Analysis of Saturation Effects of Distributed Acoustic Sensing and
Detection on Signal Clipping for Strong Motions
Abstract
Distributed Acoustic Sensing (DAS) systems are increasingly used for
seismic monitoring due to their affordability and high spatial
resolution. However, DAS systems are susceptible to signal clipping due
to their limited dynamic range, leading to substantial data loss during
strong ground motion and near-fault observations. In our study, we
investigated the effects of DAS signal clipping using collocated DAS
arrays with a cable loop in Hualien City using seismic data from the
2022 Taitung earthquake sequence. The two DAS arrays are connected to
different interrogators, allowing for direct comparisons of clipped and
unclipped signals. We observed that signal clipping in DAS is influenced
by earthquake magnitude, interrogator type, and cable coupling. Our
comparison also reveals that the signal clipping results in an amplitude
increase on all frequencies in the spectra. This drastic change in the
spectral domain led us to develop a frequency-based approach using
spectral coherence estimation on collocated channels to detect clipped
signals. The results demonstrate the coherencegrams from a single
interrogator with a cable loop can be employed to detect time intervals
in which signal clipping occurs. This finding is crucial for ensuring,
even during large earthquakes, DAS signal integrity and, furthermore,
enhancing the dependability of methods relying on high-quality data in
near-real-time, such as earthquake early warning systems. We also
recommend setting the sampling rate of interrogators based on the
slowest strong-motion related S-phase.