Pemanfaatan EQTransformer & GaMMA Dalam Deteksi Kejadian Seismik dan Pengembangan Katalog Gempa Bumi. Studi Kasus Halmahera Barat, Indonesia

Navisa, Siti (2025) Pemanfaatan EQTransformer & GaMMA Dalam Deteksi Kejadian Seismik dan Pengembangan Katalog Gempa Bumi. Studi Kasus Halmahera Barat, Indonesia. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5017211051-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Wilayah Jailolo, Halmahera Barat merupakan salah satu daerah dengan aktivitas seismik tinggi yang disebabkan oleh lokasinya yang dekat dengan 2 zona subduksi serta adanya busur pegunungan Halmahera. Gempa bumi bermagnitudo kecil (M < 3) atau mikro-seismik sering kali tidak tercatat dalam katalog global seperti ISC, sehingga deteksi otomatis berbasis machine learning menjadi penting untuk melengkapi informasi aktivitas seismik lokal. Penelitian ini bertujuan untuk mengembangkan alur kerja otomatis menggunakan pendekatan machine learning guna membangun katalog gempa bumi mikro secara efisien, sekaligus meningkatkan resolusi spasial-temporal data seismik di wilayah Jailolo. Penelitian ini mengintegrasikan model deep learning EQTransformer untuk picking otomatis fase P dan S; algoritma GaMMA untuk asosiasi fase dan pembentukan event; serta metode NonLinLoc untuk penentuan lokasi hiposenter. Data yang digunakan merupakan seismogram dari jaringan 7G GeoForschungsZentrum (GFZ) pada periode Agustus 2016 hingga Juli 2017. Proses ini menghasilkan katalog yang mencakup 1.492 kejadian gempa bumi, dengan dominasi kedalaman <50 km dan Magnitudo Lokal (ML) rata-rata 1.11, yang sebagian besar tidak tercatat dalam katalog ISC. Hasil evaluasi menunjukkan bahwa 80,6% katalog ISC tercatat dari hasil asosiasi, yang mengindikasikan bahwa skema pembelajaran mesin ini mampu mendeteksi peningkatan aktivitas mikro-seismik secara signifikan, termasuk fase awal dari aktivitas swarm earthquake yang memuncak pada akhir 2017, meskipun masih diperlukan validasi manual atau integrasi dengan katalog global untuk menjangkau cakupan magnitudo yang lebih luas.
=================================================================================================================================
The Jailolo region in West Halmahera is one of the areas with high seismic activity, attributed to its proximity to two subduction zones and the presence of the Halmahera volcanic arc. Low-magnitudo earthquakes (M < 3), or microseismic events, are often not recorded in global catalogs such as the ISC, highlighting the importance of automated detection based on machine learning to supplement local seismic information. This study aims to develop an automated workflow using a machine learning approach to efficiently build a micro-earthquake catalog while enhancing the spatial and temporal resolution of seismic data in the Jailolo area. The methodology integrates the deep learning model EQTransformer for automatic P- and S-phase picking, the GaMMA algorithm for phase association and event building, and the NonLinLoc method for hypocenter location determination. The data used consist of seismograms from the 7G GeoForschungsZentrum (GFZ) network, covering the period from August 2016 to July 2017. This process resulted in a catalog of 1,492 earthquake events, predominantly at depths <50 km and with an average local magnitudo (ML) of 1.11, most of which are absent from the ISC catalog. Evaluation shows that 80.6% of ISC-listed events were successfully associated, indicating that this machine learning-based framework effectively detects significant increases in microseismic activity, including the early phase of a swarm earthquake that peaked in late 2017, although manual validation or integration with global catalogs is still required for broader magnitudo coverage.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Katalog Gempa bumi, Jailolo, EQTransformer, GaMMA, NonLinLoc. Earthquake Catalog, Jailolo, EQTransformer, GaMMA, NonLinLoc.
Subjects:	Q Science > QC Physics > QC20.7.F67 Fourier transformations Q Science > QE Geology > QE538.8 Earthquakes. Seismology Q Science > QE Geology > QE539 Microseisms.
Divisions:	Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Geophysics Engineering > 33201-(S1) Undergraduate Thesis
Depositing User:	Siti Navisa
Date Deposited:	29 Jul 2025 00:49
Last Modified:	29 Jul 2025 00:49
URI:	http://repository.its.ac.id/id/eprint/120112

Actions (login required)

View Item