Focal mechanisms of microearthquakes ($ML<4$) are important in understanding tectonics of seismogenic structure, especially in areas with high seismic potential. However, one of the major difficulties in obtaining a reliable focal mechanism for a microearthquake is deriving sharp P‐ and S‐wave arrivals from a waveform with a low signal‐to‐noise ratio compared with those from larger earthquakes. Another challenge is the uneven distribution of seismic stations, which results in high uncertainty of focal mechanisms of microearthquakes. To address these problems, we used waveform inversion to improve the accuracy of determining focal mechanisms. High‐quality waveform data recorded by 12 temporal broadband stations deployed in the Ma River fault zone, northern Vietnam, are used to conduct a waveform inversion to obtain reliable fault‐plane solutions. In addition, a frequency–wavenumber integration method and a cut‐and‐paste method are applied to estimate the optimum focal mechanisms. This procedure is shown to effectively increase the accuracy of focal mechanisms obtained for microearthquakes. Our results are congruent with local geological structures and indicate a complex fractured crust in the Ma River fault zone. By studying local stress states, we claim that the Ma River fault zone is a wrench fault system with high seismic potential.

Seismic monitoring relies on seismography. However, the high cost of seismic equipment has presented a challenge to increasing the density of seismic networks in previous decades. Due to the large station spacing and inferior coverage of stations, this situation has led to a loss of detail in many research results. Along with the improvement of technology, the problem of increasing the density of seismographic observations is no longer an impossible issue. This makes it feasible to deploy a dense seismic network for monitoring earthquakes. This study deployed a linear dense array across the Meishan Fault in west-southern Taiwan for the purpose of analyzing the shallow fault zone structure. While the 1906 Meishan earthquake occurred in a period when historic records were available, the surficial geology surveys of the Meishan Fault are challenging because farming and construction engineering have obscured the outcrop. Early surveys of the Meishan Fault were mainly seismic surveys. In recent decades, over thirty profiles have been completed. However, the reflection seismic records had poor signal-to-noise ratios because the Meishan Fault is buried under thick sediments. Thus, the shallow structure of the Meishan Fault is still not known in detail. This study applied double-beamforming tomography to a dense seismic array to obtain high-resolution images of the Meishan Fault zone. The result shows that there is a south-dipping interface near the fault trace as indicated by the Central Geological Survey of Taiwan. In addition, we observed velocity transitions of perturbation profiles that may be caused by a branch fault, the Chentsoliao Fault. This study demonstrates that the ambient noise double beamforming method is an effective tool for imaging the detailed shallow structure along with the dense seismic array.