Introduction
Garnet peridotites form a significant part of the Earth’s upper mantle. Their occurrence as tectonic slivers inside high (HP) and ultrahigh-pressure (UHP) continental terrains implies that past events of subduction and plate collision caused the mechanical coupling of crust and mantle. This makes orogenic garnet peridotites relevant markers of deep dynamic processes allied with mountain building and plate convergence. In the last decade, an increasing number of work has shown that HP to UHP garnet peridotites may not belong to the lithospheric mantle of the subducted plates. Rather, they have been attributed to domains of the mantle wedge overlying the downgoing slab (Brueckner, 1998; Nimis and Morten 2000; Brueckner and Medaris 2000; Zhang et al., 2000).
Figure 1. Tectonic sketch showing tectonic engagement of slices of mantle wedge inside the subducting continental crust.
Incorporation of large mantle "tectonic xenoliths" into the crust is also determined by density contrast between crust and mantle. Redrawn after Brueckner (1988).
Figure 1 shows that during subduction and buoyancy-driven exhumation of the continental crust, the top of the slab can incorporate slices of the overlying mantle either by tectonic erosion, or by density contrast (Brueckner, 1998). Once engaged in the crust, these large "tectonic mantle xenoliths" are carried to subcrustal levels and exhumed at the surface. Garnet peridotites from the Western Gneiss Region of Norway are examples of mantle wedge of such an origin, captured by the subducted Baltoscandian crust during the Caledonian orogeny (Beyer et al., 2006; Lapen, 2009; Spengler et al., 2009; Van Roermund 2009a). Moreover, several studies have documented that continental subduction can be traced to exceptional depths of 180-200 km, as suggested by the occurrence in UHP eclogites and gneiss of microdiamonds and of garnet with pyroxene exsolutions, derived from precursor majoritic garnet (Dobrzhinetskaya et al., 1995; Sobolev and Shatzky, 1990; Ye et al., 2000; Mposkos and Kostoupolos, 2001; Van Roermund 2009a). Slices of the overlying mantle wedge, as deep as 200 km can thus be trapped and carried to the surface to disclose information on major processes allied with mechanical coupling of crust and mantle during deep continental subduction (Scambelluri et al., 2008; Spengler et al., 2009). In addition, mantle wedge peridotites may preserve relicts of their history pre-dating the engagement in the crust, thus enabling a reconstruction of long term mantle dynamics (e.g. Nimis and Morten, 2000; Spengler et al., 2006; Ye et al., 2009, Van Roermund 2009b).
Petrologic and geochemical studies of orogenic garnet peridotites from HP and UHP settings have shown that these rocks can contain hydrous minerals (amphibole, phlogopite) associated with carbon-bearing phases (dolomite, magnesite, graphite/diamond) and can be enriched in incompatible elements and volatiles (Obata and Morten, 1987; Rampone and Morten, 2001; Van Roermund et al., 2002; Zhang et al., 2007; Sapienza et al., 2009; Malaspina et al., 2009a; Ye et al., 2009). This reflects modal metasomatism due to interaction with agents released by the associated (underlying) subducted crustal slabs. HP and UHP orogenic garnet peridotites are thus proxies of deep fluid-mediated chemical exchange between crust and mantle. At subduction zones this exchange occurs over large scales and represents a major driving force to the chemical differentiation of the Earth. The geologic relevance of orogenic garnet peridotites has recently caused an increase in the scientific interest of these rocks, especially if one considers that the chances to investigate deep Earth’s mantle materials is quite limited. So far, knowledge on mantle wedge environments is mainly based on studies of peridotite xenoliths. These are essentially fist-sized samples hosted by lavas erupted in supra subduction settings: although providing first order information about the mantle in these regions, they represent limited rock volumes which can be affected by late stage reaction and chemical exchange with the host lava. Such xenoliths frequently contain metasomatic amphibole, phlogopite, pyroxene, carbonate (Vidal et al. 1989; Maury et al. 1992; Szabo et al. 1996; Laurora et al. 2001) and, occasionally, aqueous fluid or silicate glass inclusions with amphibole, phlogopite and locally carbonate daughter crystals (Trial et al. 1984; Mc Inness and Cameron, 1994; Schiano et al. 1995; Andersen and Neumann 2001; Demeny et al., 2004; Ducea et al., 2005). Most of these xenoliths are spinel-facies peridotites from the shallow mantle, roughly corresponding to fore-arc depths: interactions between subduction fluids/melts and mantle rocks at sub-arc and deeper levels have been poorly constrained so far. Different from xenoliths, the orogenic garnet peridotites form up to kilometre-scale mappable bodies preserving long-lived structural and petrologic histories and are potential witnesses of the deep tectonic and geochemical interplay between subducting plates and overlying lithospheric and asthenospheric mantle.
We will illustrate this concept presenting the textures and composition of the orogenic peridotites from the Ulten Zone (Eastern Italian Alps). They were part of a Variscan mantle wedge tectonically coupled with eclogitized continental crust (Nimis and Morten, 2000; Tumiati et al. 2003). In the Ulten Zone peridotites, anhydrous spinel-facies assemblages are partly replaced by HP garnet + amphibole ± dolomite parageneses as the result of cooling at increasing pressure and infiltration of slab fluids at eclogite-facies conditions (Obata and Morten 1987; Rampone and Morten, 2001; Scambelluri et al., 2006).
The main focus of this paper is to provide an extensive review on the following major aspects of the Ulten Zone peridotites: (1) fluid infiltration at eclogite-facies conditions with formation of hydrous and carbonated parageneses in the garnet-facies stability; (2) the mineralogical and geochemical imprint produced by the crust-derived metasomatic agents in the mantle rocks; (3) the redox state of these peridotites during slab fluid infiltration and the possible speciation of the incoming fluid phase. These features are finally compared with those of UHP mantle rocks from Western Norway (Bardane locality) crystallized at subduction depth of about 200 km. We show that in both cases the mantle wedge peridotites were flushed and metasomatized by incompatible element-enriched C-O-H fluids released by the subducted crust. Fluid presence or absence, respectively, determined rock recrystallization and/or preservation of former structures and compositions. Also, the metasomatic imprint imparted to these mantle rocks requires an involvement of the continental crust in deep subduction. We finally discuss the possible processes acting along a 100-200 km depth window during subduction.