Search for: All records

Abstract Relative plate motion in subduction zones transitions from frictional slip to viscous flow with increasing depth and temperature. The frictional‐viscous transition can control the depth extent of megathrust earthquakes and episodic tremor and slip (ETS). Pore fluid pressure is a critical control on the transition, but models for its depth dependence are lacking. Here, we present a steady‐state modeling framework to calculate the fluid pressure and shear stress profile along the subduction interface. We consider fluid production from dehydration reactions in the subducting oceanic lithosphere, calculated from thermodynamic equilibrium models. These fluids are channeled updip, though in some models we allow for fluid loss into the overriding plate. The fluid pressure is calculated from Darcy's law, with permeability depending on effective stress, temperature, and slip rate. We allow for both rate‐state frictional sliding and thermally activated linear viscous flow, and solve for the partitioning of deformation between them. We apply the modeling framework to the Cascadia subduction zone. Our results show nearly uniform effective stress in the seismogenic zone, below which it decreases with depth and fluid pressure approaches lithostatic pressure. The frictional‐viscous transition spans a wide range of depths, and mixed frictional‐viscous deformation is predicted at the ETS source depth. A model with fluid leak‐off into the overlying plate produces a more heterogeneous effective stress, with a local minimum near the major dehydration depth. Our results provide important insight into earthquake hazards and the mechanism of ETS in Cascadia, and the modeling framework is applicable to global subduction zones. 

Abstract The subduction interface geometry is particularly important for estimating interplate coupling and surface geodetic motion, which has significant implications for seismic hazard mapping. Several published Cascadia subduction interface geometries derived from different seismic data sets vary significantly from one another. However, results from deformation models that use the different interface geometries are rarely systematically compared. Here, we assess the impact of subduction interface geometry on surface motion predictions, slip inversion results, and interface coupling estimates from four published Cascadia subduction interface geometries. We isolate the effect of the interface geometry on the predicted surface motion by applying uniform unit slip or Gaussian slip patterns to each interface geometry and calculate the predicted displacements at locations of GNSS stations. The forward model‐predicted horizontal displacements can differ by >20% and show azimuthal differences up to 10°; such differences correlate spatially to geometric differences amongst the interface realizations. Inversions of surface displacements estimated using a Gaussian distribution of slip, mimicking an earthquake, recover the applied slip distribution with differing spatial patterns and residuals of up to 38% of the maximum applied slip. Block models that use the four interface realizations produce coupling estimates on the interface with regions of significant coupling (>50%) that differ noticeably in down‐dip extent and lateral continuity. The results we present suggest that models utilizing interface geometry as an input, such as earthquake and tsunami models, should consider comparing models with differing interface geometries to critically evaluate model uncertainty stemming from this fundamental input. 

The CRESCENT Ground Failure Viewer and underlying database are currently prepopulated with curated, open-source landslide inventories from Oregon, Washington, California, and Canada. These foundational datasets provide a harmonized regional baseline and serve as the starting point for an expanding, community-driven collection of landslide and ground-failure inventories across Cascadia and surrounding regions. We have added a series of filters, including shaking intensity metrics, mean annual precipitation, and basic landslide types that allow users to export and download referenced data. The goal of this phase was to get the infrastructure in place for a community-based portal to visualize ground failure data in context of Cascadia, giving users a way to filter and explore the data across the region (From BC to California) in a unified way.\n\nAlongside the viewer, we provide a jupyter-book to detail the processing of the available datasets and demonstrate transparently how data has been unified. This Jupyter-Book also serves as community resource to prepare and submit new data to the database.\n\nThe next goal is to allow members of the Cascadia Ground Failure community to share peer-reviewed datasets through this system, expanding open sharing and analysis of landslides and liquefaction hazards in the region. To facilitate the submission of new data, we also developed a jupyter-book going over the processing of the currently available datasets and guiding users through the steps. \n\nWe acknowledge the contributions of Kate Mickelson (DNR), Bill Burns (DOGAMI), Margaret Parks (CGS), Marc-André Brideau (BCGS), Alex Grant (USGS), Lydia Staisch (USGS), and Will T. Struble (U Houston) for providing data and comments during the development phase. 

The Earthquake Catalog Repository and Viewer is a community-oriented, curated collection of earthquake catalogs developed within the CRESCENT cyberinfrastructure framework to support research, visualization, and analysis of seismicity. The repository standardizes diverse earthquake catalogs into a consistent, machine-readable format, enabling interoperability, reproducibility, and transparent comparison across datasets. The accompanying interactive web-based Earthquake Catalog Viewer (https://eqcat.cascadiaquakes.org) allows users to explore, visualize, and compare catalog data in real time. Comprehensive documentation, metadata standards, and submission guidelines are provided in the project’s Jupyter Book (https://cascadiaquakes.github.io/earthquake_catalog_repository/), which also describes the structure and provenance of included catalogs. By integrating standardized data curation with interactive visualization tools, this resource facilitates community contribution and broadens access to high-quality earthquake catalog data for scientific research, hazard analysis, and education. 

Bugfixes from v1.0.1, including allowing variable-elevation contours 

On 29 July 2025, an Mw 8.8 earthquake struck off Kamchatka, Russia, generating a Pacific-wide tsunami and marking the largest earthquake since the launch of the surface water and ocean topography (SWOT) satellite in 2022. We analyze tsunami observations from SWOT together with three nearby deep-ocean assessment and reporting of tsunamis (DART) buoys to resolve the source of the event. SWOT provided the first high-resolution spaceborne track of a great subduction-zone tsunami, capturing waveforms that reveal complex propagation, dispersion, and scattering. Inversion of the DART time series using Gaussian unit sources shows that the rupture extended ∼400 km along strike, with peak uplift of ∼4 m, significantly different from the published finite-fault model. A blended source that combines the DART-inverted uplift with subsidence from the seismic–geodetic model best matches both datasets and reproduces the SWOT observations. Comparison with reconstructions of the 1952 Mw 9.0 Kamchatka earthquake indicates that the 2025 rupture likely reactivated significant portions of the megathrust that broke in 1952 but occurred farther down-dip and with little to no near-trench slip, consistent with its smaller tsunami impact. These findings highlight the hazard implications of short recurrence intervals of great earthquakes and show how rupture style governs tsunami severity. They also demonstrate the value of satellite altimetry for improving tsunami source characterization, post-event forecasting, and understanding of hydrodynamic processes. 

We present the first global-scale database of 4.3 billion P- and S-wave picks extracted from 1.3 PB continuous seismic data via a cloud-native workflow. Using cloud computing services on Amazon Web Services, we launched ~145,000 containerized jobs on continuous records from 47,354 stations spanning 2002-2025, completing in under three days. Phase arrivals were identified with a deep learning model, PhaseNet, through an open-source Python ecosystem for deep learning, SeisBench. To visualize and gain a global understanding of these picks, we present preliminary results about pick time series revealing Omori-law aftershock decay, seasonal variations linked to noise levels, and dense regional coverage that will enhance earthquake catalogs and machine-learning datasets. We provide all picks in a publicly queryable database, providing a powerful resource for researchers studying seismicity around the world. This report provides insights into the database and the underlying workflow, demonstrating the feasibility of petabyte-scale seismic data mining on the cloud and of providing intelligent data products to the community in an automated manner.