Relocation of clustered seismic events on the Korean Peninsular using double-difference technique and cross-correlation

Abstract

Earthquake location is one of the starting points for earthquake research. This study relocated seismic events in and around the Korean Peninsula using the cross-correlation and the double-difference of travel time residual for surface wave. The analyzed events were from the Baekryeong Island earthquake sequence, the Boryeong offshore earthquake sequence, the earthquake swarm around Gyeongju, and the seismic events in the Dogye area. During the course of the analysis based on waveform similarity, unreported events were newly detected and waveforms were classified. Epicentral area size shrank by a factor of hundreds after relocation, and the spatial distribution of relocated epicenters coincided with the locations of clusters classified according to waveform similarity. When examining each of the four regions, 24 unreported fore- and aftershocks were newly detected, and 55 relocated epicenters were in accordance with the result of waveform classification for the Baekryeong Island earthquake sequence. For the Boryeong offshore earthquake sequence, it was confirmed that 149 relocated epicenters form a clear lineament that proceeded northeast during a two- months period. The segment that could cause the largest event in the sequence was identified, and the Lg-Pg time increase with increasing epicentral distance validated the relocation result. Around the city of Gyeongju, 70 earthquakes with similar waveforms were newly detected from the continuous data recorded during a two- year period. Relocated epicenters including catalogued and detected events were concentrated east of the Yeonil Tectonic Line. The fault plane solution derived from first motions for the largest event displayed the same trend as the tectonic line. For the seismic events in the Dogye mining area, six event groups were identified by relocation and waveform classification. Additionally, more than 44,000 events were newly detected by waveform cross-correlation from the continuous data recorded during one year, and the signals produced by mining activities were distinguished.