Ya-Ju Hsu

We present a new regional Ryukyu Area Centroid Moment Tensor (RACMT) catalog that documents detailed transitions in deformation pattern and geodynamic processes across the Ryukyu subduction system. Regional CMT solutions for events with magnitudes greater than 4.0 are determined through full‐waveform inversion with systematically explored parameters to ensure robustness. The resulting focal mechanisms are broadly consistent with GCMT, while centroid depths for shallow earthquakes are significantly improved. Stress regimes and deformation patterns are characterized using K‐means clustering–based stress inversion and analysis of principal stress axes. Within the subducting Philippine Sea Plate (PSP), slab bending produces normal faulting in the eastern segment, transitioning westward to strike‐slip faulting driven by compression parallel to oblique convergence at the western Ryukyu Trench. In the overriding plate, two trench‐parallel seismic belts exhibit mutually perpendicular extension, corresponding to opening of the Okinawa Trough and stretching of the Ryukyu Arc. Furthermore, widespread lateral extensional deformation is identified throughout much of the deep PSP. These observations are consistent with enhanced trench retreat and slab rollback toward the western Ryukyu margin. We propose that efficient toroidal mantle flow around the western slab edge makes a major contribution to the westward increase in back‐arc opening rate, fore‐arc stretching, and widespread lateral extension within the deep PSP.

Understanding deformation patterns and rupture sources near the junction of the Ryukyu trench and the Longitudinal Valley in northeastern Taiwan is critical for seismic hazard assessment. We model the coseismic slip of the 2024 Mw 7.3 Hualien earthquake using onshore and seafloor geodetic data together with relocated seismicity. The rupture involves a deep 70° east-dipping Longitudinal Valley fault and a 35° west-dipping offshore fault, with maximum slips of 2.3 and 1.1 m, respectively. These faults released comparable amounts of geodetic moment, each equivalent to an Mw 7.1 event. The 2024 event and preceding offshore damaging earthquakes suggest that structures offshore Hualien accommodate a substantial portion of plate convergence. This interpretation is further supported by the sharp decrease in Global Navigation Satellite System secular velocities and shortening rates, as well as the seaward propagation of the Central Range fault from southern to northern Longitudinal Valley.

Interferometric Synthetic Aperture Radar (InSAR) deformation monitoring in vegetated mountainous areas often faces challenges, including temporal decorrelation and atmospheric noise. Extracting true surface deformation trends from time-series signals and validating their accuracy remain critical issues. This study focuses on the Lantai area in Northeast Taiwan, utilizing Sentinel-1 satellite imagery from 2021 to 2025. We employed the Small Baseline Subset (SBAS-InSAR) technique to resolve surface deformation, with a specific focus on optimizing time-series signals and performing validation analysis using in-situ GNSS data.To optimize the time-series deformation signals, this study compared the efficacy of three smoothing methods on the raw SBAS results: Mean Filter, Median Filter, and Gaussian Filter. The results indicate that while the Mean Filter is computationally efficient, it tends to cause boundary blurring and time delays. The Median Filter effectively removes sudden noise spikes but performs less effectively in smoothing subtle continuous changes. In contrast, the Gaussian Filter successfully suppresses noise while preserving waveform continuity, making it the most suitable method for analyzing long-term deformation trends in this study area.Regarding accuracy validation, the study compared the optimized InSAR time-series deformation with data from continuous GNSS monitoring stations. The comparison reveals that, due to the 12-day satellite revisit cycle and dense vegetation, the InSAR results exhibit a noticeable short-term drift effect. However, over the five-year observation period, the overall cumulative deformation trends between InSAR and GNSS show good consistency. This research confirms that with appropriate filter optimization, Sentinel-1 time-series InSAR technology can be effectively applied to broad-area surface deformation screening in mountainous regions, providing reliable long-term trend data for landslide potential zoning.

Taiwan sits at the junction of the Ryukyu and Manila subduction zones, where a rapid convergence rate of ~90 mm/yr drives intense seismic and tsunami hazards. However, land-based geodetic networks provide insufficient resolution for monitoring offshore deformation. To address this, we have developed and deployed GNSS-Acoustic (GNSS-A) systems to monitor seafloor deformation. A total of six GNSS-A sites were established along the southern Ryukyu subduction zone near Taiwan, with three additional sites located near the northern tip of the Manila Trench. GNSS-A data in the southernmost Ryukyu margin reveal an eastward increase in convergence rate, from 92 mm/yr offshore Hualien to 123 mm/yr near the Gagua Ridge, indicating the potential to generate Mw 7.5–8.4 earthquakes. The 2024 Mw 7.3 Hualien earthquake ruptured a deep 70° east-dipping Longitudinal Valley fault and a 35° west-dipping offshore fault. At seafloor site ORY2, ~ 40 km east of the epicenter, we recorded coseismic displacements of 9.1±12.1 cm eastward and 12.3±11.4 cm southward motions, along with 52.9±13.5 cm uplift. These observations are consistent with coseismic dislocation modeling results. Additionally, multiple slow slip events on fault systems in eastern Taiwan appear to have preceded the 2024 Mw 7.3 Hualien earthquake.Offshore southern Taiwan, geodetic data reveal N–S-oriented extension in the Tainan Basin and NE–SW extension between the northern Manila Trench and the North Luzon Trough. These strain axes align with the focal mechanisms of the 1994 M 6.5 and 2006 Mw 7.0 earthquakes. Notably, deformation and seismicity patterns shift distinctly across the Eurasian Plate–South China Sea continent–ocean boundary near 20°N. Together, these integrated observations provide new insights into fault segmentation, strain accumulation, and regional seismic and tsunami hazards.

Studying deep-seated catastrophic landslides is challenging due to their complex geological conditions, slow and imperceptible movements, limited real-time monitoring, and difficulties in accurately assessing sliding behavior. We examine slope instability in a deep-seated landslide-prone area located in Lantai, Taiwan, by analyzing the temporal evolutions of slope movement and rainfall. Using a non-invasive automatic monitoring system that integrates the Global Navigation Satellite System, seismic velocity changes derived from ambient noise interferometry, as well as rainfall and groundwater-level records, we investigate rainfall-induced accelerated slope deformation in 2015–2024. Our findings show that acceleration events are triggered when rainfall lasts over 30 h and cumulative rainfall exceeds 200 mm at the study site. Multiple sliding events show a consistent temporal sequence, in which rainfall triggers an immediate dv/v decrease of 2–4%, followed by GNSS horizontal displacement typically in the range of ~10–30 mm within 1–2 days, and a delayed groundwater-level rise of about ~3 m occurring 4–5 days after peak GNSS displacement rate. This time lag suggests the initial slope acceleration is associated with pore-pressure increases within the landslide mass. The one-dimensional hydrological model successfully simulates rainfall infiltration into the groundwater system, enabling prediction of delayed groundwater responses. This integrated monitoring-modeling framework improves hazard assessment and early warning for rainfall-induced deep-seated landslides and is broadly applicable worldwide.

The 4 February 2018 Mw 6.2 earthquake struck the Hsincheng Ridge, offshore Hualien, in eastern Taiwan. Using a kinematic inversion of Global Navigation Satellite System (GNSS) time series combined with a forward calculation of dilatation static offsets, we infer that the earthquake has ruptured an asperity located at a depth of 10–15 km on a 25 km × 15 km northeast (NE)-dipping subhorizontal fault plane with a dominant right-lateral strike-slip mechanism. The total rupture duration inferred from simulations of dilatation time series is about 15 s, which corresponds to a rupture velocity of approximately 1.7 km·s−1. The aftershock sequence occurring within the first 2 days following the mainshock is characterized by abundant seismicity (1138 events with ML≥1.7) comprising two clusters located on both sides of the mainshock hypocenter at midcrustal depth (about 7–13 km). The seismicity in the NE cluster is likely controlled by afterslip, whereas aftershocks in the southwest cluster likely represent time-dependent failures in response to coseismic static stress excess. Furthermore, we find that the 4 February 2018 event has generated maximum static and dynamic Coulomb stress excess of 0.1 MPa in the hypocenter region of the 6 February 2018 Mw 6.4 mainshock, which suggests that the earthquake could have possibly advanced the occurrence of the 6 February earthquake. Finally, we report preliminary observations of a possible small-scale interplay between seismic and aseismic deformation (afterslip) in the vicinity of the hypocenter of the 6 February earthquake. The investigation of the precursory phase, if any, would require further high-precision geodetic and seismological measurements.

Kinematic coupling models inverted from geodetic data are widely used to evaluate how slip deficit is distributed along subduction megathrusts during the interseismic period, and are central to earthquake and tsunami hazard assessment. Yet, existing coupling models differ widely in methodology and inputs, lack common community standards, and are scattered across publications and repositories. Here, we introduce the “Coupling Cloud” (https://couplingcloud.ucsd.edu), an open, extensible, community-driven platform that curates, standardizes, documents, and disseminates more than 95 kinematic coupling models from 56 publications across 21 subduction margins. The platform provides interactive 2D and 3D plate-interface viewers to inspect coupling models together with associated information such as slab geometry, uncertainty estimates and metadata. All datasets can be downloaded directly in standardized formats: surface-projected coupling values as NetCDF, plate-interface dislocation geometries as VTU, and model metadata as YAML files. We demonstrate the advantages of centralized and standardized coupling data through a Cascadia subduction zone example, where synthesizing eight full-margin models reveals along-strike patterns that are not apparent when models are examined individually. Consolidating coupling models within a coherent, version-controlled framework enables systematic cross-margin comparison and FAIR-compliant data sharing, opening the door to more comprehensive assessment of megathrust mechanics.

Understanding the role of aseismic slip in earthquake cycles is essential for assessing seismic hazards and short-term forecasting. Eastern Taiwan’s double-vergence suture zone, where the Philippine Sea Plate subducts beneath the Eurasian Plate, experiences frequent M ≥ 6 earthquakes and widespread aseismic slip, making it an ideal natural setting to study earthquake triggering processes. Here we demonstrate how aseismic deformation contributed to the April 3, 2024 Mw7.3 Hualien earthquake by analyzing a 24-year catalog of repeating earthquake sequences (RESs) and earthquake swarms. We find that six out of nine swarms in the epicentral area, northern Longitudinal Valley, were accompanied by increasing aseismic slip rates, as revealed by RESs on the west-dipping Central Range Fault (CRF). A notable aseismic slip episode in 2021 indicated by GNSS signals, the accelerated RESs-derived slip rate, and a four-month-long swarm sequence with high diffusivity (~5.2 m²/s), suggests joint contributions from over-pressured fluids and deep fault creep. Following this episode, a sequence of M6+ events occurred in 2022, and both seismicity and aseismic slip gradually increased again starting in 2023. Coulomb stress modeling indicates that cumulative aseismic and seismic slips since 2021 generated up to ~30 kPa positive stress on the eventual 2024 rupture, promoting fault weakening and shallower seismicity. This study provides compelling evidence for aseismic-slip-induced stress triggering of a major earthquake and highlights the importance of integrating aseismic processes into earthquake hazard models for collisional fault systems.

While slow slip events (SSEs) are recognized as a ubiquitous phenomenon in subduction zones, accurately tracking the spatiotemporal distribution of shallow SSEs remains challenging. Here, we investigate the potential occurrence of shallow SSEs in the southern Ryukyu forearc at the western boundary of the Philippine Sea plate using 25 years of Global Navigation Satellite System data. We identify trenchward transient motions near the Miyako Islands, with cumulative displacements of ∼10 mm between 2009 and 2011, indicating shallow SSEs that released a geodetic moment equivalent to M_w 6.8. These SSEs likely occurred updip of a M_w 6.1 interplate earthquake, whose associated positive Coulomb stress changes may have encouraged slow slip. Our results indicate that the shallow southern Ryukyu plate interface can accumulate elastic strain and release energy either aseismically or seismically. We therefore emphasize the importance of near‐trench seafloor geodesy to fully characterize fault slip behaviors beneath the southern Ryukyu forearc.

Eastern Taiwan is located in the transition zone where the Philippine Sea plate subducts beneath the Yangtze plate along the Ryukyu Trench and collides with the continental margin along the Longitudinal Valley suture zone. These complex tectonic interactions have led to frequent and devastating earthquakes. The GNSS-acoustic measurements in the southernmost Ryukyu margin characterize an eastward growing convergence rate from 92 mm/yr offshore Hualien to 123 mm/yr near the Gagua Ridge, suggesting a capability of hosting Mw 7.5-8.4 earthquakes. Along the Longitudinal Valley, the east-dipping Longitudinal Valley fault and the west-dipping Central Range fault form a dual-verging conjugate suture zone. The GNSS velocities relative to the Yangtze plate generally decrease northwestward from the Coastal Range, through the Longitudinal Valley, to the Central Range. Along the Coastal Range, GNSS velocities range from 67 to 72 mm/yr between Taitung and Fengbin. This rate then drops significantly to approximately 38 mm/yr at Shoufeng and further decreases to 24 mm/yr near Hualien. The shortening rate between the east coast and the Longitudinal Valley decreases northward, from 30 mm/yr between Taitung and Guangfu to approximately 10 mm/yr near Hualien. Additionally, shallow crustal earthquakes along the east coast indicate a significant clockwise rotation of P and SH axes from convergence-parallel (N120˚) south of Fengbin to about (N140˚) near Hualien. The orientations of GNSS velocity exhibit a similar clockwise rotation of 10˚ from Taitung to Hualien as well. These observations suggest a spatial change in seismotectonic stress as approaching the junction between the subduction of the Ryukyu Trench and the collision of the Longitudinal Valley suture zone. A significant portion of the accumulated strain is likely accommodated by offshore faults near Hualien, as evidenced by frequent large offshore earthquakes and interseismic subsidence along the Hualien coast. Continuous investigation of GNSS interseismic velocity, seismic activity, the long-term uplift rates of marine terraces, and coseismic uplift during historic earthquakes are crucial for revealing the long-term seismic hazard of eastern Taiwan.