This session considers all the full range of effects that may corrupt GNSS signals or induce undesirable contributions in GNSS analysis results. We focus particular attention on those effects that may have been neglected before, are of most serious magnitude, may be particularly insidious in obscuring genuine geophysical signals (such as aliasing), or degrade the stability of the reference frame.

Questions concerning the quality of IGS products and future improvements have a platform in this session, too. This includes the correlation of orbit modeling deficiencies with apparent geocenter variations or causes for persistent biases between AC products as well as further improvements of the IGS products, e.g. for applications such as precise point positioning. Of interest for users of IGS products are discussions on the procedures promising the most precise and least biased results in the vast field of applications.

It is likely that significant technique-related errors (e.g. caused by near-field and far-field multipath) are sometimes being misinterpreted as geophysical effects. Therefore, it is critical that the IGS take a stronger role in identifying the sources of GNSS errors and in finding strategies to mitigate their effects. This will require new research activities to better understand some suspected error sources. All aspects of GNSS geodesy are potentially involved, from field observations through data analysis and interpretation.

The quality and consistency of the IGS Analysis Center's contribution to the IGS combined products have steadily improved over the last years. The agreement among the Final/Rapid orbits is now at the 2 to 3 cm level or below for most of the ACs. The weekly network solutions have a consistency (stddev) of 1 to 2 mm and 5 to 6 mm for horizontal and height components, respectively. With this high level of precision smaller unmodeled or not correctly modeled effects as well as AC model or technological inconsistencies start to get visible. All those inconsistencies will have a measurable effect on the combined products, and therefore it is most important to reach the highest level of consistency in the near future, without influencing the progress at the individual ACs.

In the paper the basic IGS Products will be investigated for existing inconsistencies, especially between the orbit and the SINEX products.

Station positions obtained by Precise Point Positioning (PPP), using IGS combined orbits and clocks as well as those from individual ACs, are inspected for biases to the ITRF in realizing the IGS reference frame. The new modern mapping functions, based on numerical weather models, are evaluated for their use in PPP (station repeatability, ZPD), which give also some measure for the network processing.

The effects from various used ocean tide loading models on daily mean stations positions and geocenter are computed and compared.

When one fits a GPS spacecraft trajectory through several days of orbit positions from IGS Rapid orbit SP3 files, the orbit position residuals show discontinuities at the day boundaries between SP3 files.
The discontinuities can be of order 10 cm, especially the component in the along-track direction. The discontinuity component values for a specific spacecraft usually have the same sign for several months, that is, the variation from day to day is not random. SVN 44 (PRN 28) during the year 2005 showed an along-track discontinuity that slowly varied over the range +2 to +13 cm. IGS Final orbits show similar discontinuities at each 00 hr GPS. The biased residual discontinuities reflect a discontinuity in Rapid orbit systematic position error across day boundaries; this error is much larger than the 2 cm RMS difference between orbits from different Analysis Centers. To indicate the magnitude of systematic orbit error, the IGS should include day boundary discontinuity values in the Rapid orbit combination reports.

Precise GNSS orbits at the centimeter level are routinely generated by the IGS analysis centers using microwave phase measurements. For those GNSS satellites equipped with retroreflector arrays (i.e., all GLONASS and two of the GPS satellites) SLR observations are available in addition and well suited for validating the microwave based GNSS orbits. We present recent SLR validation results of GPS and GLONASS orbits derived from microwave phase observations. Four years (2002-2005) of SLR range residuals have been analyzed. The existing inter-technique biases of several centimeters are addressed. Dependencies of the range residuals on orbit modeling and on the observation scenario are discussed. Seasonal variations of the range residuals with maxima at eclipsing seasons indicate orbit modeling deficiencies for the GPS satellites.

In the near future, a large variety of new GNSS signals will become available. The most significant additions with respect to the current GPS signals are the introduction of two new frequency bands (E5/L5 and E6), the use of BOC or AltBOC modulations, and the fact that each carrier is modulated by at least two signal components, the so-called Pilot and Data components.

The wide offer of new signals implies that several signal processing and tracking options are available to the receiver developer. For instance, receiver A could track the Pilot component on a given carrier while receiver B tracks the Data component. The new RINEX format v3.0 takes this fact into account and defines different observable codes for all the different tracking options. Although the availability of several observable types per carrier will open up new fields of investigation, it is essential for the user to understand the difference and similarities between them. This is especially true in a heterogeneous multi-receiver network like the IGS, where all instrumental offsets must be minimized.

The paper begins with a description of all civilian Galileo and modernized GPS signal components. The operation principle of a combined GPS/Galileo receiver is described, focusing on the interoperability aspects between both constellations. The different tracking options are presented and compared, and related to the observable types as defined in RINEX v3.0.

The paper then presents the performances of the new signals with respect to tracking noise and multipath on code and phase. It is shown that all new signals significantly outperform the current GPS CA or L2C codes. Especially the unique AltBOC modulation on Galileo E5 is shown to bring unprecedented accuracy.

Finally, an overview of the current Galileo-related developments and activities within Septentrio is presented, with a particular attention to the tests currently underway with the GIOVE-A satellite.

Nearly all GPS sites display height variations with significant non-random power. Annual periods are most prominent, but other harmonics are also seen. (Annual signals are common in horizontal components, as well.) Sources of non-random temporal variations can be considered of two classes:
external, unmodeled geophysical effects that influence individual site motions (after accounting for known and tidal variations in the data analyis); and internal, unrecognized errors related to the GPS technique. There is mounting interest in using the observed non-linear GPS motions, either for individual stations or for networks, to better characterize certain geophysical processes, especially transport of fluid mass loads within the Earth system. Doing so reliably, however, requires that the technique errors be reasonably well understood and not too large.

The weekly global coordinate frames from the International GNSS Service (IGS) can be used to assess the significance of technique errors relative to geophysical effects. Many IGS sites show height variations that correlate closely with the data-quality metrics generated by the TEQC utility. The MP1 and MP2 measures of pseudorange multipath variability, for instance, often track the annual variations of heights. Abrupt changes in both height and TEQC metric variations are sometimes found to correlate with changes in station tracking hardware.

Rather than suggest that the TEQC metrics directly cause correlated position variations, or vice versa, it seems more likely that a common instrumental response may arise due to some underlying seasonal forcing. While there is a wide range of possible driving mechanisms, including mismodelings in the data analysis, an under-appreciated possibility is near-field multipath that could introduce slowly varying biases into the GPS position estimates. This hypothesis is strongly supported by the observed correlation of RMS dU variations with RMS clock jumps at day boundaries (for those IGS reference frame stations equipped with H-masers). The latter is a sensitive measure of local long-wavelength (i.e., near-field) pseudorange multipath conditions and is highly variable across the IGS network. The correlation with height variability indicates that near-field phase multipath (which is otherwise undetectable because it is well absorbed into the geodetic parameters) is probably widespread at IGS sites. This, in turn, points to basic problems in the design and installation of antenna mounts at most sites, which may give rise to artifactual signals that overwhelm large-scale geophysical loading signals. Any prospects for improved detectability in the future will require major infrastructure upgrades in the IGS reference frame network.

In order to improve the accuracy of high-rate (1-Hz) displacements for geophysical applications, it is necessary to reduce systematic errors such as multipath. Modified sidereal filtering (MSF), where time series of high-rate GPS time series are shifted by the satellite orbital repeat time, has been shown to significantly reduce multipath errors. This study investigates the frequencies and repetition of multipath in high-rate GPS time series in order to maximize the effectiveness of MSF. Specifically, we examine the interplay of GPS satellite orbital repeat periods (theoretical, average, and true repeat values) and multipath frequencies on the most effective time shift for MSF. Implementation strategies are discussed and shown for a group of 1-Hz GPS receivers operating in southern California. This technique is flexible and significantly reduces positioning noise at both short and long periods (20-1000 sec).

In this presentation we compare four different mapping functions which are used for mapping the atmospheric zenith delay to the direction of line-of-sight. Two of them are based on data from numerical weather models (NWM): the Isobaric Mapping Functions (IMF) are determined from the height of the 200 hPa pressure level for the hydrostatic part and from temperature and humidity for the wet part which can be easily extracted from NWM on a global grid, and the Vienna Mapping Functions 1 (VMF1) which are rigorously determined from the refractivity profiles downloaded with best resolution at selected sites. On the other hand, we compare them with the widely used Niell Mapping Functions (NMF) and the Global Mapping Functions (GMF): both of them depend only on the station coordinates and the day of year. The GMF are based on three years of global NWM data using spherical harmonics to be consistent with VMF1. We validate these four mapping functions with radiosonde data. Then results are shown from a global VLBI solution (1984 to 2005) with the software package OCCAM and from a global GPS network over a 12 month period (April 2004 to March 2005) with the GAMIT software package. The analyses reveal a very good agreement between the terrestrial reference frames determined with VMF1 and GMF, and between VMF1 and IMF in their ability to account for distinct weather conditions. Finally, we discuss these four mapping functions in terms of availability and applicability for certain purposes, e.g. for the determination of the terrestrial reference frame or for geophysical studies.