Earthquake Science最新文献

The ability to estimate earthquake source locations, along with the appraisal of relevant uncertainties, is paramount in monitoring both natural and human-induced micro-seismicity. For this purpose, a monitoring network must be designed to minimize the location errors introduced by geometrically unbalanced networks. In this study, we first review different sources of errors relevant to the localization of seismic events, how they propagate through localization algorithms, and their impact on outcomes. We then propose a quantitative method, based on a Monte Carlo approach, to estimate the uncertainty in earthquake locations that is suited to the design, optimization, and assessment of the performance of a local seismic monitoring network. To illustrate the performance of the proposed approach, we analyzed the distribution of the localization uncertainties and their related dispersion for a highly dense grid of theoretical hypocenters in both the horizontal and vertical directions using an actual monitoring network layout. The results expand, quantitatively, the qualitative indications derived from purely geometrical parameters (azimuthal gap (AG)) and classical detectability maps. The proposed method enables the systematic design, optimization, and evaluation of local seismic monitoring networks, enhancing monitoring accuracy in areas proximal to hydrocarbon production, geothermal fields, underground natural gas storage, and other subsurface activities. This approach aids in the accurate estimation of earthquake source locations and their associated uncertainties, which are crucial for assessing and mitigating seismic risks, thereby enabling the implementation of proactive measures to minimize potential hazards. From an operational perspective, reliably estimating location accuracy is crucial for evaluating the position of seismogenic sources and assessing possible links between well activities and the onset of seismicity.

I provide some science and reflections from my experiences working in geophysics, along with connections to computational and data sciences, including recent developments in machine learning. I highlight several individuals and groups who have influenced me, both through direct collaborations as well as from ideas and insights that I have learned from. While my reflections are rooted in geophysics, they should also be relevant to other computational scientific and engineering fields. I also provide some thoughts for young, applied scientists and engineers.

Among electromagnetic methods of short-term earthquake prediction, an approach is being actively developed based on the phenomenon of magnetic ultra-low-frequency (ULF) power depression occurring a few days before an earthquake. In particular, a nighttime geomagnetic power depression in the band 0.03–0.05 Hz was observed approximately 5 days before the catastrophic Tohoku 2011 earthquake. To verify the reliability of this method, we performed an extended analysis using data from magnetometer arrays JMA, MAGDAS, PWING, and INTERMAGNET. The selected stations included sites close to the epicenter (<300 km) and remote points (∼10000 km). The band-integrated spectral power of nighttime magnetic noise decreased significantly from March 6–9, several days before the earthquake. However, such variations occur simultaneously not only at nearby stations but also at distant stations. During this event, the ULF power depression was caused by low global geomagnetic activity, as evidenced by the planetary index SME. Thus, the depression of geomagnetic ULF noise cannot be considered a reliable short-term precursor.