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.