2007 GRL Gorczyk GA
2007 GRL Gorczyk GA
2007 GRL Gorczyk GA
1029/2007GL031059, 2007
Click
Here
for
Full
Article
L21309 1 of 5
L21309 GORCZYK ET AL.: UHP MANTLE ROCKS L21309
3. Numerical Model
[7] To simulate oceanic crust formed at slow spreading
ridge, we assume an oceanic lithosphere containing blocks
of 1.0 1.5 km of basaltic (upper 2 rows) and gabbroic Figure 2. Proposed P-T paths for Cuaba garnet-bearing
(lower row) layers, within a serpentinite matrix at every ultramafic rocks and associated eclogites (modified from
0.5 km (Figure 3). Free slip conditions are placed at all Abbott et al. [2006]). Path 1: incorporation of the Cuaba
boundaries except for the lower boundary, which is open in peridotites into the oceanic eclogitic unit at UHP conditions.
both downward and upward directions. The plate velocity is Path 2: incorporation of the Cuaba peridotites into the
assumed to be constant at 2 cm/a (Figure 3). Thus, initial oceanic eclogitic unit at HP conditions. The P-T paths of the
bending of the slab is geometrically set, during the run the low pressure eclogites and associated blueschists (HP rocks)
slab bending is not prescribed what allows the evolution of observed in the Northern Serpentinite Mélange are also
unrestricted slab movement. The dip angle of the subduct- reported (after Krebs et al. [2007]).
2 of 5
L21309 GORCZYK ET AL.: UHP MANTLE ROCKS L21309
Figure 3. Two-dimensional numerical setting employed in this study showing initial distribution of different rock
lithologies. The low-spreading oceanic crust (7 km thick) is composed by blocks of basalt and gabbro (1 km 0.5 km)
embedded in a serpentinite matrix. The upper mantle is composed of dry peridotites. The upper plate is composed of a
classical oceanic crust (sediment, basalt, and gabbro) above an initial dry peridotitic mantle. Initial temperature field
(isotherms in °C) and boundary conditions are from Gorczyk et al. [2007]. Numerical grid steps: 2 km in the area of interest,
10 km outside area of interest. See text for detailed explanations.
al., 1995] at the ridge and beneath the abyssal sediments far Northern Serpentinite Mélange. However, the blueschists
from the ridge [Carlson, 2001]. established for the natural rocks were exhumed farther from
[9] The numerical model shows that subduction of oceanic the trench (in internal part of the serpentinite mélange) than
lithosphere formed at a slow-spreading ridge produces a wide in the model (Figure 4, red P-T path). Also, the exhumation
subduction channel composed of serpentinites, metabasalts, of blueschists in the model is later than the exhumation age
metagabbros and minor metasediments. The numerical mod- observed in the field. This late exhumation of blueschists
el predicts two kinds of serpentinites: (1) incoming hydrated seems to be controlled by the rapid uprising of the deep part
abyssal peridotites and (2) hydrated, fore-arc mantle perido- of the partially hydrated and molten mantle wedge that push
tites (Figure 4). The maximum depth of circulating material the serpentinites and associated blueschist blocks to the
in the serpentinite subduction channel reaches a depth of surface (Figure 4). The final exhumation of the blueschists
60 km (2 GPa) due to antigorite P-T stability field. As is accompanied by extension at shallow surface level similar
antigorite present a negative-sloped reaction boundary in the to that what has already been described in the eastern part of
P-T field, antigorite is no more stable at temperature superior the Dominican Republic [Goncalvez et al., 2000].
than 600-700°C [Bromiley and Pawley, 2003] or 720– 740°C [12] The end of the exhumation is not simulated, but the
[Ulmer and Trommsdorff, 1999] at 2 GPa and started to location of the Cuaba unit and its intense vertical foliation
dehydrate. formed by metamorphic minerals of amphibolite facies
[10] During the steady state subduction, no garnet-bearing clearly suggests that the surface exposure of the deepest
peridotites are exhumed. Progressive hydration of the man- rocks was controlled by local vertical motion along the
tle wedge weakens its rheology allowing the upwelling of Septentrional Fault.
the asthenospheric mantle wedge and subsequent retreat of [13] Abbott et al. [2006] estimated that the Cuaba garnet-
the subducting slab. Studies show that the slab retreat would bearing peridotites and clinopyroxenite equilibrated at the
occur if the dimensionless ratio (RH2O) between plate highest P-T conditions, about 1550°C and 4 GPa, followed
convergence rate and upward water propagation velocity by cooling and rapid decompression (Figure 3). The model
is higher than 4 [Gorczyk et al., 2007]. The slab retreat shows a similar P-T path.
would induce exhumation of a deep-seated melange from a [14] Our model is in contrast to the previously proposed
depth of 100 – 150 km that consists of UHP mafic rocks model involving the incorporation of garnet-bearing perido-
(subducted oceanic lithosphere), anhydrous peridotites, hy- tites in a slice of continental crust at great depth [Brueckner
drated and partially molten peridotites of the mantle wedge. and Medaris, 2000; Ernst, 1999] before their exhumation
This deep-seated mélange does not reach the surface but driven by buoyancy forces [Ernst, 1999, 2005]. The absence
stops at about 20 km depth beneath the already exhumed of UHP continental rocks in the area studied and the
serpentinite mélange. intimate association of the Cuaba garnet-bearing peridotites
with dense eclogitized oceanic crust require an alternative
4. Discussion explanation. The exhumed garnet peridotites are not incor-
porated within a UHP oceanic crustal slice, but were
[11] The lithology of the Northern Serpentinite Mélange incorporated at a shallower level when the deep-seated
is similar to that predicted in the numerical model. Further- melange ‘‘collided’’ with the Northern Serpentinite Mélange
more, the maximum pressure and temperature of the mod- as observed in the numerical model (Figure 4). This
elled subduction channel is similar to those recorded from interpretation is compatible with the fact that the Cuaba
the Northern Serpentinite Mélange. The oldest eclogites in eclogites associated with UHP rocks do not show any
the serpentinite mélange record anti-clockwise P-T path evidence of UHP metamorphism, the maximum P-T esti-
[Krebs et al., 2007], typical for the initial phase of subduc- mates being 1.8 GPa at 730°C [Abbott et al., 2006].
tion as the mantle wedge cools gradually with progressive [15] Although some difference exists with the exhuma-
descending slab [Gerya et al., 2002]. In the model the blue tion model proposed for UHP continental unit, the common
P-T (Figure 4) path is adequate for oldest eclogites form the
3 of 5
L21309 GORCZYK ET AL.: UHP MANTLE ROCKS L21309
Figure 4. Final evolution of the 2-D model of the subduction of a slow spreading ridge in five time-step (total duration:
22 Ma). The colors of the geometrical shapes refer to the diagram with P-T paths and do not correspond to rock type. In the
color code description ‘‘m.’’ stands for partially molten. The model reproduces the occurrence in the Rio San Juan complex
of HP rocks in the front of the Northern Serpentinite Mélange and garnet bearing peridotites at its rear. The model predicts
that the exhumation of HP rocks occurred within the serpentinite channel coming from the subducted oceanic lithosphere.
The garnet bearing peridotite derived from the mantle wedge and exhumed when the buoyant partially hydrated mantle
(composed of partially dehydrated oceanic serpentinite and partially hydrated mantle wedge) is pushed toward the surface
when the slab retreat and the asthenosphere upraise. Notice the occurrence of molten peridotite that could correspond to the
observed Rio Boba intrusive (Figure 1).
point is that the driving force for the exhumation remains garnet-bearing peridotites involving continental rocks
the buoyancy. The buoyancy forces developed by the worldwide.
contrast between the dry mantle peridotites and the partially
hydrated peridotites allow the deep seated mélange to rise 5. Conclusion
quickly in extension zones induced by the slab retreat. The
numerical model may be compared with exhumation of [16] The numerical model shows that garnet-bearing
peridotites may be exhumed within oceanic subduction
4 of 5
L21309 GORCZYK ET AL.: UHP MANTLE ROCKS L21309
without involvement of buoyant continental crust. Subduc- Serpentinite Melange, Cuba: A record of tectonic instability during sub-
duction?, J. Metamorph. Geol., 20, 581 – 598.
tion of oceanic lithosphere formed at a slow spreading ridge Gerya, T. V., and D. A. Yuen (2003), Characteristics-based marker-in-cell
produces a wide serpentinite subduction channel, in which method with conservative finite-differences schemes for modeling geo-
the transient exhumation of HP eclogites occurred at the logical flows with strongly variable transport properties, Phys. Earth
Planet. Inter., 140, 295 – 320.
onset of subduction while transient exhumation of blues- Gerya, T. V., B. Stöckhert, and A. L. Perchuk (2002), Exhumation of high-
chists started when the deep-seated subduction melange pressure metamorphic rocks in a subduction channel: A numerical simu-
reached shallow crustal levels in an extensional context. lation, Tectonics, 21(6), 1056, doi:10.1029/2002TC001406.
In the deeper part, of the subduction zone, (>60 km) the Gerya, T. V., J. A. D. Connolly, D. A. Yuen, W. Gorczyk, and A. M. Capel
(2006), Seismic implications of mantle wedge plumes, Phys. Earth
progressive hydration of mantle peridotites modifies the Planet. Inter., 156, 59 – 74.
rheology of the mantle wedge and triggers the upwelling of Goncalvez, P., S. Guillot, C. Nicollet, and J. M. Lardeaux (2000), Thrusting
the hydrated peridotites and partially molten peridotites and sinistral wrenching in a pre-Eocene carribean accretionary wedge
(Samana peninsula -Dominican Republic), Geodin. Acta, 13, 119 – 132.
followed by upwelling of hot asthenosphere. The buoyant Gorczyk, W., A. P. Willner, T. V. Gerya, J. A. D. Connolly, and J.-P. Burg
deep-seated melange rises quickly (5 cm/yr) in the exten- (2007), Physical controls of magmatic productivity at Pacific-type con-
sion zone induced by the slab retreat and exhumes along a vergent margins: Numerical modeling, Phys. Earth Planet. Int., 163(1-4),
209 – 232, doi:10.1016/j.pepi.2007.05.010.
cooling or near adiabatic decompression path up to inter- Hattori, K. H., and S. Guillot (2007), Geochemical character of serpenti-
mediate crustal level. In nature the final exhumation, is nites associated with high-to ultrahigh-pressure metamorphic rocks in the
controlled by local vertical motion along major strike-slip Alps, Cuba, and the Himalayas: Recycling of elements in subduction
fault what was not modelled. zones, Geochem. Geophys. Geosyst., 8, Q09010, doi:10.1029/
2007GC001594.
[17] The proposed model explains the observations of the Krebs, M., W. V. Maresh, H.-P. Schertl, A. Baumann, G. Draper, B. Ildeman,
Northern Serpentinite Mélange in Cuba and Hispaniola and C. Münker, and E. Trapp (2007), The dynamics of intra-oceanic subduction
it is an alternative to the classical model [e.g., Medaris, zones: A direct comparison between fossil petrological evidence (Rio San
Juan Complex, Dominican Republic) and numerical simulation, Lithos,
1999] of exhumation of UHP continental units. doi:10.1016/j.lithos.2007.09.003.
Lewis, J. F., G. Draper, J. A. Proenza, J. Espaillat, and J. Jiménez (2006),
[18] Acknowledgments. This work was financially supported by Ophiolite-related ultramafic rocks (serpentinites) in the Caribbean region:
ETH research grants TH-12/04-1, TH-12/05-3, SNF research grant A review of their occufsrrence, composition, origin, emplacement and
200021-113672/1 (WG, TVG) and the CNRS INSU program DYETI Ni-laterite soil formation, Geol. Acta, 4, 237 – 263.
(SG). We would like to thank Diane Seward for correcting English. Medaris, L. G. (1999), Garnet peridotites in Eurasian high-pressure and
Constructive review by L. Dobrzhinetskaya is greatly appreciated. ultrahigh-pressure terranes: A diversity of origins and thermal histories,
Int. Geol. Rev., 41, 799 – 815.
References Meschede, M., and W. Frisch (1998), A plate-tectonic model for the Me-
sozoic and Early Cenozoic history of the Caribbean plate, Tectonophy-
Abbott, R. N., G. Draper, and S. Keshav (2005), UHP magma paragenesis, sics, 296, 269 – 291.
garnet peridotite, and garnet clinopyroxenite: An example from the Do- Pindell, J., L. Kennan, W. V. Maresch, K. P. Stanek, G. Draper, and R. Higgs
minican Republic, Int. Geol. Rev., 47, 233 – 247. (2005), Plate-kinematics and crustal dynamics of circum-Caribbean
Abbott, R. N., G. Draper, and B. N. Broman (2006), P-T path for ultrahigh- arc-continent interactions: Tectonic controls on basin development in
pressure garnet ultramafic rocks of the Cuaba Gneiss, Rio San Juan Proto-Caribbean margins, Spec. Pap. Geol. Soc. Am., 394, 7 – 52.
complex, Dominican Republic, Int. Geol. Rev., 48, 778 – 790. Ranalli, G. (1995), Rheology of the Earth, 413 pp., Chapman and Hall,
Bromiley, G. D., and A. R. Pawley (2003), The stability of antigorite in the London.
systems MgO-SiO2-H2O (MSH) and MgO-Al2O3-SiO2-H2O (MASH): Saumur, B. M., K. H. Hattori, and S. Guillot (2007), Protrusion of fore-arc
The effects of Al3+ substitution on high-pressure stability, Am. Mineral., mantle serpentinites together with HP and UHP rocks along major strike-
88, 99 – 108. slip faults in the northern subduction complex of Dominican Republic,
Brueckner, H. K., and L. G. Medaris (2000), A general model for the paper presented at Subduction Zone Geodynamics Conference, Geosci.
intrusion and evolution of ’mantle’ garnet peridotites in high-pressure Montpellier Res. Group, Montpellier, France, 4 – 7 June.
and ultra-high-pressure metamorphic terranes, J. Metamorph. Geol., 18, Ulmer, P., and V. Trommsdorff (1999), Phase relations of hydrous mantle
123 – 133. subducting to 300 km, in Mantle Petrology: Field Observations and
Cannat, M. (1993), Emplacement of mantle rocks in the sea-floor at mid- High-Pressure Experimentation, edited by Y. Fei, C. M. Bertka, and
ocean ridges, J. Geophys. Res., 98, 4163 – 4172. B. Mysen, Spec. Publ. Geochem. Soc., 6, 259 – 281.
Cannat, M., et al. (1995), Thin crust, ultramafic exposures, and rugged Van den Beukel, J. (1992), Some thermomechanical aspects of the subduc-
faulting patterns at Mid-Atlantic Ridge (22° – 24°N), Geology, 23, 49 – tion of continental lithosphere, Tectonics, 11, 316 – 329.
52. Zhang, R. Y., J. G. Liou, J. S. Yang, and T. F. Yui (2000), Petrochemical
Carlson, R. L. (2001), Lower crustal water contents, P-wave velocities, and constraints for dual origin of garnet peridotites from the Dabie-Sulu UHP
modal mineralogy of oceanic diabase and gabbro, Eos Trans. AGU, 82, terrane, eastern-central China, J. Metamorph. Geol., 18, 149 – 166.
1154.
Draper, G., and F. Nagle (1991), Geology, structure and tectonic develop-
ment of the Rio San Juan Complex, northern Dominican Republic, Spec. T. V. Gerya, ETH Zürich, ETH Hönggerberg, HPP, CH-8093 Zürich,
Pap. Geol. Soc. Am., 262, 77 – 95. Switzerland.
Ernst, W. G. (1999), Metamorphism, partial preservation, and exhumation W. Gorczyk, ETH Zürich, LEB, Leonhädstrasse 19, CH-8092 Zürich,
of ultrahigh-pressure belts, Isl. Arc, 8, 125 – 153. Switzerland. (weronika.gorczyk@erdw.ethz.ch)
Ernst, W. G. (2005), Alpine and Pacific styles of Phanerozoic mountain S. Guillot, Laboratoire de Geodynamique des Chaı̂nes Alpines, CNRS,
building: Subduction-zone petrogenesis of continental crust, Terra Nova, University of Grenoble, 1381 rue de la Piscine, BP53, F-3841 Grenoble,
17, 165 – 188. France.
Garcia-Casco, A., R. L. Torres-Roldan, G. Millan, P. Monie, and J. Schneider K. Hattori, Department of Earth Sciences, University of Ottawa, 140
(2002), Oscillatory zoning in eclogitic garnet and amphibole, Northern Louis Pasteur, Ottawa, ON, Canada K1N 6N5.
5 of 5