Search

Developing on-site earthquake early warning systems has been a challenging problem because of time limitations and the amount of information that can be collected before the warning needs to be issued. A potential solution that could prevent severe disasters is to predict the potential strong motion using the initial P-wave signal and provide warnings before serious ground shaking starts. In practice, the accuracy of prediction is the most critical issue for earthquake early warning systems. Traditional methods use certain criteria, selected through intuition or experience, to make the prediction. However, the criteria thresholds are difficult to select and may significantly affect the prediction accuracy. This paper investigates methods based on artificial intelligence for predicting the greatest earthquake ground motion early, when the P-wave arrives at seismograph stations. A neural network model is built to make the predictions using a small window of the initial P-wave acceleration signal. The model is trained by seismic waves collected from 1991 to 2019 in Taiwan and is evaluated by events in 2020 and 2021. From these evaluations, the proposed scheme significantly outperforms the threshold-based method in terms of its accuracy and average leading time. [ABSTRACT FROM AUTHOR]

The earthquake early warning (EEW) system is difficult to provide consistent magnitude estimate in the early stage of an earthquake occurrence because only few stations are triggered and few seismic signals are recorded. One of the feasible methods to measure the size of earthquakes is to extract amplitude parameters using the initial portion of the recorded waveforms after P-wave arrival. However, for a large-magnitude earthquake (M w > 7.0), the time to complete the whole ruptures resulted from the corresponding fault may be very long. The magnitude estimations may not be correctly predicted by the initial portion of the seismograms. To estimate the magnitude of a large earthquake in real-time, the amplitude parameters should be updated with ongoing waveforms instead of adopting amplitude contents in a predefined fixed-length time window, since it may underestimate magnitude for large-magnitude events. In this paper, we propose a fast, robust and less-saturated approach to estimate earthquake magnitudes. The EEW system will initially give a lower-bound of the magnitude in a time window with a few seconds and then update magnitude with less saturation by extending the time window. Here we compared two kinds of time windows for measuring amplitudes. One is P-wave time window (PTW) after P-wave arrival; the other is whole-wave time window after P-wave arrival (WTW), which may include both P and S wave. One to ten second time windows for both PTW and WTW are considered to measure the peak ground displacement from the vertical component of the waveforms. Linear regression analysis are run at each time step (1- to 10-s time interval) to find the empirical relationships among peak ground displacement, hypocentral distances, and magnitudes using the earthquake records from 1993 to 2012 in Taiwan with magnitude greater than 5.5 and focal depth less than 30 km. The result shows that considering WTW to estimate magnitudes has smaller standard deviation than PTW. The magnitude estimations using 1-s time widow have larger uncertainties. Progressively adopting peak displacement amplitudes (P d ) from 2- to 10-s WTW is suggested for EEW systems. [ABSTRACT FROM AUTHOR]