Impact of datum definition on transient velocities from GNSS displacement time series in Networks mode: A Case Study of Cascadia Subduction Zone and Hikurangi margin.

Abstract

In earthquake research, the discovery and ongoing investigation of interseismic transient processes has revealed that faults are non-steady between large earthquakes. These transients are typically identified in continuously operating tectonic GNSS stations, whereby an acceleration away from the average interseismic rates of displacement can be identified with a variety of time series analysis methods. However, the features of these transients can vary depending on the processing strategy employed to derive displacement time series from the raw GNSS observables. In the processing strategy, the definition of the geodetic datum is necessary to determine global terrestrial reference frames (TRFs), providing an accurate and stable absolute reference of Earth's locations. It is essential for comprehending the dynamic changes in Earth's geometry driven by factors like tidal and non-tidal loading, plate tectonic seismic activity, and ongoing climate change. Therefore, just as geodesy aims for accuracy and stability in the TRF, the datum definition—i.e., the realization of the TRF-defining parameters origin, orientation, and scale—may emerge as a critical factor in processing GNSS networks for geodynamic purposes. The purpose of this study is to assess up to what extent the transient velocities obtained from GNSS-derived displacement time series change under different regional and global datum definitions for the Cascadia subduction zone and Hikurangi margin; regions with very well-known catalog of interseismic transient tectonic events. In our study, we process data from Cascadia to produce network solutions both NNR (No-Net-Rotation) and NNR+NNT (No-Net-Translation) constraint for regional and global datum definition, respectively. We employed dual-frequency ionosphere-free linear combination observations from 125 GNSS stations for the time between 2015 and 2020. The same GNSS processing strategy is then followed for the Hikurangi subduction zone using a set of 72 stations from the GeoNet project as well as the same control stations used for Cascadia spanning from 2002 to 2010. For the Cascadia displacement time series, we find variations in transient velocities under different datum definitions emphasizing the need for a comprehensive understanding of its impact on dynamic geophysical processes. Processing and analyses of the New Zealand data is ongoing and results will be presented, along with recommendations for both regions on how to reduce the occurrence of likely non-tectonic transients in the displacement time series. Ultimately, our results may have implications for improving the estimate of the slip budget at plate boundaries that is released aseismically.