Thorsten Wolfgang Becker

Despite being one of the most recent major orogenic events, the timing of the Arabia‐Eurasia collision is still debated. Much of the discrepancy arises due to uncertainties over the amount of continental subduction of the Arabian passive margin and the timing of a hard collision of cratonic Arabian lithosphere with Eurasia. Subducted continental lithosphere and underthrusted cratons can be imaged using high resolution seismic tomography, which can be used to improve understanding of the tectonic history of the region. In this study, we perform a full waveform inversion of seismic wave velocity beneath the Turkish‐Iranian plateau, using regional earthquakes, which provide good horizontal and vertical resolution in the upper 500 km of the mantle. We use a normalized cross‐correlation misfit function, which fits both the phase and relative amplitude of multi‐path arrivals, allowing us to extract additional information from complex triplicated waves that provide relatively good coverage of the upper mantle. Our results reveal that 250 km thick Arabian lithosphere extends horizontally beneath Eurasia. When reconstructed back in time, these results indicate that continental underthrusting and collision initiated at ∼22 Ma, which is consistent with accelerated deformation observed at this time in the geological record.

The horizontal surface motions of the Earth's crust induced by glacial isostatic adjustment (GIA) are highly sensitive to lateral variations in Earth properties. We construct 3D GIA models by incorporating both varying elastic lithospheric thickness and lateral viscosity variations in the asthenosphere in a global, compressible, self‐gravitating, viscoelastic Earth model. This allows us to investigate how geological structures such as the North American continental craton with thick lithosphere and stiff upper mantle and oceanic plate regions with thinner lithosphere and weak asthenosphere influence mantle flow during and after the last deglaciation and how this relaxation is expressed in surface motions. While vertical crustal motions are always in phase with the ice loading change, horizontal motions can be out of phase with the glaciation history and vertical motions. The delay in response to deglaciation and the timing of the direction reversal after deglaciation is complete are controlled by the interplay between elastic lithospheric flexure and the mantle viscous drag on the base of the lithosphere. The weak asthenosphere channels the mantle return flow more horizontally and the stiff cratonic root directs the flow more vertically, forming a “box‐like” circulation beneath the former ice cover. Such complex interactions may be better constrained with joint geodetic and seismic anisotropy analysis and provide new perspectives for GIA‐ice sheet interactions as well as the interpretation of GIA associated seismotectonics.

A widely accepted global model of glacial isostatic adjustment (GIA) is based on ICE‐6G ice loading and 1D‐viscosity model VM5a, a viscosity profile which remains at odds with mantle convection constraints such as from geoid modeling. We explore 3D‐viscosity variations in models of GIA and show that a model with lateral viscosity variations (LVVs) scaled from seismic tomography and lower mantle viscosity compatible with mantle convection constraints (∼30 times upper mantle viscosity) significantly reduces the misfit to relative sea level (∼40%), (60%), and polar motion (75%) while incurring only a small misfit penalty (∼5%) for GNSS‐uplift compared to VM5a, which has lower mantle viscosity ∼6 times upper mantle viscosity. Regional results exhibit additional tradeoffs but indicate that comparable solutions might be found with additional model refinement. The origin of a longstanding disagreement on lower mantle viscosity may thus be resolved by accounting for LVVs in models of GIA.

Anatolian tectonics are associated with slab retreat in the west and gravitational potential energy and continental collision in the east, leading to westward motion of Anatolia relative to Eurasia, partially accommodated on the North and East Anatolian transform faults. We construct a three‐layer, azimuthal seismic anisotropy model using new Rayleigh wave measurements and compare results against geodetic strain‐rates and new receiver function analysis for the crust, as well as new and existing shear wave splitting for the mantle to link surface and deep dynamics. In the crust, anisotropy is primarily controlled by recent tectonics, especially along major fault zones. In the lithospheric mantle, anisotropy reflects ongoing tectonic deformation in western Anatolia, whereas eastern Anatolia mainly preserves prior deformation signatures. In the asthenosphere, anisotropy reflects mantle convection, including the effects of Hellenic slab retreat. Our model provides new constraints on lithospheric deformation and geodynamics in the Eastern Mediterranean.

Key Points: The editors thank the 2025 peer reviewers!

The shallowing of subducting slabs within the upper ~200 km of the mantle, commonly referred to as flat slab subduction is associated with extensive petrological and structural modification of the continental lithosphere. Anomalously buoyant oceanic lithosphere, upper-plate overthrusting, and interactions with cratonic keels have all been proposed as mechanisms promoting shallow slab geometries, yet the dynamics governing the initiation and duration of flat slab subduction remain to be fully understood. Here, we investigate self-consistent flat-slab subduction dynamics using the finite element code ASPECT with adaptive mesh refinement and a free surface boundary condition. We explore the influence of the overriding plate structure, including the presence of continental keels, as well as the role of heterogenous subduction interface strength on shallow slab dynamics. Our results show that flat slab geometries develop when a weak, sediment-rich subduction interface is combined with a positively buoyant overriding continental lithosphere. Substantiating previous studies, we further find that the presence of a strong cratonic keel near the continental plate margin enhances shallow slab underthrusting and encourages flat slab configurations. Importantly, we show that the timing of interface weakening, such as due to influx of sediments, exerts a first-order control on the onset and the longevity of slab flattening. As the slab flattens, pronounced subsidence, extension and transient marine inundation develop within the foreland region of the upper plate, superimposed on broader, large-scale subsidence of the continental interior. Regional uplift and subsidence are thus not solely linked to flat slab emplacement and removal, but also reflect evolving slab dynamics within the shallow upper mantle. Our results provide new constraints on the geodynamic controls of flat slab evolution and their role in driving continent-scale deformation and sediment redistribution.

Earth science data are typically highly heterogeneous which leads to mixed determined inverse problems and poses challenges to extract process-level information. For example, ocean sediment cores from the International Ocean Discovery Program (IODP) contain hundreds of millions of measurements across multiple geophysical properties, but usable datasets are only 5-10% complete due to missing data. We present a semi-supervised variational autoencoder with masked encoding that simultaneously imputes missing measurements and predicts lithology, enabling more complete utilization of legacy IODP archives. We train a masked variational autoencoder on the LILY database (89 km of core, 34 million observations, 42 IODP missions) to learn joint distributions across bulk density, magnetic susceptibility, RGB reflectance, and natural gamma ray attenuation. The model uses selective masking during training to learn imputation strategies for missing modalities. Crucially, the learned latent representations are constrained to recover lithological labels from unseen cores without retraining. We demonstrate that the model both captures the nonlinearities contained in the training data and is able to reconstruct the test data (R2_avg=0.86) and that data lithology (AUC_avg=0.9), while also providing descriptive embedding vectors (ARI=0.2). Additionally, the underlying data contains strong non-linear relationships that are not captured by simpler models on reconstruction (e.g., a typical LASSO-based regression (R2=0.24)). Our work represents a step towards scalable cross-modal assimilation and representation of existing earth datasets.

The Alpine‐Himalayan belt is one of Earth's most dynamic and complex regions, characterized by intense tectonic deformation and seismicity. Comprehensive analyses of continental‐scale crustal deformation and seismic hazards along this extensive orogenic belt require the compilation of large geodetic data sets. In this study, we integrate 42 published Global Navigation Satellite System (GNSS) velocity fields, building an internally consistent data set for the entire belt, spanning from Iberia to Southeast Asia and comprising 11,177 horizontal and 3,940 vertical velocities. We use this unified GNSS velocity field to estimate surface strain rates and their posterior uncertainties in the eastern Mediterranean region and the India‐Asia collision zone. Our results show large‐scale agreement between the orientation and style of geodetic and seismic strain rate tensors across the belt. Additionally, our analyses substantiate previously documented azimuthal alignments between principal strain rate directions and seismic anisotropy orientations, often used as a proxy for finite strain in the convecting mantle. These correlations are particularly apparent in the Aegean, North Anatolia, Tibet, Tian Shan, Altai, Sayan, and Baikal regions, underscoring the need for future research on the relationship between mantle flow and lithospheric deformation.