High-resolution Velocity Model Research Articles

Smoothness: The key factor in well-log information-assisted PINNtomo

Seismic traveltime tomography is popularly used to provide a smooth background model for subsequent depth migration or full waveform inversion (FWI). The core of the traveltime tomography process requires an eikonal solver for seismic traveltime calculation. Most of the conventional eikonal solvers, suffer from first-order convergence errors and difficulties in dealing with irregular topography, which will lead to inaccurate seismic tomography results. To solve these issues, physics-informed neural networks (PINNs) have been developed and demonstrated remarkable success in simultaneous traveltime calculation and velocity construction, which is referred to as PINNtomo. It combines the physics constraint (eikonal equation) and data constraint (traveltime fitting) to form the loss function of the networks. Though PINNtomo can manage to recover a smooth background velocity, it will fail in the inversion for a complicated high-resolution velocity model. We take advantage of the well-log velocity as a prior information to provide an additional model constraint in the loss function to improve the performance of PINNtomo. We found that the well-log information will provide high-resolution components in the inverted velocity model. However, it will destroy the lateral structure of the true model. We propose to use a smoothed version of the well-log velocity as the prior model constraint in PINNtomo. As a result, the accuracy of the inverted background velocity will be improved, which will be beneficial to the subsequent FWI.

Waveform inversion with structural regularizing constraint based on gradient decomposition

Abstract Full waveform inversion (FWI) can simultaneously update low-to-medium wavenumber velocity components and high-wavenumber velocity components. However, if seismic data lack large-offset data and effective low-frequency components, FWI updates will be dominated by high-wavenumber velocity perturbation. Meanwhile, providing that the initial model is inaccurate, inversion will have the problem of local minima. In this study, FWI is developed with structural regularizing constraint based on gradient decomposition (RGDFWI). By correlating the separated forward wavefield and backward wavefield with specific propagating direction, FWI gradient is decomposed into tomography-mode gradient and migration-mode gradient. We propose an optimized strategy taking full advantage of the two modes of FWI gradient. On the one hand, we use tomography-mode gradient to enhance low-to-medium wavenumber updates. On the other hand, we use migration-mode gradient to apply structural regularizing constraint by estimating structure dip and adding sparsity constraint in Seislet domain. During the inversion process, high-wavenumber structural information constrains and guides low-wavenumber model updates. The results of two numerical tests, Marmousi model test and Overthrust model test, validate the optimized strategy, which can produce a better initial velocity model for FWI. The inversion finally generates a high-precision and high-resolution velocity model.

A high-resolution seismic velocity model for East Asia using full-waveform tomography: Constraints on India-Asia collisional tectonics

East Asia's complex tectonic history makes it a unique place for plate-tectonic studies. While the crust and mantle structures beneath this region have been extensively investigated by receiver function and many different kinds of tomography studies, discrepancies among seismic velocity models have complicated our understanding of East Asia's tectonic evolution. Here, we use full-waveform adjoint tomography to study the seismic P- and S-wave velocity structures of the crust and mantle beneath East Asia. We use a large waveform dataset from 141 earthquakes recorded by 4640 broadband seismic stations. Our inversion optimizes the normalized correlation coefficient between synthetic and observed seismograms within individual time windows of different seismic phases, from P arrivals to the slowest surface waves. This approach allows us to fit regional multipathed waveforms and provides very well-resolved seismic P- and S-wave velocity structures from the crust to depths of about 800 km. Furthermore, we are able to obtain this resolution over most of East Asia allowing us to compare the amplitude and shapes of anomalies across the entire region. The observed structures in our new full-waveform tomography model (FWEA23) correlate well with tectonic units, revealing sharp contrasts across tectonic boundaries. Our model reveals narrow and strong linear fast anomalies associated with subduction zones in the western Pacific and Southeast Asia. Compared to these oceanic slabs, we observe different and more complicated mantle structures beneath the India-Asia continental collision zone. While our model spans a wide region in East Asia, we focus our interpretations on the seismic velocity anomalies beneath Tibet and surrounding regions, as our model provides significant implications for India-Asia continental collision tectonics. Our analysis suggests that cratonic Indian lithosphere collided with Eurasian plate around 25–20 Ma, subsequently underthrusting beneath Tibet horizontally with little deformation. This resulted in thickening, destabilization, and delamination of the rheologically weaker and less buoyant mantle lithosphere beneath north-central Tibet. Moreover, our model reveals isolated lithospheric fragments within the mantle transition zone and lower mantle beneath southern Tibet and northern India, suggesting that non-cratonic Greater India continental lithosphere has broken off and sunk into the deeper mantle. The lack of continuity among these lithospheric fragments suggests a convective destabilization process, distinct from typical oceanic plate subduction. Our results emphasize the important role of rheological properties of continental lithospheres in shaping the dynamics of continental collision.

Seismicity Migration and the Upper Crustal Structure in the Xinfengjiang Reservoir

Abstract After the impoundment of the Xinfengjiang Reservoir (XFJR) in Guangdong, China, numerous earthquakes occurred in the area, including a magnitude 6.1 event in 1962. Analysis of historical earthquakes indicates that M ≥ 4 earthquakes began occurring in the northwestern XFJR in 2012, and seismicity has gradually migrated from the southeastern to the northwestern reservoir (NWR). However, the mechanisms governing the migration of seismicity and the current upper-crustal structure beneath the reservoir area remain unclear. In our study, we conducted tomographic imaging by combining waveform data from short-period and permanent stations to construct a 3D velocity model. Our high-resolution velocity models revealed a horizontal fractured zone at ∼5 km depth that extends from the southeastern to northwestern XFJR, and a steep fault that extends to about 9 km depth. These two fractured zones may interact with each other, allowing for fluid infiltration and contributing to earthquake triggering via pore pressure diffusion in the XFJR areas. Furthermore, the calculation of Coulomb stress changes indicated that microearthquakes in the southeastern XFJR may contribute to the seismicity in the NWR. However, the influences of M ≥ 4 earthquakes in the northwestern XFJR on subsequent M ≥ 4 earthquakes in the southeastern XFJR vary differently. Our results provide crucial insights for understanding the migration of microearthquakes in the XFJR area.

Resolution enhancement for a seismic velocity model using machine learning

Summary To address complex subsurface structures, a high-resolution velocity model must be constructed. Conventionally, algorithms such as full waveform inversion (FWI) have been used to derive accurate high-resolution velocity structures, but obstacles such as high computational costs remain. Therefore, we propose a high-resolution U-NET (HR U-NET) machine learning model to derive a high-resolution velocity model from a low-resolution velocity model. The low-resolution velocity model and migration data obtained through the corresponding velocity information were used as input data for training. In addition, we tried to improve the accuracy of the high-resolution velocity model by using prior information containing accurate velocity values. A prior model generated through geophysical logging data and a weight model including the reliability information of the prior model were also utilized. Therefore, the HR U-NET model was trained using the low-resolution velocity model, the migration data, the prior model, and the weight model. Numerical experiments conducted using synthetic and field data demonstrated that the proposed model could accurately construct a high-resolution velocity model and verified that the prior model and weight model play important roles in the training process. Additionally, we confirmed that the proposed method derived almost similar results using only 8.2 per cent of the computational cost of the conventional inversion method. In other words, there is an advantage that it is possible to predict high-resolution velocity information more efficiently in terms of computational cost.

Uppermost mantle structure and dynamics of the Tanlu fault zone: New insights from Pn anisotropic tomography

Whether the segmentation of the Tanlu fault zone and mechanism of strong crustal earthquakes are related to the uppermost-mantle structure is still unclear. Here a new high-resolution uppermost-mantle velocity and anisotropy model of the Tanlu fault zone and surrounding areas is determined by using 13,045 Pn arrival-time data. These data are handpicked from seismograms recorded at 120 newly deployed TanluArray portable seismic stations and 337 Chinese provincial seismic stations. The pattern of Pn velocity and anisotropy agrees well with the surface geological and tectonic features. Obvious low-velocity (low-V) anomalies are observed in tectonically active areas, such as the Taihang mountain orogenic belt, the eastern Dabie mountain orogenic belt, and the Sulu fold belt, whereas areas with stable tectonics, such as the North China Basin and the South Yellow Sea Basin, exhibit high-velocity (high-V) anomalies. The fast propagation directions (FPDs) of Pn waves in the study area exhibit a complex distribution, but the FPDs show a rotating feature around the margin of the South Yellow Sea Basin. The block feature of the high-V anomalies beneath the North China Basin and segmentation of the low-V anomalies along the Tanlu fault zone may be related to lithospheric delamination and thermal erosion caused by hot and wet upwelling flows in the big mantle wedge. The 1668 Tancheng earthquake (M8.5) occurred near a low-V anomaly, and recent moderate and strong earthquakes (M > 4.0) took place along the boundaries of the velocity anomalies. These results offer new perspectives on the mantle dynamics and mechanism of strong earthquakes in and around the Tanlu fault zone.

Machine learning prediction of the mass and the velocity of controlled single-block rockfalls from the seismic waves they generate

Abstract. Understanding the dynamics of slope instabilities is critical to mitigate the associated hazards, but their direct observation is often difficult due to their remote locations and their spontaneous nature. Seismology allows us to get unique information on these events, including on their dynamics. However, the link between the properties of these events (mass and kinematics) and the seismic signals generated is still poorly understood. We conducted a controlled rockfall experiment in the Riou Bourdoux torrent (southern French Alps) to try to better decipher those links. We deployed a dense seismic network and inferred the dynamics of the block from the reconstruction of the 3D trajectory from terrestrial and airborne high-resolution stereophotogrammetry. We propose a new approach based on machine learning to predict the mass and the velocity of each block. Our results show that we can predict those quantities with average errors of approximately 10 % for the velocity and 25 % for the mass. These accuracies are as good as or better than those obtained by other approaches, but our approach has the advantage in that it does not require the source to be localised, nor does it require a high-resolution velocity model or a strong assumption on the seismic wave attenuation model. Finally, the machine learning approach allows us to explore more widely the correlations between the features of the seismic signal generated by the rockfalls and their physical properties, and it might eventually lead to better constraints on the physical models in the future.

Scale-Aware Edge-Preserving Full Waveform Inversion with Diffusion Filter for Crosshole Sensor Arrays.

Full waveform inversion (FWI) is recognized as a leading data-fitting methodology, leveraging the detailed information contained in physical waveform data to construct accurate, high-resolution velocity models essential for crosshole surveys. Despite its effectiveness, FWI is often challenged by its sensitivity to data quality and inherent nonlinearity, which can lead to instability and the inadvertent incorporation of noise and extraneous data into inversion models. To address these challenges, we introduce the scale-aware edge-preserving FWI (SAEP-FWI) technique, which integrates a cutting-edge nonlinear anisotropic hybrid diffusion (NAHD) filter within the gradient computation process. This innovative filter effectively reduces noise while simultaneously enhancing critical small-scale structures and edges, significantly improving the fidelity and convergence of the FWI inversion results. The application of SAEP-FWI across a variety of experimental and authentic crosshole datasets clearly demonstrates its effectiveness in suppressing noise and preserving key scale-aware and edge-delineating features, ultimately leading to clear inversion outcomes. Comparative analyses with other FWI methods highlight the performance of our technique, showcasing its ability to produce images of notably higher quality. This improvement offers a robust solution that enhances the accuracy of subsurface imaging.

Crustal Deformation in Eastern Himalayan Syntaxis Constrained by Ambient Noise Tomography

ABSTRACT As the leading edge of the Indian–Asian collision, the eastern Himalayan syntaxis region has experienced extensive tectonic activities, resulting in complex crustal uplift and deformation in the corner area of the southeastern pathway for the extrusion of Tibetan plateau materials. Despite considerable efforts, the corresponding deformation mechanisms remain uncertain. This study presents a new 3D high-resolution azimuthal anisotropic shear-wave velocity model in the crust and uppermost mantle derived from ambient noise dispersion data. Results show that the upper crustal anisotropy aligns with the geological boundaries and major faults nearby, suggesting shape-preferred orientations. The upper crustal low velocity and weak anisotropy beneath the core of the eastern Himalayan syntaxis (EHS) are closely associated with the high fragmentation of shallow rocks and the upwelling of hot materials during the ongoing subduction of the Indian plate. Our model also reveals relatively complex anisotropic patterns in the midlower crust. The eastern Lhasa terrane, in particular, exhibits low velocity and strong anisotropy with a northwest–southeast-oriented fast axis, supporting the local scale midlower crustal “channel flow” model. In addition, a conspicuous, elongated low-velocity zone along the northwest–southeast direction is observed in the midlower crust and uppermost mantle beneath the Bangong–Nujiang suture. The anisotropy in this region increases with depth, and the fast directions are consistently parallel to the northeast subduction of the Indian plate. We infer that this low-velocity zone may result from partial melting under local compression driven by the Indian–Asian collision. On the basis of newly revealed anisotropic model and previous studies, we construct a new dynamic model, which reveals that the migration of mechanically weak material in the midlower crust and the significant contribution of the northeast subduction of the Indian plate jointly control the crustal deformation of the EHS region.

The structure of the mantle transition zone and its implications for mantle upwelling beneath the central Tibetan plateau from triplicated P-waveforms modeling

The Tibetan plateau is considered as an ideal place for studying the continental collisions, tectonic movements and regional geodynamics. Although multiple episodes of igneous activities in Tibetan plateau have been suggested derived from mantle upwelling, their origins and corresponding geodynamic models remain vigorously debate. This study presents a high-resolution upper mantle velocity model of P-wave and its lateral variation beneath the central Tibetan plateau based on triplicated waveform modeling. The results show that, compared with IASP91 model, the 660-km discontinuity is deepened approximately 17–28 km from the Qiangtang terrane to the Beishan block. The lower mantle transition zone (MTZ) is characterized by high P-wave velocity anomalies (HVA) in range of 0.3%–0.9%. We infer that they are mainly related to the hydrous materials with low temperature in the lower MTZ. Besides, the depression of the 410-km discontinuity is about 6–15 km. A remarkable low velocity layer (LVL) is detected atop the 410-km discontinuity, exhibiting the thickness of 31–40 km and low P-wave velocity anomaly of −2.5% to −1.6%. This is against completely temperature-dominated origin of the LVL, and suggests that the material composition has also contributes. Additionally, we propose a new geodynamic model to demonstrate the asthenospheric upwelling beneath the Qiangtang and Songpan-Ganzi terranes, which suggests that the passively ascending water-enriched materials from the lower MTZ and their dehydration are responsible for the observed LVL atop the 410-km discontinuity, as well as the origin of the Cenozoic magmatic intrusion and volcanic eruption of the northern Tibet volcano in Pleistocene.

High-Resolution 3D QP and QS Models of the Middle Eastern Boundary of the Sichuan–Yunnan Rhombic Block: New Insight into Implication for Seismogenesis

Abstract The eastern boundary of the Sichuan–Yunnan rhombic block (EB-SYRB) has complex structures and strong seismicity. Although multiple 3D high-resolution velocity models have been constructed for this region, its seismogenic environment has been controversial. Seismic wave attenuation (inversely proportional to Q) describes the anelastic properties of the Earth’s medium, and is more sensitive to changes in subsurface fluid and temperature than seismic wave velocity. Based on the data of a long-term dense array in downstream of the Jinsha River, this article uses local earthquake tomography to obtain 3D QP and QS models of the middle EB-SYRB with the highest resolution to date, improving the lateral resolution of the Q model from 100 km to 5–10 km and the depth resolution from 10 to 2 km. Combined with the existing high-resolution velocity and resistivity models and geochemical observation results, we can comprehensively understand the medium structure and the seismogenesis in the study area. The results show that the high-attenuation characteristics in the shallow layer of the Xiaojiang fault zone and the Zemuhe fault zone (within a depth of ∼5 km) are consistent with the topographic relief and the distribution of hot springs, which reveals the Quaternary sedimentary characteristics of the basins and the presence of shallow fluids in the fault zone. The columnar high-attenuation anomaly beneath Huize reveals the fluid channel created by deep melting. The Ludian earthquake sequence occurred in a prominent low-attenuation area, which is favorable for stress accumulation and has a seismogenic environment for strong earthquakes. The high attenuation near the southwest end of the Huize fault is closely related to the Huize earthquake cluster, which is driven by fluids in the upper crust.

Role of mantle indentation in collisional deformation evidenced by deep geophysical imaging of Western Alps

In collision belts, the first-order role of the mantle in localizing deformation has remained elusive, as the resolution of geophysical imaging remains too low to constrain crustal geometry. To address this issue, we geologically interpret a recent high-resolution shear-wave velocity model from ambient-noise tomography of Western Alps. We show that the lower crustal Alpine geometry is highly variable at depth, evolving from a preserved European crustal slab in the South to a smooth crustal root in the North. Moho morphology is controlled by numerous pre-existing major faults reactivated during the Alpine orogeny. Two mantle indenters located above the subducted European plate at different depths appear to control the locus of active deformation. The rigid nature of Adria mantle explains the localization of brittle deformation that is transferred towards the upper crust. The strain-field partitioning results in a combination of strike-slip with either shortening or extension controlled by the anticlockwise rotation of Adria.

Imaging high-resolution 3D shallow crustal structure by ambient noise method in the Shiziyang Waterway area, Guangzhou, South China

The Shiziyang Waterway, located in the northern continental margin of the South China Sea, has developed a complex fault system owing to the coupling action of the oceanic crust and continental crust. However, the geological structure in the shallow crust beneath the Shiziyang Waterway and the relationship between faults and waterway development are still unclear, resulting in difficulty in assessing the earthquake risk to the surrounding cities. In this study, based on a high-density short-period seismic array that was recently deployed around the Shiziyang Waterway, a high-resolution shear wave velocity model is inverted to as shallow as 4 km by the ambient noise tomography method. The results showed that at depths ranging from 0.2 to 4 km, three sets of low-velocity zones (LVZs), trending east–west, northeast, and northwest, are revealed, which are speculated to be the Shougouling fault (SGLF), Xinhui-Shiqiao fault (XSF), and Shiziyang fault bundle (SZYFb), respectively. The shear-wave velocity anomalies illustrate that the SGLF stopped extending east at the north endpoint of the Wenchong-Zhujiangkou fault (WZF). The two-stage differential vertical movements of the WZF, Nangang-Humen fault (NHF), and Hualong-Haioudao fault (HHF) dominated the opening of the Shiziyang Waterway. This study provides new seismological evidence of the opening process for the Shiziyang Waterway and reveals that internal fault activity dominated the development and evolution of the Pearl River Delta.

3-D crustal shear wave velocity and azimuthally anisotropic structure of the Pearl River Delta onshore-offshore area and its tectonic implications

The Pearl River Delta (PRD) onshore-offshore area lies in the transition zone between South China and northern continental margin of the South China Sea. To better understand the tectonic and deformation patterns at depth, we developed a 3-D high-resolution crustal shear wave velocity (Vs) and azimuthal anisotropy model by ambient noise tomography in this region. The data sets comprise ∼30 days of seismic ambient noise recorded by 124 land seismic stations and 28 Ocean Bottom Seismometers (OBS). Our results reveal a significant lateral contrast in azimuthal anisotropy in the crust. The fast axes of the azimuthal anisotropy in land show an approximately NW-SE direction, which may be mainly controlled by the regional stress field. In the offshore PRD, however, the fast axes of anisotropy strike nearly in the NE-SW direction, likely dominated by the shearing of the littoral fault zone. Combined with other previous results, the anisotropic model reveals three major crustal deformation processes in this region since the Yanshanian. Furthermore, an upper-middle crustal frozen mafic magma chamber with high Vs (∼3.85–4.0 km/s) is imaged beneath Shenzhen, Zhuhai, and Hong Kong. It is interpreted as a source of intense magmatic activities in and around Hong Kong during the late Mesozoic. Our new model provides important constraints for understanding the crustal architecture and the multistage deformation processes in and around the PRD onshore-offshore area.

High-resolution seismic velocity structure using induced inter-event interferometry at the geothermal sites, Geysers, Eastern California

The possibility of retrieving Rayleigh wave empirical Green's functions (EGFs) using virtual seismometers provides an opportunity to calculate high-resolution velocity models without a set of dense stations and/or expensive seismic imaging/monitoring. The high-resolution tomographic map can give an overview of the fracturing within the saturated rock. Theoretically, the cross-correlations of earthquakes contain significant information about the EGFs between event-pairs. In this work, we developed an application of virtual seismometers on a very small scale in superficial layers where anthropogenic seismicity occurred. To test this method's capabilities, we applied it to investigate the NW-part of The Geysers (TG) geothermal field. We used 1276 micro-earthquakes, occurred in a cuboid with the horizontal side of ∼1 × ∼2 km and the vertical edge at a depth from 0.9 to 2.2 km, to obtain the high-resolution (100 × 100 m) velocity structure. A group of strict criteria was imposed separately for single event waveform and event pairs to stabilize the tomographic results. After retrieving the inter-event EGF signal and calculating the Rayleigh wave group velocity dispersion curves, we obtained the group velocity maps for each horizontal layer. The tomographic maps as a function of depth indicated a low-velocity anomaly related to existing inactive faults, topography variation of the top of the steam zone, and low-velocity anomalies possibly stimulated by fluid injection. The topography of the top of the steam varies from ∼1.0 km to 1.3 km of maps, and also anomalies connected with fluid migration appear at the depth of injection wells (1.5–1.9 km).

Second-order optimization for multiparameter reflection waveform inversion in acoustic VTI media

SUMMARY The seismic wave velocity in subsurface formations typically varies both spatially and directionally, resulting in a heterogeneous and anisotropic medium. To successfully image complex structures and potential reservoirs using multichannel seismic data, high-resolution interval velocity models are required. However, most controlled-source seismic recordings lack wide-angle signals, such as diving and refracted waves that propagate through deep formations. Consequently, geophysicists rely on the traveltimes and waveforms of reflections or back-scatterings in finite-frequency and finite-offset seismograms to estimate velocity and account for possible anisotropy. Recently, wave-equation based reflection waveform inversion (RWI) has emerged as an active research topic due to its ability to recover intermediate-wavelength model components that cannot be retrieved in reflection traveltime tomography. Nevertheless, the widely used gradient-type local optimization suffers from multiparameter trade-off and slow convergence of RWI in anisotropic media. Analysis of the Fréchet derivative reveals that the sensitivity of reflections to changes in long-to-intermediate wavelengths of the anisotropic velocities is jointly controlled by the specular reflection on the interface and the radiation pattern of scattered fields concerning the model parameters along the wave paths. Anatomy of the approximate Hessian reveals characteristics of parameter coupling and spatial resolution in the context of multiparameter RWI, inspiring the development of a matrix-free Gauss–Newton algorithm based on the second-order adjoint-state method for large-scale applications. Synthetic and real data examples demonstrate that the proposed approach can appropriately handle overburden heterogeneities and anisotropies, and thus improve imaging of underlying complex structures such as dipping and small-scale faults in the deep parts.

Rayleigh wave dispersion-based estimation of high-resolution 2D-subsurface shear wave velocity models: application of updated approaches

Rayleigh wave dispersion-based estimation of high-resolution 2D-subsurface shear wave velocity models: application of updated approaches

Binary structure constrained gravity inversion based on seismic first arrival travel time data

Gravity exploration method is one of the important methods for deep mineral resource exploration, but gravity data inversion has limited resolution ability in the vertical direction. In order to improve the vertical resolution of gravity data inversion, we propose a binary structure constrained gravity inversion method based on seismic first arrival travel time data. This method effectively reconstructs a density model with high vertical resolution by transferring the structural information of a high-resolution velocity model reconstructed by seismic data inversion to gravity data inversion through the binary structure constrained technique. This strategy eliminates the need to integrate both gravity and seismic methods into a single inversion framework, avoiding both the difference in convergence speeds between the two methods, as well as getting rid of the complexity associated with calculating structural coupling terms. Theoretical simulations show that the fuzzy c-means cluster analysis technique can accurately extract the target region of the velocity model reconstructed by seismic data inversion. Under the constraint of seismic structural information, the resolution of reconstructed density model is much higher than that of separate gravity data inversion, which proves that high resolution seismic information can improve the vertical resolution of gravity data inversion. Compared with the traditional cross-gradient joint inversion, the binary structure constrained gravity inversion method can further improve the resolution of the density model, especially in the reconstruction of the anomaly interface, which verifies that the method has certain effectiveness.

Upper-plate conduits linked to plate boundary that hosts slow earthquakes

In shallow subduction zones, fluid behavior impacts various geodynamic processes capable of regulating slip behaviors and forming mud volcanoes. However, evidence of structures that control the fluid transfer within an overriding plate is limited and the physical properties at the source faults of slow earthquakes are poorly understood. Here we present high-resolution seismic velocity models and reflection images of the Hyuga-nada area, Japan, where the Kyushu-Palau ridge subducts. We image distinct kilometer-wide columns in the upper plate with reduced velocities that extend vertically from the seafloor down to 10–13 km depth. We interpret the low-velocity columns as damaged zones caused by seamount subduction and suggest that they serve as conduits, facilitating vertical fluid migration from the plate boundary. The lateral variation in upper-plate velocity and seismic reflectivity along the plate boundary correlates with the distribution of slow earthquakes, indicating that the upper-plate drainage system controls the complex pattern of seismic slip at subduction faults.

The high-resolution community velocity model V2.0 of southwest China, constructed by joint body and surface wave tomography of data recorded at temporary dense arrays

The high-resolution community velocity model V2.0 of southwest China, constructed by joint body and surface wave tomography of data recorded at temporary dense arrays