Abstract Accurate retrieval of coseismic deformation fields from Interferometric Synthetic Aperture Radar is critical for constraining fault slip distributions and assessing seismic hazards. However, extracting pure deformation signals from single‐frame interferograms remains challenging due to the complex coupling of atmospheric turbulence, stratified delays, Digital Elevation Model residuals, orbital errors, and decorrelation noise, particularly in regions lacking external auxiliary data or extensive temporal archives. To address this, we propose Coseis‐Diff, an error‐correction model that requires no external auxiliary data at inference time, built upon a residual denoising diffusion framework. We employ a diffusion transformer backbone, which leverages global self‐attention mechanisms to overcome the limited receptive field of traditional Convolutional Neural Networks (CNNs). This architecture enables the unified correction of multi‐scale errors while preserving the topological continuity of deformation fringes. Evaluation on physics‐based synthetic data sets demonstrates that Coseis‐Diff achieves higher fidelity than CNN baselines. Furthermore, validation on real‐world earthquakes ( M w 5.7–7.05) confirms that the model effectively separates spatially correlated artifacts and reliably recovers deformation signals even in severe low‐coherence zones. Comparisons against an independent Okada dislocation model and GNSS coseismic displacements further show that the recovered phase unwraps more reliably and agrees more closely with these independent observations than the raw interferogram. By replacing the reliance on external corrections or time‐series stacking with physics‐based learned priors, Coseis‐Diff provides a rapid, high‐precision solution for emergency response and source mechanism inversion in data‐scarce environments.
Abstract Uplift and widespread Late Miocene‐present magmatism in Anatolia are commonly attributed to lithospheric thinning, possibly facilitated by the arrival of hot mantle plume material. There are various lines of evidence for thin present‐day lithosphere, but the spatial extent and degree of lithospheric thinning and the underlying mantle temperatures remain debated. Here we apply a Bayesian thermo‐compositional approach that jointly inverts surface‐wave dispersion, elevation and surface heat flow for temperatures of the lithosphere and underlying mantle to 360 km depth. Our results reveal that thermal lithospheric thickness across the study region is generally thin, 40–85 km, with thicker lithosphere (>70 km) beneath the Central Pontides and northernmost Arabia and exceptionally thin (40–55 km) lithosphere beneath the Eastern Pontides and the region between the Central and Eastern Anatolian Plateaus. Furthermore, we find a pronounced west‐to‐east increase in mantle potential temperature of ∼140C across 35E. The sharp mantle temperature gradient implies that any inflow of hot plume material is largely confined to the east and therefore is unlikely to be a primary driver of Anatolia's westward motion. The pervasively thin but variable lithospheric thickness indicates widespread, yet heterogeneous lithospheric removal, with the influx of hot mantle material potentially enhancing thinning beneath eastern Anatolia.
Abstract The multi‐segment ruptures of the 2023 7.8 and 7.6 Kahramanmaraş earthquakes caused extensive deformation across the entire width of the East Anatolian Fault Zone (Türkiye). To understand how this earthquake doublet could affect such a large region, we model coseismic slip for both earthquakes using surface deformation measurements. We use a Bayesian framework to obtain uncertainties and constrain the significance of our findings. Our model suggests that most of the coseismic slip distributes above 20 km depth with a shallow slip deficit ratio varying along strike from 10% to 80%. Slip is segmented along strike, with higher values on relatively straight segments and decreases at geometric complexities. We also construct postseismic InSAR displacement time series and find that segments with the most afterslip are located at the northeastern end of both ruptures. Aftershocks are usually located along the fault planes, except at the southwestern tip of both ruptures where they spread in wide fan‐shaped zones. These observations can be related to the relative geometric complexity of the southwestern termination of both earthquakes, where extensive crustal damage could have stopped the rupture. Finally, by visualizing the localization of interseismic deformation, we depict the role of strain partitioning between subparallel strike‐slip faults in the unusually large spatial extent of the Kahramanmaraş doublet. These findings highlight the seismic hazard posed by similar tectonic structures in other broad plate boundaries.
Abstract On 14 August 2021, the southern peninsula of Haiti was struck by an M w 7.2 earthquake that killed more than 2,200 people. The 2021 epicenter is located 95 km west of the M w 7.0 2010 Haiti earthquake and 7 km north of the main, geomorphic trace of the EW‐striking, left‐lateral strike‐slip Enriquillo‐Plantain Garden fault zone (EPGFZ) that forms a transpressional plate boundary between the Caribbean plate and the Gonâve microplate. Using a compilation of earthquake focal mechanism solutions, aftershocks, interferograms and coseismic displacements from ALOS‐2 and Sentinel‐1 Interferometric Synthetic Aperture Radar (InSAR), we propose that the 2021 earthquake ruptured a 2.5‐km wide band of steeply north‐dipping, secondary faults that bypassed the geomorphic prominent and curved 60‐km long restraining‐releasing paired bend on the EPGFZ. The initiating sub‐event of the 2021 mainshock is a left‐lateral and oblique thrusting event; that was followed by a left‐lateral westward‐propagating strike‐slip sub‐event along the proposed 65°N–77° N‐dipping Ravine du Sud bypass fault zone. We test our hypothesis by comparing two finite fault models that are constrained by the InSAR and global navigation satellite system displacements both on and off the bypassed restraining‐releasing paired bend on EPGFZ. We then compare our proposed model for fault evolution of the EPGFZ with similar examples of strike‐slip fault systems at a variety of map scales that are characterized by the same type of fault‐straightening process as we propose here for the 2021 Haiti earthquake.
Abstract The emplacement of igneous sill intrusions induces rapid heating of host rocks, generating fluids that migrate through structures known as hydrothermal vent complexes (HTVCs). While sill geometry and permeability are considered important factors controlling fluid migration around igneous sill intrusions, the precise mechanisms controlling vent distribution remain unclear. To improve the understanding of the primary controls on the final distribution of HTVCs, we are using a 3D seismic data set from the Mid‐Norwegian margin to analyze HTVCs associated with the ∼56 Ma Infinity Sill. The sill covers 237 km 2 and is associated with 79 HTVCs, 25 of which are classified as mini‐vents. The HTVCs have outer diameters ranging between ∼0.1–2.7 km, are separated by spacings between ∼1.2–4.8 km, have conduit heights ranging between ∼0.6–1.1 km, and variable upper terminations. Our analysis reveals a large‐scale fluid migration pattern dominated by venting along the sill margins and above the sill interior, controlled by factors such as the fluid generation potential around the sill, the density of local highs in the sill interior as well as the mode and heights of sill transgression. Sill geometry is thus found to affect the formation, size and spacing of HTVCs, and is interpreted to constitute a primary control on the final distribution and morphological characteristics of HTVCs. Our findings contribute to a better understanding of subsurface fluid dynamics associated with sill intrusions emplaced in organic‐rich strata and have broader implications for understanding the climatic effects of hydrothermal venting at the Paleocene‐Eocene transition.
