Understanding microseismicity behavior and their response to earth processes by improving earthquake catalogs

Abstract

Natural earthquakes occur on faults ranging from 0 to 700 km beneath Earth's surface in different tectonic settings, such as along major subduction zones in Japan and the arc-continent collisional environment in Taiwan. Recent studies suggest that earthquake activities can be affected by various Earth processes, including extreme weather events on the Earth’s surface, large earthquakes, water/snow/glacier loading and unloading, erosion and sedimentation, etc. The Gutenberg–Richter magnitude-frequency statistics suggest that the number of earthquakes decays as a power law with the increase of earthquake magnitude, which means most earthquakes are of small magnitudes, i.e., microseismicity. Studying the behavior of microseismicity and their response to the Earth’s surface process can help us to better understand fault structures at depth as well as the physics of earthquake nucleation, and to mitigate seismic as well as other natural hazards. However, the understanding of microseismicity may be limited by the incompleteness of standard earthquake catalogs, especially during the noisy period following extreme weather events and large earthquakes. During my Ph.D. study, I have developed/applied machine-learning and template-matching tools to improve earthquake catalogs by detecting microearthquakes and calculating their focal mechanisms. Based on the improved high-resolution catalogs, I then perform a detailed analysis of the microseismicity behavior and their response to Earth processes. Specifically, I build a deep-learning Network for Polarity Classification (NPC) to automatically determine P-wave first-motion polarity. The outputs of NPC can directly be used to build focal mechanism catalogs for several times more microearthquakes than those listed in the standard catalogs. Next, I use template-matching and deep-learning methods to build a more complete earthquake catalog in Taiwan before and after the 2009 typhoon Morakot, which brought the highest rainfall in southern Taiwan in the past 60 years and triggered numerous landslides. I then use the new catalog to investigate the impact of this wet typhoon on microseismicity. I observe no other significant seismicity changes that can be attributed to surface changes induced by typhoon Morakot, but a clear reduction in seismicity rate near the typhoon’s low-pressure eye center in northeastern Taiwan during the typhoon passed by. I also relocate earthquakes in this new catalog and use it to study the spatiotemporal variations of mainshock-aftershock sequences and the subsurface faults structure in Taiwan. Last, I perform a systematic detection of intermediate-depth earthquakes (IDEQs) in the Japan subduction zone using the template-matching technique. I obtain a more complete IDEQ catalog before and after the 2011 magnitude (M) 9 Tohoku-Oki earthquake (TOEQ) and ten M5+ IDEQ mainshocks in Japan. The newly built template-matching catalog does not show any significant increase in IDEQs in the two months prior to TOEQ. But following the TOEQ, I find a significant increase in the rate of IDEQs in both upper and lower planes of the double seismic zone beneath 70 km depth. These results suggest that like seismic activity at shallow depth, IDEQs in the double seismic zone also respond to stress perturbations generated by the 2011 M9 TOEQ, highlighting a sustained seismic hazard associated with these intraslab events in the next decades.