An Innovative Gap-Filling Approach for Ambient Seismic Noise Data

Over recent decades, the use of seismic ambient noise has become widespread for estimating velocity and attenuation structures, and variations in these parameters on Earth. However, working with software that requires continuous data can prove challenging, as any gaps in the data render the software unusable. Substituting zero values for these gaps fundamentally alters the frequency content of the time series. The process of filling gaps is crucial to ongoing work with the software, yet producing a time series similar to the gap is complex. Through investigation, it was found that the time series after the gap shares remarkable similarities with the gap. Our goal was to fill in the gap in the time series by making corrections to the data that came after it. To do this, we used a moving periodogram with various windows on the time series before and after the gap. We then stacked these periodograms to create a reference periodogram that accurately indicated the frequency content of the gap. We used this reference periodogram to improve the frequency content of the primary signal.