Offset Detection in Geodetic Coordinate Time Series
Chair: Simon Williams (England)
Terms of Reference
The accuracy and validity of geodetic positioning time series are often degraded by the presence of step discontinuities (offsets) that may either be known (e.g. documented equipment changes or earthquakes) or unknown, and with amplitudes that are, at best, known imprecisely. Undetected offsets can have an adverse effect on estimated velocities. Accurate velocities are required for many geophysical studies such as plate tectonics, intra-plate deformations, global reference frames and regional and global sea level. For example vertical land movements at tide gauges need to be obtained with a precision and accuracy of 0.1-0.2 mm/yr for sea level change studies. As the length of time series continue to increase the number of offsets is likely also to increase and the cumulative effect of even the smallest of offsets can seriously alter our velocity estimates. This, coupled with the huge growth in the number of sites, particularly GNSS, necessitates the automation of site velocity estimation and therefore offset detection. Offset detection is an issue in many different scientific studies, where it is often called data segmentation or homogenization, such as climate/meteorology, economics, image processing and bio-statistics. However what works in one discipline may not be suitable in another. The aim of this working group is to encourage cooperation between different groups in the geodetic community to contribute, investigate and disseminate different offset detection methods and provide a realistic benchmark dataset(s) on which to test their efficacy.
- Encourage cooperation between different groups in the geodetic community to contribute, investigate and disseminate different offset detection methods;
- Identify and provide offset detection methods for the use of the community, including code;
- Provide a realistic benchmark dataset(s) on which to test their efficacy (successor to the DOGEx);
- Evaluate the validity of alternative velocity estimation methods that may be less biased by undetected offsets particularly in the context of a DOGEx follow on;
- Provide guidelines and advice on offset detection in geodetic coordinate time series;
- Foster and establish interactions with other areas of science for which offset detection is also an issue to identify different approaches to the problem.
- Simon Williams (UK), Chair
- Machiel Bos (Portugal)
- Norman Teferle (Luxembourg)
- Matt King (Australia)
- Xavier Collilieux (France)
- Jarir Saleh (USA)