Meeting the Requirements of Early Warning Systems in Real-Time GNSS Operations

Poster presentation given at the 2018 Seismology of the Americas Meeting, 14-17 May, 2018, Miami, Florida, USA.

Abstract

UNAVCO streams data from ~800 real-time, GNSS sites (RT-GNSS) from a combination of NSF-sponsored networks that together span large segments of the North American-Pacific plate boundary, the EarthScope Plate Boundary Observatory (PBO), the Caribbean plate boundary (COCONet) and Mexico’s Pacific plate boundary (TLALOCNet). Raw data streams are transmitted from the remote GNSS sites to UNAVCO’s data operations center where Precise Point Position (PPP) estimates are generated and distributed in real-time. Recent work has shown that GNSS-defined peak ground displacements (PGD) provide a magnitude scaling relation that, unlike estimates based on the first seconds of the P-wave, does not saturate above M7.  The inclusion of RT-GNSS PGD data can therefore greatly enhance the accuracy of early warning systems by providing improved magnitude estimates of large earthquakes. To take full advantage of this, not only must the data be complete and low latency, but it is essential that the PPP estimates have sufficiently low noise-levels such that long-period surface waves can be detected in real-time. Here, we show that we can combine the ambient noise levels in the GNSS PPP solutions with the GNSS-derived PGD scaling relationship to assess the ability of the PBO, TLALOCNet and COCONet networks to unambiguously detect the long-period surface waves generated by large earthquake events. This enables implementation of tools for evaluating and continuously monitoring the capability of magnitude-threshold detection level for RT-GNSS networks.  We find that with the current network configuration, RT-GNSS PGD would be detectable for most UCERF3-defined faults in California but this capability is decreased for events along sections of the Cascadia and Mexican Pacific plate boundaries.