Longterm Propagation of Geostationary Satellites

From: Gerhard HOLTKAMP (grd.holtkamp@t-online.de)
Date: Tue Jan 16 2007 - 15:43:35 EST

  • Next message: Kevin Fetter: "progress spacecraft launch on Dec 18"

    Recently when trying to calculate some GEO events a few months into the future 
    I chanced upon the following: Using SDP4 the GEO positions were offset a lot 
    further from their nominal positions as when using SGP4 (or a simple 
    Keplerian propagation which gave results quite similar to SGP4). Upon closer 
    investigation it turned out that ironically this meant that SDP4 was the 
    better propagator (as it should) for these orbits! 
    I used Elsets from August 2006 and November 2006 and propagated them to the 
    second week of January 2007 for which I could use recent Elsets. The 
    discrepancies were largest where according to the Handbook of Geostationary 
    Orbits the natural drift rates (due to unsymmetric terms in the geopotential) 
    are greatest (35 East, 116 East, 56 West, 155 West) while near the 
    equilibrium points (75 East, 162 East, 11 West, 105 West) SDP4 in general 
    stayed closer to the nominal positions than SGP4. In other words SDP4 
    represents the effects of the geopotential more accurately than SGP4. Of 
    course it doesn't know about the frequent longitude station keeping maneuvers 
    performed on operational satellites that keep them at their nominal positions 
    and that's why predictions a few months into the future will go wrong. 
    However when I checked a few retired orbits (for which no orbit maintenance 
    is performed anymore) the predictions of SDP4 over that period were much 
    better than the ones by SGP4.
    What it all means is that longterm propagations with SDP4 can be done only on 
    those satellites for which no maneuvers are performed anymore. Otherwise use 
    an artificial Elset for the satellites you are interested in with their 
    nominal positions at the epoch in question and redo the calculations with 
    actual Elsets a few days before the event happens to get more accurate 
    With regard to the Classified list however you should use SGP4 rather than 
    SDP4 as to the best of my knowledge those elements are generated with the SGP 
    model (correct me if I'm wrong).
    Gerhard HOLTKAMP
    Darmstadt, Germany
    Subscribe/Unsubscribe info, Frequently Asked Questions, SeeSat-L archive:  

    This archive was generated by hypermail 2b29 : Tue Jan 16 2007 - 16:22:29 EST