Earthquake nucleation in the lower crust by local stress amplification

The study addresses how earthquakes nucleate within the dry, strong lower continental crust, a setting typically dominated by viscous deformation. Lower crustal seismicity is observed in intraplate regions and beneath cratons, but mechanisms for nucleation at depths ≥25–30 km remain debated. Geological evidence (pseudotachylytes) indicates that lower crust can be strong and seismic under granulite facies conditions, aligning with seismological observations of deep continental earthquakes. Proposed mechanisms include downward propagation of ruptures from the upper crust, thermal runaway plastic instabilities, dehydration or eclogitisation reactions, and local stress redistributions; however, these require specific conditions not universally present. The purpose of this study is to document natural examples demonstrating that earthquakes can nucleate in dry, strong lower crust without syn-deformational reactions or loading from shallower seismicity, instead arising from local stress amplification due to deformation along networks of intersecting viscous shear zones. Understanding this mechanism has important implications for seismic hazard in intracontinental regions lacking obvious upper crustal seismogenic faults or fluids.

Prior work shows: (1) Pseudotachylytes provide geological evidence for seismic slip at lower crustal conditions and imply high mechanical strength of dry lower crust. (2) Seismology documents lower crustal earthquakes in regions like East Africa and the India–Tibet collision zone, supporting a dry, metastable, strong lower crust. (3) Proposed nucleation mechanisms include downward propagation of ruptures from the upper crust, stress pulses, thermal runaway instabilities, fluid-related dehydration and reaction-induced embrittlement, and eclogitisation; these often require active reactions or external loading. (4) Shear zone mechanics and inclusion models indicate that strong inclusions within viscously deforming matrices can experience significant stress amplification, especially with high viscosity contrasts and interacting inclusions. (5) Previous studies in other granulite terranes (e.g., Musgrave Ranges) hypothesized cyclic pseudotachylyte generation associated with shear zone activity. The present work builds on these by providing field-based, in-situ evidence that stress amplification within strong blocks bounded by actively creeping lower crustal shear zones can nucleate earthquakes independent of fluids or shallow seismic triggering.

Field area and geological context: The Nusfjord anorthosite (Lofoten, Norway) is dissected by a network of narrow mylonitic shear zones active at 650–750 °C and 0.7–0.8 GPa (granulite–amphibolite facies conditions). Shear zones occur in three main orientation sets (Sets 1–3) within an approximately 1 km wide high strain corridor. The shear zones commonly exploit precursor dykes and older pseudotachylyte-bearing faults. Anorthosite blocks between shear zones remain largely undeformed internally except for microcracks, indicating high internal viscosity and lack of pervasive dislocation creep.

Identification of pseudotachylyte types: Type-1 pseudotachylytes are commonly mylonitized and incorporated within shear zones. Type-2 pseudotachylytes are pristine, undeformed veins interpreted as coseismic melts, occurring on small faults confined within anorthosite blocks bounded by shear zones (often near intersections of Set 1 and Set 2). Field features (en echelon arrays, pull-apart jogs, chilled margins, equant clasts) and microlitic/spherulitic vein-core textures confirm a melt origin and pristine character; local millimetre-scale viscous overprint at some vein margins is noted but accounts for negligible displacement.

Kinematic relationships and contemporaneity: Type-2 pseudotachylytes neither cut nor are cut by the bounding shear zones and are locally dragged into shear zone foliations near boundaries, indicating coseismic formation during ongoing viscous creep on the shear zones at lower crustal conditions. Mineral assemblages in host, vein, and sheared margins are consistent with granulite facies conditions contemporaneous with shear zone activity.

Measurement of displacement and fault geometry: Fault displacements were measured from dilational pull-apart apertures in pseudotachylyte veins to isolate single-event slip and avoid confounding viscous overprint. Where multiple offset dikelets exist, slip vectors were constrained from marker separations. Apparent fault lengths were mapped in outcrop; most ruptures are <~15 m long. Vertical extents are unobserved; circular and elliptical fault geometries were considered to bound rupture areas.

Seismic source parameter calculations: Inputs were fault length (L), area (A), and single-event displacement (S). For circular faults, diameter equals measured length (A = π(L/2)^2). For elliptical faults, the semi-major axis b (vertical) was assumed up to 10 times the measured semi-minor axis a (horizontal; a = L/2). Seismic moment M0 = μ A S with μ = 38 GPa for anorthosite. Moment magnitude Mw = (log10 M0 − 6.07) / 1.6. Static stress drop Δσ = μ S / (C r), where r is radius (circular) or semi-minor axis a (elliptical); C is a geometric coefficient for transverse faults computed from complete elliptic integrals given the ellipse aspect ratio. Both circular and elliptical cases were evaluated to provide bounds on Mw and Δσ.

Data synthesis: Field mapping constrained the spatial relation of type-2 pseudotachylytes to shear zones and intersection geometries. Stereonet analyses described fault and shear zone orientations. Results were compared with published earthquake scaling relationships and stress drop datasets from shallow and lower crustal seismicity, and with other pseudotachylyte-based estimates.

• Geometry and context: Pristine, coseismic type-2 pseudotachylyte veins occur on short faults confined within anorthosite blocks bounded by actively creeping lower crustal shear zones, especially near shear zone intersections. Veins show classic melt features and minimal viscous overprint.
• Coeval formation: Dragging of veins into shear zone foliations and consistent granulite facies assemblages demonstrate that seismic slip occurred contemporaneously with viscous shear at 650–750 °C and 0.7–0.8 GPa.
• Rupture dimensions and slip: Apparent rupture lengths are typically <~15 m. Single-event displacements from pull-aparts range from 1 to 26 cm. Slip/length ratios are 1e-3 to 1e-1, far exceeding typical km-scale earthquake ratios (1e-7 to 1e-4), suggesting premature rupture arrest by bounding shear zones and block geometry.
• Source parameters: Assuming circular ruptures, Mw ranges from 0.2 to 1.8 with static stress drops between 0.1 and 4.2 GPa. For elliptical ruptures with vertical extent up to 10× the measured length, Mw ranges 0.8–2.6 and stress drops 0.06–2.5 GPa. Even lower-bound values are high relative to seismological estimates for both shallow and lower crustal earthquakes and to most pseudotachylyte-derived values elsewhere.
• Mechanical implications: High stress drops imply failure shear strengths >1 GPa for intact, dry anorthosite, consistent with laboratory strengths and the inability of coarse-grained, dry plagioclase-rich rocks to flow viscously without grain-size reduction, mineralogical change, or fluid influx. The observations indicate local stress amplification within strong blocks due to strain incompatibility across a network of viscously creeping shear zones with very high viscosity contrasts.
• Process model: Earthquakes nucleated as transient brittle failures within strong blocks where stresses accumulated elastically as surrounding shear zones creeped, with geometry and interaction of multiple shear zones focusing stresses; repeated events promoted strain compatibility across the network.

The findings demonstrate that lower crustal earthquakes can nucleate in situ within strong, dry anorthosite blocks as a direct mechanical consequence of local stress amplification produced by concurrent viscous creep along intersecting shear zones. This mechanism addresses the longstanding question of how seismic failure occurs within a dominantly viscous regime at lower crustal depths without invoking fluids, thermal runaway, or downward rupture propagation from the upper crust. The high observed slip-to-length ratios and very large stress drops are consistent with ruptures constrained by block size and shear zone geometry, leading to premature arrest and concentration of stress.

The proposed cycle comprises: (1) progressive viscous slip on narrow, weak shear zones surrounding strong blocks, (2) buildup of elastic stress within blocks due to strain incompatibility and high viscosity contrasts, (3) localised seismic failure when block shear stress exceeds failure strength, producing pseudotachylyte, and (4) partial relaxation of stress and improved strain compatibility allowing continued viscous creep. Over time, cumulative seismic failure across multiple blocks facilitates sustained creep across the kilometre-wide high strain zone.

This model provides a simpler, widely applicable explanation for observed lower crustal seismicity in intraplate regions lacking major overlying active faults, fluid evidence, or reaction-induced weakening, such as the northern Central Alpine foreland. It remains compatible with scenarios where deep events are enhanced as aftershocks to upper crustal mainshocks, but does not require such triggering. The study underscores the importance of shear zone network geometry and differential creep rates in controlling deep earthquake nucleation and suggests that many granulite terranes with similar rheological and structural configurations may host low-magnitude lower crustal seismicity.

This work provides direct field evidence that earthquakes can nucleate within dry, strong lower crustal rocks due to local stress amplification caused by concurrent viscous creep on a network of intersecting shear zones. Short, block-confined ruptures produced pristine pseudotachylyte veins with single-event displacements of centimeters and exceptionally high static stress drops (up to several GPa), implying failure strengths of order 1 GPa for intact anorthosite. The mechanism does not require fluids, mineral reactions, or downward rupture propagation, offering a broadly applicable explanation for intraplate lower crustal seismicity.

Main contributions: (1) Documentation of coseismic pseudotachylytes formed at granulite facies conditions contemporaneous with viscous shear; (2) Quantification of source parameters for small lower crustal events showing unusually high stress drops; (3) A process model linking shear zone network geometry, viscosity contrasts, and strain incompatibility to seismic nucleation within strong blocks.

Future research: Constrain stress amplification magnitudes and spatial patterns via microstructural paleopiezometry and targeted numerical modeling tailored to the Nusfjord geometry and rheology; determine event recurrence and timescales; better resolve 3D rupture areas and aspect ratios; expand surveys to other granulite terranes to test the generality of the mechanism.

• Rupture geometry uncertainties: The vertical extent of faults is unobserved; circular and elliptical assumptions provide bounds but still introduce uncertainty in rupture area, Mw, and Δσ. Stress drops may be elevated because rupture areas are limited by block size.
• Stress amplification quantification: Absolute magnitudes are difficult to estimate due to differences between available models and natural complexity (geometry, rheology, deformation mechanisms). Direct constraints require further microstructural and numerical studies.
• Spatial and temporal scope: Evidence derives from a specific locality (Nusfjord anorthosite); recurrence intervals and temporal evolution of the seismic-aseismic cycle are not resolved.
• Minor viscous overprint: Localised millimetre-scale shearing at some pseudotachylyte margins indicates limited post-seismic viscous deformation, though assessed to be negligible for displacement estimates.