Antipodal Earthquakes Identified by Data Mining and Cross-Correlation Analysis of P-wave

Abstract

This thesis examines whether moderate-magnitude earthquakes (approximately M4) in the Vanuatu region and the Reguibat Shield of northwest Africa exhibit reproducible antipodal P-wave similarity. Although antipodal seismic focusing is well established, direct earthquake-to-earthquake waveform comparison using modern global datasets has received limited attention. To address this gap, three near-antipodal earthquake sequences from 2018, 2020, and 2021 were analysed using vertical-component P-wave onsets recorded at regional broadband seismic stations. Waveform data were obtained from the IRIS Data Management Center and processed using both standard and epicentral-distance-tuned band-pass filters. A unified analysis workflow was applied, and waveform similarity was quantified using cross-correlation coefficients (ρ), median coherence, and polarity behaviour. Both positive and negative correlation values were considered meaningful, representing direct and polarity-reversed but coherent waveform similarity. Measurable antipodal similarity was observed in all sequences, with the strongest results along the second antipodal links (AL2) of the 2020 and 2021 triplets (ρ ≈ −0.79 and −0.85). Polarity behaviour varied systematically across earthquake-to-earthquake links and stations, particularly where waveform coherence was higher. Overall, the results indicate that antipodal P-wave onset similarity can be detected for moderate-magnitude earthquakes using consistent signal processing and statistical evaluation. The Vanuatu–Reguibat Shield corridor provides a useful observational setting for examining long-distance seismic waveform similarity without requiring explicit modelling of deep-Earth seismic phases.