Myron G. Best. Igneous and metamorphic 2003 Blackwell Science

Подождите немного. Документ загружается.

feldspar at higher grades. Excess Al over that needed

for micas and feldspars stabilizes additional aluminous

minerals, including: at low grades, chlorite and garnet;

at medium grades, garnet, staurolite, and either kyanite

or andalusite; and at high grades, cordierite and

sillimanite.

Chlorite Zone. The lowest grade of metamorphism of

shale protoliths (brieﬂy outlined in Section 14.1.1) cre-

ates a mineral assemblage in aphanitic slates of white

mica, chlorite, and quartz together with possible stilp-

nomelane, Fe–Ti oxides, and pyrite. Tourmaline, zir-

con, and monazite are common accessory minerals.

Calcite and/or ankerite occur in more calcareous pro-

toliths, pyrophyllite in highly aluminous ones, and

paragonite or albite where protoliths are relatively rich

in Na and Al. Graphite and/or other carbonaceous

material is common (Wada and Tomita, 1994). K-feldspar

is stabilized in low-Al protoliths transitional to felds-

pathic and lithic sandstones.

In the following paragraphs, pelitic mineral assem-

blages are tracked from this chlorite-zone assemblage

along an intermediate-P metamorphic ﬁeld gradient

through the greenschist, amphibolite, and granulite

facies.

Biotite Zone. With increasing T from the chlorite

zone, at least three reactions can cause the appearance

of biotite, commonly as small porphyroblasts. The

ﬁrst two reactions are discontinuous and the third is

continuous (Figure 18.14a, b). The ﬁrst discontinuous

reaction occurs in low-Al rocks lacking white mica,

Metamorphism at Convergent Plate Margins: P–T–t Paths, Facies, and Zones

577

Orthopyroxene

Plagioclase

K-feldspar

Quartz

Fe–Ti oxides

Ankerite-

Calcite

Epidote

FACIES Greenschist Amphibolite Granulite

Melt

Albite

Oligoclase Andesine

ZONE

APPROX.

Chlorite

T (°C) 300

Phengite

425 500 550 600 650 750 780 800

Biotite Garnet Staurolite Kyanite Sillimanite Sil Kfs Crd  Grt  Kfs Opx

MINERAL

Muscovite

Chloritoid

Chlorite

Mg-rich

Stilpnomelane

Biotite

Garnet Mn-rich

Staurolite

Kyanite Metastable

Sillimanite

Cordierite

18.13 Generalized mineral compatibilites in typical Barrovian zones in pelitic rocks metamorphosed under intermediate-P conditions. Width

of bands for more prominent minerals, such as micas, garnet, and so on, shows schematically their increased modal amount as produced

by discontinuous and continuous reactions, or their consumption by these reactions. Dashed lines indicate possible mineral stabilities in

nonspeciﬁc modal proportions, such as for quartz. Temperatures are approximate. Note possible coexistence of two sodic plagioclases

in the garnet zone. Some petrologists place the transition between the amphibolite and granulite facies at the low-T boundary of the

sillimanite  K-feldspar zone, whereas others place it at the high-T boundary.

To K-feldspar

Phengite  chlorite →

biotite  muscovite  quartz  H

Chlorite  muscovite → garnet  Mg-richer chlortite 

biotite  H

Prl

(a) CHLORITE ZONE (b) BIOTITE ZONE (c) GARNET ZONE

Phe Kfs

Chl

Prl Phe Kfs A

Chl

Cld

Grt

Prl

Chl

Cld

Grt

0 0.1mm 0 0.1mm 0 1mm

 Muscovite

 quartz  H

Muscovite

Biotite

Chlorite

Muscovite

QuartzBiotite

Chlorite

Quartz

18.14 Prograde regional metamorphism of pelitic rocks as occurs in typical Barrovian zones. T and grade increase from left to right. The top

panel across the pages consists of drawings of thin sections from the lowest grade chlorite zone on the left to the highest grade granulite

facies on the right. Note changes in scales of drawings. Thin sections were provided through the courtesy of William D. Carlson and

Douglas Smith from the University of Texas at Austin collections. Beneath each drawing are one or two composition diagrams showing

stable mineral assemblages for the corresponding zone or facies. The ﬁlled circle represents average low-Al shale, open circle Al-rich

shale, and open square Al-poor lithic/feldspathic sandstone and granitoid protoliths from Blatt and Tracy (1996) and Spear (1993). Be-

neath the composition diagrams is the typical reaction producing the characteristic mineral assemblage of the zone or facies. Only in (d)

is a reaction texture preserved to suggest the nature of reactant and product phases. (a) Chlorite zone. Slate from the type area of George

Barrow’s study in the Dalradian Supergroup collected along River North Esk about 1.3 km north of Dalbog in northeast Scotland con-

sists of quartz  chlorite  phengite  graphite(?). Al-rich protoliths can contain pyrophyllite and granitoid protoliths K-feldspar 

chlorite and possible phengite. (b) Biotite zone. Weakly crinkled porphyroblastic slate collected along River North Esk about 350 m

northwest of the chlorite-zone sample in (a). Porphyroblasts of biotite and coarse ﬂakes of graphite lie in matrix of chlorite  muscovite

 quartz. In the smaller AFM diagram projected from muscovite, biotite and chlorite are complete solid solutions between F and M

components. (c) Garnet zone. Locality of schist unknown. High-relief garnet poikiloblast contains abundant inclusions, partly of quartz,

forming a curved pattern. Weakly crenulated schistose matrix of muscovite  quartz  biotite. The two AFM diagrams represent, on

left, compatibility of biotite and chlorite at a T below the garnet isograd (garnet not stable) in average low-Al pelite. Right-hand diagram

represents the equilibria at the isograd as the three-phase compatibility triangle garnet  chlorite  biotite expands toward more

magnesian compositions, eclipsing average low-Al pelite. Chloritoid will appear in Al–Fe-rich pelites that are common in New England.

(d) Staurolite zone schist. Locality unknown. Euhedral staurolite poikiloblast that contains abundant inclusions of quartz, deﬁning sieve

texture, partially surrounds anhedral garnet; this may be a reaction texture in which garnet is consumed with production of staurolite

and quartz. Matrix consists of quartz  chlorite  muscovite  biotite. (e) Kyanite zone. Schist from Simplon Road near border of

Switzerland in the Italian Alps. Anhedral garnet porphyroblast with kyanite, quartz, and lesser biotite, muscovite, and graphite. (f) Silli-

manite zone. Schist from near Orange, Massachusetts, made of the assemblage sillimanite  biotite  garnet  quartz  muscovite 

graphite(?). All grains except quartz are relatively euhedral. Note that most pelitic rocks (circles) contain sillimanite  garnet  biotite

in contrast to the more variable assemblages in pelites of the staurolite and especially the kyanite zone. (g) Second sillimanite zone, or

sillimanite  alkali feldspar zone. Gneiss from the Adirondack Mountains along Route 8 east of Graphite, New York. Rock is composed

of aligned ﬂakes of biotite within an aggregate of quartz  alkali feldspar  garnet  sillimanite  Fe–Ti oxides. (h) Granulite-facies

pelitic rock, Tanil Nadu, Maduri block, India. Thin section provided courtesy of Ayati Ghosh. Rock is composed of quartz  perthite 

cordierite  garnet  sillimanite  spinel  Fe–Ti oxides (see, for example, Mukhopadhyay and Holdaway, 1994). Note distinctive

pleochroic haloes around minute zircon inclusions in cordierite.

namely, chlorite  K-feldspar → biotite  H

O, and

the second in calcareous pelites, ankerite  muscovite

→ biotite  calcite. The third reaction is the one most

commonly accounting for the biotite isograd in average

pelitic rocks. It is a continuous reaction destabilizing

2

–Mg-rich phengitic white micas in favor of more

typical muscovite

18.1 phengite  chlorite →

biotite  muscovite  quartz  H

(It should be noted here that these and other reactions

listed below are generalized because of the highly vari-

able chemical compositions of the solid solutions. Ex-

act stoichiometrically balanced reactions are not easily

written.) In the AKF diagrams (Figure 18.14a, b), at in-

creasing T, the continuous shrinking of the two-phase

stability ﬁeld of phengite  chlorite allows an average

pelite to be encompassed by an expanding three-phase

ﬁeld of biotite  phengitic muscovite  chlorite. In the

AFM diagram (Figure 18.14b), pelites richest in Al do

not record the stable appearance of biotite as do rocks

that contain less Al. Hence, the bulk chemical com-

position of the rock controls stabilization of biotite

and, hence, manifestation of the biotite isograd. It

should also be noted that, at the low temperatures of

the incoming of biotite, there may be manifestations of

the sluggish kinetics of mineral reactions. For example,

Dempster and Tanner (1997) argue that the biotite

isograd in the Central Pyrenees was controlled by

deformation, allowing its appearance at a lower T than

would otherwise have occurred.

Garnet Zone. Typically very small garnets can locally

appear in chlorite zone pelites that have unusually high

concentrations of Mn or of Ca and Fe

3

in strongly

oxidized rocks. Stabilized garnets are essentially spes-

sartine and andradite, respectively. More Mn is parti-

tioned into garnet than other maﬁc phases. However,

the appearance of Fe

2

-rich almandine porphyroblasts

Metamorphism at Convergent Plate Margins: P–T–t Paths, Facies, and Zones

579

Grt

Chl

Garnet  chlorite  muscovite → staurolite  biotite  quartz  H

O Staurolite  chlorite  muscovite  quartz → kyanite  biotite  H

(d) STAUROLITE ZONE

Grt

Chl

Grt

Chl

Grt

Chl

(e) KYANITE ZONE

01mm 0 0.5mm

Quartz

Garnet

Muscovite

Biotite

Kyanite

Staurolite

Chlorite

Quartz

 Muscovite  H

 quartz

 Muscovite  H

 quartz

18.14 (Continued).

of sufﬁcient size to be readily visible in thin section

and even hand sample delineates the lower bound

of the garnet zone in the Barrovian sequence. These

garnets, nonetheless, commonly contain appreciable

concentrations of the spessartine end-member in their

cores (to as much as 30 mole % according to Spear,

1993, p. 354), as well as much of the CaO in the rock.

The exact reaction governing the ﬁrst appearance of

this critical index mineral is somewhat uncertain but

one possibility is the continuous reaction

18.2 chlorite  muscovite → garnet

 Mg-richer chlorite  biotite  H

The two AFM diagrams in Figure 18.14c depict this re-

action and show that as chlorites become more Mg-rich

with increasing T garnet is stabilized in more alumi-

nous pelites. Chloritoid is stable in highly aluminous

rocks in the garnet zone.

Within the garnet zone, plagioclases become more

calcic, not only in pelites but in other compositional

groups as well. The stabilized oligoclase–andesine con-

trasts with the albite  epidote that coexists at lower

temperatures. In some terranes, albite and oligoclase

stably coexist (Figure 18.13) because of a so-called

peristerite gap—a sort of solvus—in the otherwise con-

tinuous solid solution series in plagioclases.

Staurolite Zone. Although many protoliths will con-

tain biotite, chlorite, and muscovite upon metamor-

phism, only ones with sufﬁcient concentrations of Al

580 Igneous and Metamorphic Petrology

Staurolite  muscovite  quartz →

sillimanite  biotite  garnet  H

Muscovite  quartz →

K-feldspar  sillimanite + H

Sillimanite  biotite  quartz →

cordierite  K-feldspar  H

Sillimanite  biotite  quartz →

cordierite  garnet  K-feldspar  H

Bt Bt

Cordierite with

pleochroic

halo around

zircon

Muscovite

Quartz

Sillimanite

Cordierite

Perthite

Garnet

Cordierite

Garnet

Cordierite

Sillimanite Sillimanite Sillimanite

Garnet

Biotite

Quartz

0 0.5mm 0 0.25mm 0 0.5mm

AAA

 Muscovite  quartz

 H

 K-feldspar  quartz

 H

(f) SILLIMANITE ZONE (g) SILLIMANITE  K-FELDSPAR ZONE (h) GRANULITE FACIES

18.14 (Continued).

and Fe

2

will contain garnet and staurolite under

suitable conditions. In highly aluminous rocks, stauro-

lite can form by the reaction chloritoid  quartz →

staurolite  garnet  H

O. A possible corresponding

reaction texture seen in some staurolite-bearing rocks

consists of garnet poikiloblasts that enclose chloritoid;

the armoring garnet prevented further reaction with

neighboring quartz. A more common discontinuous

reaction, resulting in a tie-line switch in the AFM dia-

gram, is suggested by staurolite poikiloblasts that are

crowded with inclusions of quartz (Figure 18.14d)

18.3 garnet  chlorite  muscovite →

staurolite  biotite  quartz  H

The staurolite isograd mapped on the ground is not

merely the ﬁrst appearance of staurolite, as this mineral

can be stabilized in maﬁc aluminous rocks just below

the isograd. Rather, the isograd is the ﬁrst appearance

of the stable pair staurolite  biotite.

Though incompatible as a pair in the staurolite

zone, chlorite and garnet individually form stable

three-phase assemblages with staurolite  biotite in

relatively magnesian and Fe-rich rocks, respectively.

Chlorite coexistent with staurolite  biotite is more

magnesian than the chlorite coexistent with garnet

in the garnet zones. This contraction in chlorite sta-

bility with increasing T is caused by the continuous

reaction

18.4 chlorite  muscovite →

staurolite  biotite  quartz  H

The contraction of chlorite stabilities toward the Mg

end-member and the accompanying expansion of the

two-phase ﬁeld of staurolite  biotite can be seen in

the AFM diagrams for the staurolite zone (Figure

18.14d).

Kyanite Zone. The ﬁrst appearance of kyanite with

increasing grade involves a similar reconﬁguration of

tie-lines in the AFM diagram as for the appearance of

staurolite. In this change in topology, the staurolite

 chlorite compatibility is broken by another discon-

tinuous tie-line switching reaction (Figure 18.14e)

18.5 staurolite  chlorite  muscovite  quartz

→ kyanite  biotite  H

Hence, the kyanite isograd marks the ﬁrst appearance

of stably coexistent kyanite  biotite in the same way

that the staurolite isograd marks the ﬁrst appearance of

stably coexistent staurolite  biotite. Continued pro-

duction of this Al

SiO

polymorph through the kyanite

zone is accomplished by the continuous reaction

18.6 chlorite  muscovite  quartz →

kyanite  biotite  H

In an AFM diagram, this reaction is manifest by ex-

pansion toward M of the bundle of kyanite–biotite

tie-lines as the three-phase triangle kyanite  biotite 

chlorite also shifts continuously toward M (Figure

18.14e). Consequently, the range of stable chlorite solid

solutions is reduced to only the most Mg-rich com-

positions. Many pelitic rocks in the kyanite zone will,

therefore, consist of kyanite  biotite  staurolite 

muscovite  quartz  plagioclase. Apparently stable

assemblages in some terranes consist of staurolite 

biotite  garnet  kyanite ( muscovite  quartz 

plagioclase). This four-phase assemblage should not be

allowed in the three-component AFM diagram. A pos-

sible explanation depends on unusually large concen-

trations of components not ordinarily accounted for in

the diagram (Section 15.3), such as Ti stabilizing bio-

tite, Ca and Mn stabilizing garnet, or Zn stabilizing

staurolite. Some staurolites have as much as 2.7 wt.%

ZnO.

Sillimanite Zone. The highest grade zone seen by Bar-

row in Scotland was marked by the appearance of the

high-T Al

SiO

polymorph, sillimanite. In most rocks,

kyanite grains are not pseudomorphed by sillimanite.

Rather, ﬁne needles and bundles of sillimanite called

ﬁbrolite commonly nucleate within or on muscovite or

especially biotite (Figure 16.2). With increasing T into

the zone more robust prisms of sillimanite develop

(Figure 18.14f).

In the Barrovian terrane of Scotland, the prevail-

ing P (about 6 kbar) during metamorphism allowed

the almost coincidental destruction of staurolite at

the sillimanite isograd according to the discontinuous

reaction

18.7 staurolite  muscovite  quartz →

sillimanite  biotite  garnet  H

In Figure 18.14f, tie-lines linking staurolite to garnet,

biotite, and sillimanite all vanish as a single new

three-phase triangle representing coexistent sillimanite

 biotite  garnet replaces the three lower grade,

three-phase compatibilities. Continued contraction of

chlorite stability (Guidotti et al., 1991) with increasing

T by the continuous reaction 18.6 ﬁnally eliminates

even the most magnesian chlorites (Figure 18.13).

Sillimanite + Alkali Feldspar Zone. As temperatures

rise in many Barrovian terranes to above 750°C, mus-

covite solid solutions react with quartz to form a

potassic alkali feldspar plus Al

SiO

(reaction 14.1 and

Figure 14.31). At lower pressures, andalusite is stable,

whereas higher pressures produce kyanite. However, in

the majority of pelitic sequences, the typical Al

SiO

polymorph is sillimanite and its ﬁrst appearance with

alkali feldspar in the ﬁeld occurs above the upper, or

second, sillimanite isograd.

Metamorphism at Convergent Plate Margins: P–T–t Paths, Facies, and Zones

581

Three additional consequences of muscovite break-

down are petrologically signiﬁcant.

1. Many petrologists (Miyashiro, 1994; Blatt and

Tracy, 1996) regard the compatibility of sillimanite

 alkali feldspar as characteristic of the granulite

facies that is dominated by nonhydrous minerals.

2. With the demise of muscovite, the remaining phyl-

losilicate is only biotite. (Chlorite disappeared in the

lower part of the sillimanite zone.) Consequently,

pelitic rocks tend to be less schistose and more

gneissic.

3. Contributing to the gneissic aspect is the fact that

in many terranes this breakdown of muscovite

more or less coincides with the beginning of partial

melting, possibly prompted by the released water.

Many, if not most, migmatites (Section 15.2.3) are

believed to have formed by in situ dehydration

partial melting. (Production of partial melts was

discussed in Section 11.6 where it should be noted

that the temperatures of melting are relatively high

because of the high pressures of the experiments

and possible overstepping required.)

Partial melting has sometimes been viewed as the

culmination of metamorphism or as ultrametamor-

phism, but it can be debated whether either is appro-

priate. This phenomenon does, nonetheless, represent

a turning point in metamorphism because of the

afﬁnity of the melt phase for any water in the rock—it

is rather effectively dehydrated. Thereafter, any further

reactions at increasing T are dominated by nonhydrous

reactant and product phases.

Graphically, the demise of muscovite and stabilization

of alkali feldspar creates a fundamental change in the

AFM diagram because the point of projection is now

from feldspar in the AFMK tetrahedron (Figure 15.27a).

More magnesian pelites in the sillimanite  alkali

feldspar zone may contain cordierite—the most Mg-

rich of the common aluminous maﬁc minerals (Section

16.5.2). Mineral compatibilities involving cordierite,

shown in Figure 18.14g, suggest it appears by the con-

tinuous reaction

18.8 sillimanite  biotite  quartz →

cordierite  K-feldspar  H

Higher Grade Pelites. At still higher temperatures, the

stability of cordierite may expand toward more Fe-rich

compositions while garnet becomes more Mg-rich,

eliminating the compatibility of sillimanite  biotite

according to the continuous reaction

18.9 sillimanite  biotite  quartz →

cordierite  garnet  K-feldspar  H

This new equilibria involving the rare compatibility of

garnet  cordierite, shown in Figure 18.14h, occurs in

some high-grade migmatites where dehydration partial

melting allowed the liberated water to dissolve in the

melt. According to Yardley (1989), the appearance of

cordierite  garnet  K-feldspar deﬁnes the beginning

of the granulite facies.

Yet higher temperatures in the granulite facies can

cause breakdown of biotite to produce orthopyroxene

 K-feldspar that can coexist with combinations of

garnet, cordierite, sillimanite, and quartz (Figure 18.15;

see also the charnockite in Figure 15.20). Extreme

metamorphic temperatures of about 1000°C at 8–10

kbar are recorded in very unusual granulite-facies

rocks of the Precambrian Enderby Land in Antarctica

and the Archean Narryer complex of Western

Australia (Section 19.2.4) that contain the exotic

volatile-free Mg–Fe–Al silicates sapphirine, spinel, and

osumilite in addition to the above phases (Spear, 1993).

18.3.2 P T t Paths and Chronology of Barrovian

Metamorphism

One of the most thoroughly documented determina-

tions of P–T paths in a Barrovian terrane was made by

St-Onge (1987) in pelitic rocks of the early Proterozoic

Wopmay Orogen, Northwest Territories, Canada (Fig-

ure 18.16). GASP and garnet–biotite thermobarometry

(Section 16.11.2) was applied to six rock samples that

contained the basic assemblage garnet  biotite  pla-

gioclase  quartz  sillimanite  kyanite. Zoned poik-

iloblastic garnets 5–7 mm in diameter are crowded

with inclusions of the same minerals as occur in the

matrix, justifying use of these thermobarometers to

monitor the progress of metamorphism. Six to 17 de-

terminations of P and T were made for each of the six

samples by microprobe analyses of garnet and neigh-

boring inclusions from poikiloblast core outwards,

582 Igneous and Metamorphic Petrology

Orthopyroxene

Sillimanite

Crd

Grt

 K-feldspar

 quartz

 H

18.15 AFM diagram depicting pelitic equilibria with orthopyroxene

in the high-T part of the granulite facies. Under some condi-

tions cordierite  garnet (dashed line) is a possible compat-

ibility rather than sillimanite  orthopyroxene. Open square

is charnockite in Figure 15.20.

resulting in six separate P–T paths recording garnet

growth during declining P of about 2 kbar but T in-

crease of 25–75°C. The garnet-forming reaction was

probably staurolite  muscovite  quartz → garnet 

biotite  Al

SiO

 H

O. The six measured paths are

not nested, as predicted for samples at different crustal

depths in tectonothermal models of thickening crust

experiencing thermal relaxation (Figures 14.36 and

18.1c). This is explained by St-Onge in terms of large

amplitude folding that resulted in variable uplift rates

and an irregular crustal T distribution. Nonetheless,

the overall pattern of paths is consistent with the posi-

tion of the samples in the tilted 30-km-thick section in

the Wopmay Orogen.

The time of the Grampian metamorphism in the

Caledonide Orogen of Scotland that created the classic

Barrovian and Buchan zones (Figure 14.30) lasted from

about 480 to 465 Ma during collision between an island

arc and the continental margin of Laurentia (Oliver et

al., 2000). Laurentia (Figure 19.27) lay to the north of

the inferred oceanic trench now marked by blueschist-

facies rocks and ophiolite (Figure 14.30). Constraining

data include K–Ar, Rb–Sr, and Sm–Nd ages on the

metamorphic sole of the ophiolite (Section 18.5),

on metamorphic minerals in pelitic zones, and on

syn- and post-tectonic granitoids. A similar time of

metamorphism has been documented in the northern

Appalachian Mountains in eastern USA that were

connected with the Caledonide Orogen in northwest-

ern Europe before the opening of the Atlantic Ocean.

Dating shows the peak metamorphism in the Barrovian

and Buchan terranes was contemporaneous and within

about 8 My of the time of their exhumation, as indi-

cated by the age of erosional debris that contains meta-

morphic index minerals. These chronologic data show

that collisional orogeny, metamorphism, and exhuma-

tion can all occur within only several million years.

18.3.3 Buchan Metamorphism

Some regional metamorphic terranes preserve mineral

assemblages in pelitic rocks that equilibrated at dis-

tinctly lower P than typical Barrovian assemblages.

Metamorphic zones in the Buchan area of Scotland

north-east of Aberdeen (Figure 14.30) have long served

as the type terrane for such rocks, hence they are called

Buchan zones. Other well known Buchan zones occur

in New England, USA (Blatt and Tracy, 1996) and in

the Proterozoic of Australia (e.g. Johnson and Vernon,

1995; Warren and Ellis, 1996). There is a clear sim-

ilarity with pelitic mineral assemblages in contact

Metamorphism at Convergent Plate Margins: P–T–t Paths, Facies, and Zones

583

P (kbar)

020km

020miles

700500 600

T (°C)

Kyanite

Sillimanite

Archean

Chl

(b)

(a) (c)

Batho-

lith

Ksp  Sil

18.16 Barrovian pelitic rocks and their P–T paths, Wopmay Orogen, Northwest Territories, Canada. Redrawn from St-Onge (1987). (a) Index

map showing sample location. (b) Down-plunge cross-section constructed from the geologic map of the structurally tilted metamorphic

terrane that is 30 km thick. Four metamorphic zones are shown; the sampled muscovite  Al

SiO

( garnet  biotite  plagioclase 

quartz) zone is dark shaded. Syntectonic 1.9 Ga batholith is composed of tonalite and quartz diorite. Teeth are on the hanging-wall of a

late-metamorphic fault. (c) P–T paths for six samples whose locations are indicated in (b). For clarity, as many as 17 data points deﬁning

each of the other curves are omitted. See also Figures 16.35 and 18.19.

metamorphic aureoles (Figure 14.25) developed around

magmatic intrusions in the shallow (10 km) crust.

The occurrence of abundant synmetamorphic intru-

sions in Buchan terranes lends support to the notion

that the heat for metamorphism was magmatic; thus,

the terranes are often considered to be “regional con-

tact aureoles” possessing tectonite fabric instead of the

hornfelsic fabric of typical aureoles.

The lower pressures of equilibration of Buchan alu-

minosilicates lead to signiﬁcant contrasts with higher P

Barrovian assemblages (Figure 18.17; see also Figures

14.33 and 14.34a and Yardley, 1989). These contrasts

include:

1. Andalusite and, at higher grades, sillimanite are

stabilized rather than kyanite. Muscovite break-

down can occur in the stability ﬁeld of andalusite,

producing andalusite  K-feldspar compatibility.

2. Garnet and staurolite are less common or even absent

but cordierite is widespread and forms at much

lower temperatures than in some rare Barrovian rocks.

3. Migmatites are less common and developed only at

highest temperatures above the upper sillimanite

isograd where sillimanite  K-feldspar are compatible.

The Buchan biotite zone has coexisting muscovite

 chlorite  quartz as in Barrovian pelites. However,

the lower pressures are manifest in relatively less Al in

chlorites and less phengite end-member in muscovite.

A more striking contrast is found at the upper limit of

the biotite zone where, instead of the garnet isograd in

Barrovian terranes, the next index mineral to appear is

cordierite, manifest as ill-deﬁned “spots” in slaty rocks.

The reaction producing this isograd is

18.10 chlorite  muscovite →

cordierite  biotite  quartz  H

This reaction changes the basic topology of the AFM

diagram in Figure 18.17 and leads to stabilization of

the assemblage andalusite  cordierite  biotite as this

three-phase triangle sweeps into the AFM diagram

from the AM side, eclipsing bulk pelite compositions.

Changing biotite Fe/Mg ratio in the assemblage ac-

counts for most of the sweep. Like the garnet isograd

in Barrovian terranes, the position of the Buchan

cordierite isograd is composition dependent.

The andalusite isograd can be produced by two

possible reactions in rocks that are not enriched in Fe.

If chlorite has not been consumed by the cordierite-

producing reaction 18.10 above, then andalusite

appears as a result of

18.11 chlorite  muscovite  quartz →

andalusite  cordierite  biotite  H

On the other hand, if no chlorite is available, then an-

dalusite is produced at slightly higher temperatures by

18.12 cordierite  muscovite  quartz →

andalusite  biotite  H

P T t Paths. In contrast to Barrovian terranes, Buchan

terranes reveal no clear evidence of crustal thickening.

In fact, Wickham and Oxburgh (1985) describe

Buchan-style metamorphism in the Pyrenees along the

border of France and Spain that they believe occurred

in a continental rift setting where crust was thinned

and probably underplated by maﬁc magma. H and O

isotopic ratios indicate equilibration with advecting

seawater to depths of 12 km during the high-T, low-P

metamorphism. They suggest a modern analog might

be the Salton Sea area just north of the actively

opening Gulf of California near the international bor-

der of California and Mexico (Figure 13.29 and Special

Interest Box 18.1 and Figure 18.18).

The counter- (anti-) clockwise P–T path of Buchan

metamorphism (Figure 18.19) is inconsistent with the

crustal thickening that typiﬁes Barrovian terranes in

584 Igneous and Metamorphic Petrology

Biotite Biotite

CordieriteCordierite

AndalusiteAndalusite

Staurolite

Chlorite Garnet

Garnet

Increasing T

 Muscovite  quartz

 H

18.17 AFM diagrams depicting pelitic mineral assemblages during prograde Buchan metamorphism. With increasing T, staurolite and chlorite

(ellipse) are consumed and the compatibility ﬁeld of stable andalusite  biotite  cordierite sweeps toward more Fe-rich compositions,

encompassing most pelites (circles) and creating a cordierite isograd. Redrawn from Spear (1993).

Metamorphism at Convergent Plate Margins: P–T–t Paths, Facies, and Zones

585

Depth (km)

Prehnite

Epidote

300°C 320°C

Biotite

Chlorite

Actinolite

Prehnite

Epidote

Chlorite

ActinoliteClinopyroxene

150°C 230°C

Chlorite

Illite and

white mica

Illite and

white mica

Calcite

Epidote

Wairakite

Chlorite

Calcite

01km

02miles

18.18 Cross-section and mineral assemblages in the Cerro Prieta geothermal ﬁeld, Mexico. Top of diagram shows isotherms and distribution

of calcite-cemented impervious caprock (shaded) overlying geothermal reservoir; constraints are provided by numerous well cores

and cuttings. Mineral assemblages are plotted in terms of A  Al

 Fe

, C  CaO, F  FeO  MgO  MnO. Ellipse represents

composition of carbonate-cemented quartzo-feldspathic sandstone protolith. See Special Interest Box 18.1. Redrawn from Schiffman

et al. (1984).

586 Igneous and Metamorphic Petrology

P (kbar)

Depth (km)

400 500 600 700 800 900

T (°C)

And

Sil

Chl

Grt



Crd

18.19 Petrogenetic grid for intermediate grade pelitic rocks

comparing Barrovian and Buchan P–T paths and mineral

assemblages. Depending on rock composition and formative

reaction, cordierite is stable at higher temperatures and lower

pressures than dotted line labeled Crd. Reactions and assem-

blages in intermediate-P Barrovian terranes involve staurolite,

garnet, and kyanite, whereas in low-P Buchan regional ter-

ranes and contact aureoles cordierite and andalusite are dom-

inant phases. Compare Figure 14.1. Redrawn from Dymoke

and Sandiford (1992) and Spear (1993).

“

”

Buchan

And

Sil

Granulite

facies

Granitic

Diapir

(a)

(b)

(c)

(d)

Fertile source rock

18.20 Geologic model and schematic P–T diagrams for Buchan meta-

morphism. Redrawn from Warren and Ellis (1996). (a)–(c) show

progressive steps in growth of hypothetical buoyant granitic

magma diapirs from a fertile source layer near the base of the con-

tinental crust, with heat possibly inserted from an underlying

maﬁc magma underplate. These three diagrams were inspired

by laboratory experiments on “gravity tectonics” by Ramberg

(1981). Note the deepening of a reference volume (open square)

of the wallrock surrounding the ascending diapirs and the sub-

horizontal elongation of a reference volume (open circle and

ellipses) of the source layer. (d) Schematic diagram showing

P–T paths of reference volumes from upper model. The wall-

rock follows a counter-clockwise path similar to Buchan terranes,

while the extended and ﬂattened source layer follows a retro-

grade path similar to foliated and lineated granulite-facies rocks.

collisional orogens. The prograde part of the Buchan

path is nearly isobaric through the andalusite ﬁeld and

to peak T in the sillimanite ﬁeld; this trajectory is read-

ily explained in terms of heating by copious volumes of

intruded granitoid magma (Barton and Hanson, 1989),

these likely being a product of introduction of mantle-

derived maﬁc magma into the lower crust to drive par-

tial melting. But the subsequent upturn of the path to

higher P followed by isobaric cooling in an counter-

clockwise direction has proven to be a major hurdle in

interpretations of Buchan metamorphism.

Warren and Ellis (1996) have proposed a novel ex-

planation based on return ﬂow of country rocks sur-

rounding ascending granitoid diapirs (Figure 18.20; see

also Figure 9.15). Field relations around plutons com-

monly indicate the country rock moved downward in

response to the pluton making room for itself in the

shallow crust. Where the volume of intruded plutons is

large relative to intervening country rock the down-

warping in rim synclines can be considerable, if the re-

sults of laboratory models can be extrapolated to the

real crust. Following initial heating of the wallrock in a

regional contact aureole, continued rise of the diapiric

magma causes increased depression of the synclinal

wallrock to its maximum P. After the thermal pertur-

bation, possibly induced by maﬁc underplating at the

base of the crust, has died, the metamorphosed coun-

try rocks thermally relax along an essentially isobaric

trajectory back to a normal geotherm. The lack of

overall crustal thickening that results in signiﬁcant

isostatic rebound and unrooﬁng precludes develop-

ment of a clockwise path.

An intriguing facet of the crustal model in Figure

18.20 is the subhorizontal extension that takes place in