IGS Models for GNSS-Specific Effects           (updated 23 July 2007)
------------------------------------

* antenna models -- applied for both satellite transmit antennas and
ground tracking antennas (including radomes) to account for physical
locations of phase centers and for deviations from spherical wavefronts

  --> background info: IGS Mail #5189, 17 Aug 2005
      http://igscb.jpl.nasa.gov/mail/igsmail/2005/msg00111.html

  --> current model values: in antex format for all IGS antenna models
      ftp://igscb.jpl.nasa.gov/igscb/station/general/igs05.atx

  --> relation to other techniques: strictly limited to GNSS

  --> impact: strongly affects network scale, scale-rate, and deformation,
      as well as cm-level position of tracking sites, including local ties


* satellite orbit models -- dynamical effects of satellite motion are
handled variously by different groups and can be complex, esp for solar
radiation pressure effects and eclipsing events; situation will likely grow
more so if future models include albedo and similar effects

  --> Conventions: should conventional model and parameterization strategies
      be documented (probably difficult to do) ?

  --> relation to other techniques: similar situation for SLR, DORIS, and
      other satellite systems

  --> impact: to better inform users of products


* satellite attitude models -- to describe and parameterize the behavior
of satellite yaw variations

  --> current treatments: poorly documented

  --> relation to other techniques: situation probably similar for SLR,
      DORIS, and other satellite systems

  --> impact: can strongly influence estimated orbital parameters, esp
      near eclipses


* satellite polarization models -- variation in apparent rotation of
right-hand circular (RHC) polarization phase angle with changing relative
satellite-ground geometry follows treatment of Wu et al. (Manuscripta
Geodaetica,18, 91-98, 1993)                                     

  --> Conventions: should conventional model be documented ?

  --> relation to other techniques: strictly limited to GNSS

  --> impact: strongly affects carrier phase solutions, esp for 
      undifferenced processing


* ionospheric delay modeling -- lowest-order effect accounted for by
simple linear combination of dual-frequency observations

  --> Conventions: should conventional model be documented ?

  --> neglected higher-order effects: models currently under development
      and testing, but results not fully consistent

  --> Conventions: should higher-order effects be documented ?

  --> relation to other techniques: analogous situation for VLBI, DORIS,
      and any other radiometric systems

  --> impact: neglected effects cause small internal deformations in frame


* inter-modulation delay biases -- relative signal delays exist for
different modulations and frequencies in both the satellite transmit
and the ground receiver systems (also known as "differential code biases"
for P1-P2 biases)

  --> Conventions: should IGS methods and limitations to account for
      these biases be documented ?

  --> complications: the expected large growth in new GNSS signals will
      make this problem increasingly complex and difficult; also,
      increasingly different types of receiver models provide different
      sets of GNSS observables which are not necessarily consistent in a
      global network

  --> relation to other techniques: situation is analogous for VLBI
      receiver systems, in principle, but attempts to calibrate delay
      biases in VLBI are not normally attempted

  --> impact: neglect of these effects mostly affects clock estimates
      and limits accuracy of time transfers, but can also limit the
      ability to resolve carrier phase ambiguities correctly and is
      linked to ionospheric TEC measurements


* SP3 orbit frame -- by convention, SP3 orbits are given in a crust-fixed
(rotating) frame, which requires that subdaily motions in inertial space
be projected into tabulated satellite satellite positions for ease of use
for terrestrial positioning directly without the need for any inertial
transformation

  --> Conventions: should conventional models be better documented ?
      (this has been done already in the newly updated section on
      ocean tidal loading)

  --> tracking of model effects: current method to encode effects that
      have been modeled in IGS analyses is obscure and not well suited for
      future expansion; if Conventions are complete and consistent with
      IGS practice, then tracking in SP3 headers might not be so necessary

  --> relation to other techniques: situation could be similar for SLR,
      DORIS, and other satellite systems if they want to use SP3-type
      orbit files

  --> impact: strict analysis consistency is required by users to achieve
      full accuracy using IGS products (e.g., in PPP mode)


* relativistic effects for GPS clocks -- IGS follows GPS ICD-200 for
consistency, whereby 1st-order effects due to time dilation & gravitational
potential shifts have already been accounted for by offsets applied in the
oscillator settings aboard the spacecraft (assuming nominal orbital
elements), and 2nd-order effects due to non-circular orbits are handled in
analyses by applying a periodic time correction:  -2(R <dot> V)/c^2

  --> Conventions: should conventional ICD-200 model be documented ?

  --> neglected geopotential effects: show up as periodic and drift
      variations in observed satellite clocks (neglect of the Earth's
      oblateness (J2) causes drift of ~0.2 ns/day plus 6-hr & 14-day
      periodics according to J. Kouba, 2004)

  --> Conventions: should neglected effects be documented for users ?

  --> relation to other techniques: in principle, analogous effects
      could apply to other geodetic satellite systems if they have
      onboard clocks, but this does not seem relevant at present

  --> impact: for high-accuracy (sub-ns) applications IGS products
      could be misinterpreted, and extrapolation/interpolation of
      tabulated products is adversely affected