GNSS related periodic signals in coordinate time-series from Precise Point Positioning

In Global Navigation Satellite System (GNSS) coordinate time series unrecognised errors and
un-modelled (periodic) effects may bias non-linear motions induced by geophysical signals.
Hence, understanding and mitigating these errors is vital to reducing biases and on revealing
subtle geophysical signals. To assess the nature of periodic signals in coordinate time series
Precise Point Positioning (PPP) solutions for the period 2008 to 2015 are generated. The solu-
tions consider Global Positioning System (GPS), GLObalnaya NAvigatsionnaya Sputnikovaya
Sistema (GLONASS) or combined GPS+GLONASS (GNSS) observations. We assess the pe-
riodic signals of station coordinates computed using the combined International GNSS Service
(IGS) and four of its Analysis Centers (ACs) products. Furthermore, we make use of different
filtering methods to investigate the sources of the periodic signals. A faint fortnightly signal
in our PPP solution based on Jet Propulsion Laboratory (JPL) products and the existence of
an 8-day period for those ACs generating combined GPS+GLONASS products are the main
features in the GPS-only solutions. The existence of the 8-day period in the GPS-only solution
indicates that GPS orbits computed in a combined GNSS solution contain GLONASS-specific
signals. The GLONASS-only solution shows highly elevated powers at the 3rd draconitic harmonic ( ~ 120-day period), at the 8-day period and its harmonics (4 days, 2.67 days) besides the
well-known annual, semi-annual and other draconitic harmonics. We show that the GLONASS
constellation gaps before December 2011 contribute to the power at some of the frequencies.
However, the well known fortnightly signal in GPS-only solutions is not discernible in the
GLONASS-only solution. The combined GNSS solution contains periodic signals from both
systems, with most of the powers being reduced when compared to the single-GNSS solutions.
A 52% reduction for the horizontal components and a 36% reduction for the vertical compo-
nent are achieved for the fortnightly signal from the GNSS solution compared to the GPS-only
solution. Comparing the results of the employed filtering methods reveals that the source of
most of the powers of draconitic and fortnightly signals are satellite-induced with a non-zero
contribution of site-specific errors.