Evaluating and correcting short-term clock drift in data from temporary seismic deployments

Abstract

Temporary seismic network deployments often suffer from incorrect timing records and thus pose a challenge to fully utilize the valuable data. To inspect and fix such time problems, the ambient noise cross-correlation function (NCCF) has been widely adopted by using daily waveforms. However, it is still challenging to detect the short-term clock drift and overcome the influence of local noise on NCCF. To address these challenges, we conduct a study on two temporary datasets, including an ocean-bottom-seismometer (OBS) dataset from the southern Mariana subduction zone and a dataset from a temporary dense network from the Weiyuan shale gas field, Sichuan, China. We first inspect the teleseismic and local event waveforms to evaluate the overall clock drift and data quality for both datasets. For the OBS dataset, NCCF using different time segments (3, 6, and 12-h) beside daily waveforms data is computed to select the data length with optimal detection capability. Eventually, the 6-h segment is the preferred choice with high detection efficiency and low noise level. For the land dataset, higher drift detection is achieved by NCCF using the daily long waveforms. Meanwhile, we find that NCCF symmetry on the dense array is highly influenced by localized intense noise for large interstation distances (>1 ​km) but is well preserved for short interstation distances. The results have shown that the use of different segments of daily waveform data in the OBS dataset, and the careful selection of interstation distances in the land dataset substantially improved the NCCF results. All the clock drifts in both datasets are successfully corrected and verified with waveforms and NCCF. The newly developed strategies using short-segment NCCF help to overcome the existing issues to correct the clock drift of seismic data.