P-wave arrival time picking, Initial-motion polarity determination, Minimum entropy criterion, Transitive probability matrix, Geography. Anthropology. Recreation, Geodesy, QB275-343, Geology, QE1-996.5

Abstract In this study, we have developed and implemented a new technology capable of probabilistically selecting phase arrival times and determining the initial polarity of seismic waveforms without the requirement of prior information. In this new method, the arrival time is determined through an eigen-equation associated with the probability distribution of the noise level, which is then used to calculate the probability of the polarity. We have tested this method using synthetic waveforms as well as records from well-established databases. The results demonstrate a high degree of concurrence with manually picked arrival times and polarities (98% accuracy) in the local seismic catalog. This suggests that the proposed method can provide consistent and unified judgments in phase picking tasks. In comparison, this method has shown comparable reliability to existing neural-network-based AI methods while maintaining greater portability due to its lack of dependence on training data. Graphical abstract