Real-time GNSS for geosciences: Initial assessment of new data products from GFZ

Global navigation satellite systems (GNSS) are used for various applications in the Earth and atmospheric sciences, navigation, surveying and mapping, as well as in early warning systems for geo-hazards. Solutions are often required to not only be of high accuracy, integrity, and continuity, but also to be available in real-time with a delay of only a few seconds. A prerequisite for real-time precise point positioning (PPP) are precise satellite orbit and clock products. While satellite orbits can be predicted with high precision, at least for a few hours, the satellite clocks have to be estimated using real-time GNSS data from a global network of reference stations and distributed via real-time data streams to the user. GFZ is operating a real-time GNSS analysis center, which is contributing to the Real-Time Service (RTS) of the International GNSS Service (IGS). In this contribution, we introduce the new GFZ in-house real-time GNSS network analysis software that is currently being developed and provide an initial assessment of the generated products.

In the first development stage that is presented in this contribution, the generated products contain satellite orbits and satellite clocks referring to the ionosphere-free code observations. The orbits are taken from the predicted part of the operational GFZ IGS ultra-rapid GPS, GLONASS, and Galileo solution, which are updated every three hours. The associated satellite clocks are estimated every five seconds using a recursive least-squares estimator from globally recorded real-time dual-frequency code and phase observations, together with receiver clock parameters, tropospheric zenith delay parameters, inter-system biases, and carrier-phase ambiguities.

Important aspects are the data cleaning to obtain high-quality results and an efficient implementation of the estimation filter to satisfy the delay requirements of the products – less than five seconds for the IGS. For the data cleaning and cycle-slip detection, the concept of single-receiver, single-channel integrity is used, in which the uniformly most powerful invariant test statistics are evaluated separately for each satellite-receiver link using its code and phase observations of two consecutive epochs. For the estimation filter, a sequential Kalman filter implementation using the standard covariance form is used. ‘Sequential’ refers to the strategy that the scalar observations of the same epoch are processed sequentially one at a time, leading to a more efficient operation of the filter compared to the case that the entire vector of measurements is processed at once. With this strategy, the processing time per epoch is around two seconds.

An initial evaluation of this real-time satellite orbit and clock product will be presented by means of a direct comparison to post-processed multi-GNSS reference products and a comparison of PPP analyses using in addition also broadcast navigation data and real-time products of other analysis centers.