ELEMENTS, VOL. 9, PP. 427–432 DECEMBER 2013427

1811-5209/13/0009-427$2.50 DOI: 10.2113/gselements.9.6.427

Garnet: Witness to the Evolution of Destructive Plate Boundaries

INTRODUCTIONGarnet plays a special role in revealing the thermal and mechanical processes controlling the evolution of Earth’s crust at plate boundaries. In particular, after geologists recognized that specifi c indicator minerals and preserved garnet compositions can be used as reliable proxies for metamorphic grade, garnet gradually assumed a central position for inferring depths, temperatures, and durations of metamorphism, metasomatism, deformation, and melting. Its unique compositional and mechanical charac-teristics provide sensitive records of metamorphic condi-tions, and these records are commonly far better preserved and easier to interpret than in other minerals. Furthermore, its stability over a wide range of temperatures and crustal depths (FIG. 1) permits its use in remarkably diverse tectonic settings. Here, we illustrate garnet’s power by concentrating on three evolutionary stages of destructive tectonic plate boundaries: subduction of oceanic crust, subduction of continental crust as an early phase of collision, and crustal thickening during continent–continent collision. Metamorphic garnet growth occurs at key stages in the evolution of each of these systems and has proven invalu-able for understanding each process. In the absence of garnet (or any understanding of its properties), our knowl-edge of the processes and rates of evolution of plate bound-aries would be far poorer, particularly as garnet growth can often be placed within the context of other events in the

evolving rock package (e.g. rapid burial or heating, growth or decay of other metamorphic phases, and dehydration).

DETERMINING METAMORPHIC PRESSURE AND TEMPERATURE Conditions of thermodynamic equilibrium and mass balance interrelate mineral compositions, mineral abundances, pressure (P), and temperature (T ) (e.g. Spear 1988), and garnet composi-tions in equilibrium with other minerals are commonly employed for quantitative calculation of metamorphic conditions. Garnet’s prominent role arises crystallo-

graphically because its large cubic site favors Fe rather than Mg and can accept Ca and Mn (FIG. 1). A high Fe/Mg ratio imparts temperature sensitivity to numerous Fe–Mg-exchange thermometers involving minerals such as chlorite, biotite, hornblende, and pyroxene, whose octahedral sites prefer Mg. Garnet’s high density means that reactions involving low-density plagioclase are highly sensitive to pressure. The analysis of appropriate coexisting mineral compositions has thus been the basis of metamor-phic P–T determination for more than 30 years. Many calibrations of compositional relationships between garnet and coexisting phases as functions of P and T have been proposed and refi ned as new natural and experimental constraints have emerged. It is beyond the scope of this article to list these, but the reader is directed to Spear (1993) for more information.

Consider the implications of such mineral equilibria. If garnet’s abundance and chemistry in a rock refl ect a partic-ular P and T, then changes in P and T should drive changes in garnet’s modal abundance and composition. Reactions cannot proceed at true thermodynamic equilibrium, so in nature, oversteps in P and T (and thus time, t) are required to approach equilibrium at the thin section scale (for more information on kinetic controls on metamorphic equilibra-tion, see Ague and Carlson 2013 this issue). If changing Pand T cause garnet to grow, and intracrystalline diffusion is not too fast, then garnet crystals may encode a series of near-equilibrium compositions from their core to their rim (which can be visualized by assuming equilibrium crystal growth along an arbitrary P–T vector in FIGURE 1). Such simple growth zoning is extremely common at interme-diate metamorphic grades in a variety of tectonic settings, where garnet crystals generally grow upon heating and/or loading, and it usually manifests itself as decreasing Mn and increasing Mg from core to rim (FIG. 2). Ca and

Thanks to its unique chemical and mechanical properties, garnet records evidence of rocks’ paths through the crust at tectonic plate boundaries. The compositions of garnet and coexisting mineral phases

permit metamorphic pressure and temperature to be determined, while garnet’s compositional zoning allows the evolution of these parameters to be constrained. But careful study of garnet reveals far more, including the dehydration history of subducted oceanic crust, the depths reached during the earliest stages of continental collision, and the mechanisms driving heat and mass fl ow as orogens develop. Overall, chemical and textural characterization of garnet can be coupled with thermodynamic, thermoelastic, geochrono-logic, diffusion, and geodynamic models to constrain the evolution of rocks in a wide variety of settings.

KEYWORDS: metamorphism, subduction zones, orogenesis, thermobarometry, P–T path, Dora-Maira

Mark J. Caddick1 and Matthew J. Kohn2

1 Department of Geosciences, Virginia Tech 4044 Derring Hall (0420), Blacksburg, Virginia 24061, USAE-mail: caddick@vt.edu

2 Department of Geosciences, Boise State University 1910 University Drive, Boise, Idaho 83725-1535, USAE-mail: mattkohn@boisestate.edu

Oscillatory and sector zoning in garnet from

central Nepal.

ELEMENTS DECEMBER 2013428

Fe may increase or decrease, depending on the exact P–T path, the reactivity of other minerals such as plagioclase, and changes in the mineral assemblage. Decreasing core-to-rim Mn and heavy rare earth elements (HREEs) also refl ect these elements’ typically low bulk-rock abundances and strong preferential incorporation into garnet over many other major phases, resulting in progressive fraction-ation from the reactive matrix during garnet growth (e.g. Hollister 1966).

If rocks continue to heat up, intracrystalline diffusion eventually fl attens and destroys prograde zoning, usually at upper amphibolite to granulite facies conditions (Yardley 1977). Even at lower temperatures and during rapid metamorphism, compositions associated with specifi c P–T conditions of growth can be lost through diffusion long before peak metamorphic conditions are reached, while still retaining broadly “prograde” zoning patterns (e.g. Florence and Spear 1991; Caddick et al. 2010). During cooling, reverse profi les can develop (i.e. increasing Mn and decreasing Mg towards the rim; Tracy et al. 1976; FIG. 2C, D), especially if garnet partially dissolves and previously seques-tered components are liberated. The repercussions of such changes for thermobarometry are clear: erroneous peak-metamorphic P–T conditions will be inferred if original equilibrium compositions are assumed to be preserved yet have actually been modifi ed. However, numerous strategies have been devised in attempts to counter this modifi ca-tion (e.g. Pattison et al. 2003) or to use it as a way of constraining rates of metamorphic processes (e.g. Spear and Parrish 1996).

If chemical zoning in garnet progressively records its passage through P–T space, surely we must be able to retrieve infor-mation about this journey from the zoning. Indeed, thermo-dynamic inversion of preserved growth zoning in garnet is a powerful method for inferring a rock’s P–T history, which in turn provides insight into tectonic processes (e.g. Spear et al. 1984). That is, if changes in P and T (ΔP and ΔT) cause compositional changes in a zoned garnet, one can invert these changes to infer the ΔP and ΔT. Different

practitioners apply somewhat different methods for this inversion, specifi cally either employing thermodynamic interrelationships alone (which requires more composition measurements but fewer assumptions) or including rock-specifi c mass balance constraints (which requires fewer composition measurements but more assumptions). Spear (1993) details these different methods, which are further illustrated below.

Forward modeling of garnet composition by calculating equilibria for a specifi c rock composition (e.g. FIG. 1) has become an increasingly important way of deciphering metamorphic histories due to the acquisition of thermo-dynamic data and the development of models for an ever-expanding range of minerals, fl uids, and melts (see, for example, reviews in volume 6, issue 5 of Elements). The sequestration of garnet-forming elements into the cores of growing crystals, however, will progressively deplete the “residual” rock, effectively requiring stepwise recalcula-tion of the phase equilibria in many cases. Increasingly sophisticated methodology now records fractionation of both porphyroblast-forming phases and evolved fl uids, and it can predict the suite of mineral phases that might be trapped as inclusions along specifi c P–T paths and the composition of those phases (e.g. Konrad-Schmolke et al. 2005). Prediction of changing garnet composition through P–T evolution has revealed the metamorphic histories of rocks from a variety of tectonic settings, which we summa-rize here with respect to three stages in the evolution of a destructive plate boundary.

SUBDUCTION OF OCEANIC CRUSTGarnet plays a signature role in unraveling subduction geodynamics and elucidating fl uid-mediated mass transfer processes across the slab–mantle interface, with implica-tions for arc magmatism and mantle geochemistry (e.g. Bebout 2007). Deciphering P–T evolution in subduction zones is required to estimate paleosubduction rates and exhumation processes, and currently, there is particular interest in understanding the maximum depths that rocks

X almandine ( )

0.35

0.40

0.45

0.50

0.55

0.60

0.65

0.70

FeFe + Mg + Ca + Mn

(garnetnot stable)

X spessartine ( )

0.05

0.150.20

0.10

0.250.300.350.400.45

MnFe + Mg + Ca + Mn

(garnetnot stable)

(garnet not stable)

(garnetnot stable)

0.05

0.10

0.15

0.20

0.25

0.30

0.05

0.10

0.15

0.20

0.25

0.30

0.35

X pyrope ( )MgFe + Mg + Ca + Mn

X grossular ( )CaFe + Mg + Ca + Mn

450 500 550 600 650 700 750 800 850 900Temperature, °C

450 500 550 600 650 700 750 800 850 900Temperature, °C

2.0

400 450 500 550 600 650 700 750 800 850 900Temperature, °C

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

Pre

ssu

re, G

Pa

2.0

400 450 500 550 600 650 700 750 800 850 900Temperature, °C

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

Pre

ssu

re, G

Pa

FIGURE 1 Calculated composition of garnet in equilibrium

with minerals and melt in an “average pelitic rock” over a wide P–T range encompassing most common metamorphic conditions in the crust. The panels represent the relative molar proportions (X) of Fe, Mg, Mn, and Ca in the cubic site, assuming that these are the only elements permitted. The molar proportions were calculated using an update (ds55) of Holland and Powell’s (1998) thermodynamic data set and appro-priate mineral solution models.

ELEMENTS DECEMBER 2013429

can reach while still retaining the possibility of exhuma-tion. Whereas estimating temperature has proven relatively straightforward from phase equilibria, trace element thermometers, and major element exchange thermome-ters, estimating pressures accurately has proven elusive. For example, estimates of maximum pressure for blueschist–eclogite rocks of Sifnos island, Greece, range from 1.4 to 2.4 GPa (55–85 km depth; FIG. 3A, C), but maximum temperature is relatively well constrained at ~500 to 550 °C. It is possible that this wide P range partly refl ects tectonic juxtaposition of rocks with diverse subduction histories; multiple high-P deformation and metamorphism stages have been interpreted in these rocks from microstructurally distinct mineral-growth episodes (Forster and Lister 2005), and pervasive overprinting of much of the island is associ-ated with the development of extensional shear zones. However, such interpretation (and a broader discussion of

the importance of mélange zones in facilitating exhuma-tion) requires very accurate constraint of the evolving metamorphic conditions of individual metamorphic rocks.

Fortunately, garnet is abundant in subducted oceanic crust and its sedimentary cover (e.g. FIG. 3B). High P in subduc-tion zone settings initiates growth at temperatures as low as 300 ºC, and low peak T prevents substantial diffusive modifi cation of its chemical record of subduction processes. In contrast to other mineral equilibria, which may refl ect resetting during exhumation, garnet-based equilibria for Sifnos consistently indicate maximum P values of ≥1.7 GPa, and textural evidence suggests that this refl ects overgrowth following an earlier lawsonite-blueschist metamorphic stage (Forster and Lister 2005). Although exhumation-related breakdown of garnet did occur in some Sifnos lithologies, many samples preserve strong evidence for near-peak conditions in which garnet composition has not been strongly modifi ed after apparent equilibration. Indeed, recent studies (g to h in FIG. 3C) that combined garnet and coexisting mineral compositions with phase equilibria and microtextural constraints (e.g. the presence of reaction products from lawsonite breakdown) suggest that some of Sifnos’s blueschists reached ~2.2 GPa, some 30 km deeper than paths based primarily on pyroxene and phengite compositions (e.g. a to c in FIG. 3C).

In addition to garnet’s chemical advantages, novel appli-cation of Raman spectroscopy to quartz inclusions in garnet from Sifnos’s blueschist–eclogite unit confi rms maximum pressure of approximately 2.0 GPa (Ashley et al. 2012). In the latter case, the physical properties of garnet, especially its high bulk modulus, protect inclusions from post-entrapment pressure variation. In particular, core-to-rim inclusion suites suggest minimal compression during garnet growth over a temperature interval of approximately 100 ºC. Available evidence based on multizone Sm–Nd analyses of garnet further implies that this heating required <1 My (Dragovic et al. 2012). The tectonic confi guration that produces this phase of heating and then exhumation

FIGURE 2 (A–F) X-ray maps of Himalayan garnet crystals. Concentric decreases in Mn and increases in Mg from

core to rim in LHS and THS garnet grains refl ect simple growth zoning. Reverse zoning in GHS garnet refl ects resorption and diffu-sional reequilibration. The numbers are percent spessartine (Mn) and pyrope (Mg) components in core and rim. Scale bars are 500 µm. LHS = Lesser Himalayan Sequence; THS = Tethyan Himalayan Sequence; GHS = Greater Himalayan Sequence.

FIGURE 3 (A) Geological sketch map of the island of Sifnos, Greece, after Matthews and Schliestedt (1984).

(B) Large garnet porphyroblasts in an amphibole-rich matrix, Syros, Greece. (C) Summarized P–T constraints from Sifnos, Greece, modifi ed after Ashley et al. (2012); see Dragovic et al. (2012) for the cited litterature.

A B

C

E

D

F

2.5

2.0

1.5

1.0

0.5

0

Pres

sure

(GPa

)

400 500 600 700300

a

b

b

h

h

Dep

th (k

m),

assu

min

g de

nsity

= 3

200

kg/m

3

50

40

30

10

0

20

60

70

g

f

c

e

d

Schliestedt (1984)b) Schliestedt and

Schliestedt et al. (1987)c) Avigad et al. (1992)d) Schmädicke and Will

(2003)e) Trotet et al. (2001)f) T range of Spear et al.

(2006)g) Groppo et al. (2009)h) Dragovic et al. (2012)i) Ashley et al. (2012)

i

0 2 4 km

Upper MarbleUnitBlueschist-Eclogite UnitMain Marble UnitGreenschist Unit

N

Sifnos

Milos

Paros

Naxos

AmorgosIos

Thera0 30 km

A C

B

2.5 cm2.5 cm

A

B

C

ELEMENTS DECEMBER 2013430

requires further investigation, but both the chemical and physical properties of garnet play an important role in establishing important constraints on subduction pressures and temperatures, and on the duration of their evolution.

Subduction zone metamorphic fl uids may infl uence or control volatile fl ux into the mantle, arc-magma genesis, physical properties including seismicity, and the global circulation of H2O (e.g. Kerrick and Connolly 2001; Hacker 2008). Although the production of fl uid is typically diffi -cult to detect, most dehydration reactions produce both H2O and garnet. Thus, the P–T–t history of garnet growth directly records part of the history of fl uid generation (e.g. Baxter and Caddick 2013), with the breakdown of specifi c hydrous parent phases (e.g. lawsonite) representing partic-ularly important volatile sources. Models for subduction zone metamorphism of both hydrated basalts and pelitic rocks suggest that extensive garnet growth between 1.4 and 3.0 GPa is coupled with signifi cant fl uid production. As such, the ~100 °C of heating at ~65 km depth inferred above for Sifnos probably generated a distinct pulse of hydrous fl uid (Dragovic et al. 2012), consistent with the short-lived channelized fl uid pulses interpreted for other subduction zones based on geochemical and isotopic traces (John et al. 2012), predicted from phase equilibria (Schmidt and Poli 2003), and witnessed by the growth of garnet.

SUBDUCTION OF CONTINENTAL MATERIALThe early stages of orogenesis can involve subduction of continental material to ≥100 km depth, but the scarce rocks exhumed from these conditions invariably contain an enigmatic metamorphic record of their journey. Indeed, these rocks often contain no strong evidence for ultrahigh- pressure (UHP) conditions, because key indicator minerals either never transformed to high-pressure polymorphs (e.g. diamond or coesite) or experienced almost complete inver-sion to low-pressure forms (graphite and quartz). However, as in the case of oceanic subduction, garnet plays a key role in evaluating geodynamic scenarios that permit deep subduction and subsequent exhumation, primarily by eluci-dating depths, temperatures, and burial/exhumation rates.

The preservation of coesite as inclusions in garnet led to the fi rst unambiguous evidence of UHP conditions in rocks from the Dora-Maira massif, Italy (Chopin 1984; reviewed in Elements volume 9, number 4). These samples, which contain abundant matrix quartz, characteristically exhibit coesite inclusions surrounded by polygonal quartz haloes and cracks radiating into the Mg-rich host garnet (FIG. 4). Clearly, these rocks equilibrated at coesite-stable pressures but failed to preserve such evidence except where the inclusions were shielded by garnet. As in the Sifnos example, the physical properties of garnet allow large pressure gradients to be preserved, although it is clearer here that these supported gradients must be of the order of several giga pascals over millimeter length scales. P–T histories for these southern Dora-Maira rocks now indicate that maximum pressures reached or exceeded 3.5 GPa (ca 120 km depth), at temperatures in the range of 700–800 °C (FIG. 5D). Perhaps surprisingly, given the high temperature from which they cooled and depth from which they exhumed, southern Dora-Maira garnet crystals still preserve some evidence of zoning that is interpreted to be due to growth during burial. This interpretation concurs with other evidence implying rapid exhumation at rates of >3 cm y-1 (e.g. Rubatto and Hermann 2001).

Garnet from northern Dora-Maira tells a different story through its preserved compositional zoning (FIG. 5A, B). Forward modeling of mineral abundance and composition for an appropriate rock composition suggests that the outer-

most parts of these crystals record growth during both decompression and heating (Gasco et al. 2011; FIG. 5C, D). There is little evidence that these samples reached the high P and T seen by the coesite-bearing samples from farther south, confi rming earlier suggestions that the Dora-Maira massif represents at least three tectonic slices, each of which experienced different early-Alpine conditions before later juxtaposition.

Garn

et

Quar

tzQu

artz

Kyanite

Coesitewith polygonalquartz haloes

Garn

et

Quar

tzQu

artz

Kyanite

Coesitewith polygonalquartz haloes

Garn

et

Quar

tzQu

artz

Kyanite

Coesitewith polygonalquartz haloes

Garn

et

Quar

tzQu

artz

Kyanite

Kyanite KyaniteKyanite Kyanite

Coesitewith polygonalquartz haloes

A B

Aragonite

Calcite

CoesiteQuartz

Diamond

GraphiteGraphite

500475 400 500 600 800700525 625

3

2

0

P, GPa

T,°C500475 525 625

11

0.060.07

0.090.08

24 1417

19

0.060.07

0.090.08

24 1417

19

Ca Fe

corecore

mantlemantle

rimrim

500 μm

Highabundance

LowabundanceLowabundance

XMg GrtGross mol %

OutermantleOutermantle

InnermantleInnermantle

0.8

1.0

1.21.4

1.61.8P, GPa

0.8

1.0

1.21.4

1.61.8P, GPa

TT,°C

FIGURE 4 Garnet from a southern Dora-Maira (Case Ramello) whiteschist in plane-polarized light (A) and under

crossed polarizers (B). Highlighted inclusions of coesite are rimmed by a thin, polycrystalline quartz halo (palisade texture). In the rock matrix (upper left of each image), quartz is the only SiO2 polymorph present. COURTESY OF R. TRACY

FIGURE 5 (A, B) Strongly zoned garnet from a chloritoid–garnet schist (northern Dora-Maira massif), from Gasco et al.

(2011). (C) Modeled garnet compositions for this sample (from Gasco et al. 2011). Blue lines = XMg; red lines = mol% grossular; yellow fi elds represent the compositions of the inner and outer garnet “mantle” labeled in panel A; underlying gray shading denotes thermodynamic variance (darker = higher variance). (D) Resultant P–T path for this sample (blue line) compared to the paths for coesite-bearing southern Dora-Maira samples like that in Figure 4. The green and red paths are from Castelli et al. (2007) and Rubatto and Hermann (2001), respectively.

A B

C

A

D

B

ELEMENTS DECEMBER 2013431

Continent–Continent CollisionGarnet-based P–T paths are particularly useful because the shape of the path distinguishes tectonic processes (Spear et al. 1984; Thompson and England 1984). For example, recent debate about Himalayan geodynamics centers on the magnitude of lower-crustal fl ow. Large-scale fl ow, or “channel fl ow,” transports heat and mass laterally in response to focused erosion at the orogenic front (FIG. 6A). In contrast, classic models of orogenic (critical) wedges commonly infer progressive in-sequence thrusting, poten-tially interspersed with extensional events, generally in response to distributed erosion; this maintains a regular, material-specifi c geometry without requiring lower-crustal fl ow (FIG. 6B). These two end-members predict distinctly different metamorphic P–T paths (Kohn 2008). Channel fl ow’s profound lateral heat transport virtually contact metamorphoses over- and underlying rocks at nearly constant pressure. In contrast, thrust duplexing rapidly loads rocks as a higher thrust is emplaced, then soon after exhumes them as the next lower thrust activates. Thus garnet above and below a fl owing channel should record isobaric heating over many tens of degrees Celsius, whereas, depending on thrusting and erosion rates, thrust-duplex garnet should record either loading or possibly exhumation with slight heating (FIG. 6).

This geodynamics debate has focused on the metamor-phic core of the Himalaya, specifi cally migmatitic rocks of the Greater Himalayan Sequence (GHS), which are sandwiched between mid-amphibolite facies rocks of the Lesser Himalayan Sequence (LHS), below, and the Tethyan Himalayan Sequence (THS), above (FIG. 6). Thus, the GHS represents the weak high-T channel, whereas the THS and LHS represent the stiff over- and underlying sheets, respectively. P–T paths from the central and eastern Himalaya have now been calculated from zoned LHS and THS garnet crystals (Kohn 2008), and these consistently show minimal heating during substantial loading (FIG. 6D). While some lateral heat transport is permissible, these paths imply that channel fl ow must have been far less effective than implied by present models, at least at these localities. Approximately equivalent LHS and THS rocks from a section in the western Himalaya preserve greater evidence of synburial heating (Chambers et al. 2009), with interpreted P–T paths being more consistent with channel fl ow predictions and implying along-strike variability in the orogen. The degree to which one model or another is preferred requires acquisition and integration of numerous other data sets that can be compared to predictions of geodynamic models. Of these, temperature–time histories are particularly distinctive and informative, but are beyond the scope of this review (see for example Kohn 2008).

The dating of garnet growth is discussed elsewhere in this issue, but trace element compositions and zoning within garnet porphyroblasts can also be used to link the systematics of garnet growth and dissolution with that of accessory phases, such as monazite (e.g. Pyle and Spear 1999), amenable to (U–Th)–Pb dating. Thus, the growth of accessory phases can be directly linked to stages of a P–T history, constraining rates of metamorphic evolution (e.g. Foster et al. 2004). In the absence of such information, garnet found in orogenic belts is typically considered to preserve the story of continental collision, subject to the potential loss of compositional information through diffu-sive modifi cation, described above. However, mounting geochronological evidence from both garnet and its included phases suggests that garnet is readily inherited from earlier metamorphic events and may not “reset” its composition during later heating (e.g. Argles et al. 1999).

This inheritance may arise through direct polymetamor-phism of the rocks of interest or the presence of detrital garnet in sediments that themselves experience a relatively simple metamorphic history. Either case demands partic-ular caution when using garnet for thermobarometry as its composition may preserve a record of its growth rather than that of equilibrium with minerals in the last metamor-phic episode. However, with careful teasing out of appro-priate crystals and compositions, metamorphic P–T–t paths based on garnet zoning will continue to provide important insights and constraints into geodynamic processes.

CONCLUDING REMARKSGarnets preserve diverse records of orogenic processes in their chemical zoning. Petrologists commonly employ thermodynamic modeling and interpretation of major and minor element zoning in garnet as their fi rst line of geochemical inquiry in tectonic studies. Such a scien-tifi cally mature approach elucidates many aspects of orogenesis, from oceanic subduction through all stages of

10

400 500 600 7002

4

6

8

Pre

ssur

e (k

bar)

Temperature (°C)

ChannelFlow

LHS

500400 600

HigherLHS

THS

A

C D

B

Duplex,LHS & THS

Obs. P-T paths

THS

Model P-T paths

LowerLHS

Critical Wedge

Channel flow

MCTSTDS

LHS

THS

GHS

THS

LHS

GHS

FIGURE 6 (A) Channel fl ow geodynamic model (based on Godin et al. 2006). Focused erosion at the orogenic front

couples to the weak, hot channel, causing fl ow towards the foreland. STDS = South Tibetan Detachment System, which is a major, orogen-parallel, normal-sense shear zone separating the Tethyan Himalayan Sequence (THS) from the Greater Himalayan Sequence (GHS). MCT = Main Central Thrust, a major thrust-sense shear zone separating the GHS from the Lesser Himalayan Sequence (LHS). (B) Critical wedge geodynamic model. The THS and LHS experience in-sequence thrusting, producing duplexes. THS duplexes (blue arrows) predate STDS, which refl ects an orogenic collapse structure, and predate LHS duplexes (red arrows). (C) Predicted P–T paths for different geodynamic models (from Kohn 2008). The thin vertical line represents the upper tempera-ture limit of panel D. The grey lines represent the andalusite–kyanite–sillimanite fi eld boundaries. (D) Measured P–T paths based partly on observed garnet zoning (e.g. FIG. 2) from the THS in Bhutan and LHS in central Nepal (Kohn 2008; Corrie et al. 2012).

C

A

D

B

ELEMENTS DECEMBER 2013432

REFERENCESA gue JJ, Carlson WD (2013)

Metamorphism as garnet sees it: The kinetics of nucleation and growth, equilibration, and diffusional relax-ation. Elements 9: 439-445

Argles TW, Prince CI, Foster GL, Vance D (1999) New garnets for old? Cautionary tales from young mountain belts. Earth and Planetary Science Letters 172: 301-309

A shley KT, Caddick MJ, Bodnar RJ (2012) Improving pressure estimation in high pressure terranes with Raman spectros-copy: New calibration and an imple-mentation on blueschists from Sifnos, Greece. GSA Annual Meeting, Charlotte, abstract 148-7

B axter EF, Caddick MJ (2013) Garnet growth as a proxy for progressive subduction zone dehydration. Geology 41: 643-646

B ebout GE (2007) Metamorphic chemical geodynamics of subduction zones. Earth and Planetary Science Letters 260: 373-393

C addick MJ, Konopásek J, Thompson AB (2010) Preservation of garnet growth zoning and the duration of prograde metamorphism. Journal of Petrology 51: 2327-2347

Ca stelli D, Rolfo F, Groppo C, Compagnoni R (2007) Impure marbles from the UHP Brossasco-Isasca Unit (Dora-Maira Massif, western Alps): evidence for Alpine equilibration in the diamond stability fi eld and evaluation of the X(CO2) fl uid evolution. Journal of Metamorphic Geology 25: 587-603

Ch ambers J and 6 coauthors (2009) Empirical constraints on extrusion mechanisms from the upper margin of an exhumed high-grade orogenic core, Sutlej valley, NW India. Tectonophysics 477: 77-92

Ch opin C (1984) Coesite and pure pyrope in high-grade blueschists of the Western Alps: a fi rst record and some conse-quences. Contributions to Mineralogy and Petrology 86: 107-118

Co rrie SL, Kohn MJ, McQuarrie N, Long SP (2012) Flattening the Bhutan Himalaya. Earth and Planetary Science Letters 349-350: 67-74

Dr agovic B, Samanta LM, Baxter EF, Selverstone J (2012) Using garnet to constrain the duration and rate of water-releasing metamorphic reactions during subduction: An example from Sifnos, Greece. Chemical Geology 314-317: 9-22

Fl orence FP, Spear FS (1991) Effects of diffusional modifi cation of garnet growth zoning on P-T path calcula-tions. Contributions to Mineralogy and Petrology 107: 487-500

Fo rster MA, Lister, GS (2005) Several distinct tectono-metamorphic slices in the Cycladic eclogite–blueschist belt, Greece. Contributions to Mineralogy and Petrology 150: 523-545

Fo ster GL, Parrish RR, Horstwood MSA, Chenery S, Pyle JM, Gibson HD (2004) The generation of prograde P–T–t points and paths; a textural, compositional, and chronological study of metamor-phic monazite. Earth and Planetary Science Letters 228: 125-142

Ga sco I, Gattiglio M, Borghi A (2011) Lithostratigraphic setting and P-T metamorphic evolution for the Dora Maira Massif along the Piedmont Zone boundary (middle Susa Valley, NW Alps). International Journal of Earth Sciences 100: 1065-1085

Go din L, Grujic D, Law RD, Searle MP (2006) Channel fl ow, ductile extrusion and exhumation in continental colli-sion zones: an introduction. Geological Society Special Publication 268: 1-23

Ha cker BR (2008) H2O subduction beyond arcs. Geochemistry Geophysics Geosystems 9: Q03001

Ho lland TJB, Powell R (1998) An inter-nally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16: 309-343

Ho llister LS (1966) Garnet zoning: An interpretation based on the Rayleigh fractionation model. Science 154: 1647-1651

John T and 8 coauthors (2012) Volcanic arcs fed by rapid pulsed fl uid fl ow through subducting slabs. Nature Geosciences 5: 489-492

Kerrick DM, Connolly JAD (2001) Metamorphic devolatilization of subducted oceanic metabasalts: implica-tions for seismicity, arc magmatism and volatile recycling. Earth and Planetary Science Letters 189: 19-29

Ko hn MJ (2008) P-T-t data from central Nepal support critical taper and repudiate large-scale channel fl ow of the Greater Himalayan Sequence. Geological Society of America Bulletin Special Paper 120: 259-273

Ko nrad-Schmolke M, Handy MR, Babist J, O’Brien PJ (2005) Thermodynamic modelling of diffusion-controlled garnet growth. Contributions to Mineralogy and Petrology 149: 181-195

Ma tthews A, Schliestedt M (1984) Evolution of the blueschist and green-schist facies rocks of Sifnos, Cyclades, Greece. Contributions to Mineralogy and Petrology 88: 150-163

Pattison DRM, Chacko T, Farquhar J, McFarlane CRM (2003) Temperatures of granulite-facies metamorphism: Constraints from experimental phase equilibria and thermobarometry corrected for retrograde exchange. Journal of Petrology 44: 867-900

Py le JM, Spear FS (1999) Yttrium zoning in garnet: coupling of major and accessory phases during metamorphic reactions. Geological Materials Research 1: 1-49

Ru batto D, Hermann J (2001) Exhumation as fast as subduction? Geology 29: 3-6

Schmidt MW, Poli S (2003) Generation of mobile components during subduc-tion of oceanic crust. Treatise on Geochemistry 3: 567-591

Sp ear FS (1988). The Gibbs method and Duhem’s theorem: The quantitative relationships among P, T, chemical potential, phase composition and reaction progress in igneous and metamorphic systems. Contributions to Mineralogy and Petology 99: 249-256

Sp ear FS (1993) Metamorphic Phase Equilibria and Pressure-Temperature-Time Paths. Mineralogical Society of America, Chantilly, VA, Monograph 1, 799 pp

Spear FS, Parrish RR (1996) Petrology and cooling rates of the Valhalla Complex, British Columbia, Canada. Journal of Petrology 37: 733-765

Sp ear FS, Selverstone J, Hickmott D, Crowley PD, Hodges KV (1984) P-T paths from garnet zoning: A new technique for deciphering tectonic processes in crystalline terranes. Geology 12: 87-90

Th ompson AB, England PC (1984) Pressure–temperature–time paths of regional metamorphism II. Their infer-ence and interpretation using mineral assemblages in metamorphic rocks. Journal of Petrology 25: 929-955

Tr acy RJ, Robinson P, Thompson AB (1976) Garnet composition and zoning in the determination of temperature and pressure of metamorphism, central Massachusetts. American Mineralogist 61: 762-775

Yardley BWD (1977) An empirical study of diffusion in garnet. American Mineralogist 62: 793-800.

continental collision. Recent work on major and minor element geochemistry now emphasizes length scales and mineralogical controls on thermodynamic and chemical equilibrium, with implications for biases to P–T path calcu-lations. Other important aspects of garnets, including their elastic properties and trace element and isotopic zoning, remain less well studied and are promising areas for future investigations into orogenic processes.

ACKNOWLEDGMENTSWe thank Frank Spear and Bradley Hacker for insightful and helpful reviews, and gratefully acknowledge support from National Science Foundation grants EAR-1250470 to MJC and EAR-1048124 and EAR-1321897 to MJK. Kyle Ashley is thanked for Figure 3.