 |
| Titel |
Impacts of GNSS position offsets on global frame stability |
| VerfasserIn |
Jake Griffiths, Jim Ray |
| Konferenz |
EGU General Assembly 2015
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 17 (2015) |
| Datensatznummer |
250114572
|
| Publikation (Nr.) |
 EGU/EGU2015-15362.pdf |
|
|
|
|
|
| Zusammenfassung |
| Positional offsets appear in Global Navigation Satellite System (GNSS) time series for a
variety of reasons. Antenna or radome changes are the most common cause for
these discontinuities. Many others are from earthquakes, receiver changes, and
different anthropogenic modifications at or near the stations. Some jumps appear for
unknown or undocumented reasons. Accurate determination of station velocities,
and therefore geophysical parameters and terrestrial reference frames, requires
that positional offsets be correctly found and compensated. Williams (2003) found
that undetected offsets introduce a random walk error component in individual
station time series. The topic of detecting positional offsets has received considerable
attention in recent years (e.g., Detection of Offsets in GPS Experiment; DOGEx), and
most research groups using GNSS have adopted a mix of manual and automated
methods for finding them. The removal of a positional offset from a time series is
usually handled by estimating the average station position on both sides of the
discontinuity. Except for large earthquake events, the velocity is usually assumed
constant and continuous across the positional jump. This approach is sufficient in the
absence of time-correlated errors. However, GNSS time series contain periodic
and power-law (flicker) errors. In this paper, we evaluate the impact to individual
station results and the overall stability of the global reference frame from adding
increasing numbers of positional discontinuities. We use the International GNSS Service
(IGS) weekly SINEX files, and iteratively insert positional offset parameters. Each
iteration includes a restacking of the modified SINEX files using the CATREF
software from Institut National de l’Information Géographique et Forestière (IGN).
Comparisons of successive stacked solutions are used to assess the impacts on the
time series of x-pole and y-pole offsets, along with changes in regularized position
and secular velocity for stations with more than 2.5 years of data. Our preliminary
results indicate that the change in polar motion scatter is logarithmic with increasing
numbers of discontinuities. The best-fit natural logarithm to the changes in scatter
for x-pole has R2 = 0.58; the fit for the y-pole series has R2 = 0.99. From these
empirical functions, we find that polar motion scatterÂincreases from zero when
the total rate of discontinuities exceeds 0.2 (x-pole) and 1.3 (y-pole) per station,
on average (the IGS has 0.65 per station). Thus, the presence of position offsets
in GNSS station time series is likely already a contributor to IGS polar motion
inaccuracy and global frame instability. Impacts to station position and velocity
estimates depend on noise features found in that station’s positional time series. For
instance, larger changes in velocity occur for stations with shorter and noisier data
spans. This is because an added discontinuity parameter for an individual station
time series can induce changes in average position on both sides of the break. We
will expand on these results, and consider remaining questions about the role of
velocity discontinuities and the effects caused by non-core reference frame stations. |
|
|
|
|
|
|