JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
PAGES 255–278
2003
Tectonometamorphic Evolution of the
Eastern Tibet Plateau: Evidence from the
Central Songpan–Garzê Orogenic Belt,
Western China
M.-H. HUANG1, I. S. BUICK1∗ AND L. W. HOU2
1
DEPARTMENT OF EARTH SCIENCES, LA TROBE UNIVERSITY, BUNDOORA, VIC. 3086, AUSTRALIA
2
SICHUAN EXPLORATION BUREAU OF GEOLOGY AND MINERAL RESOURCES, CHENGDU, 610081, P.R. CHINA
RECEIVED SEPTEMBER 3, 2001; REVISED TYPESCRIPT ACCEPTED AUGUST 1, 2002
The Songpan–Garzê Orogenic Belt (northeastern Tibet Plateau)
experienced polyphase deformation and metamorphism that is best
developed in the Danba Domal Metamorphic Terrane (DDMT),
in which three tectonometamorphic events can be identified. The first
event (D1–M1) is characterized by prograde sericite- to kyanite-grade
Barrovian metamorphism during late Indosinian (>205–190 Ma)
crustal thickening and shortening. A subsequent early Yanshanian
(>165 Ma) sillimanite- to migmatite-grade event (M2) developed
during predominantly east–west compression (D2). A final greenschist-facies event (M3) is best developed in shear zones of probable
Himalayan age. P–T conditions during M1 varied from >3–5
kbar and >410–530°C (biotite zone) to 5·3–8 kbar and
570–600°C (staurolite and kyanite zones), and during M2 from
4·8–6·3 kbar and 640–680°C (sillimanite zone) to 5·8–6·2
kbar and 660–725°C (migmatite zone). Clockwise P–T–t segments were inferred for the staurolite, kyanite and sillimanite zones.
Muscovite-dehydration melting during M2 was largely responsible
for the generation of migmatites and locally voluminous pegmatites.
The polyphase tectonometamorphic evolution of the eastern Tibet
Plateau, as documented in the Danba area, resulted from interactions
between the Indian, Tibet, and the South and North China Blocks.
The eastern Tibet Plateau experienced limited uplift during the
Mesozoic, followed by large-scale uplift and rapid cooling during
the Tertiary Himalayan Orogeny.
KEY WORDS: Barrovian-type metamorphism; migmatites; P–T–t path;
partial melting; Tibet Plateau
∗Corresponding author. Telephone: 61-3-9479-2647. Fax: 61-3-94791272. E-mail: [email protected]
INTRODUCTION
The Songpan–Garzê Orogenic Belt (SGOB) occupies
the northeastern portion of the Tibet Plateau (Fig. 1a),
which comprises the tectonically distinct Tethyan–
Himalayan domain between India and Eurasia and is
characterized by a polyphase continent–continent collisional history (Dewey et al., 1988). The Tibet Plateau
has been the subject of a number of studies, which have
focused on the Himalayas themselves and on the major
strike-slip faults within, or fault zones along, the margin
of the Tibet Plateau (e.g. Dirks et al., 1994; Roger et al.,
1995; Arne et al., 1997; Meng & Zhang, 1999). In
comparison, little work has focused on the interior of the
SGOB. As the SGOB resulted from the closure of the
Palaeo-Tethys and subsequent continental collision and
convergence associated with interaction between the Tibet, Indochina, South China and North China Blocks
(Xu et al., 1992), the multiphase deformational and metamorphic evolution of the SGOB is critical for understanding the Mesozoic–Tertiary tectonometamorphic
evolution of the Tibet Plateau.
Metamorphic studies of the SGOB are hampered by its
generally low grade (lower greenschist facies). However,
poorly documented medium-pressure Barrovian-type
metamorphic complexes as high as upper amphibolite
grade occur locally, and are associated with basementcored structural domes (Mattauer et al., 1992). The best
examples are the Danba Domal Metamorphic Terrane
(DDMT) in the central SGOB (Calassou, 1994; Figs 1
Journal of Petrology 44(2)  Oxford University Press 2003;
all rights reserved
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
SGOB, the DDMT, and to place metamorphic and
tectonic constraints on the evolution of the eastern Tibet
Plateau.
GEOLOGICAL SETTING
Songpan–Garzê Orogenic Belt
Fig. 1. (a) Tectonic setting of the Songpan–Garzê Orogenic Belt
(SGOB), and location of the Danba Domal Metamorphic Terrane
(DDMT). GTS, Gangdese Thrust System. (b) Schematic tectonic
map of the SGOB, showing the distribution of the granitoids, domal
metamorphic terranes and F1 fold axial traces (modified from Xu et al.,
1992). ICB, Indochina Block.
and 2) and the Xuelongbao dome in the Longmenshan
of the eastern SGOB (Dirks et al., 1994; Fig. 1a). Metamorphism in these medium- to high-grade regions is
generally considered to have occurred at the deepest
structural levels of the SGOB during the Indosinian
(>230–190 Ma) to Yanshanian (>190–65 Ma; Ren et
al., 1987; Fig. 3) Orogenies (e.g. Xu et al., 1992). However,
their P–T conditions, P–T–t–deformation path, and uplift
mechanism (see Hou, 1996) are generally not well constrained. In this study, we present new data to determine
the P–T–t history in the highest-grade portion of the
The SGOB is bounded to the north, south, SE and west
by the North China, Indochina, South China and the
Tibet Blocks, respectively (Fig. 1a). It was formed by
shortening and closure of a large sedimentary basin
during late Triassic subduction of the South China Block
northwards under the North China Block (Laurasia), and
of the Tibet (North and South Tibet) Block eastwards
under the South China Block (Dewey et al., 1988; Mattauer et al., 1992; Xu et al., 1992). The sedimentary basin
comprised a thick (5–10 km) sequence of Triassic flysch
that was conformably deposited on a 4–6 km thick Sinian
(Neoproterozoic)–Palaeozoic succession of sedimentary
rocks and basalts (Zhang & Luo, 1988), which themselves
unconformably overlie crystalline pre-Sinian (Archaean–
Mesoproterozoic) basement of the South China Block.
Two major episodes of granitic magmatism have been
recognized in the SGOB (Fig. 1b). The first comprises
calc-alkaline mica granites with U–Pb zircon ages of
>209–190 Ma (Mattauer et al., 1992). These granites
intrude the Triassic sequence, are pre- to syn-tectonic
with respect to penetrative deformation of presumed
Indosinian age and are characterized by contact aureoles
of 1–5 km width. The second comprises numerous elongate Miocene-age biotite granite plutons that are aligned
along the sinistral Xianshuihe strike-slip fault zone.
The composite Neoproterozoic–Triassic sequence was
multiply metamorphosed and deformed during late- to
post-Triassic collisions involving the Tibet, North China
and South China Blocks. During the Indosinian Orogeny,
the Triassic series was detached and thrust southwards
onto the South China Block, whereas the Sinian–Palaeozoic supracrustal sequence was deformed within a largescale intracontinental top-to-the-south high-strain zone
that resulted in considerable crustal thickening (Mattauer
et al., 1992). In general, the Triassic sediments experienced
very low- to low-grade greenschist-facies metamorphism,
but the Sinian–Palaeozoic sequences (e.g. in the Danba
area) experienced higher-grade medium-pressure metamorphism at this time (Mattauer et al., 1992).
During the Himalayan collision between India and
Asia, NW-striking strike-slip faulting and thrusting
strongly modified the tectonic grain of the terrane. Strikeslip movement is represented by the lithospheric-scale
ductile sinistral Xianshuihe and Red River faults (Figs
1a and 3) that crosscut the SGOB. The Xianshuihe fault
has been active since at least 16–12 Ma (Mattauer et al.,
1992; Roger et al., 1995), following the extrusion of the
256
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
Fig. 2. (a) Geological map showing the stratigraphy and metamorphic isograds of the DDMT (modified from Zhou et al., 1981; Hou et al.,
1996). (b) Detailed metamorphic map around Danba Township in (a). Also shown are sample locations (Χ) with sample numbers corresponding
to Table 1. CND, Cunnuchan dome; GCD, Gongcai dome; GZD, Gezong dome; QGD, Qinganlin dome; Ser, sericite; Mig, migmatite; XF,
Xianshuihe strike-slip fault. Mineral abbreviations after Kretz (1983).
Indochina Block from the South China Block along the
Red River fault between 50 and 20 Ma (Tapponnier et
al., 1990; Fig. 3). NE-trending Longmenshan foreland
thrusts were formed or reactivated with lateral movement
at this stage (Xu et al., 1992; Arne et al., 1997).
Local geology of the Danba region
The DDMT comprises a complex of pre-Sinian basement
orthogneisses (Gongcai gneiss suite exposed in the Cun-
nuchan (CND), Gongcai (GCD), Gezong (GZD) and
Qinganlin (QGD) domes; Fig. 2) and the overlying metamorphosed and deformed Sinian–Mesozoic cover. The
Gongcai gneiss suite is dominated by migmatized plagioclase-rich orthogneisses with inferred emplacement ages
between 784 ± 24 Ma (Gongcai orthogneiss; Xu et al.,
1996) and 864 ± 26 Ma (Gezong orthogneiss; Xu et al.,
1996; Fig. 3). The cover can be divided into four main
lithological associations in ascending order (Fig. 2): (1)
lowermost Sinian massive dolomitic marbles with a thin
257
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
Fig. 3. Sequence of regional events, and timing of deformation events constrained by geochronological data in the SGOB and DDMT. It
should be noted that the Yanshanian Orogeny brackets a long time period, but it occurred at different stages in different places in China or
east Asia (Ren et al., 1987). The bold bars represent the age data of Huang et al. (2002).
basal layer of micaschist and gneiss; (2) Silurian
(Ordovician)–Devonian metapelites intercalated with
quartzite, amphibolite, calc-silicate and marble; (3)
Carboniferous–Permian calcite marbles and metabasalts
with minor metapsammitic rocks; (4) Triassic flysch,
which comprises metasandstones, slates and phyllites.
Large-scale structural evolution
At least three deformation events can be distinguished
based on the sedimentation record in the Triassic basin
and overprinting fabrics (Figs 3–5). Among these events,
D1 can be divided into an early episode (D1a) and a major
episode (D1b).
For D1a, the lack of sediments younger than Norian–
Rhaetian (>210–205 Ma) indicates that sedimentation
in the SBOB was terminated in the late Triassic, presumably as a result of tectonism. This is consistent with
the emplacement ages of >206–190 Ma syn-tectonic
granitoids within the SGOB (Mattauer et al., 1992). The
S1a fabric appears to be sub-parallel to bedding (S0).
Around Indosinian granites at high stratigraphic levels
the S1a fabric defined by sericite + quartz + graphite
is preserved as inclusion trails in contact metamorphic
andalusite or staurolite, e.g. around the 204 Ma Keerying
granite (Xu et al., 1992), >70 km to the north of Danba
Town (Fig. 1b). Therefore, D1a occurred between >210
and 204 Ma.
D1b represents the most significant deformation event
in the DDMT. F1b folding intensifies with increasing
structural depth. Within the Triassic sequence, D1b is
characterized by extensive upright, generally east–westtrending chevron folds (Fig. 5; Calassou, 1994), with a
variably developed axial planar slaty cleavage. D1b folding
is ubiquitous in the Triassic flysch sequence across the
SGOB. It also intensifies towards lower structural levels
and changes from upright chevron folds to isoclinal folds
with a penetrative axial planar foliation (S1b; Calassou,
1994; Dirks et al., 1994; Worley et al., 1997) that is
defined by the preferred orientation of muscovite, biotite,
staurolite and kyanite depending on grade. Between the
base of the Triassic and the pre-Sinian basement, D1b
produced widespread high-shear zones of centimetre to
hundred metre width and up to kilometre-scale layerparallel, isoclinal recumbent F1b folds that developed
within a high-strain zone of kilometre to tens of kilometre
scale (Mattauer et al., 1992; Dirks et al., 1994; Figs 5 and
6a and b). The contact between the basement orthogneiss
and Sinian rocks forms a major D1b shear zone at the
base of this high-strain zone. Regionally, F1b fold axes
trend approximately east–west, and the L1b mineral lineation is approximately north–south-oriented (plunging
0–20° NNW–NNE; Fig. 5). Within the high-strain zone,
mica fish, asymmetric pressure shadows, and rotated
garnet and kyanite (Fig. 6c–e) are commonly developed.
The consistent asymmetry of kinematic indicators within
the north-dipping S1b foliation is in accordance with topto-the-south shearing during D1b (Fig. 5). Within the
DDMT, the asymmetry of overturned F1b folds and the
overturned southern margins of the CND, GCD and
QGD (Fig. 2) are further evidence supporting northover-south shearing and compression. Geochronological
258
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
Fig. 4. Relationships between mineral growth and deformation in the Danba area. The thickness of the black bar represents the relative
abundance of mineral, and the dashed line represents the extent of inferred mineral growth.
studies suggest that D1b occurred between >205 and
>190 Ma (Huang et al., 2002).
Similar observations have been made around the
Jianglang dome in the southern SGOB (Xu et al., 1992)
and the Xuelongbao dome in the central Longmenshan
(Dirks et al., 1994; Fig. 1b), where Barrovian-type metamorphic (chlorite to kyanite) zones were developed during
early shearing. This suggests that non-coaxial D1 deformation is regionally distributed throughout the SGOB.
The D2 event is expressed by north–south-trending,
open to tight folds with amplitudes ranging from centimetres to tens of kilometres (Fig. 6a and b). F2 folds
that overprint F1 fold structures have a variably developed, sub-vertical axial planar cleavage (S2; Fig. 5). In
the M2 sillimanite and migmatite zones in the northern
DDMT, S2 is defined by preferentially oriented sillimanite
and mica and represents the highest-grade fabric. In
contrast, to the south in the M1 chlorite–sericite to kyanite
zones S2 occurs as a fracture foliation that overprints the
higher-grade S1b at a high angle. On a map-scale, F2
folds deform M1b metamorphic isograds, and refold F1.
The outcrop pattern in the centre of the DDMT (Fig.
5) is due to interference of initially recumbent east–westtrending F1 folds and north–south-trending upright F2
folds [Type 2 interference pattern of Ramsay (1967)].
On a larger scale, D2 is responsible for the arcuate trend
of S1 throughout the SGOB (Fig. 1b; see Xu et al., 1992).
U–Pb (titanite and monazite) and Sm–Nd (garnet–whole
rock) geochronological data suggest that M2 took place
at >168–158 Ma (Huang et al., 2002). Together, field
and geochronological data strongly suggest that M1 and
M2 are separate events.
D3 is is characterized by NW–SE-trending sinistral
strike-slip shear zones, and NW–SE-trending and NEdipping thrusts and mesoscopic folds (Figs 1b, 5, and 6b
and f ). D3 quartz–sericite–chlorite mylonite zones of tens
of metres width occur along the strike-slip zones and
thrusts. F3 folds were found only in some incompetent
lithologies, and exhibit a steeply NE-dipping (>60°) crenulation cleavage or retrogressive schistosity (S3; Fig. 5).
The thrusts and strike-slip faults have the same grade
and orientation, suggesting that D3 occurred in a transpressional setting. D3 is thought to have formed as a
response to Tertiary Himalayan movement, and may
have been synchronous with the well-dated (Miocene)
Xianshuihe strike-slip fault (Roger et al., 1995).
Intrusive rocks
There are five major suites of granitoids in the Danba
area (Fig. 2a), at least one of which clearly pre-dates the
main Barrovian (M1–D1) metamorphic event. For the
sake of simplicity, contact aureoles around the granitoids
are not shown in Figs 2 and 5.
(1) The Manai granite (conventional U–Pb zircon ages
of >206 Ma, Xu et al., 1992; >197 ± 6 Ma, Calassou,
1994; Fig. 3) mainly comprises biotite granite and is
intensely deformed. It has been interpreted as a synorogenic (D1a) granite emplaced during the Indosinian
Orogeny (Calassou, 1994) before high-grade regional
metamorphism.
(2) The Manai syenite shows the same consistent
north–south-trending composite S1–S2 foliation as the
259
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
Fig. 5. Schematic tectonic map and cross-section showing the D1b, D2 and D3 structures of the DDMT. Lower hemisphere stereoplots refer to
orientations of the L1b lineation, and poles to S1b, S2 and S3 throughout this area. Data were incorporated from Zhou et al. (1981), Calassou
(1994), Hou et al. (1996) and this study. Symbols are as in Fig. 2.
260
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
Fig. 6. Structural relationships in the DDMT. (a) Overprinting relationships (recumbent F1b folds refolded by F2 folds) in thinly layered kyanitezone quartzites, 4 km NW of Danba Township. (b) Line drawing of S1b foliation refolded by F2 open folds in kyanite-zone metapelites. The two
generations of pegmatite veins (Pg1 and Pg2) that both crosscut the F1b and F2 folds, and that locally have undergone D3 ductile shearing with a
mylonitic foliation (S3Z) should be noted. (c) Micaschist from the staurolite zone in the major D1b high-strain zone, showing rotated garnet and
biotite. The asymmetry of biotite ‘fish’, and inclusion trails (tightly folded S1) defined by graphite in garnet indicate a top-to-the-south shear
sense. (d) Close-up view of crenulation cleavage S1b and folded S0/1a enclosed in garnet porphyroblast in (c). (e) Micaschist at the eastern contact
between Sinian metasediments and GCD basement, showing rotated garnet and kyanite. A top-to-the-south shear sense is also indicated by
asymmetric inclusion trails. (f ) Tight F3 folds of the S2 foliation were defined by sillimanite and mica. This is strongly folded, and a weak S3
axial planar foliation defined by sericite and minor chlorite is locally developed.
Manai granite and with it probably formed part of an
Indosinian composite intrusion.
(3) The Bianer suite consists of monzonite and minor
quartz diorite. Contact metamorphism around this suite
overprinted probable D1 fabrics.
(4) The north–south-trending Mongou biotite granite
intrudes the Manai granite. The contact metamorphic
assemblages on the western side of the Mongou granite
have been replaced by M2 assemblages, whereas on the
eastern side they overprint M1b assemblages. The Bianer
suite and Mongou pluton have not yet been dated, but
both appear to cut across D1 folds (Fig. 5). In other areas
to the west of the DDMT, this generation of granitoid
is cut by D3 faults. Therefore, it appears to be broadly
coeval with D2.
(5) The Zheduoshan granite suite is dominated by
261
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
with the mineral paragenesis Ser + Qtz ± Chl. Chlorite
is associated with prograde assemblages in the principal
S1b crenulation cleavage that overprints S1a, which is itself
defined by quartz, graphite and sericite.
biotite–muscovite granite and was emplaced at 16–13
Ma into the sub-vertical Xianshuihe strike-slip fault zone
(Roger et al., 1995).
METAMORPHIC ZONES, MINERAL
ASSEMBLAGES AND
MICROTEXTURES
Biotite zone
Assemblages in metapelites and amphibolites are the
focus for discussion of this study, and their distribution
as a function of metamorphic grade is summarized in
Fig. 4. Mineral abbreviations are after Kretz (1983).
The DDMT can be divided into a series of metamorphic zones based on mineral assemblages in metapelitic rocks (Fig. 2). The majority of zones are typical
of medium-pressure Barrovian-type metamorphism (Ser–
Chl, Bt, Grt, St and Ky zones). All isograds are commonly
parallel to, but locally crosscut at a high angle, the grossscale lithostratigraphy. The discordance between M1 isograds and the stratigraphy implies that large-scale folding
had already taken place before the isograds were set.
The Barrovian metamorphic zones are roughly symmetric with respect to the north–south-oriented basement
domes, with metamorphic grade decreasing towards the
east and west. In addition, M2 sillimanite (Sil) and migmatite zones (Mig; defined by partial melting in orthogneisses and metapelites) can be differentiated in the
granitic basement and closely adjacent metapelitic rocks
in the northern DDMT (Fig. 2). The sillimanite-in isograd
occurs around the domes in the northern DDMT, and
crosscuts the M1b garnet, staurolite and kyanite isograds,
and the closure of an F1b fold, developed in Silurian–
Devonian sedimentary units (Figs 2 and 5). Locally, the
sillimanite-in isograd is also folded by D2 structures,
suggesting that M2 occurred syn- to late-D2. In the NW
portion of the DDMT sillimanite-zone rocks are in fault
contact with much lower-grade equivalents (Fig. 2). In
early maps of the DDMT (Hou et al., 1996) this fault
was interpreted as a D3 thrust. However, given the overall
transpressive nature of NW–SE-trending D3 faults, its
orientation within the D3 stress field suggests that this is
either a D3 normal fault or a late fault unrelated to D3.
Metapelitic rocks exhibit abundant textural information that provides good constraints on the timing
relationships between deformation and metamorphic
mineral growth (Fig. 4), as described below.
Sericite–chlorite zone
This zone corresponds to the structurally highest Triassic
sequence, and is consistent with the regional low-grade
metamorphism observed in the Triassic elsewhere in the
SGOB. Metapelites from this zone are slates and phyllites,
In the eastern and western DDMT, this zone is restricted
to the Triassic metasediments, whereas in the southern
DDMT it includes the marble-dominated Silurian–
Permian sequence (Fig. 2a). The typical mineral assemblages are Bt + Ms + Qtz ± Chl ± Pl in metapelites
and Act + Ep (Czo) + Chl + Ab + Qtz ± Bt in
metabasic rocks.
In metapelites, biotite and muscovite define L1b and
the principal foliation (S1b). Biotite commonly occurs as
elongated porphyroblasts up to 4 mm in length that
contain inclusions of graphite, quartz and rare chlorite.
These inclusion trails define open to tight internal crenulations of S0/1a, or a straight internal foliation that is at
a low angle to the external S1b foliation, suggesting that
biotite formed during D1b.
Garnet zone
This zone is restricted largely to the upper Devonian–
Permian sequence. Assemblages in metapelites and amphibolites are Bt + Ms + Grt + Qtz ± Chl ± Pl and
Hbl + Pl + Czo (Ep) + Qtz + Ttn ± Chl ± Bt,
respectively.
In metapelites, biotite occurs as porphyroblasts up to 7
mm long, and in the matrix. The former show crenulated
inclusion trails similar to those in the biotite zone and
suggest growth early during D1b. The porphyroblasts are
typically deformed (mica fish), and indicate a top-tothe-south sense of shear during D1b. Garnet occurs as
subhedral, inclusion-rich porphyroblasts (2–4 mm diameter) in which quartz defines an internal foliation that
is generally continuous with, but locally discontinuous at
variable angles to, the external foliation (S1b). This suggests
that garnet growth was early to peak-D1b.
Staurolite zone
This zone is developed in Silurian–Devonian sequences
in the eastern and southern DDMT. Metapelitic rocks
are medium- to coarse-grained schists. The common
assemblages are St + Grt + Bt + Ms + Qtz ± Pl ±
Ilm in metapelites, and Hbl + Pl + Czo (Ep) + Qtz
+ Ttn ± Grt ± Bt ± Ilm in amphibolites.
Garnet in metapelites occurs in three forms. First, it
forms 3–15 mm diameter, equant, euhedral to subhedral
poikiloblasts with inclusion-free rims of 0·1–0·3 mm width
that are commonly included in staurolite. Inclusions
262
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
of fine-grained quartz, plagioclase, biotite and minor
ilmenite in the garnet cores define a sigmoidal to linear
internal foliation (Si) that is locally discordant with respect
to the external foliation Se (S1b). This garnet type is
interpreted to have started growing at an early stage of
cleavage development of S1b in a single continuous prograde growth period (see Bell & Rubenach, 1983). In highstrain zones, inclusion-rich garnet may show evidence for
syn-D1b growth and rotation (Fig. 6c–e). The second
garnet type occurs as inclusion-poor, elongated, anhedral
grains aligned parallel to S1b. This garnet type is finer
grained than the first, but contains coarser inclusions of
quartz and biotite. It is interpreted to represent mineral
growth at the peak of M1b–D1b. Lastly, garnet may occur
as fine euhedral crystals (0·2–0·6 mm) that overgrow S1b
(Fig. 7a). This generation is interpreted as a late growth
possibly related to D2, as discussed below for texturally
similar garnet in the kyanite zone.
Staurolite is generally restricted to the top unit of the
upper Silurian pelitic suite. It forms coarse, euhedral
tabular poikiloblasts (up to 10 cm) elongate in L1b. Porphyroblasts contain inclusions of quartz, biotite, euhedral
garnet and, rarely, ilmenite. These define internal foliations with similar relationships to the external S1b
foliation as seen at lower grades in biotite (Fig. 7b), and
which imply syn-D1b staurolite growth.
biotite content increases upgrade from the sillimanite-in
isograd.
Sillimanite is primarily fibrolitic and occurs in S2 (Fig.
7c), where it defines a sub-horizontal, L2 north–southoriented lineation (Fig. 5). Locally, sillimanite also occurs
in S1b, where it is inferred to overgrow M1 minerals
mimetically. Fibrolitic or prismatic sillimanite also forms
elongated pods in quartz or pegmatite veins (Fig. 6f ).
Two texturally distinct generations of garnet occur
in this zone. The first occurs as euhedral–subhedral
poikiloblasts, similar to those in the staurolite and kyanite
zones, although with lower abundance. These poikiloblasts are truncated and replaced by sillimanite intergrown with biotite (Fig. 7d). The second generation
is less common, and occurs as fine-grained, anhedral,
elongated grains (Fig. 7e). These grains are aligned
parallel to the sillimanite foliation. These relationships
suggest that the first garnet type is syn-D1 (M1) and the
second may be syn-D2 (M2). Rarely, euhedral garnet also
overgrows sillimanite, representing relatively late growth.
Staurolite is locally present as an isolated relict phase
undergoing replacement by sillimanite, biotite and garnet.
Immediately adjacent to the sillimanite-in isograd, relict
M1 kyanite locally coexists with M2 sillimanite.
Migmatite zone
Kyanite zone
This zone is restricted to the Silurian sediments around
the GCD (Fig. 2). The most common assemblages include
Ky + St + Grt + Bt + Ms + Qtz ± Pl ± Ilm and
St + Grt + Bt + Ms + Qtz ± Pl ± Ilm in metapelites.
Amphibolites display similar assemblages to those in the
staurolite zone.
Garnet and staurolite show similar textures to those in
the garnet and staurolite zones, but are coarser grained.
Kyanite is oriented in S1b (Fig. 6e), or rotated relative to
S1b, suggesting that it grew at the peak of M1b–D1b.
Texturally late, fine-grained garnet from a metapelite in
this zone has yielded an Sm–Nd (garnet–whole-rock) age
of >160 Ma, i.e. an M2 age (Huang et al., 2002), implying
a syn-M2 growth.
Sillimanite zone
In the sillimanite zone, the representative parageneses
are Sil + Bt + Ms + Qtz ± Grt ± Ilm ± Pl in
metapelites, and Hbl + Pl + Qtz ± Grt ± Czo ± Bt
± Ttn in amphibolites. In metapelites, muscovite is
much less common than biotite, and its abundance
decreases progressively northwards through the sillimanite zone. Biotite occurs as porphyroblasts rich in
graphite inclusions and as flakes in the matrix. The modal
The migmatite zone is variably developed in the dome
cores and overlying metasediments in the northern part
of the DDMT (Fig. 2). Amphibolites exhibit similar
assemblages to those in the sillimanite zone. Based on
the structural definitions of Mehnert (1968) and McLellan
(1983), two distinct migmatite types (stromatic migmatites
and metapelitic lenticular migmatites) occur. The stromatic migmatites can be further divided into those that
develop in basement quartzofeldspathic orthogneisses and
pelitic migmatites, respectively. The stromatic basement
migmatites (Fig. 7f ) will not be discussed in detail here.
Migmatitic metapelitic rocks experienced the highest
metamorphic grade in the DDMT and occur nowhere
else in the SGOB.
Muscovite-poor stromatic metapelitic migmatites are
locally present in Sinian–Silurian metapelites around
the CND (Fig. 2a). The leucosomes in these rocks are
concordant with S2 or a composite S1–S2 sillimanite–
biotite foliation (Fig. 7g). They make up 2–20 vol. % of
the outcrop and are characterized by the assemblage Sil
+ Bt + Pl + Qtz ± Ms ± Kfs. The assemblage
Sil–Kfs occurs mainly in metapelites around the basement
of the CND. In these migmatites muscovite appears as
relics within K-feldspar, plagioclase and quartz, and
plagioclase is locally surrounded by K-feldspar. The
mesosomes comprise Sil + Bt + Qtz ± Grt ± Pl ±
Ms. K-feldspar-present and -absent leucosomes probably
263
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
Fig. 7. Fabrics and reaction textures. (a) Texturally late garnet overprinting S1b in metapelites from the staurolite zone. Cross-polarized light.
(b) Inclusion trails defined by quartz and minor biotite in staurolite showing an early fabric (S0/1a) and crenulations. The weak internal crenulation
cleavage is parallel to S1b in the matrix. Cross-polarized light. (c) S2 foliation defined by oriented sillimanite and biotite that overprint sillimanitefree S1b. Cross-polarized light. (d) Subhedral garnet corroded by sillimanite and biotite. It should be noted that sillimanite intergrown with biotite
occurs between muscovite and garnet. Plane-polarized light. (e) Elongated garnet parallel to S2 defined by mica and euhedral garnet. Planepolarized light. (f ) Stromatic migmatites from the pre-Sinian basement of the GCD, showing distinct leucosomes (Pl + Qtz + Kfs) and
melanosomes (Hbl + Bt). (g) Migmatitic metapelites in the DDMT showing concordant leucocratic segregations. Leucosomes (Leuc) consist of
Pl + Qtz + Bt ± Ms ± Sil, and are surrounded mainly by Sil + Bt melanosome. (h) Relict kyanite and muscovite in microcline in metapelitic
lenticular migmatite immediately overlying the Gongcai dome core. Plane-polarized light.
264
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
formed by different melting reactions, as discussed below.
Metapelitic migmatites characterized by distinctive
leucocratic lenses occur within the lower Sinian sequence
on the southern side of the GCD. These lenses are up
to 6 cm in length and 4 cm in width, and consist of Pl
+ Kfs + Qtz + Ms + Bt ± Grt with minor kyanite
and/or sillimanite (Ky + Sil >1–3%). Between the lenses
are discontinuous mica-rich segregations that contain Ms
+ Bt + Qtz + Pl + Ky + Sil ± Grt. Microcline
occurs as 1–20 mm crystals rich in inclusions of relict,
isolated plagioclase, muscovite and kyanite (Fig. 7h).
Locally, the metapelitic lenticular migmatites preserve
enclaves of kyanite-bearing micaschists, or can be traced
laterally into kyanite-bearing micaschists.
In the sillimanite and migmatite zones, peraluminous
granitic pegmatites are extremely abundant. Minor pegmatite veins are also observed in the staurolite and kyanite
zones and basement cores. Pegmatites mainly comprise
Pl (Ab) + Mc + Ms + Bt + Qtz. Some additionally
contain large beryl, garnet and tourmaline crystals. They
are generally oriented parallel to, but locally crosscut, S2.
At least two generations of pegmatites can be identified,
and all have been deformed by D3 (e.g. Fig. 6b). This
suggests that pegmatites were emplaced post-D1 to preD3 and were probably related to M2 metamorphism.
Localized M3 metamorphism
Retrograde textures related to D3 are developed to variable extents in all zones. Where D3 folds are present,
pre-existing minerals are deformed and kinked. In all
retrogressed metapelites, biotite and garnet are altered
along margins and/or microfractures to chlorite, whereas
kyanite and feldspar are commonly altered to sericite.
Chlorite and sericite occur as retrograde phases that
define the mylonitic foliation in all of the NW-trending
D3 shear zones.
MINERAL CHEMISTRY
Mineral compositions from 15 metapelites and 11 amphibolites were analysed using an automated CAMECA
SX-50 electron microprobe at the University of Melbourne, with 15 kV accelerating potential, 25 nA beam
current and 1–5 m beam diameter. Data were reduced
by the Cameca PAP matrix correction program. Natural
minerals were used as standards. Structural formulae
were calculated after Droop (1987). Mineral assemblages
are given in Table 1, and a set of representative electron
microprobe analyses used for pressure–temperature calculations is given in Table 2. Further mineral analyses
are given in Electronic Appendices 1a–e, which can be
downloaded from the Journal of Petrology web site, at http:/
/www.petrology.oupjournals.org.
Garnet
Garnet in amphibolites is commonly unzoned, with the
exception of garnet from garnet–amphibole–biotite schist
9819 (Table 2; Electronic Appendix 1a), which preserves
growth zoning (discussed below). All garnet in amphibolites from the staurolite, kyanite and migmatite
zones is almandine rich (Alm0·57–0·76Prp0·09–0·11Grs0·11–0·28
Sps0·01–0·07), with Fe/(Fe + Mg) ranging between 0·81
and 0·94.
Garnet in metapelites is almandine rich, and shows
variable pyrope, grossular and spessartine contents
(Alm0·62–0·84Prp0·06–0·155Grs0·03–0·22Sps0·01–0·19) (Table 2; Electronic Appendix 1b). The Fe/(Fe + Mg) of M1 garnet
rims decreases slightly from the garnet (0·89) to migmatite
zones (0·85). Fine-grained, late-M2 garnet shows lower
pyrope contents and higher Fe/(Fe + Mg) (0·89–0·93).
Most garnet porphyroblasts in metapelites are compositionally zoned. Three patterns of chemical zonation
have been observed: (1) normal zoning (e.g. staurolite
zone sample D976; Fig. 8a) characterized by bell-shaped
zoning profiles for spessartine and grossular, as well as
Fe/(Fe + Mg) decreasing from core to rim; (2) completely
flat compositional profiles, which occur in all garnets <3
mm diameter in the central part of the migmatite zone
(e.g. sample 9821; Fig. 8b); (3) zoning profiles intermediate
between (1) and (2), in which steep compositional gradients near grain rims have been partially homogenized to
form plateaux. Coarse-grained garnets (>5 mm diameter)
with these profiles occur locally in the sillimanite (e.g.
sample D972334; Fig. 8c) and migmatite zones (e.g.
sample 9819; Fig. 8d). Type 1 zoning patterns are typical
of prograde garnet growth at medium to low grades
(Tracy, 1982; Spear et al., 1991), and in the DDMT are
likely to have developed during prograde M1. These
garnets additionally may show narrow rims with higher
Fe/(Fe + Mg), which probably resulted from limited
late retrogression or diffusion during cooling (Tracy,
1982). We interpret Type 2 profiles to have developed
by homogenization of original M1 growth zoning by
volume diffusion during M2 (see Florence & Spear, 1989),
and Type 3 profiles to represent partial homogenization.
Staurolite
Staurolite has high Fe/(Fe + Mg) values at all metamorphic grades (0·79–0·85; Table 2; Electronic Appendix
1c). It contains variable amounts of Zn with an apparent
trend towards higher contents with increasing grade
[0·020–0·038 cations per formula unit (p.f.u.) in the
staurolite zone and 0·04–0·059 cations p.f.u. in the kyanite and sillimanite zones]. Most euhedral staurolite
porphyroblasts are slightly zoned in composition. For
example, sample D976 in the staurolite zone shows a
265
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
Table 1: Mineral assemblages of the analysed samples
No.
Grade Sample
Rock-type
Qtz
Pl
Chl
Ser
Ms
Bt
Grt
St
Ky
Sil
Amp Ep/Czo
Ttn
Ru
Ilm
Mag
Metapelite
1
Bt
D973
Metapelite
+
+
+
+
+
+
+
2
Bt
9836
Metapelite
+
+
+
+
+
+
+
3
Grt
D9733
Metapelite
+
+
+
+
+
+
r
4
St
D976
Metapelite
+
+
+
+
+
+
5
St
VIII6
Metapelite
+
+
+
+
+
+
6
St
985539
Metapelite
+
+
+
+
+
7
St
92D30
Metapelite
+
+
+
+
+
8
Ky
9863b
Metapelite
+
+
+
+
+
9
Ky
D979
Metapelite
+
+
+
+
+
10
Ky
D9723
Metapelite
+
+
+
+
+
11
Ky
9813–2
Metapelite
+
+
+
+
+
12
Sil
D972334 Metapelite
+
+
13
Sil
D9741
Metapelite
+
+
14
Sil
9827
Metapelite
+
15
Mig
9814
Mig
9821
r
+
+
+
m
+
m
+
m
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
m
Migmatitic
+
+
+
+
+
+
pelite
+
+
m
+
+
+
pelite
16
r
m
Migmatitic
m
Amphibolite
17
Grt
D974
Metabasalt
+
+
18
Grt
983450
Metabasalt
+
+
+
+
+
+
19
St
D978
Amphibolite +
+
+
20
St
D9742
Amphibolite +
+
21
Ky
D97123
Amphibolite +
+
22
Ky
D9720
Amphibolite +
+
23
Sil
9822
Amphibolite +
+
24
Mig
9819
Amphibolite +
+
25
Mig
9821-2
Amphibolite +
+
26
Mig
9817-4
Amphibolite +
+
27
Mig
9820
Metagabbro∗ +
+
m
m
m
m
+
m
m
+
+
m
m
+
+
+
m
m
+
+
+
m
m
+
+
+
+
m
m
+
+
+
r
r
+
+
+
+
+
+
m
m
m
m
∗Protolith—mafic dyke intruding the basement rocks of the CND about 3 km from the eastern border.
Amp, amphibole; m, minor phase; r, retrograde phase; Ser, sericite. Other abbreviations are after Kretz (1983).
core to rim decrease in Mg from 0·39 to 0·29 p.f.u. and
an increase in Fe/(Fe + Mg) from 0·81 to 0·85.
Plagioclase
In amphibolites, plagioclase varies from oligoclase (An24)
in the garnet zone to andesine (An43) in the sillimanite
zone (Table 2; Electronic Appendix 1a). Plagioclase
inclusions in amphibole and garnet are generally more
anorthitic than that in the matrix, but are generally
unzoned. In sample 9819, which contains garnet porphyroblasts up to 22 mm diameter, matrix plagioclase is
abnormally albitic (An18).
The composition of plagioclase in metapelites is variable (An21–85), but no systematic change with increasing
grade was found (Table 2; Electronic Appendix 1c).
Individual plagioclase grains commonly show a core to
rim decrease in XAn. Plagioclase enclosed in garnet is
more anorthitic than matrix plagioclase, consistent with
closed-system garnet growth at the expense of plagioclase
(see Spear et al., 1991). Opposite plagioclase compositional
trends occur in sillimanite-grade metapelites where garnet
is replaced by sillimanite, biotite and new plagioclase
(e.g. sample D9741 plagioclase cores: An21; rims: An28).
In some sillimanite-grade, garnet-free metapelites, plagioclase occurs as elongate porphyroblasts intergrown with
266
267
9821
Mig
9821-2
9817-4
9820
Mig
Mig
Mig
c
r
c
r
c
r
c
r
∗
c
r
r
r
∗
c
r
c
r
c
r
∗
c
r
c
r
c
r
0·874
0·869
0·812
0·944
0·877
0·889
0·868
0·917
0·863
0·906
0·918
0·868
0·859
0·851
0·929
0·842
0·817
0·844
0·836
0·822
0·819
0·890
0·839
0·855
0·849
0·863
0·846
0·856
0·565
0·570
0·647
0·737
0·763
0·732
0·765
0·765
0·652
0·772
0·617
0·661
0·714
0·776
0·844
0·618
0·691
0·721
0·766
0·703
0·705
0·747
0·695
0·700
0·744
0·755
0·729
0·746
0·082
0·086
0·149
0·044
0·107
0·092
0·116
0·059
0·123
0·085
0·055
0·100
0·116
0·136
0·065
0·116
0·154
0·133
0·150
0·152
0·155
0·092
0·133
0·119
0·132
0·120
0·134
0·125
0·278
0·278
0·172
0·175
0·114
0·103
0·102
0·099
0·064
0·060
0·191
0·224
0·071
0·054
0·057
0·153
0·105
0·053
0·029
0·046
0·045
0·031
0·033
0·037
0·044
0·049
0·062
0·052
XGrs
0·071
0·063
0·031
0·038
0·013
0·073
0·016
0·189
0·043
0·037
0·135
0·014
0·099
0·034
0·034
0·117
0·052
0·092
0·055
0·101
0·097
0·131
0·139
0·144
0·081
0·076
0·074
0·075
XSps
0·81
0·79
0·81
0·85
XFe
XPrp
XFe
XAlm
St
Grt
0·06
0·06
0·02
0·03
XZn
0·24
0·23
0·38
0·34
0·21
0·29
0·38
0·44
0·35
0·18
0·28
0·33
0·39
0·33
0·22
0·17
0·27
0·21
0·67
0·55
0·60
0·55
0·28
0·25
0·31
0·39
0·32
0·12
XAn
Pl
0·76
0·76
0·62
0·65
0·79
0·71
0·62
0·56
0·64
0·82
0·72
0·67
0·60
0·65
0·78
0·83
0·73
0·78
0·33
0·44
0·40
0·45
0·72
0·74
0·69
0·31
0·62
0·88
XAb
0·117
0·171
0·171
0·120
0·135
0·120
0·105
0·107
0·110
0·103
0·089
0·094
0·098
0·084
0·099
XTi
Bt
0·532
0·556
0·563
0·551
0·535
0·500
0·459
0·493
0·507
0·462
0·514
0·445
0·460
0·501
0·508
XFe
0·065
0·062
0·054
0·067
0·080
0·064
0·047
0·067
0·086
0·090
0·086
0·060
0·158
0·070
0·420
XF
3·079
3·061
3·041
3·017
3·080
3·057
3·170
3·209
3·083
3·143
3·070
3·125
3·175
XSi
Ms
0·054
0·119
0·129
0·090
0·204
0·163
0·060
0·138
0·171
0·237
0·102
0·084
0·079
XPa
0·024
0·019
0·001
0·005
0·040
0·005
0·002
0·010
0·030
0·030
0·020
0·016
0·180
XF
7·525
6·432
6·458
6·497
6·456
6·436
6·346
6·192
6·159
6·360
6·404
6·455
XSi
Amp
0·674
2·339
2·137
2·005
2·020
2·505
2·448
2·779
2·855
2·594
2·387
2·432
XAl
0·195
0·656
0·521
0·452
0·470
0·510
0·458
0·466
0·544
0·346
0·344
0·377
XNa
0·489
0·559
0·449
XFe
4·867∗ 0·278∗
5·368∗ 0·344∗
5·483
5·481
5·118
XAl
Chl
XSi, XTi, XAl, XZn and XF represent the molar fractions of Si, Ti, Al, Zn and F per formula unit (p.f.u.) in minerals, respectively. Amp, amphibole; Pa, paragonite; c,
core (for garnet and plagioclase); r, rim (for garnet and plagioclase).
∗Retrograde.
D97123
D9720
9822
9819
Ky
Ky
Sil
Mig
Amphibolites
Grt
D974
St
983450
St
D978
St
D9742
9814
Mig
D972334
Sil
9827
D9723
Ky
Sil
9863b
D979
Ky
Ky
D9741
985539
St
Sil
D976
St
Metapelites
Bt
D973
Bt
9836
Grt
D9733
Grade Sample
Table 2: Summary of representative microprobe analyses used for P–T calculations
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
Fig. 8. Compositional profiles of garnet from: (a) a staurolite-zone metapelite; (b) a migmatite-zone metapelite; (c) a sillimanite-zone metapelite
containing kyanite and staurolite; (d) a migmatite-zone garnet–cummingtonite–biotite schist.
sillimanite and biotite and is very rich in anorthite (e.g.
An85). These anorthite-rich porphyroblasts are interpreted
as resulting from the complete breakdown of garnet
during M2.
Biotite and muscovite
Most biotite, especially that from sample D976 and from
the kyanite zone, displays low K contents (K + Na
+ Ca = 0·83–0·98 cations p.f.u.; Table 2; Electronic
Appendix 1d). This may reflect minor late alteration. At
the lower grades, biotite porphyroblasts are slightly zoned,
with Fe and Ti increasing from core to rim, whereas in
the migmatite zone biotite porphyroblasts are compositionally homogeneous. The Mg content decreases
and Ti content increases with increasing metamorphic
grade (from 1·35 and 0·094 cations p.f.u. in the biotite
zone to 0·965 and 0·171 cations p.f.u. in the migmatite
zone). Biotite included in garnet has lower Fe/(Fe +
Mg) than other types of biotite. The F and Cl contents
of biotite are 0·047–0·158 and 0–0·003 anions p.f.u.,
respectively. The Si content of muscovite varies between
3·02 and 3·16 cations p.f.u. based on 11 oxygens (Table
2; Electronic Appendix 1d), with the values in the sillimanite and migmatite zones slightly lower than in the
other zones. Muscovite contains 5–20 mol % paragonite
component.
Amphibole
According to the nomenclature of Leake (1978), amphiboles in the amphibolites analysed are classified as
ferro-tschermakitic hornblende, tschermakitic hornblende and tschermakite (Table 2; Electronic Appendix
1a). Sample 9820 from the basement has the lowest Al
content (2·005 p.f.u.). Amphibole in sample 9819 is
268
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
cummingtonite, which could have been derived from
originally very Fe-rich marl as the rock contains abundant
magnetite and as calcite was found within coexisting
garnet. No obvious correlation exists between amphibole
chemistry and metamorphic grade.
Chlorite
Primary chlorite in metapelites from the biotite and
garnet zones has total Al contents of 5·44–5·52 cations
p.f.u. Chlorite from the biotite zone has lower Fe/(Fe
+ Mg) (0·44–0·45) than that from the garnet zone
(Table 2; Electronic Appendix 1e). Retrograde chlorite
in amphibolites has more scattered compositions, with
Si, Al and Fe contents ranging from 5·15 to 5·72, 4·72
to 5·54 and 2·57 to 3·27 cations p.f.u., respectively. Fe/
(Fe + Mg) of the retrograde chlorite lies in the range
0·28–0·35.
METAMORPHIC P–T CONDITIONS
Selected metapelites and amphibolites (Fig. 2a; Tables 1
and 2) were used to calculate P–T conditions for the
major metamorphic zones. Where garnet displays growth
zoning, the compositions just inboard of narrow retrograde rims were used to estimate the peak temperatures
and pressures. For those garnets in the sillimanite and
migmatite zones that show flat compositional profiles,
the average compositions were used. The average compositions of matrix biotite were used for geothermobarometry. The average or most albitic rim
compositions of plagioclase were used for geobarometric
calculation (Spear et al., 1991). Mineral end-member
activities are listed in Table 3, and results of P–T calculations are summarized in Tables 4 and 5.
with the software AX98 (T. J. B. Holland, personal
communication, 1998; Table 3). Amphibolites were not
used for average P–T calculations because of their relatively high variance. Two types of calculations were
made: (1) average pressures (avP) or average temperatures
(avT ) were calculated for a fixed range of temperatures
or pressures, respectively; (2) average pressures and temperatures (avPT ) were calculated simultaneously. Average P, T and average PT results are given in Table 4.
Results satisfy the least-squares test with fit <1·4.
Even though no suitable conventional geothermobarometers are available for P–T calculations for
the biotite-zone micaschists (Bt–Ms–Chl–Pl–Qtz),
THERMOCALC calculations for two metapelites
(sample D973 and 9836) in this zone gave avPT of 490
± 40°C and 4·9 ± 3·4 kbar and 480 ± 70°C and 3
± 3 kbar, respectively (Table 4). M1 average temperatures (avT ) and pressures (avP) change from
>560–580°C and >4·8–6·4 kbar (garnet zone) to
580–590°C and 5·3–7·4 kbar (staurolite and kyanite
zones). M2 average temperatures (avT ) and pressures
(avP) change from >640–680°C and 4·8–6·3 kbar (sillimanite zone) to >660–700°C and 6–6·2 kbar (migmatite zone; Table 4). Average pressure–temperature
calculations from all metapelites except sample D9741
are very close to the individual average pressure and
temperature estimates (Table 4). Sample D9741 in the
sillimanite zone yielded higher avPT values of 717 ±
78°C and 6·3 ± 2·6 kbar. The large uncertainties are
interpreted to result from late incomplete re-equilibration,
as this sample shows moderate retrogression to chlorite
and sericite. AvPT values from sample 9827 in the
sillimanite zone are of lower pressure (4·8 ± 3 kbar) and
higher temperature (678 ± 284°C) than the average
pressure and temperature. These higher 2 uncertainties
could be related to the limited number of mineral endmembers available in the mineral assemblage (Sil–
Grt–Bt–Ms–Qtz) for this rock.
Average P–T calculations
Average P–T conditions of mineral assemblages were
calculated based on the equilibrium thermodynamic
method of Powell & Holland (1988) and Holland &
Powell (1998). Calculations were carried out using version
2.6 of the program THERMOCALC (Powell et al., 1998)
together with the Holland & Powell (1998) internally
consistent thermodynamic dataset. Many of the metapelitic rocks contain minor graphite, and so would have
coexisted with a fluid of mixed CO2 and H2O with
XH2O buffered to high values but <1 (Ohmoto & Kerrick,
1977). Because of a lack of assemblages to constrain
oxygen fugacity, a reference value of XH2O = 0·9 was
used for all calculations. If XH2O = 1 was used, estimated
pressures and temperatures decreased only by 0·2–0·4
kbar and 5–10°C. End-member activities were calculated
Conventional geothermobarometry of
amphibolites
As amphibolites lacked sufficient minerals to yield linearly
independent sets of reactions, conventional geothermometry and geobarometry were applied to calculate
P–T conditions (Table 5). Using the garnet–hornblende
geothermometer of Graham & Powell (1984), temperatures were estimated at 586 and 630°C for the
kyanite-zone samples D97123 and D9720, respectively.
The former result is consistent with that of THERMOCALC calculations, but the latter is slightly higher. Temperatures for samples 9819 from the migmatite zone and
9822 from the sillimanite zone (596 and 584°C) are much
lower than those for the adjacent metapelite (sample
269
0·41
4600
230
1390
254
250
370
590
985539
9863b
D979
D9723
270
300
9821
230
0·43
0·38
0·41
0·31
0·29
0·4
0·27
27
4000
0
24000
7100
1400
1110
340
8300
0·031
0·044
0·04
0·044
0·052
0·062
0·061
0·047
0·056
0·051
0·055
0·052
0·325
0·503
0·77
0·31
0·95
0·65
0·94
0·513
0·434
0·79
0·685
0·532
0·55
pa
0·02
0·018
0·011
0
0·014
0
0·019
0·011
0·03
0
0·023
0·037
0·037
cel
0·31
0·31
0·25
0·73
0·39
0·42
0·74
0·48
0·6
0·22
0·53
an
0·78
0·78
0·83
0·51
0·74
0·72
0·5
0·69
0·61
0·87
0·68
ab
192
280
64
(×10−5)
mst
0·391
0·352
0·5
fst
0·028
0·047
clin
0·016
0·015
daph
order Al ( M4) with random mixing of Al and Si on 2-sites; Wclin-daph = 2·5, Wclin-ames = 18 and Wames-daph =
Staurolite (after Worley & Powell, 1998):
Chlorite ( T. J. B. Holland, personal communication, 1998):
End-member abbreviations are after Holland & Powell (1998), except for pyrope, grossular and spessartine.
20·5 kJ
1-binary, ideal mixing
4-site Fe–Mg mixing amst = [1 – Fe/(Fe + Mg + Zn)]4 and afst = [Fe/( Fe + Mg + Zn)]4
Plagioclase (Holland & Powell, 1992):
A1–M1 ordered, ideal site-mixing; Wpa = 9, Wpe = 10, Wpo = 3, Wao = 6, Wae = −1 and Woe = 15 kJ
0·044
0·056
ames
ideal mixing
Wprp.adr = 73, Walm.adr = 60 and Wsps.andr = 60 kJ
2-site non-ideal mixing and regular solution gammas; Wprp.alm = 2·5, Wgrs.prp = 41·4–0·0188T,
0·71
0·69
0·77
0·67
0·72
0·66
0·69
0·67
0·69
0·67
0·73
0·6
0·65
mu
Chl
Muscovite ( T. J. B. Holland, personal communication, 1998):
0·033
0·044
0·04
0·049
0·047
0·064
0·056
0·042
0·047
0·047
1·695
7·96
(×10−3)
naph
St
Biotite (T. J. B. Holland, personal communication, 1998):
0·049
0·053
0·063
0·052
0·046
0·03
0·038
0·042
0·047
0·047
0·055
0·069
east
Pl
NUMBER 2
Garnet ( T. J. B. Holland, personal communication, 1998):
Activity model
47
130
0·26
0·32
35
640
0·041
0·03
ann
Ms
VOLUME 44
aqtz = aky = asill = 1
310
220
9827
490
9814
31
92
D972334 420
D9741
14000
0·37
0·053
280
270
280
D9733
D976
0·067
0·063
D973
1400
phl
(×10−7)
(×10−6)
(×10−5)
sps
grs
prp
alm
Bt
Grt
End-member activity
9836
Sample
Table 3: End-member activities used for average P–T calculations of metapelites in the DDMT
JOURNAL OF PETROLOGY
FEBRUARY 2003
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
Table 4: Summary of average P–T results for metapelites
Grade Sample
Bt
Bt
Grt
St
St
Ky
Ky
Ky
Sil
Sil
Sil
Sil
Mig
Mig
avP (kbar)
avT (°C)
avPT
°C:
500
550
575
600
650
700
725
kbar:
5
5·5
6
6·5
7
7·5
P
°C:
400∗
425∗
450∗
475∗
500∗
525∗
550∗
kbar:
3·5∗
4·0∗
4·5∗
5·0∗
5·5∗
6·0∗
(kbar) (°C)
D973
3·1
3·4
3·9
4·3
4·8
5·3
5·9
484
491
499
506
513
519
4·9
490
2
4·5
3·5
3·1
2·7
2·8
3
3·6
58
30
60
62
64
66
3·4
40
fit
1·9
1·5
1·2
1
0·9
0·9
1·1
1
9836
2·6
2·9
3·3
3·8
4·2
4·8
5·3
383
391
401
410
418
427
3
480
16
16
16
16
16
18
3
70
2
3·7
2·9
2·6
2·6
2·7
2·8
3·2
fit
1·5
1·2
0·9
0·7
0·8
0·9
1·1
4·8
5·6
6·4
D9733
1·5
0·9
1·5
0·9
1·4
0·9
1·5
0·9
1·6
1
T
1·1
1·7
0·9
563
568
573
577
581
5·4
569
30
32
32
34
38
2·4
42
2
1·8
1·6
2
fit
1·5
1·4
1·6
D976
4·5
5·6
6·6
568
577
587
595
605
615
6·1
589
24
24
24
24
24
26
2·2
44
7·7
615
1·4
1·4
1·5
1·5
1·7
1·6
2
1·3
1·1
1·2
fit
1·1
0·8
0·8
985539
5·9
6·7
7·3
542
556
567
580
590
600
50
46
42
40
38
40
2
1·5
1·4
1·4
fit
1·5
1·4
1·3
3·3
5·3
9863b
0·8
1·6
0·7
1·4
0·7
1·3
0·8
1·2
0·8
1·2
0·9
0·9
3
1·3
90
1·4
6·3
623
633
642
651
660
6·7
655
24
24
24
24
2·4
48
2
2·2
1·7
1·2
24
fit
1·9
1·5
0·8
1
6·1
7·5
9·5
569
576
582
589
595
601
7·5
602
28
26
24
22
18
18
2·2
34
D979
4·6
2
2·5
1·8
1·3
1·3
fit
2
1·4
1
0·9
D9723
4·9
6
7·1
9·4
1·7
0·9
1·5
0·9
1·4
0·8
1·3
0·9
1·1
0·9
1
1·2
560
568
577
584
592
600
6
576
24
24
24
24
24
24
2
40
2
1
1·1
1·1
2
fit
0·8
0·6
0·9
1·3
4·9
5·5
6·7
579
597
615
634
652
670
5·3
588
12
16
20
30
0·6
26
D972334 (St, Ky & Sil)
0·7
0·7
0·5
0·5
0·7
0·9
0·7
2
0·3
0·2
0·4
10
10
fit
1·1
1
1·9
1
1
D972334 (no Ky & St)
4·5
5·5
6·2
8·4
572
581
589
597
605
612
5·8
590
2
0·6
0·7
0·7
0·8
26
26
26
26
26
26
2·2
48
fit
0·8
0·6
0·6
0·9
5·6
6·2
6·7
D9741
0·8
0·7
1·3
0·7
1·7
0·7
2·2
0·8
2·7
1·1
0·9
0·7
6·9
640
652
665
677
689
701
6·3
717
28
28
30
30
30
32
2·6
78
2
2·3
2·1
2·1
2·2
fit
1·2
0·9
0·7
0·7
†
†
†
†
679
694
710
728
746
761
4·8
678
232
232
264
264
228
230
3
284
9827
†
†
0·9
0·8
0·8
0·8
0·9
1
0·9
2
†
†
†
†
†
†
fit
†
†
†
†
†
†
9814
4
5·7
7·5
634
644
655
665
676
685
6·2
663
2
0·6
0·4
0·5
10
9
10
10
10
10
1·2
29
fit
1·2
0·6
0·8
5·2
6·8
7·5
659
671
684
697
709
720
6·1
692
32
32
32
30
30
30
2·4
70
9821
4·4
2
1·7
1·2
1·2
1·2
fit
1·5
0·8
0·6
0·7
0·9
0·8
0·9
1
0·7
0·8
1·1
0·6
0·7
1·3
0·6
0·6
1·5
0·7
0·7
1·7
0·9
0·8
0·8
0·7
0·7
Average P–T calculations for the peak metamorphism were made using version 2.6 of THERMOCALC (Powell et al., 1998)
consistently at XH2O = 0·9. 2, standard deviation; fit, 2 test.
∗PT range only for biotite zone samples (D973 and 9836). †Not available.
271
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
Table 5: Summary of P–T results (in kbar and °C) for amphibolites calculated using conventional
geothermobarometry
Grade
Sample
TG+P
TH+B
(P = 6 kbar)
(P = 6 kbar)
TP
TB
PP
PK+S
m
m
m
r
m
m
300
7·0
Grt
D974
565
528
Grt
983450
575
525
8·0
St
D978
587
570
6·1
St
D9742
577
595
Ky
D97123
586
550
540
8·1
7·2
Ky
D9720
630
585
575
6·2
6·8
600
5·5
5·2 (T = 650°C)
Sil
9822
584
587
Mig
9819
596
579
Mig
9821-2
Mig
9817-4
674
Mig
9820
723
290–350
7·0
4·2 (T = 650°C)
626
330–345
5·8 (PA+S)
Temperatures were calculated using garnet–hornblende Mg–Fe exchange geothermometer (TG+P , Graham & Powell, 1984),
amphibole–plagioclase geothermometer (TH+B, Holland & Blundy, 1994), plagioclase–hornblende geothermometer (TP ,
Plyusnina, 1982) and chlorite geothermometer (TB, Bevins et al., 1991); pressures calculated using the garnet–plagioclase–hornblende–quartz geothermobarometer (PK+S, Kohn & Spear, 1989), plagioclase–hornblende geothermobarometer
(PP , Plyusnina, 1982) and Al-in-hornblende barometer (PA+S, Anderson & Smith, 1995). m, matrix; r, retrogression.
9821; >690°C), and hence probably suggest that garnet
is not in equilibrium with amphibole. Garnets in sample
9819 range from 1·5 to 3 cm in diameter (see above),
which is far larger than the distance at which diffusive
Fe–Mg exchange occurs at >650°C (Florence & Spear,
1989). Garnet from this sample shows prograde zoning
that is inferred to have developed during M1 and yielded
Sm–Nd ages of >204–200 Ma (Huang et al., 2002).
These two temperature estimates may be considered as
supporting evidence for remnants of an early metamorphic phase, and could approximately represent the
pre-M2 conditions. Therefore, we suggest that these rocks
give M1b rather than M2 temperatures, probably indicating they have not been re-equilibrated during M2.
Temperatures calculated from the amphibole–plagioclase
thermometer of Holland & Blundy (1994) are 565°C
(garnet zone), 577–587°C (staurolite zone) and
585–587°C (kyanite zone). These are in the range of those
calculated from THERMOCALC, except for migmatitezone samples 9819 and 9821-2. Temperature estimates
for the metagabbro sample 9820 (Table 5) indicate that
the CND basement was metamorphosed at temperatures
in excess of 720°C during M2. Plots of plagioclase–calcic
amphibole pairs after Plyusnina (1982) indicate a reasonable temperature range of 530°C for the garnet zone to
560–570°C for the kyanite zone. Temperatures obtained
based on an x–T plot of retrograde (M3) chlorite (Bevins
et al., 1991, and references therein) from two amphibolites
in the staurolite (D9742) and migmatite (9821-2) zones
yielded temperatures of 290–350°C (Table 5).
Garnet–plagioclase–amphibole–quartz assemblages in
amphibolites were also used to calculate pressures. According to the calibrations of the plagioclase–hornblende
geothermobarometer (Plyusnina, 1982) and garnet–
plagioclase–hornblende–quartz geobarometer (Kohn &
Spear, 1989), a cluster of higher pressures of >6·6–8
kbar was obtained for the garnet to kyanite zones, and
slightly lower pressures of >5·5 kbar for the sillimanite
zone compared with pressures derived from the metapelites. However, these results are probably within uncertainty of calculations made from metapelites (1–2
kbar). The range of 6·6–8 kbar probably represents the
maximum pressure for the M1 metamorphism. For sample
9820 (CND basement), a pressure of 5·8 kbar was obtained from the intersection between the thermometer
of Holland & Blundy (1994) with the Al-in-hornblende
geobarometer of Anderson & Smith (1995). This geobarometer is applicable to limited rock types with assemblages of Hbl + Kfs + Pl + Qtz + Bt + Ttn +
Czo. Sample 9820 lacks K-feldspar, but contains the
other minerals. In the absence of K-feldspar, the geobarometer can still be applied; however, the pressure
estimate is the maximum, as discussed by Anderson &
Smith (1995).
272
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
DISCUSSION
P–T–t paths
The P–T–t history of the DDMT may be constrained
by phase relations based on the observed mineral parageneses and microtextures. As M1a phases are only
rarely included as relics in the M1b minerals, and M1a
mineral parageneses indicate lower greenschist-facies
metamorphism, it is impossible to precisely constrain the
P–T conditions or evolution of the M1a event. Given the
mineral assemblage Ser–Chl–Ab–Qtz commonly seen in
the sericite–chlorite zone, temperatures are likely to be
lower than those in the biotite zone ( T Ζ475°C), and
could be in the range 300–420°C on the basis of metamorphic gradients and reported temperatures for chlorite-zone metapelites elsewhere in the world (Turner,
1981). Similarly, the late retrograde conditions of M3 are
not well constrained owing to the high variance of M3
mineral assemblages. Temperatures for this event could
only be roughly estimated at 290–350°C according to
the chlorite solid solution geothermometer of Bevins et
al. (1991).
Phase relationships during M1b and M2 in the biotite
to sillimanite zones can be expressed by using a petrogenetic grid for metapelites in the KFMASH system (e.g.
Spear & Cheney, 1989; Powell et al., 1998; Fig. 9).
Addition of minor components such as MnO and CaO
to the KFMASH system could affect the stability of some
minerals in the grid. However, in the present case the
KFMASH grid provides an adequate theoretical framework in which to examine phase relationships. Figure 9
shows a partial KFMASH petrogenetic grid relevant to
the biotite to sillimanite zones, with some additional
experimental constraints relevant to melting in the migmatite zone.
In most of the biotite-zone metapelites, chlorite occurs
in association with biotite and muscovite. However, in
the garnet zone chlorite is minor or absent. This suggests
that garnet has probably formed by the continuous
KFMASH reaction
Chl + Ms + Qtz = Grt + Bt + H2O
(1)
which migrates to more Mg-rich compositions with rising
temperature. The Fe-end-member reaction for (1) occurs
at >500°C (Fig. 9) and may provide a minimum temperature constraint for the garnet zone. This is consistent
with P–T constraints from THERMOCALC calculations
(Table 4).
The first appearance of staurolite is probably related
either to the KFMASH discontinuous reaction
Grt + Chl + Ms + Qtz = St + Bt + H2O (2)
or the continuous KFMASH reaction
Chl + Bt + Qtz = St + Ms + H2O.
(3)
Both reactions explain the lack of chlorite in the staurolite
zone. Reaction (2) occurs at 570–580°C at 6–7 kbar (Fig.
9), and provides a minimum temperature constraint for
the staurolite zone.
In the kyanite zone, kyanite commonly coexists with
staurolite, suggesting that the first appearance of kyanite
is related to
St + Ms + Chl + Qtz = Ky + Bt + H2O (4)
or a related continuous reaction. This reaction occurs at
590–600°C at 6–7 kbar (Fig. 9), which explains overlapping temperature estimates calculated for the staurolite
and kyanite zones. The lack of evidence for breakdown
of muscovite + staurolite within the kyanite stability
field suggests that temperatures were less than >650°C
(Fig. 9). The paragenesis Ky + St + Grt + Ms +
Bt + Qtz + Ilm suggests a typical medium-pressure
Barrovian-type metamorphism, with the pressure never
exceeding 10 kbar (Spear, 1993). The zoning pattern
of garnet compositions suggests that prograde M1 in
individual zones involved heating and burial (Tracy,
1982), which is consistent with a clockwise P–T–t path.
In the southern part of the sillimanite zone, staurolite is
locally pseudomorphed by sillimanite, garnet and biotite,
which suggests that the discontinuous KFMASH reaction
St + Ms + Qtz = Sil + Grt + Bt + H2O (5)
has been crossed. According to the grid in Fig. 9, the
reaction predicted by THERMOCALC is
St + Bt + Qtz = Sil + Grt + Ms + H2O. (6)
However, based on the decrease in muscovite abundance
the reaction is likely to really be (5). The topology of
reaction (6) is different from that of reaction (5) because
of problems with thermodynamic data used for staurolite
by THERMOCALC (Powell & Holland, 1990). Crossing
of reaction (5) suggests a minimum temperature of
>650°C within the sillimanite zone at 6 kbar (also see
Spear & Cheney, 1989), consistent with the P–T estimates
(Fig. 9). In some metapelites (e.g. D972334), polymorphic
replacement of kyanite by sillimanite may also have taken
place:
Ky = Sil.
(7)
Reaction textures, which are generally lacking in the
sericite–chlorite to kyanite zones, are very common in
metapelites from the sillimanite and migmatite zones.
These reaction textures are inferred to reflect the overprint of M1 by M2, rather than a continuous metamorphic
evolution, because geochronological evidence suggests
that: (1) peak metamorphism in the M1 zones occurred
at >200–190 Ma, whereas it occurred at >168–158
Ma in the M2 zones; (2) locally >200 Ma M1 ages are
preserved in the M2 zones (Huang et al., 2002).
273
JOURNAL OF PETROLOGY
VOLUME 44
NUMBER 2
FEBRUARY 2003
Fig. 9. P–T diagram showing the P–T estimates and metamorphic reactions used to constrain the metamorphic evolution of the DDMT.
Clockwise P–T paths [bold arrows shown in inset (a)] are inferred for rocks of the staurolite, kyanite and sillimanite zones. Dotted fields represent
P–T results for different zones (see text). The small bold arrows inferred from textures refer to the apparent P–T–t caused by the M2 overprint
on M1 assemblages. Curves (see text): 1, 2, 4, 6, 7 and relevant grids from Powell et al. (1998); 9 and 10, muscovite melting reactions from White
et al. (2001); 11, water-saturated solidus for system Ab–An–Or–Qtz–H2O from Johannes (1985).
Within the M2 sillimanite and migmatite zones, M1
garnet is commonly replaced or corroded by sillimanite,
and separated from muscovite by fibrolite mats and
biotite (Fig. 7d). Garnet is locally absent in biotite and
sillimanite-rich rocks. These textures suggest that the
pressure-sensitive divariant KFMASH reaction
Grt + Ms = Sil + Bt + Qtz
(8)
has occurred. This reaction texture can be examined in
a KFMASH P–T pseudosection for a relevant metapelitic
bulk composition (Powell et al., 1998). As the Grt–
Ms–Sil–Bt field is sub-horizontal in P–T space the progress of this reaction typically indicates decompression
that for most pelitic bulk compositions occurs at pressures
of Ζ6 kbar. Pseudosection calculations using the bulk
compositions of representative metapelites from the
Danba area indicate that the divariant reaction (8) occurs
at between 4·2 and 5 kbar. Therefore, the occurrence of
garnet-free sillimanite–biotite-bearing metapelites in the
sillimanite and migmatite zones is consistent with peak
M2 pressures being <6 kbar. Anorthite-rich plagioclase
(bytownite) intergrown with sillimanite and biotite occurs
only in some garnet-free metapelites (see above), and
thus may have been produced by the CNKFMASH
(CaO–Na2O–K2O–FeO–MgO–Al2O3–SiO2–H2O) analogue of reaction (8).
The observed garnet breakdown textures reflect the
higher temperature and lower pressure for M2 with
respect to M1, even though M2 occurred at the structurally
lowest levels in the cover sequence. This feature is consistent with the observation that sillimanite-grade metamorphism was coeval with limited uplift of the DDMT.
Overall, relationships between mineral growth and deformation phases, mineral compositions, and P–T estimates for the DDMT define clockwise P–T–t paths for
both M1 and M2 (Fig. 9). Because M1 and M2 have
recently been dated at >210–190 and 168–158 Ma,
respectively (Huang et al., 2002), reaction textures in
the sillimanite and migmatite zones probably reflect a
discontinuous P–T–t evolution between M1 and M2, as
indicated by the crosscutting nature of M2 isograds (Fig.
2a). Calassou (1994) suggested that the metamorphic
zones in the DDMT were formed progressively during
crustal thickening. If this were the case, the sillimaniteand migmatite-zone rocks would give higher pressures
and temperatures than the kyanite-zone rocks and reaction (6) would be crossed from the kyanite to migmatite
zone. However, these were not observed. The higher
274
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
temperatures and lower pressures for the sillimanite and
migmatite zones compared with the kyanite zone also
suggest that M1 and M2 represent discrete events.
Partial melting
Migmatites, which are widespread in the northern
DDMT, show evidence of having originated by partial
melting. As can be seen from the P–T calculations
(Tables 4 and 5), the migmatite zone shows maximum
temperatures of >660–720°C. The peak temperatures
experienced by the basement rocks in the northern
DDMT therefore probably exceeded the wet solidus
of the granite system Ab–An–Or–Qtz–H2O ( Johannes,
1985; Fig. 9). It is possible that some pegmatites, particularly those observed in the basement, may be in part
derived from the crystallization of the segregated granitic
melts that formed during D2.
Immediately above the orthogneiss basement, leucosomes of Pl + Qtz ± Kfs in metapelitic stromatic
migmatites are rimmed by fibrolitic sillimanite, suggesting
that they formed via the melting reactions
Ms + Qtz + H2O Pl = Sil + melt
(9)
Ms + Qtz + Pl = Kfs + Sil + melt.
(10)
and/or
For normal plagioclase-bearing metapelites at 5–6 kbar,
reactions (9) and (10) occur at >650°C and
>670–680°C, respectively (White et al., 2001), in the
K2O–Na2O–Al2O3–SiO2–H2O end-member system.
However, they are shifted to higher temperatures by a
few tens of degrees in the plagioclase-free or Ca-bearing
systems. Taking into account the P–T conditions for the
migmatite zone for K-feldspar-free metapelitic migmatites (>690–700°C), the temperatures for the Sil–
Kfs–Pl assemblages could be in excess of >720°C, which
is consistent with these leucosomes having formed via
reaction (10).
The general spatial restriction of pegmatites to the
sillimanite and migmatite zones supports a genetic relationship between them and M2 metamorphism. The
great decrease in muscovite abundance in metapelites,
especially in the migmatite zone, is consistent with partial
melting by reactions (9) and (10) as the most likely
mechanisms for pegmatite production. Whereas reaction
(9) may have been responsible for volumetrically minor
plagioclase-rich, K-feldspar-free pegmatites, reaction (10)
appears to be the most plausible model for the more
volumetrically important plagioclase–K-feldspar pegmatites. A local origin within the metasedimentary
sequence for these pegmatites via reaction (10) is also
consistent with: (1) stable isotope evidence (M.-H. Huang
& I. S. Buick, unpublished data, 2000); (2) experimental
constraints on melting of similar real micaschists (Patiño
Douce & Harris, 1998), which produce identical assemblages via reaction (10) to those seen in the DDMT
at >720–750°C; (3) melts formed via water-present
reactions such as reaction (9) not being able to segregate
easily from their source rocks (McLellan, 1983).
Tectonic implications
Shortly after the deposition of the Triassic (Rhaetian–
Norian) sediments, the DDMT experienced M1 metamorphism, resulting in the development of the main
prograde metamorphic zones. Kyanite-zone metamorphism reached maximum P–T conditions of >6–8
kbar and >570–600°C (Tables 4 and 5; Fig. 9). These
pressures imply a burial depth of 22–29 km, suggesting
that the Sinian–Triassic sedimentary pile (originally 9–14
km thick; Hou et al., 1996) was doubly thickened during
M1. The normal growth zoning of syn-D1/M1 garnet from
the garnet to kyanite zones and trends in composition for
included and matrix plagioclase are consistent with both
an increase in pressure and a rise in temperature during
garnet growth (Spear, 1993). These observations suggest
that M1 followed a thickening–heating P–T–t path. Inasmuch as M1b occurred immediately after intrusion of
Indosinian S-type granitoids, advective heat transferred
by magma may have been an additional heat source for
M1b.
The M2 metamorphic event is characterized by sillimanite-bearing assemblages and anatexis, and is constrained to have occurred at somewhat lower pressures
(4·8–6·3 kbar) but higher temperatures ([620–725°C)
than the kyanite zone. High-grade, garnet-free M2 metapelitic rocks experienced pressures of Ζ5 kbar, based on
pseudosection constraints. P–T estimates suggest that the
lower crust in the northern part of the DDMT underwent
limited (1–2 kbar) syn-D2 decompression after D1 thickening. Decompression was the result of superposition
of D2 and D1 antiformal structures. The attainment of
temperatures as high as >720°C during decompression
suggests that syn-M2–D2 exhumation may have also required advection of heat (Sandiford et al., 1995). The
degree to which the lower-grade (sericite–chlorite to
kyanite zones) rocks at higher structural levels underwent
M2 metamorphism is uncertain. However, texturally late,
unzoned, fine-grained garnet that locally overprints S1b
in the staurolite and kyanite zones may represent a partial
M2 overprint, as indicated by garnet Sm–Nd ages as
young as >160 Ma (Huang et al., 2002). M2 was followed
by greenschist-facies retrogression ( M3). The low M3
temperatures suggest that M3 occurred after the region
had been exhumed to a shallow level.
Regionally, the metamorphic history reconstructed for
the DDMT can be correlated to some extent with that
275
JOURNAL OF PETROLOGY
VOLUME 44
of other higher-grade metamorphic domains within the
SGOB. In the Xuelongbao region (Fig. 1b), Barroviantype metamorphic (chlorite to kyanite) zones were developed within the Palaeozoic–Mesozoic metasediments
in response to SW-directed crustal thickening of Indosinian age (210–196 Ma, Dirks et al., 1994). Similar
peak P–T conditions were obtained for this region, except
for higher pressures (>10 kbar) for the kyanite zone
(Worley et al., 1997). Another Barrovian-type metamorphic terrane also occurs in Palaeozoic metasediments
in the Jianglang dome, where the metamorphic zones
were considered to form by heating associated with
crustal thickening and nearly horizontal layer-parallel
shearing (Xu et al., 1992).
Barrovian-type metamorphism in the DDMT occurred
at >210–190 Ma (Huang et al., 2002). These ages are
close to the ages of >220–206 Ma of collision-type
granites in the Qinling orogen on the northern margin
of the SGOB (see Meng & Zhang, 1999). Therefore, it
is likely that Barrovian-type metamorphism ( M1) took
place throughout the SGOB during a late stage of the
Indosinian Orogeny (210–190 Ma) in response to the
subduction of the South China Block under the North
China Block (Laurasia). The collision between the two
blocks caused crustal thickening by folding and top-tothe-south thrusting, and led to crustal thickening (England
& Thompson, 1984) within a large-scale intracontinental
high-strain zone.
The recognition of M2 as temporally distinct from
M1 on the basis of mapping of isograds and recent
geochronology (Huang et al., 2002) suggests that the
Yanshanian Orogeny at >165 Ma had a significant
effect on the tectonic evolution of the SGOB. The
structural observations indicate that the eastern Tibet
Plateau underwent east–west compression during the
early Yanshanian Orogeny. This is consistent with the
late Jurassic collision of South and North Tibet along
the Lancang River suture (Peng & Hu, 1993). In the
Danba area, east–west compression during the Yanshanian Orogeny caused north–south-oriented folding,
partial exhumation of the pre-Sinian basement (Fig.
10b), and high-temperature metamorphism and partial
melting. The orientation of granitoid intrusions and the
basement was strongly controlled by the east–west-directed shortening. Limited uplift of portions of the DDMT
took place at this stage as a result of F2–F1 fold interference
to produce numerous basement-cored domes. On a
regional scale, the arcuate trend of S1 throughout the
SGOB is probably related to Yanshanian east–westdirected shortening. The 1–2 kbar decompression from
M1 to M2 suggests that during Mesozoic time the eastern
Tibet Plateau experienced only limited uplift.
During the Himalayan Orogeny, convergence between
India and Asia, which started at >50 Ma (e.g. Dewey
et al., 1988), resulted in DDMT in the development of
NUMBER 2
FEBRUARY 2003
NW-striking thrusts and sinistral strike-slip fault zones
within an overall transpressional setting. Movement along
one such strike-slip zone, the Xianshuihe strike-slip fault
in the SW of the DDMT, occurred at >16–13 Ma
(Roger et al., 1995; Figs 1b and 3).
Some previous workers in the SGOB have interpreted
metamorphism and differential exhumation of amphibolite-grade rocks as occurring mainly in the Mesozoic
(e.g. Xu et al., 1992; Dirks et al., 1994). Available geochronology (Huang et al., 2002) suggests that the whole
of the DDMT cooled slowly throughout the Jurassic–
Cretaceous post-D2, suggesting that there was little differential uplift at this stage. D3 may have played a role in
exhuming the DDMT, particularly as M2 isograds in the
DDMT are themselves domed (Figs 2 and 5). However,
Rb–Sr biotite ages across the Danba area (Huang et al.,
2002) suggest that the whole DDMT cooled through
>350°C at >30–24 Ma. This suggests that little differential exhumation could have occurred during the Oligocene to early Miocene. As D3 appears to begin at >16
Ma we cannot rule out further differential exhumation
post-30–24 Ma associated with D3 transpression, but it
would have to have occurred at a lower temperature
than the closure temperature of biotite for the Rb–Sr
system. Based on lower-temperature chronometers (apatite fission track), Arne et al. (1997) showed that differential exhumation of Barrovian terranes in the
Longmenshan occurred as a result of transpression at
10–20 Ma. However, similar data are lacking from
DDMT. The available data from the Longmenshan
suggest, but do not prove, that differential exhumation
in the DDMT was also associated with mid- to lateMiocene transpression.
CONCLUSIONS
Field and petrological data indicate that the DDMT
experienced polyphase metamorphism and deformation
during the Indosinian, Yanshanian and Himalayan Orogenies, which shaped the regional tectonics of the SGOB.
Its evolution is divided into four distinct phases:
(1) low-grade metamorphism M1a developed during
the initial episode of crustal thickening (D1a) immediately
before the extensive emplacement of late-Indosinian
granitoids.
(2) Progressive M1b metamorphism developed during
a major crustal thickening episode (D1b) during a late
stage of the Indosinian Orogeny. M1 biotite- to kyanitezone assemblages formed in response to burial and heating concomitant with top-to-the-south thrusting as a result
of collision between the North China Block (Laurasia) and
South China Block.
(3) Sillimanite-bearing assemblages record a discrete
overprint (M2) related to migmatization and limited
276
HUANG et al.
EVOLUTION OF EASTERN TIBET PLATEAU
differential uplift owing to heating and F2 folding during
early Yanshanian compression. Migmatites and pegmatites within metapelitic rocks were formed predominantly through muscovite-dehydration melting. The
DDMT is a regional, composite structural dome, and its
doming and uplift is partly caused by interference between
D1 and D2 anticlinal structures.
(4) M3 greenschist-facies assemblages developed
mainly within NW-oriented thrusts and strike-slip shear
zones in a transpressional setting associated with the
uplift and large-scale lateral movement of the DDMT
during the Himalayan collision between India and Asia
(D3). Further differential exhumation of the DDMT probably occurred at this time, but low-temperature thermochronological data to constrain the timing and extent of
exhumation are currently lacking.
ACKNOWLEDGEMENTS
This project was funded by the Australian Research
Council. We wish to thank Rulong Yu and Xiaofang Fu
from the Sichuan Exploration Bureau of Geology and
Mineral Resources for their great help in the field in
China. Alexander Priymak, Jorg Metz and Roland Maas
are acknowledged for assistance with microprobe work
at the University of Melbourne. Breton Worley made
valuable comments on an earlier version of this paper.
Thorough and constructive reviews by Roger Gibson,
Moonsup Cho and Christopher McFarlane are also gratefully acknowledged. M.-H. also acknowledges the financial support of a La Trobe University Postgraduate
Scholarship.
SUPPLEMENTARY DATA
Supplementary data for this paper is available on Journal
of Petrology online.
REFERENCES
Anderson, J. L. & Smith, D. R. (1995). The effects of temperature and
fO2 on the Al-in-hornblende barometer. American Mineralogist 86,
549–559.
Arne, D., Worley, B., Wilson, C., Chen, S. F., Foster, D., Luo, Z. L.,
Liu, S. G. & Dirks, P. (1997). Differential exhumation in response
to episodic thrusting along the eastern margin of the Tibetan Block.
Tectonophysics 280, 239–256.
Bell, T. H. & Rubenach, S. E. (1983). Sequential porphyroblast
growth and crenulation cleavage development during progressive
deformation. Tectonophysics 92, 171–194.
Bevins, R. E., Robinson, D. & Rowbotham, G. (1991). Compositional
variations in mafic phyllosilicates from regional low-grade metabasites and application of the chlorite geothermometer. Journal of
Metamorphic Geology 9, 711–721.
277
Calassou, S. (1994). Étude tectonique d’une chaı̂ne de décollement: A)
tectonique et Tertiaire de la chaı̂ne de Songpan–Garzê; B) géométrie
et cinématique des déformations dans les prismes d’accretion sedimentaire: modélization analogique. PhDThèse, Université de Montpellier II.
Dewey, J. F., Shackleton, R. M., Chang, C. F. & Sun, Y. Y. (1988).
The tectonic evolution of the Tibet Plateau. Philosophical Transactions
of the Royal Society of London, Series A 327, 379–413.
Dirks, P. H., Wilson, C. J. L., Chen, S., Luo, Z. & Liu, S. (1994).
Tectonic evolution of the NE margin of the Tibetan Plateau: evidence
from the central Longmen Mountain, Sichuan Province, China.
Journal of Southeastern Asian Earth Sciences 9, 181–192.
Droop, G. T. R. (1987). A general equation for estimating Fe3+
concentrations in ferromagnesian silicates and oxides from microprobe analyses, using stoichiometric criteria. Geological Magazine 51,
431–435.
England, P. C. & Thompson, A. B. (1984). Pressure–temperature–time
paths of regional metamorphism. I. Heat transfer during the evolution
of regions of thickened continental crust. Journal of Petrology 25,
894–928.
Florence, F. P. & Spear, F. S. (1989). Volume diffusion in garnet:
effects on P–T path calculations. EOS Transactions, American Geophysical
Union 70, 492.
Graham, C. M. & Powell, R. (1984). A garnet–hornblende geothermometer: calibration, testing, and application to the Pelona
Schist, Southern California. Journal of Metamorphic Geology 2, 13–21.
Holland, T. J. B. & Blundy, J. (1994). Non-ideal interactions in calcic
amphiboles and their bearing on amphibole–plagioclase thermometry. Contributions to Mineralogy and Petrology 116, 433–447.
Holland, T. J. B. & Powell, R. (1992). Plagioclase feldspars: activity–
composition relations based on Darken’s Quadratic Formalism and
Landau theory. American Mineralogist 77, 53–61.
Holland, T. J. B. & Powell, R. (1998). An internally consistent thermodynamical data set for phases of petrological interest. Journal of
Metamorphic Geology 16, 309–343.
Hou, L. W. (1996). Type and origin of the core complexes and the
domal deformational–metamorphic bodies in the western margin of
Yangtze Craton. Acta Geologica Sichuan 16, 6–11 (in Chinese).
Hou, L. W., Den, Z. M., Fu, X. F., Li, R. Q . & Yang, X. J. (1996).
China Geological Map, 1:50,000 Danba sheet and explanatory note. Chengdu:
Geological Publishing House (in Chinese).
Huang, M.-H., Maas, R., Buick, I. S. & Williams, I. (2002). Crustal
response to collisions between the India, Tibet, South China and
North China Blocks: geochronological constraints from the Songpan–
Garzê Orogenic Belt, west China. Journal of Metamorphic Geology (in
press).
Johannes, W. (1985). The significance of experimental studies for the
formation of migmatites. In: Ashworth, J. R. (ed.) Migmatites. Glasgow:
Blackie, pp. 36–85.
Kohn, M. J. & Spear, F. S. (1989). Empirical calibration of geobarometers for the assemblage garnet–hornblende–plagioclase–quartz. American Mineralogist 74, 77–84.
Kretz, R. (1983). Symbols for rock-forming minerals. American Mineralogist 68, 277–279.
Leake, B. E. (1978). Nomenclature of amphiboles. American Mineralogist
63, 1023–1052.
Mattauer, M., Malavieile, J., Calassou, S., Lancelot, J., Roger, F., Hao,
Z. W., Xu, Z. Q . & Hou, L. W. (1992). La chaı̂ne triasique
de Songpan–Garzê (Ouest Sechuan et est Tibet): une chaı̂ne de
plissement–décollement sur marge passive. Comptes Rendus de l’Académie
des Sciences, Série II 314, 619–626.
McLellan, E. L. (1983). Contrasting textures in metamorphic and
anatectic migmatites: an example from the Scotland Caledonides.
Journal of Metamorphic Geology 1, 241–262.
JOURNAL OF PETROLOGY
VOLUME 44
Mehnert, K. R. (1968). Migmatites and the Origin of Granitic Rocks. Amsterdam: Elsevier.
Meng, Q .-R. & Zhang, G.-W. (1999). Timing of collision of the North
and South China Blocks: controversy and reconciliation. Geology 27,
123–126.
Ohmoto, H. & Kerrick, D. (1977). Devolatilization equilibria in graphitic systems. American Journal of Science 277, 1013–1044.
Patiño Douce, A. E. & Harris, N. (1998). Experimental constraints on
Himalayan anatexis. Journal of Petrology 39, 689–710.
Peng, X. & Hu, C. (1993). Tectonic evolution of the Sanjiang Belt in
eastern Xizang. Regional Geology of China 2, 140–148 (in Chinese).
Plyusnina, L. P. (1982). Geothermometry of plagioclase- and hornblende-bearing assemblages. Contributions to Mineralogy and Petrology
80, 140–146.
Powell, R. & Holland, T. J. B. (1988). An internally consistent thermodynamic dataset with uncertainties and correlations: 3. Application
methods, worked examples and a computer program. Journal of
Metamorphic Geology 6, 173–204.
Powell, R. & Holland, T. J. B. (1990). Calculated mineral equilibria
in the pelite system KFMASH (K2O–FeO–MgO–Al2O3–SiO2–H2O).
American Mineralogist 75, 367–380.
Powell, R., Holland, T. & Worley, B. (1998). Calculating phase diagram
involving solid solution via non-linear equations with example using
THERMOCALC. Journal of Metamorphic Geology 16, 577–588.
Ramsay, J. G. (1967). Folding and Fracturing of Rocks. New York: McGraw–
Hill.
Ren, J. H., Jiang, C. F., Zhang, Z. K. & Qin, D. Y. (1987). Geotectonic
Evolution of China. Beijing: Science Press.
Roger, F., Calassou, S., Lancelot, J., Malavieile, J., Mattauer, M., Xu,
Z. Q ., Hao, Z. W. & Hou, L. W. (1995). Miocene emplacement
and deformation of the Konga Shan (Xianshui He fault zone, West
Sichuan, China): geodynamical implications. Earth and Planetary Science
Letters 130, 201–216.
Sandiford, M., Fraser, G., Arnold, J., Foden, J. & Farrow, T. (1995).
Some causes and consequences of high-T, low-P metamorphism,
Mount Lofty Ranges. Australian Journal of Earth Sciences 42, 233–240.
Spear, F. S. (1993). Metamorphic phase equilibria and pressure–
temperature–time-paths. In: Monograph Series. Washington, DC: Mineralogical Society of America, 779.
278
NUMBER 2
FEBRUARY 2003
Spear, F. S. & Cheney, J. T. (1989). A petrogenetic grid for pelitic schists
in the system K2O–FeO–MgO–Al2O3–SiO2–H2O. Contributions to
Mineralogy and Petrology 101, 149–164.
Spear, F. S., Kohn, M. J., Flourence, F. P. & Menard, T. (1991). A
model for garnet and plagioclase growth in pelitic schists: implications
for the thermobarometry and P–T path determinations. Journal of
Metamorphic Geology 8, 683–696.
Tapponnier, P., Lacassin, R., Leloup, P. H., Schärer, U., Zhong, D.
L., Wu, H. W., Liu, X. H., Ji, S. C., Zhang, L. S. & Zhong, J. Y.
(1990). The Ailao Shan/Red River metamorphic belt: Tertiary shear
between Indochina and South China. Nature 343, 431–437.
Tracy, R. J. (1982). Composition zoning and inclusions in metamorphic
minerals. In: Ferry, J. M. (ed.) Characterization of Metamorphism through
Mineral Equilibria. Mineralogical Society of America, Reviews in Mineralogy
10, 355–397.
Turner, F. J. (1981). Metamorphic Petrology: Mineralogical, Field, and Tectonic
Aspects. New York: McGraw–Hill.
White, R. W., Powell, R. & Holland, T. J. B. (2001). Calculation of
partial
melting
in
the
system
Na2O–
CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O (NCKFMASH). Journal of
Metamorphic Geology 19, 139–154.
Worley, B. & Powell, R. (1998). Singularities in the system NCKFMASH
(Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O). Journal of Metamorphic Geology 12, 99–119.
Worley, B. A., Powell, R. & Wilson, C. J. L. (1997). Crenulation
cleavage formation: evolving diffusion, deformation and equilibration
mechanism with increasing metamorphic grade. Journal of Structural
Geology 19, 1121–1135.
Xu, S. J., Wang, R. C., Shen, H. Z., Zong, H., Lu, J. J., Hou, L. W.,
Fu, X. F., Huang, M. H., Yang, J. D., Wang, Y. X. & Tao, X. C.
(1996). U–Pb, Rb–Sr isotopic chronology of Jinning granites in the
Songpan–Garzê Orogenic Belt and its tectonic significances. Science
in China D39, 576–586.
Xu, Z. Q ., Hou, L. W., Wang, Z. X., Fu, X. F. & Huang, M. H.
(1992). Orogenic Processes of the Songpan–Garzê Orogenic Belt of China.
Beijing: Geological Publishing House (in Chinese).
Zhang, Y. X. & Luo, Y. N. (1988). Panxi Rift and its Geodynamics. Beijing:
Geological Publishing House (in Chinese).
Zhou, Z. X., Wang, Z. Y. & Huang, Y. C. (1981). China Geological
Map, 1:200,000 Danba sheet and explanatory note. Beijing: Geological
Publishing House (in Chinese).