Clipped seismic record recovery analysis based on the cubic spline interpolation algorithm

Abstract

The clipping of near-field seismic waveforms caused by instrument limitations results in the waste of lots of waveforms and hinders the effective advancement of seismic research projects like source parameter inversion, near-fault ground motion characteristics, and earthquake early warning (EEW). In this paper, we propose to use the cubic spline interpolation method to restore the clipping waveform to solve the near-field data loss. We evaluate the recoverability of seismic waveforms with different clipped levels through artificial clipping experiments and compare the recovery waveforms with the projection onto convex sets (POCS) method. The results show that the restoration deviation of the cubic spline interpolation method is less than 1% at the clipped point, for the records whose clipped amplitude does not exceed 50% of the peak. In addition, the restored phase feature of the cubic spline interpolation is closer to the real waveform than the POCS method. To verify the performance of this method in the realistic earthquake, we apply it to the Barkam M_w 5.9 earthquake. The results show that the peak ratio of the restored clipping records to the measured non-clipping records is close to 1, and the relative error of the response spectrum is less than 0.1. Finally, we apply this method to successfully restore about 60% of near-field clipped records from the 2022 Lushan M_w 5.8 earthquake, which provided more data support for the study of the near-field ground motion.