Search

This study analyzed time series data from six GPS stations within the Tatun Volcano Group (TVG), a long‐dormant volcanic system in northern Taiwan, using multichannel singular spectrum analysis to search for potential spatiotemporally correlated transient deformations. A notable cycle of transient deformation was identified from 2015 to 2020, characterized by ground subsidence and uplift of up to 10 mm, accompanied by asymmetric horizontal motions directed inward and outward toward Dayoukeng, the largest fumarole and hydrothermal area in TVG. Evidence from earthquake focal mechanisms and gas composition, along with preliminary source modeling, suggest that these transient phases were likely caused by the pressure change of shallow hydrothermal systems beneath Dayoukeng. Further analysis of time series data from three long‐operating GPS stations revealed similar patterns of transient motion in the area from 2006 to 2015, indicating that TVG has experienced cyclical deformation, akin to many other volcanic systems worldwide. Plain Language Summary: Monitoring ground deformation in volcanoes is crucial for understanding the underlying processes and providing early warnings of potential hazards. Previous geodetic studies of the long‐dormant Tatun Volcano Group (TVG) in northern Taiwan relied on sporadic data, leading to incomplete insights into its time‐varying deformation behavior. Continuous GPS, however, allows for accurate tracking of surface movements over time, including short‐term changes caused by intermittent volcanic activity. This first GPS study of TVG employed a method called multichannel singular spectrum analysis to identify transient signals in noisy data from six stations. Significant ground deformation was revealed between 2015 and 2020, with the ground sinking and rising by up to 10 mm, along with horizontal movements toward and away from Dayoukeng, a major steam vent area in TVG. This sinking is likely related to changes in the shallow hot water and steam systems beneath Dayoukeng, as suggested by anomalies in earthquake activity and surface gas composition. Additional GPS data from 2006 to 2015 revealed that TVG experienced recurring cycles of ground deformation, closely correlated with local seismic activity. Key Points: Data‐adaptive detection of transient deformation in noisy GPS time seriesCycle of ground deformation first detected in the Tatun Volcano area, where about 7 million people reside nearbyHydrothermal processes may have driven consistent spatiotemporal variations in seismicity and deformation in the Tatun Volcano area [ABSTRACT FROM AUTHOR]

Southwestern Taiwan exhibits multiple fold-thrust systems as a consequence of the interaction between the Philippine Sea Plate and the Eurasian Plate. A prominent geological feature of this region is the extensive layer of GuTingKeng mudstone, with a thickness of approximately 4 km, which serves as a source material for the formation of mud or shale diapirs. The 2016 Mw 6.4 Meinong earthquake, striking southwestern Taiwan at a depth of 15–20 km and inducing approximately 100 mm of uplift, has prompted investigations into the potential involvement of shallow structures (< 4 km) in this uplift. Recent studies have proposed that such shallow structures may have contributed significantly to the observed uplift during the earthquake. This study aims to elucidate the role of buried mud diapirs in the context of coseismic deformation. Here, we present a modeling approach that utilizes sill-like dislocations to simulate the deformation at the upper tip of the diapir. Our results indicate a vertical opening of approximately 60 mm at a depth of 1.4 km, which closely aligns with the spatial distribution of tomographic and gravity anomalies. We also examine how the coseismic stress changes induced by the Meinong earthquake can lead to a dilatational strain of about 1.2 microstrain within the shallow depth range of 0–4 km, resulting in extension within our modeled region. In contrast, the dilatational strain diminishes from 0.2 to − 1.2 microstrain at greater depths (4–8 km), implying compression in the subsurface beneath the diapir's top. This study discusses the potential mechanisms how fluid-rich and high-pressure mudstone may be deformed through coseismic process and how mud diapirs may contribute to additional deformation within the seismic cycle. [ABSTRACT FROM AUTHOR]

We test the feasibility of GPS-based rapid centroid moment tensor (GPS CMT) methods for Taiwan, one of the most earthquake prone areas in the world. In recent years, Taiwan has become a leading developer of seismometer-based earthquake early warning systems, which have successfully been applied to several large events. The rapid determination of earthquake magnitude and focal mechanism, important for a number of rapid response applications, including tsunami warning, is still challenging because of the limitations of near-field inertial recordings. This instrumental issue can be solved by an entirely different observation system: a GPS network. Taiwan is well posed to take advantage of GPS because in the last decade it has developed a very dense network. Thus, in this research, we explore the suitability of the GPS CMT inversion for Taiwan. We retrospectively investigate six moderate to large (M w6.0 ∼ 7.0) earthquakes and propose a resolution test for our model, we find that the minimum resolvable earthquake magnitude of this system is ∼ M w5.5 (at 5 km depth). Our tests also suggest that the finite fault complexity, often challenging for the near-field methodology, can be ignored under such good station coverage and thus, can provide a fast and robust solution for large earthquake directly from the near field. Our findings help to understand and quantify how the proposed methodology could be implemented in real time and what its contributions could be to the overall earthquake monitoring system. [ABSTRACT FROM AUTHOR]

Geoelectric self‐potential (SP) signals are sensitive to natural and anthropogenic factors. The SP spectral characteristics under the different factors in Taiwan were investigated, and the SP spectral scalings were correlated with urbanization level, seismicity, and crustal deformation. The ambient SP noise models were first established by estimating the probability density functions of the spectrograms at each frequency. The effects of the natural and anthropogenic factors on the SP signals are understood by comparing the SP noise models under various conditions, such as precipitation, urbanization, and electric trains. Results show that the SP signals in areas of high industrialization and human activity and areas close to train stations behave as white noises and exhibit a distinct spectral ripple at frequencies around 1 Hz. On the other hand, the SP spectral power law parameters, Gutenberg‐Richter b values, and dilation strain rates were estimated by using the SP, earthquake catalog, and GPS data, respectively, during 2012–2017. By investigating the correlations of the SP spectral parameters with the Gutenberg‐Richter b value, dilation strain rates, and urbanization level, the SP optimal frequency band is found between 0.006 and 1 Hz due to the high correlation between the SP and seismicity data and between the SP and dilation data and the low correlation between the SP and urbanization data. Hence, this study may help the filtering and screening of the SP data and facilitate the understanding of the mechano‐electric behavior in the crust. Plain Language Summary: Self‐potential is a naturally electric potential difference beneath the Earth's surface, measured between any two points on the ground or in boreholes. The self‐potential signals are sensitive to natural factors, such as ionospheric disturbance, solar wind, and tidal forces, and to artificial factors, such as electric trains, factories, and power pipelines. Hence, the investigation of the background noises is critical. The self‐potential noise models were established by estimating the probability density functions of the daily power spectral densities. We compared the noise models under several conditions, such as urbanization, electric trains, and rainfalls. We found that the noise models behave as white noises in areas of high industrialization and human activity. Furthermore, we also estimated the self‐potential power law parameters, seismic b values, and dilation strain rates. We then studied correlations of the self‐potential power law parameters with the urbanization levels, b values, and dilation strain rates. We found out the frequency‐dependent correlations of the spectral scalings with the other parameters and determined the self‐potential signals with a high signal‐to‐noise ratio in the frequency band of 0.006–1 Hz. Understanding such correlations may understand the feature of the mechano‐electric behavior in the crust and facilitate the development of a seismic‐electric theory. Key Points: Self‐potential (SP) signals show white noises or spectral ripples in areas of high industrialization and areas close to train stationsThe SP spectral exponent seems to positively relate to the seismic b value and dilation rate and negatively to the urbanization levelThe SP spectral exponent shows frequency‐dependent correlations with the seismic b value, dilation rate, and urbanization level [ABSTRACT FROM AUTHOR]