Pseudotachylytes (quenched frictional melts produced on a fault by seismic slip) in dry rocks exhumed from the mid-lower crust are potential indicators of earthquakes that either nucleated at, or propagated to, depths below the main shallow brittle-ductile transition zone. Establishing whether these pseudotachylytes effectively record deep-crustal earthquakes, or shallow-level earthquakes overprinting the mid-lower-crustal rocks during the exhumation path, may represent a major challenge. This challenge is mainly related to the fact that the mineral assemblage of a pseudotachylyte develops out of equilibrium during the coseismic thermal transient leading to melting and melt quenching. Here we investigate pseudotachylytes within peraluminous, sillimanite-garnet-rich, migmatitic paragneiss of the Serre Massif in Calabria (Southern Italy). These exhumed lower-crustal rocks experienced granulite-facies metamorphism (∼700–800 °C; ∼600–800 MPa), partial melting and dehydration during the late Variscan Orogeny (ca. 320–280 Ma). The crosscutting pseudotachylytes contain hercynite and sillimanite microlites, globular-shaped poikilitic cordierite and plagioclase, and rare cauliflower- to subhedral-shaped garnet. The pseudotachylytes are pristine, not affected by ductile deformation, recrystallisation or extensive alteration by fluid after their formation. A Rb-Sr isochron age of 51.4 ± 5.1 Ma is obtained for the pervasively kinked biotite in the host rock immediately adjacent to the pseudotachylyte and associated with earthquake damage, while an age of 105.3 ± 4.1 Ma is obtained for the undeformed host-rock biotite. This indicates that the granulites were cooler than the closing Rb-Sr temperature of biotite (ca. 300–400 °C) in the Cretaceous and that the studied pseudotachylytes formed by shallow seismic faulting. Therefore, sillimanite, hercynite, garnet, plagioclase, and cordierite all formed during quenching of the frictional melt well above the ambient temperature. Modelling of cordierite growth during melt quenching indicates that cordierite should have started to crystallise at T > 900 °C to achieve the grain size (up to 10 μm in diameter) observed in the pseudotachylyte. Modelling and microstructural observations allow the crystallisation sequence of microlites during melt cooling to be established. These microlites include cauliflower garnet which, in this case, did not develop in a deep-seated faulting context as commonly reported.
Read full abstract