Time-Correlated GPS Noise Dependency on Data Time Period
GPS position time series contain time-correlated noise. The estimated parameters using correlated time series data, as station velocities, are then more uncertain than if the time series data were uncorrelated. If the level of the time-correlated noise is
- PDF / 365,147 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 7 Downloads / 168 Views
19
Alvaro Santamarı´a-Go´mez, Marie-Noe¨lle Bouin, Xavier Collilieux, and Guy Wo¨ppelmann
Abstract
GPS position time series contain time-correlated noise. The estimated parameters using correlated time series data, as station velocities, are then more uncertain than if the time series data were uncorrelated. If the level of the time-correlated noise is not taken into account, the estimated formal uncertainties will be smaller. By estimating the type and amplitude of the noise content in time series, more realistic formal uncertainties can be assessed. However, time-correlated noise amplitude is not constant in long time series, but depends on the time period of the time series data. Older time series data contain larger time-correlated noise amplitudes than newer time series data. This way, shorter time series with older data time period exhibit time-correlated noise amplitudes similar to the whole time series. This paper focuses on the source of the time-correlated noise amplitude decrease from older to newer time series period data. The results of several tested sources are presented. Neither the increasing ambiguity fixation rate, nor the increasing number of tracking stations, nor the increasing number of observed satellites are likely the source of the noise reduction. The quality improvement of the equipment of both tracking network and constellation is likely the main source of the correlated noise evolution. Keywords
GPS Time series Time-correlated noise
1 A. Santamarı´a-Go´mez (*) Instituto Geogra´fico Nacional, Cerro de la Palera s/n, E-19141 Yebes, Spain IGN/LAREG, Universite´ Paris Diderot, 35 rue He´le`ne Brion, 75013 Paris, France e-mail: [email protected] M.-N. Bouin Centre National de Recherches Me´te´orologiques, Me´te´o France, 13 rue du Chatellier, F-29228 Brest, France X. Collilieux IGN/LAREG, Universite´ Paris Diderot, 35 rue He´le`ne Brion, 75013 Paris, France G. Wo¨ppelmann Fe´de´ration de Recherche en Environnement pour le De´veloppement Durable (FR 3097 CNRS), Universite´ de La Rochelle/CNRS, 2 rue Olympe de Gouges, F-17000 La Rochelle, France
Introduction
Since more than a decade it is known that GPS station position time series contain time-correlated noise (Zhang et al. 1997; Mao et al. 1999). Noise parameters describing the type and the amplitude of the time-correlated noise content in position time series are usually estimated to reevaluate realistic uncertainties of the estimated parameters, such as station velocities (Williams 2003a; Wo¨ppelmann et al. 2009), offset amplitudes (Williams 2003b), periodic signals (Bos et al. 2010), etc. The analysis of the noise properties of the station position time series can also be used to get insights into the timecorrelated noise sources of the GPS technique. This way, noise analyses have been used to estimate the amplitude of the station monument noise (King and Williams 2009), to
Z. Altamimi and X. Collilieux (eds.), Reference Frames for Applications in Geosciences, International Association of Geodesy Symposia 138, DOI 10.100
Data Loading...
