GNSS data processing in the National Multi-Hazard Early Warning System in the Sultanate of Oman

After the disastrous tsunami on December 26, 2004 in the Indian Ocean, a lot of effort has been undertaken to plan and realize tsunami early warning systems. As one of the first reactions, already in 2005 the German Indonesian Tsunami Early Warning System was initiated by the German government. Its establishment was coordinated by the German Research Centre for Geosciences Potsdam (GFZ) and it was successfully handed over to the Indonesian partner institutions in 2011.

In 2009, UNESCO signed an agreement with the Sultanate of Oman to establish a National Multi-Hazard Early Warning System (NMHEWS, http://www.unesco.org/new/en/unesco/partners-donors/the-actions/sciences/national-multi-hazard-early-warning-system-nmhews/). It is composed of several sensor networks, with seismic stations, GNSS stations, meteorological stations and tide gauges as the main components. The German company SpaceTech GmbH was commissioned to coordinate the setup of the GNSS sector, in cooperation with the GFZ and the Astronomical Institute of the University of Bern (AIUB). While SpaceTech carried out the system engineering, installation, training, initial operation and maintenance of the GNSS network with expertise contributed by GFZ, the AIUB provided the expertise for the GNSS data processing.

The GNSS network consists of 10 permanent stations, equipped with Septentrio PolaRx4 receivers and NavXperience 3G+C antennas providing real-time data. The locations of the GNSS stations in Oman:

Figure 1: Locations of the GNSS stations (red dots) of the NMHEWS. By courtesy of Luís Costa.

The purpose of the network is the mid- and long-term monitoring of the tectonic plate movement of the Arabian Peninsula which can be used for analyzing Earthquake potential and the near real-time monitoring of crustal deformation to cross validate the possibility of a tsunami after a nearby submarine Earthquake. The data is streamed to a processing facility at a tsunami warning center in Muscat, where, with a delay of less than 2 minutes, the displacement vectors of the stations are computed. Due to current limitations of the communication bandwidth, only GPS data is streamed and processed.

The data is processed in a batch mode with the Bernese GNSS Software (Version 5.2 – as distributed to the user community), using the ultra-rapid orbits and Earth rotation parameters of the Center for Orbit Determination in Europe (CODE). To bypass the need of accurate near real-time satellite clock corrections for a precise point positioning and to make use of ambiguity resolution, the processing is based on double-differenced GNSS data, where baselines are formed with an algorithm to maximize the number of common observations.

The data processing is split up into two main steps: a datum step and a near real-time step. In the datum step (repeated once per hour) a solution for the ten warning stations together with external IGS reference stations is computed to monitor the stability of the warning stations w.r.t. the global reference frame. For this, a defined amount of the most recent data is used from the warning and the reference stations and ambiguities are resolved in an extended procedure. In the near real-time step (repeated every two minutes) only data from the warning stations is processed. To speed up the preprocessing, ambiguities are resolved in a single strategy suitable for the present baseline lengths and by using again a certain number of hours of the most recent data. For the last three minutes of the data, kinematic coordinates are estimated for the warning stations introducing the resolved ambiguities and using a no-net-translation condition for each epoch w.r.t. the coordinates computed for the warning stations in the most recent datum step. The ambiguity resolution success rate, the characteristics of consecutive coordinate batch overlaps and the processing time served as quality measures to develop and optimize the processing routines.

A detailed description of the procedure and some initial results are available in:

Arnold, D., Lutz, S., Dach, R., Jäggi, A., Steinborn, J. (2016). Near real-time coordinate estimation from double-difference GNSS data. A case study for the National Multi-Hazard Early Warning System in the Sultanate of Oman. IAG Symposia Series (in press).