Costanzo Antonio
Istituto Nazionale di Geofisica e Vulcanologia, 00143 Rome, Italy.
Sensors (Basel). 2024 Dec 26;25(1):82. doi: 10.3390/s25010082.
This paper presents a new catalogue of the 2022/2023 Adriatic Offshore Seismic Sequence obtained by machine learning-based processing. The procedure performs the automatic picking and association of phases starting from the analysis of the continuous waveforms recorded by 40 seismic stations of the Italian National Seismic Network and 5 stations of the SISMIKO emergency group network. The earthquakes were detected over a 3-month period, between 1 November 2022 and 31 January 2023. This new catalogue consists of 2780 earthquakes with a magnitude equal to or greater than ML 0.4, providing more information about lower-magnitude earthquakes in particular. The results make available, on the one hand, new insights into the offshore sequence, which can contribute to confirming the attribution of the earthquakes to the Adriatic Fault System, and in particular, the mainshocks to the Cornelia fault thrust, as also hypothesised by other works in the literature. Moreover, the work provides a further contribution in showing the great potential of using machine learning-based procedures to build catalogues with a greater degree of completeness, even in very particular cases such as the one represented by the Adriatic offshore sequence, for which the minimum distance from the epicentres is high and the azimuth coverage limited.
本文展示了通过基于机器学习的处理获得的2022/2023年亚得里亚海近海地震序列的新目录。该程序从对意大利国家地震台网的40个地震台站和SISMIKO应急小组网络的5个台站记录的连续波形进行分析开始,执行相位的自动拾取和关联。地震是在2022年11月1日至2023年1月31日的3个月期间检测到的。这个新目录包含2780次震级等于或大于ML 0.4的地震,尤其提供了关于低震级地震的更多信息。一方面,这些结果为近海地震序列提供了新的见解,有助于确认地震与亚得里亚海断层系统的归属关系,特别是与科妮莉亚断层逆冲相关的主震,正如文献中的其他研究也所假设的那样。此外,这项工作还进一步证明了使用基于机器学习的程序构建具有更高完整性的目录的巨大潜力,即使在非常特殊的情况下,如亚得里亚海近海地震序列,其震中最小距离较大且方位覆盖范围有限。