,plagioclase is concentrically zoned and the An Max

composition is similar to the rim compositions thatwould typically be used for thermobarometry.

Multi-equilibrium results derived from An Max values

The An Max plagioclase composition in each samplearea was used to calculate P–T conditions for stableGarbenschiefer and partially pseudomorphed samplesvia the Average P–T routine in THERMOCALC (Powell andHolland 1994; Fig. 8). The calculated conditions spanthe range calculated by Selverstone et al. (1984), thoughshifted to higher temperatures (likely reflecting the use ofdi!erent thermodynamic datasets). Although there isconsiderable scatter in the data overall, the results showreasonable overlap for each rock type. In particular,they show considerably less variability than the resultsshown in Fig. 6, which utilized a range of di!erentplagioclase compositions. The largest error ellipse,however, remains associated with type B plagioclase.

In general, if the fit/fit cuto! ratio for the multi-equilibrium calculations exceeds 1.0, the calculated P–Tconditions are not internally consistent for the selectedactivity and enthalpy uncertainty values (Powell andHolland 1994). Fit/fit cuto! values between 1.0 and 1.5for the Greiner samples indicate that the mineral com-positions used in this study generally yield Average P–Tvalues that are not internally consistent. Increasing theuncertainties in the activity (calculated in AX) and en-thalpy input would make the calculations consistent, butwould also increase the size of the uncertainty ellipse.

GS type BGS type A

PS type C

400 8007006005002

10

8

6

4

12

Pre

ssur

e, k

bar

Temperature, oC

Fig. 8 Results of multi-equilibrium calculations (Average PT;Holland and Powell 1994) using just An Max compositions. GSGarbenschiefer with type A or B plagioclase; PS samples withpseudomorphs after hornblende that contain type C plagioclase.Ellipses for type A plagioclase are shaded for ease of reading. Notelarge ellipses associated with type B plagioclase, and small ellipsesbut scattered P–T results for samples containing pseudomorphs;the former represents disequilibrium, whereas the latter representsequilibration at di!erent sets of P–T conditions. Arrowed P–T pathis from Selverstone et al. (1984)

610

There are two potential sources for the poor con-sistency recorded by these samples: (1) The endmemberthermodynamic data of Powell and Holland (1998) andthe activity-composition models used by AX may besignificantly in error. Although possible, we considerthis explanation unlikely: other studies have applied thesame endmember data and activity-composition modelsto samples with similar assemblages and obtainedconsistent results (Puga et al. 2000; Dale and Holland2003). (2) Disequilibrium between minerals in thesamples also can lead to inconsistency. The generallyaccepted approach to evaluating the role of disequi-librium in multi-equilibrium thermobarometry is toremove activities of certain endmembers from consid-eration using statistics provided by the program (seePowell and Holland 1994, for details). However, in oursamples, removing various endmembers from consid-eration did not significantly improve the consistency ofthe results. In particular, no specific endmember couldbe identified whose activities contributed the largestportion of the inconsistency. These results indicate thatthe samples used in this study generally fail theassumption of local equilibrium, rather than recording

the metastable persistence of a single phase. Relativetiming of equilibration, likely related to the relativetiming of strain localization events, appears to haveexerted the dominant control on chemical equilibra-tion. Although all samples in the Greiner shear zoneexperienced the same overall P–T history, the calcu-lated P–T conditions were frozen in at very di!erenttimes in di!erent sample areas, depending on localvariations in bulk composition, reaction rates, fluidmobility, and mechanisms of strain accommodation(e.g., Fig. 4). In addition to choosing appropriatemineral compositions for thermobarometry withinindividual samples, it is thus also necessary to choose asample suite recording a wide range of metamorphicand deformational features when attempting to retrieveP–T information from shear zones.

Linked metamorphism and deformation in the Greinershear zone

In horizons where coarse hornblende developed rela-tively early, the interlocking hornblende grains caused

Type C plag:reaction overgrowthsand/or new grains

hbl growth

hbl to bio rxn

Type B plag:grain-boundarydiffusion creep

(b) Shearing accompanies prograde metamorphic reactions

Type A plag:growth dislocation creep

Pre

ssur

eP

ress

ure

Temperature

Temperature

(a) Shearing precedes prograde metamorphic reactions

(c) Pseudomorph reaction: strong hbl replaced by weak bio

"static" metamorphism

strain partitioned elsewhere

deformation

syn-kinematic metamorphism

calculated P-T = spurious

calculated P-T = conditionsof hbl breakdown reactionin different domains

calculated P-T = actual "peak" equilibration conditions

Fig. 9 Schematic diagramsshowing consequences ofdi!erences in relative timing ofprograde metamorphism,deformation, changes inplagioclase deformationmechanism, and hornblendebreakdown for calculated P–Tconditions. a Samples thatunderwent early shearingfollowed by growth of type Aplagioclase yield consistentinformation on thermal peak ofmetamorphism. b Samples thatexperienced strainaccommodation via grain-boundary di!usion creep nearthermal maximum yieldscattered P–T results owing tochaotic zoning produced inplagioclase. c Breakdown ofhornblende to produce biotiteand type C plagioclase occurredat di!erent conditions indi!erent samples, and hence avariety of P–T conditions willbe calculated. Once biotite wasproduced, strain was localizedback into these samples

611

later deformation to be partitioned into other domainswithin the shear zone (Ste!en et al. 2001). As the P–Tpath neared the thermal maximum, plagioclase contin-ued to grow in these horizons but there was relativelylittle synchronous strain accumulation. Hence, meta-morphic processes dominated over deformational pro-cesses, and plagioclase developed concentric growthzoning (type A; Fig. 9a). Thermobarometric calcula-tions in these strong domains yield reliable P–T condi-tions with relatively small scatter.

In horizons where grain-size reduction occurred clo-ser to the thermal maximum, the Garbenschiefer texturedeveloped later than in the example above. Grain-sizereduction resulted in strain accommodation by DPC,producing type B plagioclase (Ste!en et al. 2001). Inthese samples, there was insu"cient time for the pla-gioclase to be modified by metamorphic reactions nearTmax, and chaotic DPC zoning was preserved (Fig. 9b).Thermobarometric calculations thus return meaninglessP–T values.

Rocks exhibiting pseudomorphs after hornblendecontain type B and type C plagioclase. In these samples,the Garbenschiefer texture formed early in concert withDPC but was later destabilized bymetamorphic reactions(Fig. 9c). The destruction of the Garbenschiefer frame-work weakened the domains and allowed strain to a!ectthe plagioclase, leading to the formation of type Bplagioclase in addition to the type C plagioclase createdby the reactions that consume hornblende. In biotiteschist domains, the rocks never strengthened and defor-mation and metamorphism interacted throughout theP–T path.

In di!usion-limited systems, reaction kinetics aredominated by thermally activated di!usion rates andthis relationship explains why thermobarometric calcu-lations using rim compositions yield conditions of thethermal maximum. In our model, static metamorphismat the thermal maximum is relevant only for stableGarbenschiefer with type A plagioclase (Fig. 9a). Thesamples used by Selverstone et al. (1984) were of thistexture and produced thermobarometric results for rimcompositions that showed little variation within or be-tween samples. Thermobarometric results obtained inthis study for this texture also show relatively low scatterof approximately ±50"C and ±1.5 kbar [with theexception of the results for sample BH-00-86-1K2 inFig. 6c; sodic plagioclase in this sample is out of therange suggested for the Kohn and Spear (1989) cali-bration and is likely a remnant high-P composition].

For other textures and plagioclase types, deforma-tional and metamorphic processes were both active nearthe thermal maximum, and deformation likely had astrong e!ect on grain-boundary di!usion rates andreaction kinetics. The pressures calculated from rimcompositions in these samples span the decompressionpath calculated by Selverstone et al. (1984) from zonedminerals. Episodes of strain localization in these sampleswould have caused transient increases in di!usion ratesand rates of mineral equilibration. Because equilibration

rates were likely slower during slow strain-rate periods,the mineral compositions produced during deformationwere unlikely to be significantly reset by di!usion andreaction at the thermal maximum. Thermobarometriccalculations on many of these rocks thus yield P–Tconditions related to discrete episodes of strain locali-zation, and not necessarily to the maximum temperatureattained by the rocks.

Implications and recommendations for P–T calculationsin shear zones

Transient changes in reaction rates due to the combinede!ects of metamorphism and deformation may funda-mentally a!ect the accuracy of thermobarometricresults. In cases of ‘‘static’’ metamorphism, where dif-fusion rates depend primarily on temperature, reactionrates should co-vary with temperature. Because tem-perature changes will be limited by the thermalconductivity of the surrounding rocks, changes in reac-tion rates are predicted to occur gradually. Duringsynmetamorphic deformation, however, strain localiza-tion and/or changes in deformation mechanism relatedto high strain-rate events can strongly influence reactionrates and kinetic pathways. The combined spatial andtemporal variations in reaction rates may result insituations in which chemical equilibrium is only attainedfor certain components or phases, or only in localdomains. Preservation of the resulting compositionalheterogeneities will contribute to poor consistency and/or meaningless calculated P–T conditions (Fig. 9b).

The activation of di!erent deformation mechanismscan lead to the formation of mineral zoning patternsthat are di"cult to interpret and lend large geologicuncertainty to thermobarometric calculations. In gen-eral, fine-grained samples that experience dissolution-precipitation creep are likely to be the most problematicowing to the discontinuous zoning that can develop. Inthese samples, the most evolved mineral composition(An Max in this study) should be used in P–T calcula-tions, even if that composition is not a rim composition.In some cases, even this composition may not have beenin local equilibrium with other phases in the rock(Fig. 6).

Determination of the most-evolved mineral compo-sition requires thermodynamic modeling of the bulkrock in order to identify the key continuous equilibriathat operated and their predicted e!ects on mineralcompositions. Construction of pseudosections withTHERMOCALC or other programs can give a general senseof the pressure–temperature conditions that prevailedduring synmetamorphic deformation. However,pseudosections are generally calculated for relativelysimple chemical systems and also do not take into ac-count changes in e!ective bulk composition arising fromdevelopment of chemical zoning. As a result, the use ofpseudosections does not completely replace standardthermobarometric methods. Based on the results of this

612

study, we recommend the use of pseudosection calcula-tions primarily to predict the most evolved mineralcompositions that should develop during synmetamor-phic deformation in a rock of a particular bulk com-position. Once that general composition has beenidentified (e.g., the most An-rich plagioclase or the mostFe-rich garnet), microprobe data should be used todetermine specific mineral compositions for thermoba-rometry.

In shear zones that experience protracted deforma-tion histories, the relative timing of textural develop-ment and chemical equilibration can produce largevariations in calculated P and T within and betweensamples. In the Greiner shear zone, samples that werelittle a!ected by deformation at conditions near themaximum temperature yield P–T results that are moreinternally consistent than results from samples thatexperienced shear localization at or near the thermalmaximum. Some of the scatter likely represents realspatial variation in the P–T conditions at whichequilibration occurred within the shear zone. However,the large variation in calculated P–T conditions withinand between samples containing plagioclase thatunderwent dissolution and reprecipitation suggests thatmany of the thermobarometric results are spurious andreflect short-lived, deformation-induced chemical het-erogeneities.

Retrieval of meaningful P–T data from shearedsamples thus requires detailed assessment of the inter-actions between strain accommodation mechanisms andchemical equilibration at both the field and microscopicscales. In general, analysis of mineral deformationmechanisms should be coupled with thermodynamicmodeling to place constraints on the entire P–T intervalover which deformation occurred. Individual ther-mobarometers are unlikely to provide meaningful PTthistories if dissolution-precipitation creep of zonedminerals was activated during shearing.

Acknowledgements This research was funded by NSF grant EAR-0000965 to J. Selverstone and A. Brearley, a Kelley-Silver GraduateFellowship and a Vincent Kelley Scholarship (Ste!en) from theDepartment of Earth & Planetary Sciences at the University of NewMexico, a Mineralogy/Petrology Research Grant from the Miner-alogical Society of America (Ste!en), and a Student Research Grantin Mathematical Geology from the International Association ofMathematical Geology (Ste!en). Micah Jessup, Aaron Cavosie,Jaime Barnes, and Amanda Tyson provided able field assistance.We thank journal reviewers Ethan Baxter and, especially, Jan Tullisfor their constructive comments on the manuscript.

Appendix: Mineral activity models and bulk compositions

Activity-composition models used in thermodynamicmodel (Fig. 7) Endmember thermodynamic data fromHolland and Powell (1998), water model using CORK(Holland et al. 1998)

Bulk compositions used in thermodynamic modeling

References

Arnold J, Powell R, Sandiford M (2000) Amphibolites with stau-rolite and other aluminous minerals: calculated mineral equi-libria in NCFMASH. J Metamorph Geol 18:23–40

Baxter EF, DePaolo DJ (2004) Can metamorphic reactions pro-ceed faster than bulk strain. Contrib Mineral Petrol 146:657–670

Brodie KH (1981) Variation in amphibole and plagioclase com-position with deformation. Tectonophysics 78:385–402

Brodie KH, Rutter EH (1985) On the relationship between defor-mation and metamorphism, with special reference to thebehavior of basic rocks. In: Thompson AB, Rubie DC (eds)Metamorphic reactions. Springer, Berlin Heidelberg New York,pp 138–179

Carlson WD (2002) Scales of disequilibrium and rates of equili-bration during metamorphism. Am Mineral 87:185–204

Phase Model

Paragonite Unit activityLawsonite Unit activityQuartz Unit activityWater Unit activityKyanite Unit activitySillimanite Unit activityAndalusite Unit activityPrehnite Unit activityPumpellyite Unit activityEpidote Unit activity, clinozoisitePlagioclase DQF model (Holland and Powell 1992)Cordierite Ideal crd, fcrd, site multiplicity = 2Chloritoid Ideal mctd, fctd, site multiplicity = 1Carpholite Ideal mcar, fcar, site multiplicity = 1Staurolite Ideal mst, fst, site multiplicity = 4Spinel Ideal sp, herc, site multiplicity = 1Omphacite DQF omph, feomph, di, hed

(di 20,000 J/m, hed 20,000 J/m)Garnet Non-ideal gr, alm, py (Dale et al. 2000)Chlorite Disordered clin, daph, ames parameters

from (Holland et al. 1998)Muscovite Mu, cel, fecel, pa, Mu–cel–fel parameters from

(Coggon and Holland 2002), exchange toparagonite by DQF (pa 20,000 J/m)

Biotite Non-ideal phl, ann, east disordered(Powell and Holland 1999)

Hornblende tr, fact, ts, fets, gl, fegl, parg, feparg, DQF model(tr 46,000 J/m, fact 35,000 J/m, gl 32,000 J/m,fegl 16,000 J/m, parg 20,000 J/m,feparg 32,000 J/m). Values modified to createhornblende compositions similar to thoseobserved in natural samples.

Glaucophane Non-ideal tr, ts, fact, gl (Dale et al. 2000)

Sample SiO2 Al2O3 FeO MgO CaO Na2O K2O

99-2-1 59.58 16.77 8.87 3.81 2.55 3.22 2.8699-4-1 49.58 19.35 12.64 5.61 6.51 3.69 0.6099-5-1 65.60 14.72 7.36 2.45 3.27 5.48 0.11PJ107-H 59.87 17.38 7.76 3.66 2.72 3.90 2.09

99-2-1=biotite schist, 99-4-1=Garbenschiefer, 99-5-1=Garbens-chiefer, PJ-107H= pseudomorph-bearing sample

613

Cherno! CB, Carlson WD (1997) Disequilibrium for Ca duringgrowth of pelitic garnet. J Metamorph Geol 15:421–438 Uni-versity of Texas at Austin, Department of Geological Sciences,Austin, TX, United States

Coggon R, Holland T (2002) Mixing properties of phengitic micasand revised garnet-phengite thermobarometers. J MetamorphGeol 20:683–696

Dale J, Holland T (2003) Geothermobarometry, P–T paths andmetamorphic field gradients of high-pressure rocks from theAdula Nappe, Central Alps. J Metamorph Geol 21:813–829

Dale J, Holland T, Powell R (2000) Hornblende-garnet-plagioclasethermobarometry: a natural assemblage calibration of thethermodynamics of hornblende. Contrib Mineral Petrol140:353–362

Farver J, Yund R (1999) Oxygen bulk di!usion measurements andTEM characterization of a natural ultramylonite: implicationsfor fluid transport in mica-bearing rocks. J Metamorph Geol17:669–683

Frost R, Tracy RJ (1991) P–T paths from zoned garnets; someminimum criteria. Am J Sci 291:917–939

Ghent ED, Stout MZ (1981) Geobarometry and geothermometryof plagioclase-biotite-garnet-muscovite assemblages. ContribMineral Petrol 76:92–97

Graham C, Powell R (1984) A garnet-hornblende geothermometer:calibration, testing, and application to the Pelona Schist.J Metamorph Geol 2:13–21

Hirth G, Tullis J (1992) Dislocation creep regimes in quartz.J Struct Geol 14:145–159

Holland T, Baker J, Powell R (1998) Mixing properties andactivity-composition relationships of chlorites in the systemMgO–FeO–Al2O3–SiO2–H2O. Eur J Mineral 10:395–406

Holland T, Powell R (1991) A Compensated-Redlich-Kwong(CORK) equation for volumes and fugacities of CO2 and H2Oin the range 1 bar to 50 kbar and 100–1600"C. Contrib MineralPetrol 109:265–273

Holland T, Powell R (1992) Plagioclase feldspars: activity-com-position relations based upon Darken’s quadratic formalismand Landau theory. Am Mineral 77:53–61

Holland T, Powell R (1998) An internally consistent data set forphases of petrologic interest. J Metamorph Geol 16:309–343

Kerrich R, Beckinsale RD, Durham JJ (1977) The transition be-tween deformation regimes dominated by intercrystalline dif-fusion and intracrystalline creep evaluated by oxygen isotopethermometry. Tectonophysics 38:241–257

Kohn MJ, Spear FS (1989) Empirical calibration of geobarometersfor the assemblage garnet + hornblende + plagio-clase + quartz. Am Mineral 74:77–84

Kohn MJ, Spear FS (1991) Error propagation for barometers 2:application to rocks. Am Mineral 76:138–147

Koons PO, Rubie DC, Frueh-Green G (1987) The e!ects of dis-equilibrium and deformation on the mineralogical evolution ofquartz diorite during metamorphism in the eclogite facies.J Petrol 28:679–700 ETH-Zentrum, Dep. Erdwissenschaften,Zurich, Switzerland

Kretz R (1966) Grain-size distribution for certain metamorphicminerals in relation to nucleation and growth. J Geol 74:147–173

Kretz R (1973) Kinetics of the crystallization of garnet at twolocalities near Yellowknife. Can Mineral 12:1–20

Lammerer B (1988) Thrust-regime and transpression-regime tec-tonics in the Tauern Window (Eastern Alps). Geol Rund77:143–156

Lammerer B, Weger M (1998) Footwall uplift in an orogenicwedge: the Tauern Window in the Eastern Alps of Europe.Tectonophysics 285:213–230

Morteani G (1974) Petrology of the Tauern Window, AustrianAlps. Fortschr Mineral 52:195–220

Powell R, Holland T (1994) Optimal geothermometry and geoba-rometry. Am Mineral 79:120–133

Powell R, Holland T (1999) Relating formulations of the thermo-dynamics of mineral solid solutions: activity modeling of py-roxenes, amphiboles, and micas. Am Mineral 84:1–14

Powell R, Holland T, Worley B (1998) Calculating phase diagramsinvolving solid solutions via non-linear equations, with exam-ples using THERMOCALC. J Metamorph Geol 16:577–588

Puga E, Nieto JM, De Federico AD (2000) Contrasting P–T pathsin eclogites of the Betic Ophiolitic Association, MulhacenComplex, southeastern Spain. Can Mineral 38:1137–1161

Ratschbacher L, Merle O, Davy P, Cobbold P (1991) Lateralextrusion in the eastern Alps; 1. Boundary-conditions andexperiments scaled for gravity. Tectonics 10:245–256

Selverstone J (1985) Petrologic constraints on imbrication, meta-morphism, and uplift in the SW Tauern Window, Eastern Alps.Tectonics 4:687–704

Selverstone J (1988) Evidence for east-west crustal extension in theEastern Alps: Implications for the unroofing history of theTauern Window. Tectonics 7:87–105

Selverstone J (1993) Microscale to macroscale interactions betweendeformational and metamorphic processes, Tauern-Window,Eastern Alps. Schweiz Mineral Petrograph Mitt 73:229–239

Selverstone J, Munoz JL (1987) Fluid heterogeneities and horn-blende stability in interlayered graphitic and nongraphiticschists (Tauern Window, Eastern Alps). Contrib Mineral Petrol96:426–440

Selverstone J, Spear FS, Franz G, Morteani G (1984) High-pres-sure metamorphism in the SW Tauern Window, Austria: P–Tpaths from hornblende-kyanite-staurolite schists. J Petrol25:501–531

Spear FS, Selverstone J (1983) Quantitative P–T paths from zonedminerals: theory and tectonic applications. Contrib MineralPetrol 83:348–357

Ste!en K, Selverstone J, Brearley A (2001) Episodic weakening andstrengthening during synmetamorphic deformation of a deepcrustal shear zone in the Alps. In: Holdsworth RE, StrachanRA, Magloughlin JF, Knipe RJ (eds) The nature and tectonicsignificance of fault zone weakening. The Geological Society ofLondon, London, pp 141–156

Ste!en KJ (2004) Automation of phase classification, calculation ofmetastable chemical energy, and the interaction of deforma-tional and metamorphic processes. In: Earth and PlanetarySciences. University of New Mexico, Albuquerque, pp 508

Stunitz H (1998) Syndeformational recrystallization: dynamic orcompositionally induced. Contrib Mineral Petrol 131:219–236

Stunitz H, Tullis J (2001) Weakening and strain localization pro-duced by syn-deformational reaction of plagioclase. Int J EarthSci 90:136–148

Stuwe K (1997) E!ective bulk composition changes due to cooling:a model predicting complexities in retrograde reaction textures.Contrib Mineral Petrol 129:43–52

Thompson JB, Laird J, Thompson AB (1982) Reactions inamphibolite, greenschist and blueschist. J Petrol 23:1–27

Tracy RJ (1976) Garnet compositions and zoning in the determi-nation of temperature and pressure of metamorphism, CentralMassachusetts. Am Mineral 61:762–775

Tracy RJ (1982) Compositional zoning and inclusions in meta-morphic minerals. In: Ferry JM (ed) Characterization ofmetamorphism through mineral equilibria. MineralogicalSociety of America, Washington, pp 355–397

Tullis J, Yund R, Farver J (1996) Deformation-enhanced fluiddistribution in feldspar aggregates and implications for ductileshear zones. Geology 24:63–66

Tullis J, Yund RA (1991) Di!usion creep in feldspar aggregates:experimental evidence. J Struct Geol 13:987–1000

Wintsch RP, Knipe RJ (1983) Growth of a zoned plagioclaseporphyroblast in a mylonite. Geology 11:360–363

Wintsch RP, Yi K (2001) Dissolution and replacement creep: asignificant deformation mechanism in mid-crustal rocks.J Struct Geol 24:1179–1193

Yardley B, Rochelle CA, Barnicoat AC, Lloyd GE (1991) Oscil-latory zoning in metamorphic minerals—an indicator of infil-tration metasomatism. Mineral Mag 55:357–365

Yund RA, Tullis J (1991) Compositional changes of mineralsassociated with dynamic recrystallization. Contrib MineralPetrol 108:346–355

614