Structure of orogenic belts controlled by lithosphere age - Evgueni Burov

Aug 18, 2013 - CAN. 35±7. 52.17±10.43. 1. 65. 14. 2,000. 58. 1,935±387. CANT. 11±2 ..... 160. 200. 600 1,000. Thick-skinned belt. Thin-skinned belt. Foreland.
589KB taille 7 téléchargements 290 vues
LETTERS PUBLISHED ONLINE: 18 AUGUST 2013 | DOI: 10.1038/NGEO1902

Structure of orogenic belts controlled by lithosphere age Frédéric Mouthereau1,2 *, Anthony B. Watts3 and Evgueni Burov1,2 The structure of a mountain belt reflects the manner in which plate convergence is accommodated in Earth’s lithosphere. However, the extent to which orogenic structure is preconditioned by the thermo-mechanical conditions of the converging plates is debated1–9 . Here we re-process and analyse existing data on the amount and style of contractional deformation in 30 orogens worldwide and compare this with the lithospheric strength and age of the colliding plates. We find a correlation between orogenic deformation, and specifically the depth at which the crust decouples from the underlying plate, and the age of the lithospheric plate at the time of collision. Orogens formed from Phanerozoic lithosphere, which has high geothermal gradients and weak mantle, are characterized by several under-thrust faults that form in the mid-to-lower crust and moderate amounts of deformation, at less than 35% crustal strain. In contrast, orogens formed on older lithospheric plates, which have greater strength and higher-viscosity mantle, are characterized by a large detachment fault and large amounts of deformation, at about 70% crustal strain. We conclude that inherited lithospheric strength influences the style and amount of plate-tectonic contraction during mountain building, and thus the stability of continental subduction. Our results emphasize the influence of the deep Earth on the structural style of collisional orogens. The processes by which plate convergence in collisional orogenic belts is variably partitioned into underthrusting (simple-shear) and vertically distributed (pure-shear) crustal strain are still strongly debated4–9 . Possible models include changes in plate interactions caused by contrasts in lower plate buoyancy6,7 or episodic growth of eclogitic roots in arc regions and associated changes in slab geometry8 . Other types of models favour variable crustal and mantle mechanical properties related to the abundance of weak mineral phases or fluids in the lower crust9 and inherited continental lithospheric strength4,5 . Identification of the first-order controls on deformation styles in convergent plate boundaries, however, requires a global approach that considers the distribution of shortening in orogenic belts and the configuration and age of the basement that underlies their associated foreland basins. It has been shown that the effective elastic thickness, Te , a proxy for the long-term strength of the lithosphere, correlates with thermo-tectonic age, which we define as age since the last tectonic event. It also correlates with thermal thickness, and seismic S-wave speeds10–12 . Archean provinces (>1.5 Gyr) are generally thicker, colder and stronger (Te > 60 km) than younger Late Proterozoic and Phanerozoic provinces (80

Figure 1 | Shortening (R) in collisional orogens and global Te structure of the lithosphere (2◦ × 2◦ ). Shortening values (white vertical bars) for external domains of a collision segment (black thick solid lines) were calculated from balanced and restored cross-sections specified in Table 1. Current mean thermo-tectonic ages (Myr) of the colliding foreland continental lithosphere are shown for each thrust belt segment. Because the Te data are based on spectral and forward models, we limited the upper Te bound to 80 km because of the uncertainty of spectral results in resolving how high Te is. a

b

80

80

NEP

70

EH CANT

60

CPT

PAK

NEP

70

NWI

SEV

60

SEV

CPT RHE

RHE CAN SP

40

NP TA

30

DIN HAT SA

20 10 0

T

ZA

APN

ALB

TS

SUL

BR POT

APP ALAI/TAD

50

R (%)

R (%)

50

BOL

TA

60 Te (km)

80

100

0

SUL

SPA COL PER

Depths to decoupling levels H = 20 + 7 km

10

40

BOL

CAN BR

APP

NA

20

NA

20

ZAAPN ALAI TAD DIN ALB TS

HAT SA

Thin-skinned Thick-skinned 0

SP NP

30

PER

CANT POT

40

NWI PAK EH

Thin-skinned Thick-skinned 0

H ≥ 7 km

1,500 2,000 500 1,000 tT-S, age of the lithosphere at the time of shortening (Myr)

2,500

Figure 2 | Shortening in thrust belts plotted against proxies of the long-term lithosphere strength. a, Shortening versus Te in foreland basins. b, Shortening against age of the foreland lithosphere at the time of shortening tT−S . Grey quadrangles in b denote the range of shortening values predicted by modelling (Supplementary Information S2). Errors on shortening data account for the number of individual estimates and scattering of measurements with a confidence interval of 95%. It is 20% on average. Where the shortening value is constrained by a single restored cross-section, a consistent average error of 20% was assumed.

lithosphere and the structural styles in fold and thrust belts. Pertinent to this link is the current debate on whether continental lithospheric strength resides mainly in the crust or the mantle19 . Because continental Te exhibits both low and high values (larger than crustal thickness), then the mantle must be involved in the support of long-term geological loads, as it is in the oceans19 . The demonstrated R−tT−S relations are consistent with this proposition. Otherwise, tectonic styles as expressed by shortening estimates would be more similar between the different orogens and show little correlation with tT−S . Interestingly, continental Te also shows a bi-modal distribution19 , which may result from a jump-like increase, for a given age, in flexural strength if the crust and mantle are mechanically coupled together. Our results also confirm the

inferences from numerical modelling, which show that the stability of underthrusting during collision is favoured in the case of a stronger lithosphere20 (Supplementary Fig. S4). The dependence of the amount of shortening on the age of the lithosphere at the time of collision suggests control by subcrustal mantle properties, which are also age-dependent21 . Yet, thermal models predict saturation of the geotherm after 0.6–1 Gyr in continents12 . Hence, shortening data can not only reflect a dependence on temperature. Pre-orogenic stretching and thinning induce rheological weakening by grain-size reduction and/or phase changes in mid-crustal shear zones, as seen for example in Alpine-type margins22 . Together with melt infiltration in the mantle peridotite23 the lithosphere becomes mechanically unstable

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience © 2013 Macmillan Publishers Limited. All rights reserved.

3

NATURE GEOSCIENCE DOI: 10.1038/NGEO1902

LETTERS Young Phanerozoic lithosphere age = 50 Myr Differential stress (MPa) 250 500 0

40

Moho

40

60

60

80

rm

17.5¬32% shortening

LC

Geothe

Lithosphere

120 140

LC

200

Moho Byerle

e's law

Te = 70 km

160

600 1,000 1,400 (T°C)

200

400 1,000 1,400

(T°C)

Thin-skinned belt

Thick-skinned belt Deeply-rooted thrust ramps

35¬70% shortening

w

140

160

1,000

100 120

Te = 25 km

750

80

rm Geothe

Asthenosphere

UC

20

Depth (km)

Depth (km)

UC 20

100

Differential stress (MPa) 250 500

0

l's la

0

Peie r

0

Cratonized lithosphere age = 1 Gyr

Foreland Moho

Foreland Shallow-dipping detachment

Weak lithospheric mantle

Eclogitization

Moho Strong lithospheric mantle

Figure 3 | Thermo-mechanical properties for lithospheres with thermal ages of 50 Myr and 1 Gyr. Strength profile for a 50 Myr Phanerozoic continental margin agrees with stress estimates in sandstones from Taiwan belt (black squares) based on a recrystallized grain size piezometer24 . The strength of natural ductile shear zones in the lower crust and mantle (dark and light grey shaded areas) is after a compilation of grain size palaeopiezometers27 . The range of predicted crustal shortening values is shown for lithospheres with weak zones in the middle-lower crust (orange) and in the sedimentary cover (blue). Sketches illustrate tectonic styles expected for both types of lithosphere. UC, upper crust; LC, lower crust.

and weak during extension. Initiation of deeply-rooted tectonic inversion in the crust is further facilitated where serpentinized, frictionally weak and hydrated mantle are present. Such mechanical weakening of the extended Phanerozoic mantle lithosphere promotes inversion of deeply-rooted normal faults as crustal thrust ramps, but this process is not applicable to cratons that are characterized by an anhydrous depleted sub-continental mantle18 . Figure 3 shows two endmember rheological yield-stress envelopes calculated for pre-1 Gyr, that is cratonized lithosphere and young Phanerozoic, 50 Myr, lithosphere. Average thermal thicknesses of 180 and 100 km are assumed for the old and young lithosphere, respectively, which reflect Te values of 70 km and 25 km respectively. Such Te values are representative of the Himalaya and Taiwan forelands (Supplementary Information S3). In the case of young lithosphere, the low ductile strength is consistent with the low bulk differential stress