Clock Error Detection for Decade-Long Seismic Monitoring Network in Marapi Volcano, Indonesia

Abstract Accurate seismometer timing is crucial for seismic studies that rely on arrival time, such as earthquake relocation and subsurface velocity change monitoring—both are essential for understanding volcanic systems. To assess the timing accuracy in the decadal seismic records from the monitoring network at the Marapi volcano, Indonesia, we apply both teleseismic P-wave waveform cross-correlation function and ambient seismic noise cross-correlation function methods. These methods derive consistent clock error estimates, revealing both constant and nearly linear drifting time errors at two stations, PCAK and GGSL. Our results indicate a ∼1 s constant time error at PCAK for approximately nine months (2019–2020) and at GGSL for about one month (2018). In addition, at both stations, time drifted nearly linearly over three to five months in 2016 at a rate of ∼4 to 5 ms/day. Synthetic tests show that an ∼1 s constant time shift in current network configuration leads to an ∼1.5 km depth error for an event at ∼10 km depth beneath the crater. The linearly drifting time error causes an artificial downward migration of seismicity that should remain stationary, potentially leading to misinterpretations of magma transport and storage variations. Timing error estimations are essential for future seismic studies using this dataset, and the workflow developed here can be generalized to other long-term seismic monitoring networks.