Re: Geodetic precision

From: Chris Olsson (
Date: Sat Aug 18 2001 - 18:16:10 PDT

  • Next message: Kevin Fetter: "RE: Satellite lunar transit"

    Bruno Tilger wrote:
    >I am surprised at the large difference (between Posns 1&3) in longitude of
    The longitudinal difference between WGS72 and WGS84 at Greenwich is 0.554",
    My typo in which I inadvertently typed "E" instead of W in Pos#3 is to blame
    and I apologise to the List if any members were genuinely confused by my typing
    Of course, both the WGS72 and WGS84 positions were in the same hemisphere. 
    WGS84's zero degree meridian (ignoring a little bit of tectonic plate movement
    since 1989) lies 102.478 metres East of the Prime Meridian at Greenwich.
    The difference between WGS72 and WGS84, although quite comprehensive in view of
    the fact that the size and shape of the two spheroids are rather different and
    the fact that the two datums have a rotational difference, as well as markedly
    different geoids, is a mere handful of metres in most places around the work. 
    The greatest 3-D difference is little more than half a dozen metres or so. 
    Converting between WGS 72&84 is very easy, so there is little scope for
    The big geodetic difference is when co-ordinates are misunderstood as to their
    geodetic basis.  For example, if a Cospar site in Scotland lists its position
    to a precision of a tenth of an arc-second of Lat/Long and another Cospar site
    in Sweden lists its position to four places of decimal degrees, then there is
    an easy trap for the unwary who might make the error of presuming that both
    co-ordinates are on some kind of common ground, such as WGS84 or WGS72.
    Note that neither the Scottish nor the Swedish position is wrong in any way. 
    In fact they are quite accurate and the basis from which they were derived is
    fundamentally sound.  The potential error is in the presumption that they have
    a common basis.  In fact: the Scottish point is listed with reference to the
    Airy spheroid and the OSGB36 + OD(Newlyn) datums; and the Swedish position is
    listed -- equally legitimately -- with reference to its national mapping datum
    of Sveriges Lantmäteriverket RT90 datum på Bessel's 1841 spheroid in the
    horizontal axis and the local RH70 datum in the vertical axis.  
    Each has a sound provenance and each is capable of being used with suitable
    precision in 3D calcs upon either WGS72 or WGS84, but the important thing to
    recognise is which geodetic basis is associated with any stated position.
    An analogy might be the European price of oil.  If I simply state that the
    current (epoch: close of business on Friday evening) Rotterdam market spot
    price of Brent Crude oil is 25.12345, then my stated price is capable of
    misinterpretation unless I declare whether I am referring to Dollars or Euros
    and whether or not I confirm that I'm referring to a barrel of crude or a tonne
    or a ton or a tun or a firkin or whatever.  
    Note too that the stated Edinburgh and Malma locations are not in any way wrong
    and that they are just as legitimate as a position which is expressed in one of
    the international standards such as ED50 or WGS84.  
    It is reasonable to presume that the Swede is using Degrees, not Gon, as his
    angular measure, just as it is reasonable to presume that the Scotsman is using
    arc-seconds rather than Groats or Drams as his.
    It might be reasonable to presume that both of those well-qualified observation
    locations are using a common spatial datum.  Such a presumption would be quite
    My thesis is simple.  I am suggesting that we either adopt a single common
    standard or else -- perhaps more practically -- declare what geodetic standard
    we are referring to when we list apparently precise geographical (or geodetic)
    co-ordinates.  The differences between the actual locations of stated positions
    may be small or great.  We can only know what those discrepancies (they are
    *not* errors) are if we have some indication as to what they refer to.  
    It is sufficient to simply say that a particular position is with reference to
    WGS84 or a particular national mapping datum or whatever the basis of the
    stated position is.
    By stating the geodetic basis of stated co-ordinates, we give a positional
    indication which is almost as important as stating the epoch of an elset or a
    SatObs.  It gives a pre- and post- calculable peg upon which other spatial
    calcs can be attached and assessed.
    Cheers,       Chris Olsson
    (In the Northern and Western hemispheres, well above sea level, wrt WGS84!)
    Unsubscribe from SeeSat-L by sending a message with 'unsubscribe'
    in the SUBJECT to

    This archive was generated by hypermail 2b29 : Sat Aug 18 2001 - 18:16:38 PDT