Earthquake Science最新文献

We conducted rapid inversions of rupture process for the 2023 earthquake doublet occurred in SE Türkiye, the first with a magnitude of M_W7.8 and the second with a magnitude of M_W7.6, using teleseismic and strong-motion data. The teleseismic rupture models of the both events were obtained approximately 88 and 55 minutes after their occurrences, respectively. The rupture models indicated that the first event was an asymmetric bilateral event with ruptures mainly propagating to the northeast, while the second one was a unilateral event with ruptures propagating to the west. This information could be useful in locating the meizoseismal areas. Compared with teleseismic models, the strong-motion models showed relatively higher resolution. A noticeable difference was found for the M_W7.6 earthquake, for which the strong-motion models shows a bilateral event, rather than a unilateral event, but the dominant rupture direction is still westward. Nevertheless, all strong-motion models are consistent with the teleseismic models in terms of magnitudes, durations, and dominant rupture directions. This suggests that both teleseismic and strong-motion data can be used for fast determination of major source characteristics. In contrast, the strong-motion data would be preferable in future emergency responses since they are recorded earlier and have a better resolution ability on the source ruptures.

In this study, we analyzed 100 three-component strong ground motion records observed within 200 km of the causative fault of the 6 February 2023 M_W7.8 Pazarcık (Kahramanmaraş) Earthquake in SE Türkiye. The wavelet method was utilized to identify and analyze the characteristics of pulse-like ground motions in the near-fault region, while considering the uncertainty of the pulse orientation during the analysis. Our investigation focused on the effects of the focal mechanism and rupture process on the spatial distribution, pulse orientation, and maximum pulse direction of the observed pulse-like ground motion. We also analyzed the amplitude and period of the observed ground pulses and the effect of long-period amplification on the ground motion response spectra. Our results indicated the following: (1) A total of 21 typical ground velocity pulses were observed during this earthquake, exhibiting complex characteristics due to the influence of the strike-slip mechanism and rupture directivity. Most ground pulses (17 out of 21) were recorded within 20 km of the fault, in a wide range of orientations, including normal and parallel to the fault direction. The waveforms exhibited unidirectional features, indicating the effects of left-lateral fault slip. Distinct pulses observed more than 20 km from the fault were mainly oriented normal to the fault. The waveforms were bidirectional with double- or multi-round trips as a result of rupture directivity. (2) The amplitudes of the observed pulses ranged from 30.5 to 220.0 cm/s, with the largest peak velocity of 220.0 cm/s observed at Station 3138. The pulse periods ranged from 2.3 to 14.5 s, with the longest pulse period of 14.5 s observed at Station 3116. The amplitude and period of the pulses observed during this earthquake were comparable to those of similar-magnitude global earthquakes. The amplitude of the pulses decreased significantly with increasing fault distance, whereas the pulse period was not significantly affected by the fault distance. (3) Compared with non-pulse records, the velocity pulse records had a pronounced amplification effect on the acceleration response spectra near the pulse period, with factors ranging from 2.1 to 5.8. The larger velocity pulses also significantly amplified the velocity response spectra, particularly over the long periods. This significant amplification effect of the pulses on the response spectra leads to empirical models underestimating the long-period earthquake ground motion.

Stochastic finite-fault simulations are effective for simulating ground motions and are widely used in engineering to determine the impacts of ground motion and develop relevant predictive equations. In this study, the source, path, and site amplification coefficient of western Sichuan Province, China, and stochastic finite-fault simulations were used to simulate the acceleration time series, Fourier amplitude spectra, and 5% damped response spectra of 28 strong-motion stations with rupture distances within 300 km of the 2022 M_S6.8 Luding earthquake. The simulation results of 14 stations at rupture distances of 45–185 km match the observation. However, the simulation results of 3 near- and 6 far-field stations at rupture distances of 12–36 km and 222–286 km, respectively, were obviously deviated from the observations. Simulation results of the near-field stations are larger than the observations at high frequencies (>6 Hz). The discrepancy likely comes from the nonlinear site effect of near-field stations, which reduced the site amplification at high frequencies. Simulation result of the far-field stations is smaller than the observation at frequencies above 1 Hz. As these stations are located close to the Longmenshan Fault Zone (LFZ), thus, we obtained a new quality factor (Q) from data of historical events and stations located around LFZ. Using the new Q value, the discrepancies of the high-frequency simulation results of the far-field stations were corrected. This result indicated that the laterally varying Q values can be used to address the impact of strong crustal lateral heterogeneity on simulation.

In this study, the broadband ground motions of the 2021 M7.4 Maduo earthquake were simulated to overcome the scarcity of ground motion recordings and the low resolution of macroseismic intensity map in sparsely populated high-altitude regions. The simulation was conducted with a hybrid methodology, combining a stochastic high-frequency simulation with a low-frequency ground motion simulation, from the regional 1-D velocity structure model and the Wang WM et al. (2022) source rupture model, respectively. We found that the three-component waveforms simulated for specific stations matched the waveforms recorded at those stations, in terms of amplitude, duration, and frequency content. The validation results demonstrate the ability of the hybrid simulation method to reproduce the main characteristics of the observed ground motions for the 2021 Maduo earthquake over a broad frequency range. Our simulations suggest that the official map of macroseismic intensity tends to overestimate shaking by one intensity unit. Comparisons of simulations with empirical ground motion models indicate generally good consistency between the simulated and empirically predicted intensity measures. The high-frequency components of ground motions were found to be more prominent, while the low-frequency components were not, which is unexpected for large earthquakes. Our simulations provide valuable insight into the effects of source complexity on the level and variability of the resulting ground motions. The acceleration and velocity time histories and corresponding response spectra were provided for selected representative sites where no records were available. The simulated results have important implications for evaluating the performance of engineering structures in the epicentral regions of this earthquake and for estimating seismic hazards in the Tibetan regions where no strong ground motion records are available for large earthquakes.

Teleseismic traveltime tomography is an important tool for investigating the crust and mantle structure of the Earth. The imaging quality of teleseismic traveltime tomography is affected by many factors, such as mantle heterogeneities, source uncertainties and random noise. Many previous studies have investigated these factors separately. An integral study of these factors is absent. To provide some guidelines for teleseismic traveltime tomography, we discussed four main influencing factors: the method for measuring relative traveltime differences, the presence of mantle heterogeneities outside the imaging domain, station spacing and uncertainties in teleseismic event hypocenters. Four conclusions can be drawn based on our analysis. (1) Comparing two methods, i.e., measuring the traveltime difference between two adjacent stations (M1) and subtracting the average traveltime of all stations from the traveltime of one station (M2), reveals that both M1 and M2 can well image the main structures; while M1 is able to achieve a slightly higher resolution than M2; M2 has the advantage of imaging long wavelength structures. In practical teleseismic traveltime tomography, better tomography results can be achieved by a two-step inversion method. (2) Global mantle heterogeneities can cause large traveltime residuals (up to about 0.55 s), which leads to evident imaging artifacts. (3) The tomographic accuracy and resolution of M1 decrease with increasing station spacing when measuring the relative traveltime difference between two adjacent stations. (4) The traveltime anomalies caused by the source uncertainties are generally less than 0.2 s, and the impact of source uncertainties is negligible.

The number of dispersion curves increases significantly when the scale of a short-period dense array increases. Owing to a substantial increase in data volume, it is important to quickly evaluate dispersion curve quality as well as select the available dispersion curve. Accordingly, this study quantitatively evaluated dispersion curve quality by training a convolutional neural network model for ambient noise tomography using a short-period dense array. The model can select high-quality dispersion curves that exhibit a ≤ 10% difference between the results of manual screening and the proposed model. In addition, this study established a dispersion curve loss function by analyzing the quality of the dispersion curve and the corresponding influencing factors, thereby estimating the number of available dispersion curves for the existing observation systems. Furthermore, a Monte Carlo simulation experiment is used to illustrates the station-pair interval distance probability density function, which is independent of station number in the observational system with randomly deployed stations. The results suggested that the straight-line length should exceed 15 km to ensure that loss rate of dispersion curves remains < 0.5, while maintaining the threshold ambient noise tomography accuracy within the study area.

Waveforms of seismic events, extracted from January 2019 to December 2021 were used to construct a test dataset to investigate the generalizability of PhaseNet in the Shandong region. The results show that errors in the picking of seismic phases (P- and S-waves) had a broadly normal distribution, mainly concentrated in the ranges of −0.4–0.3 s and −0.4–0.8 s, respectively. These results were compared with those published in the original PhaseNet article and were found to be approximately 0.2–0.4 s larger. PhaseNet had a strong generalizability for P- and S-wave picking for epicentral distances of less than 120 km and 110 km, respectively. However, the phase recall rate decreased rapidly when these distances were exceeded. Furthermore, the generalizability of PhaseNet was essentially unaffected by magnitude. The M4.1 earthquake sequence in Changqing, Shandong province, China, that occurred on February 18, 2020, was adopted as a case study. PhaseNet detected more than twice the number of earthquakes in the manually obtained catalog. This further verified that PhaseNet has strong generalizability in the Shandong region, and a high-precision earthquake catalog was constructed. According to these precise positioning results, two earthquake sequences occurred in the study area, and the southern cluster may have been triggered by the northern cluster. The focal mechanism solution, regional stress field, and the location results of the northern earthquake sequence indicated that the seismic force of the earthquake was consistent with the regional stress field.

As dense seismic arrays at different scales are deployed, the techniques to make full use of array data with low computing cost become increasingly needed. The wave gradiometry method (WGM) is a new branch in seismic tomography, which utilizes the spatial gradients of the wavefield to determine the phase velocity, wave propagation direction, geometrical spreading, and radiation pattern. Seismic wave propagation parameters obtained using the WGM can be further applied to invert 3D velocity models, Q values, and anisotropy at lithospheric (crust and/or mantle) and smaller scales (e.g., industrial oilfield or fault zone). Herein, we review the theoretical foundation, technical development, and major applications of the WGM, and compared the WGM with other commonly used major array imaging methods. Future development of the WGM is also discussed.

Current popular deep learning seismic phase pickers like PhaseNet and EQTransformer suffer from performance drop in China. To mitigate this problem, we build a unified set of customized seismic phase pickers for different levels of use in China. We first train a base picker with the recently released DiTing dataset using the same U-Net architecture as PhaseNet. This base picker significantly outperforms the original PhaseNet and is generally suitable for entire China. Then, using different subsets of the DiTing data, we fine-tune the base picker to better adapt to different regions. In total, we provide 5 pickers for major tectonic blocks in China, 33 pickers for provincial-level administrative regions, and 2 special pickers for the Capital area and the China Seismic Experimental Site. These pickers show improved performance in respective regions which they are customized for. They can be either directly integrated into national or regional seismic network operation or used as base models for further refinement for specific datasets. We anticipate that this picker set will facilitate earthquake monitoring in China.

Seismic phase pickers based on deep neural networks have been extensively used recently, demonstrating their advantages on both performance and efficiency. However, these pickers are trained with and applied to different data. A comprehensive benchmark based on a single dataset is therefore lacking. Here, using the recently released DiTing dataset, we analyzed performances of seven phase pickers with different network structures, the efficiencies are also evaluated using both CPU and GPU devices. Evaluations based on F₁-scores reveal that the recurrent neural network (RNN) and EQTransformer exhibit the best performance, likely owing to their large receptive fields. Similar performances are observed among PhaseNet (UNet), UNet++, and the lightweight phase picking network (LPPN). However, the LPPN models are the most efficient. The RNN and EQTransformer have similar speeds, which are slower than those of the LPPN and PhaseNet. UNet++ requires the most computational effort among the pickers. As all of the pickers perform well after being trained with a large-scale dataset, users may choose the one suitable for their applications. For beginners, we provide a tutorial on training and validating the pickers using the DiTing dataset. We also provide two sets of models trained using datasets with both 50 Hz and 100 Hz sampling rates for direct application by end-users. All of our models are open-source and publicly accessible.