Comparison of Different Techniques for Processing GNSS Data in Real Time for Tsunami Warning

Real-time processing and dissemination of GNSS derived co-seismic position offsets for tsunami early warning has been under development at the Geological Survey of Canada (NRCan) for the past six years. Using streamed data from the Western Canada Deformation Array (WCDA), results from two variations of Precise Point Positioning (PPP) and two variations of network-based solutions are evaluated: NRCan's real-time PPP software, and two commercial software packages RTD and RTNET, licensed from Geodetics Inc. and GPS Solutions Inc. respectively. Of the two commercial packages, the former provides independent, epoch-by-epoch network solutions, while the latter has the capability to process data using both PPP and Network approaches. There are clear advantages and disadvantages to the independent PPP solutions vs. the network based solutions. PPP solutions (smoothed and un-smoothed) provide position estimates for single stations in a global framework. They are not reliant on 'fixed' reference stations that, if located within the seismogenic zone, will clearly experience co-seismic displacements. Under best-case scenarios, error levels of 1-2 cm horizontal and 3-6cm vertical are achievable for network based solutions whereas PPP solutions with ambiguity resolution are currently somewhat noisier. Both analysis methodologies are critically dependent on data continuity and robust tracking of all available satellites at each site. A clear case is made for the requirement for network based processing capacity for baselines of greater than 150km in order to ensure robust co-seismic displacement estimates during large earthquakes. Results from tests within the WCDA network using the different techniques are presented as well as recommendations for computational and operational improvements. NRCan's real-time web based interface offers the user community with a real-time display of position offset estimates for sites in southwestern coastal British Columbia.