Geodetic precision

From: Chris Olsson (
Date: Thu Aug 16 2001 - 11:18:33 PDT

  • Next message: Michael McCants: "Re: Ed Cannon's UNID probably is USA 159 r"

    As a newcomer to the List, from the background of being a Geodesist, I've been
    wondering about the apparent precision of some List members' declared
    co-ordinates and to what geodetic framework they refer.
    Let's take three specific positions, in SouthEastern London, England, as a
    working example to illustrate what I'm trying to say:
    Position #1:
    51 28' 40.125" N   0 00' 05.310" W    Height 92.603m
    Position #2:
    51 28' 38.265" N   0 00' 00.418" E    Height 47.155m
    Position #3:
    51 28' 40.028" N   0 00' 05.864" E    Height 89.563m
    At first glance, it might appear that the three positions each describe three
    different places, with Pos#1 being 124.6m from Posn#2.  The two places do not
    even appear to be in the same hemisphere!
    In fact, all three positions actually describe a single point in space, but
    only if I mention which spheroid and datum they refer to.
    Position #1 is expressed with reference to WGS84.  
    Position #2 is expressed with reference to the Airy spheroid and Britain's
    national mapping datum OSGB36 in the horizontal plane and the height refers to
    Britain's Ordnance Survey datum which is actually referred to the mean sea
    level in 1921 at a small fishing town in the South West of England called
    Position #3 is with respect to WGS72.
    Note that the actual physical location of these three positions is in fact, by
    definition, exactly located on the 0 Prime Meridian at Greenwich. All three
    sets of co-ordinates describe the location of Airy's Transit, which was adopted
    by International Convention in 1884 to be the zero degree meridian for
    astronomical and navigational purposes.
    Without saying which horizontal and vertical datum(s) a particular reported
    position refers to, the stated position is quite useless for any but the most
    vague description of geographical location.
    Might I respectfully suggest to members of SeeSat-L that whenever they list
    their observational location co-ordinates, they also mention to which geodetic
    datum those co-ordinates refer?
    It may be impressive to list position to an apparent precision of five decimals
    of a degree or to a tenth of an arc-second of Lat/Long, but unless the geodetic
    datum to which those co-ords are referred is associated with such apparent
    precision, then any accuracy which might be associated with such co-ords is
    quite wasted.
    The same applies to vertical datums.  Spheroidal height and ellipsoidal height
    are the same thing, but they are quite different to Geoidal height and are not
    necessarily co-incident with "sea level".  In any case, it is essential to
    indicate exactly what a vertical "height" refers to. That might be "sea level"
    or it might be an ellipsoidal height.  Some GPS units, for example,  output
    height in one form, others in another.
    There are several very high quality software packages out there which appear to
    be able to list the orbital height of any satellite (given good enough elsets)
    to an apparent precision of a metre.   Dr TS Kelso's TrakStar, for example,
    uses WGS72 rather than WGS84 as a geoidal model and geodetic framework,
    apparently because the TLEs which most users have access to are still based
    upon that otherwise obsolete reference frame.  
    Of course for most casual observers of satellites, using binoculars and a
    stopwatch, a positional discrepancy in an observational location of a hundred
    metres or so is of no practical consequence whatsoever.  It is only when high
    accuracy astrometry is used, such as is obtainable by CCD  cameras on good
    telecopes, that the datum becomes a significant factor.  Nevertheless, I would
    suggest that anyone who posts their geographical position of observation to
    high levels of precision (i.e. 10 to 20 metres or so) might usefully mention
    which spheroid and datum those co-ordinates refer to.
    Cheers,       Chris Olsson
    57 02' 30.9" N   3 10' 25.9" W   Ht=314m
    wrt WGS84!
    Unsubscribe from SeeSat-L by sending a message with 'unsubscribe'
    in the SUBJECT to

    This archive was generated by hypermail 2b29 : Thu Aug 16 2001 - 11:19:11 PDT