Earthquakes in the Mantle? Insights from Ultramafic Pseudotachylytes

Abstract:

Deep earthquakes in subduction/collision zones may originate from mechanical failure of ultramafic rocks at mantle depths. Fault pseudotachylytes in peridotites have been attributed to seismic slip at depths >30 km. However, the possibility of frictional melting at shallower depths still exist.

While pristine mantle rocks typically lack magnetite, postseismic serpentinization would likely involve formation of abundant multi-domain (MD) magnetite. Single-domain (SD) to pseudo-single domain (PSD) magnetite may also form in pseudotachylytes through breakdown of mafic silicates. Magnetite has a large magnetic susceptibility (Km). MD magnetite shows low magnetic remanence / magnetic saturation ratios (Mr/Ms) compared to SD-PSD magnetite. The formation of coseismic magnetite however would depend on fO2. Hence, in unserpentinized ultramafic pseudotachylytes, magnetite would form preferentially under shallow, high fO2 conditions. Coseismically deformed magnetite would result in a high anisotropy of magnetic susceptibility (AMS).

Here, we present a predictive model of the magnetic properties and magnetic fabrics of ultramafic pseudotachylytes formed under four conditions:

i) deep seismic slip and no syn- or postseismic serpentinization: low Km (<600 . 10^-6 [SI]), low Mr/Ms (<0.1), and low AMS (<1.1)

ii) deep seismic slip followed by static serpentinization: high Km (>3,000 . 10^-6 [SI]), low Mr/Ms (<0.1), low AMS (<1.1)

iii) deep or shallow seismic slip in previously serpentinized peridotites: high Km (>3,000 . 10^-6 [SI]), moderate Mr/Ms (0.1-0.5), high AMS (>1.5)

iv) shallow seismic slip with no serpentinization: moderate Km (600-3,000 . 10^-6 [SI]), high Mr/Ms (>0.5), moderate AMS (1.1-1.5)

We test these models using samples from the Balmuccia Massif (Italy) and the Schistes Lustrés (Corsica). These models may provide new constrains for ultramafic pseudotachylytes regarding their depth of formation and the timing of serpentinization.