Real-time coseismic deformations from adaptively tight integration of high-rate GNSS and strong motion records

Abstract

SUMMARY In traditional tight integration of high-rate GNSS and strong motion sensors, an appropriate process variance is crucial for obtaining accurate broad-band coseismic deformations. In this paper, instead of using a subjectively empirical value, we present an approach for determining the process variance adaptively based on the adaptive Kalman filter for real-time use. The performance of the approach was validated by the colocated stations collected during the 2010 Mw 7.2 earthquake in El-Mayor, 2016 Mw 7.8 earthquake in New Zealand and 2016 Mw 6.5 earthquake in central Italy. The results show that this method complements the advantages of GNSS and strong motion accelerometers and can provide more accurate coseismic waveforms especially during the strong shaking period, due to the ability of the method to adjust the process variance in real time according to the actual status of the station. In addition, this method is also free from the influence of the baseline shift. Testing of the new method for the integration of strong motion and multi-GNSS indicates that multi-GNSS has an obvious improvement in the precision while single GPS has a poor observation condition.