Subducting lithosphere is hidden from direct observation, but surface exposures of rocks that experienced burial to high-pressure (HP) to ultra-high pressures (UHP, 2500-3000 MPa) beneath the Alps in Early Cenozoic time (Chopin et al. 1991, Berger and Bousquet 2008) are flight recorders of the path the rocks took as they descended within the subduction channel and then returned to the surface via the same channel. So far, it is unclear what parts of the subduction channel are sampled by exhumation, how HP rocks are exhumed within the channel (Agard et al. 2009) and whether this channel can be detected in geophysical images of the tops of slabs (Kopp et al. 2009, Friederich et al. 2014). Petrological and thermo-chronological studies in the Alps reveal that subduction affected not only oceanic lithosphere, but also the distal continental margins of Europe and even of the Adriatic upper plate (Babist et al. 2006, Bousquet et al. 2012a, b). A striking and testable result of recent plate kinematic reconstructions is that most lithosphere subducted since Late Cretaceous time resides in the slabs or the mantle transition zone (Hafkenscheid et al. 2006, Lombardi et al. 2009) and that almost half of that material is continental (Handy et al. 2010).
Plate reconstructions indicate that the Alpine subduction rate exceeded the Adria-Europe convergence rate (Piromallo & Faccenna 2004, Handy et al. 2010, 2015), but these rates are poorly constrained. Retrodeforming the Alps along the EGT section (Fig. 4a) yields a convergence rate of 1-2 cm/yr (Schmid et al. 1996) which is greater than the exhumation rates of subducted rocks in the Alps (≤ 1 cm/yr, Rubatto & Hermann 2006). Estimating subduction and exhumation rates of HP and UHP rocks depends on obtaining accurate ages of their mineral assemblages; this is a vexing problem due to the sluggish kinetics of mineral reactions in the 500-600°C temperature range of subduction (Berger & Bousquet 2008). Controversy persists regarding the age of Alpine subduction, for example, in the Tauern Window (Fig. 1a) where thermo-chronology yields conflicting Eocene (Ratchbacher et al. 2004, Kurz et al. 2008) and Oligocene (Nagel et al. 2013) ages of HP metamorphism near the planned swath D and profiles (Fig. 3b). Despite more than 50 years of thermo-chronology in the Alps, the database is not uniform and is notably thin in some key areas (e.g., Tauern Window, star 2 in Fig. 4b). 4D-MB will implement in-situ thermo-chronology including the use of high-retentivity isotopic systems to obtain more robust assessments of the age, duration and rates of subduction and exhumation in the Alps.
The microfabric of exhumed HP rocks contains clues about the strain field and rheology of the subduction channel and, in the case of ultramafic rocks, of structural and seismic anisotropy of the mantle (Long & Silver 2008). Seismic anisotropy, especially the fast direction of polarized or split shear (SKS) waves, is diagnostic of flow in the asthenosphere. Recent studies of anisotropy of the Alpine-Mediterranean area (Jolivet et al. 2009, Barruol et al. 2011) reveal that while SKS-splitting directions emulate the arc of the Western Alps, they deviate by some 30° in a counterclockwise sense from the structural grain of the orogen. In the Eastern Alps, however, SKS-splitting directions are subparallel to post-20 Myr orogen-parallel crustal motion (TRANSALP section, Kummerow & Kind 2006) and oblique to the subduction direction of the Eastern Alpine slab. However, the depth interval sampled by the SKS waves is poorly defined and the physical causes of mantle anisotropy are still unclear (preferred orientation of crystallography and/or microcracks; Healy et al. 2009, Ullemayer et al. 2011). Fry et al. (2010) proposed a depth-dependent anisotropy model in which orogen-parallel fast S-wave directions are related to stacked crustal nappes, whereas fast directions in the crustal root and lithospheric mantle seem to be oriented perpendicular to the orogen, i.e., subparallel to the subduction direction. Linking SKS-splitting with asthenospheric flow is problematic given that these flow patterns change over time (Kaminski & Ribe 2002) and may leave multiple imprints (Endrun et al. 2011). This is likely in the Alps given its complex kinematic history. Interpreting seismic anisotropy reliably thus requires dense seismic networks (Rümpker et al. 2003, Ryberg et al. 2005) and numerical flow models to track variations in asthenospheric flow patterns through time. The 4D images acquired by AlpArray will go beyond existing models in which asthenospheric flow is assumed to be static over millions of years (Faccenna & Becker 2010).