RE: TLE and SGP4 accuracies

From: Ted Molczan (
Date: Thu Feb 19 2004 - 12:02:46 EST

  • Next message: Thomas Fly: "Re: TLE and SGP4 accuracies"

    Katherine Kline wrote:
    > Does anyone know of any data on how often specific 
    > constellations are updated
    suggests the updating frequency is at least 23 hours:
    "The operational element set update is initiated based on the arrival of new
    observation data into the system. If the new observation residuals calculated
    with the current operational element set exceed pre-established tolerances, an
    update of the element set is initiated. Or, if it has been more than 23 hours
    since the last update, an update is initiated."
    The paper appears to be from 1999, so the procedure may have changed since then.
    Apart from asking USSTRATCOM, the only way to be certain would be to perform a
    study of historical elsets. 
    Registered users of the NASA/OIG web site may request historical elsets of
    specific unclassified objects, subject to limitations on frequency and volume of
    Some of the same data is available in the following compilation, but there are
    gaps depending on the source of the data:
    Note that in the above compilation, elsets of U.S. classified objects in orbit
    as of June 1983 are almost exclusively the product of hobbyist trackers, so they
    would not be relevant to your question.
    > and the accuracy of NORAD's models (i.e., do they also 
    > use special perturbation models for higher fidelity)
    Special perturbations models are available within USSTRATCOM, but it is my
    impression that they are used only when the SGP* models would be inadequate; for
    example, in collision-avoidance analysis.
    > Also, I am trying to quantify the accuracy of SGP4 models... 
    > Is it largely dependent on the atmospheric density model used, or are there 
    > also errors from the propagation?
    None of the SGP models employ atmospheric density models. They model decay as a
    constant value, as I discussed in this brief report:
    SGP4's analytic propagation probably is less accurate than numerical
    integration, but probably not materially, given its simplistic treatment of
    drag, and the omission of other perturbations, such as solar radiation pressure,
    which affects objects of high area to mass ratio, and solar gravity which
    affects sun-synchronous orbits. Also, SGP4's earth gravity model includes terms
    only up to J4. The SDP4 model includes luni-solar perturbations appropriate to
    high altitude orbits.
    Ted Molczan
    Subscribe/Unsubscribe info, Frequently Asked Questions, SeeSat-L archive:

    This archive was generated by hypermail 2b29 : Thu Feb 19 2004 - 12:08:48 EST