Worst case datum error?

Discussion in 'Global Navigation Satellite Systems' started by Thomas Magma, Feb 28, 2005.

  1. Thomas Magma

    Thomas Magma Guest

    Does anyone know what the worst case position error would be when flipping
    between two different datums? I heard it could be up to several kilometres
    in certain regions around the world. Does anyone know of any reference or
    website that might have this information?

    Thanks
    Thomas
     
    Thomas Magma, Feb 28, 2005
    #1
    1. Advertisements

  2. Thomas Magma

    Sam Wormley Guest

    Thomas Magma wrote:
    > Does anyone know what the worst case position error would be when flipping
    > between two different datums? I heard it could be up to several kilometres
    > in certain regions around the world. Does anyone know of any reference or
    > website that might have this information?
    >
    > Thanks
    > Thomas
    >
    >


    The *error* translation from WGS-84 to another datum is typically less
    than two meters. If you chose datum A to use with software or maps
    using datum B, the "error" due to incompatible datums can be hundred
    of meters.

    -Sam Wormley
    http://edu-observatory.org/maps/waypoints.html
    Note datum links
     
    Sam Wormley, Feb 28, 2005
    #2
    1. Advertisements

  3. Thomas Magma

    Graham W Guest

    Thomas Magma wrote:
    > Does anyone know what the worst case position error would be when flipping
    > between two different datums? I heard it could be up to several kilometres
    > in certain regions around the world. Does anyone know of any reference or
    > website that might have this information?


    I assume you mean the largest distance between identical coordinates in
    different datums. I believe the older Japanese datum differs from WGS84
    by up to 2 km.
     
    Graham W, Feb 28, 2005
    #3
  4. Thomas Magma

    Thomas Magma Guest

    Thanks Graham,

    That is what I meant. Like if two people were told to go to a certain
    Lat/Long coordinate but used two completely different datums. How far apart
    could these two people be in the worst case scenario.

    Do you know where you acquired the info on how the "older Japanese datum
    differs from WGS84"?

    Thomas

    "Graham W" <> wrote in message
    news:...
    > Thomas Magma wrote:
    > > Does anyone know what the worst case position error would be when

    flipping
    > > between two different datums? I heard it could be up to several

    kilometres
    > > in certain regions around the world. Does anyone know of any reference

    or
    > > website that might have this information?

    >
    > I assume you mean the largest distance between identical coordinates in
    > different datums. I believe the older Japanese datum differs from WGS84
    > by up to 2 km.
     
    Thomas Magma, Feb 28, 2005
    #4
  5. Thomas Magma

    Don Robinson Guest

    Don Robinson, Mar 1, 2005
    #5
  6. Thomas Magma

    Stichting ST Guest

    On Tue, 01 Mar 2005 09:23:41 +1100, Graham W <>
    wrote:

    >Thomas Magma wrote:
    >> Does anyone know what the worst case position error would be when flipping
    >> between two different datums? I heard it could be up to several kilometres
    >> in certain regions around the world. Does anyone know of any reference or
    >> website that might have this information?

    >
    >I assume you mean the largest distance between identical coordinates in
    >different datums. I believe the older Japanese datum differs from WGS84
    >by up to 2 km.


    All datums made after say 1950 do not differ very much between each other.
    Some tenth of meters.
    Older mapdatums say Swiss CH1903 differs about 200 meter and yes indeed old
    Tokio is very far off with about 500 meters.
    If you have a program like Ozi Explorer you can experiment with many different
    mapdatums. Recalibarte and see where your waypoint moves to.
    For a map showing the differences: see this (german) link with picture;
    http://www.kowoma.de/gps/geo/mapdatum.htm

    Piet
     
    Stichting ST, Mar 1, 2005
    #6
  7. Thomas Magma wrote:
    > Thanks Graham,
    >
    > That is what I meant. Like if two people were told to go to a certain
    > Lat/Long coordinate but used two completely different datums. How far
    > apart could these two people be in the worst case scenario.
    >
    > Do you know where you acquired the info on how the "older Japanese
    > datum differs from WGS84"?
    >
    > Thomas
    >
    > "Graham W" <> wrote in message
    > news:...
    >> Thomas Magma wrote:
    >>> Does anyone know what the worst case position error would be when
    >>> flipping between two different datums? I heard it could be up to
    >>> several kilometres in certain regions around the world. Does anyone
    >>> know of any reference or website that might have this information?

    >>
    >> I assume you mean the largest distance between identical coordinates
    >> in different datums. I believe the older Japanese datum differs from
    >> WGS84 by up to 2 km.


    If you have a GPS you can find this out by simply standing on one spot and
    changing the datums, noting down the Lat/long for each datum.


    --
    Satellite photocharts of the UK & Ireland
    available, excellent detail and accurate
    calibration using Oziexplorer.
    Remove *nospam* to reply.
     
    Dennis Pogson, Mar 1, 2005
    #7
  8. Glenn Proctor, Mar 1, 2005
    #8
  9. Thomas Magma

    Graham W Guest

    Thomas Magma wrote:

    > Thanks Graham,
    >
    > That is what I meant. Like if two people were told to go to a certain
    > Lat/Long coordinate but used two completely different datums. How far apart
    > could these two people be in the worst case scenario.
    >
    > Do you know where you acquired the info on how the "older Japanese datum
    > differs from WGS84"?


    A BBC broadcast several years ago, confirmed by later reading...

    Not very helpful, sorry.
     
    Graham W, Mar 1, 2005
    #9
  10. Thomas Magma

    Graham W Guest

    Stichting ST wrote:
    > All datums made after say 1950 do not differ very much between each other.
    > Some tenth of meters.


    AGD66 and AGD72 differ from AGD94 and WGS84 by about 200 metres
     
    Graham W, Mar 1, 2005
    #10
  11. Thomas Magma

    Graham W Guest

    > Stichting ST wrote:
    >> All datums made after say 1950 do not differ very much between each
    >> other.
    >> Some tenth of meters.


    Graham W wrote:
    > AGD66 and AGD72 differ from AGD94 and WGS84 by about 200 metres


    GDA94...
     
    Graham W, Mar 1, 2005
    #11
  12. Thomas Magma

    Gary S. Guest

    On Mon, 28 Feb 2005 22:11:49 GMT, Sam Wormley <>
    wrote:

    >Thomas Magma wrote:
    >> Does anyone know what the worst case position error would be when flipping
    >> between two different datums? I heard it could be up to several kilometres
    >> in certain regions around the world. Does anyone know of any reference or
    >> website that might have this information?
    >>

    > The *error* translation from WGS-84 to another datum is typically less
    > than two meters. If you chose datum A to use with software or maps
    > using datum B, the "error" due to incompatible datums can be hundred
    > of meters.
    >
    > -Sam Wormley
    > http://edu-observatory.org/maps/waypoints.html
    > Note datum links
    >

    Correct me if I am wrong, but why would there be a constant
    discrepancy between any two systems?

    There could well be locations where they are nearly indistinguishable,
    and others where they are significantly off. Newer ones I would agree
    are going to be more similar, but a lot of basic surveying is based on
    very old datums and coordinate systems.

    Happy trails,
    Gary (net.yogi.bear)
    --
    At the 51st percentile of ursine intelligence

    Gary D. Schwartz, Needham, MA, USA
    Please reply to: garyDOTschwartzATpoboxDOTcom
     
    Gary S., Mar 1, 2005
    #12
  13. On Mon, 28 Feb 2005 22:11:49 GMT, Sam Wormley
    <> wrote:

    >Thomas Magma wrote:
    >> Does anyone know what the worst case position error would be when flipping
    >> between two different datums? I heard it could be up to several kilometres
    >> in certain regions around the world. Does anyone know of any reference or
    >> website that might have this information?
    >>
    >> Thanks
    >> Thomas
    >>
    >>

    >
    > The *error* translation from WGS-84 to another datum is typically less
    > than two meters. If you chose datum A to use with software or maps
    > using datum B, the "error" due to incompatible datums can be hundred
    > of meters.
    >
    > -Sam Wormley
    > http://edu-observatory.org/maps/waypoints.html
    > Note datum links
    >


    That is not true at all. It depends on the transformation
    method and the local datum. For example, using the DoD/NIMA
    transformation (the one built into the receivers) between
    WGS84 and NAD27 can mean 10-20 metres of error in Canada.
    The transformation to SAD69 is 50 metres out in NE Brasil.

    As to the original question, I did such an analysis for a
    GPS coursebook I developed in '99 for the government. The
    difference between WGS84 and other datums can be over 1000
    metres (mostly for Asian datums).

    WGS84 (and NAD83, and most country datums since 1980 when we
    began using VLBI, Doppler, and then GPS for long-range
    measurements) are geocentric datums - that is, they are
    "anchored" to the geophysical centre of the earth (defined
    at a certain epoch, for example NAD83 uses the BIH84
    definition of the earth), and oriented to North and a
    specific meridian. Previous datums were typically chosen as
    a "best-fit" to (sometimes very limited and innacurate)
    existing survey measurements, typically triangulation
    networks. The precise differences between coordinates on
    these datums and modern datums is very poorly known and
    typically very inconsistent over large areas.

    Sorry to be so pedantic.

    steve
     
    steve robertson, Mar 1, 2005
    #13
  14. Thomas Magma

    Thomas Magma Guest

    Thanks Steve,

    Sounds like you have already done some work in this field.

    I agree that the amount of error between two datums will depend on where you
    are on the globe.

    I thought the answer to the original question would have been better
    researched and documented. Would it not just take a computer program to
    generate all possible datum errors worldwide based on all the different
    representations of the sphere, spheroid and ellipsoid? I thought someone
    would have already have done this years ago.

    Thomas

    "steve robertson" <> wrote in message
    news:...
    > On Mon, 28 Feb 2005 22:11:49 GMT, Sam Wormley
    > <> wrote:
    >
    > >Thomas Magma wrote:
    > >> Does anyone know what the worst case position error would be when

    flipping
    > >> between two different datums? I heard it could be up to several

    kilometres
    > >> in certain regions around the world. Does anyone know of any reference

    or
    > >> website that might have this information?
    > >>
    > >> Thanks
    > >> Thomas
    > >>
    > >>

    > >
    > > The *error* translation from WGS-84 to another datum is typically less
    > > than two meters. If you chose datum A to use with software or maps
    > > using datum B, the "error" due to incompatible datums can be hundred
    > > of meters.
    > >
    > > -Sam Wormley
    > > http://edu-observatory.org/maps/waypoints.html
    > > Note datum links
    > >

    >
    > That is not true at all. It depends on the transformation
    > method and the local datum. For example, using the DoD/NIMA
    > transformation (the one built into the receivers) between
    > WGS84 and NAD27 can mean 10-20 metres of error in Canada.
    > The transformation to SAD69 is 50 metres out in NE Brasil.
    >
    > As to the original question, I did such an analysis for a
    > GPS coursebook I developed in '99 for the government. The
    > difference between WGS84 and other datums can be over 1000
    > metres (mostly for Asian datums).
    >
    > WGS84 (and NAD83, and most country datums since 1980 when we
    > began using VLBI, Doppler, and then GPS for long-range
    > measurements) are geocentric datums - that is, they are
    > "anchored" to the geophysical centre of the earth (defined
    > at a certain epoch, for example NAD83 uses the BIH84
    > definition of the earth), and oriented to North and a
    > specific meridian. Previous datums were typically chosen as
    > a "best-fit" to (sometimes very limited and innacurate)
    > existing survey measurements, typically triangulation
    > networks. The precise differences between coordinates on
    > these datums and modern datums is very poorly known and
    > typically very inconsistent over large areas.
    >
    > Sorry to be so pedantic.
    >
    > steve
     
    Thomas Magma, Mar 1, 2005
    #14
  15. Thomas Magma

    jeff feehan Guest

    Stichting ST wrote:
    > On Tue, 01 Mar 2005 09:23:41 +1100, Graham W <>
    > wrote:
    >
    >
    >>Thomas Magma wrote:
    >>
    >>>Does anyone know what the worst case position error would be when flipping
    >>>between two different datums? I heard it could be up to several kilometres
    >>>in certain regions around the world. Does anyone know of any reference or
    >>>website that might have this information?

    >>
    >>I assume you mean the largest distance between identical coordinates in
    >>different datums. I believe the older Japanese datum differs from WGS84
    >>by up to 2 km.

    >
    >
    > All datums made after say 1950 do not differ very much between each other.
    > Some tenth of meters.
    > Older mapdatums say Swiss CH1903 differs about 200 meter and yes indeed old
    > Tokio is very far off with about 500 meters.
    > If you have a program like Ozi Explorer you can experiment with many different
    > mapdatums. Recalibarte and see where your waypoint moves to.


    i have done this, and my waypoints don't move at all, no matter which datum
    i select. i figured something was wrong, or that maybe once a datum was selected,
    it can't be changed without starting from scratch.

    jeff


    > For a map showing the differences: see this (german) link with picture;
    > http://www.kowoma.de/gps/geo/mapdatum.htm
    >
    > Piet
     
    jeff feehan, Mar 9, 2005
    #15
  16. Thomas Magma

    Stichting ST Guest

    On Wed, 09 Mar 2005 08:30:03 -0500, jeff feehan <> wrote:

    >Stichting ST wrote:
    >> On Tue, 01 Mar 2005 09:23:41 +1100, Graham W <>
    >> wrote:
    >>
    >>
    >>>Thomas Magma wrote:
    >>>
    >>>>Does anyone know what the worst case position error would be when flipping
    >>>>between two different datums? I heard it could be up to several kilometres
    >>>>in certain regions around the world. Does anyone know of any reference or
    >>>>website that might have this information?
    >>>
    >>>I assume you mean the largest distance between identical coordinates in
    >>>different datums. I believe the older Japanese datum differs from WGS84
    >>>by up to 2 km.

    >>
    >>
    >> All datums made after say 1950 do not differ very much between each other.
    >> Some tenth of meters.
    >> Older mapdatums say Swiss CH1903 differs about 200 meter and yes indeed old
    >> Tokio is very far off with about 500 meters.
    >> If you have a program like Ozi Explorer you can experiment with many different
    >> mapdatums. Recalibarte and see where your waypoint moves to.

    >
    >i have done this, and my waypoints don't move at all, no matter which datum
    >i select. i figured something was wrong, or that maybe once a datum was selected,
    >it can't be changed without starting from scratch.
    >
    >jeff


    Maybe I was not clear in my saying.
    In Ozi you should not do a 'display different mapdatum' but do
    (menu:file)'check and recalibrate map.'
    Give the map itself a new datum.
    Things maybe further complicated by the fact that waypoints imported from
    files are normally considered WGS84 and wil be converted to the new mapdatum
    on import.
    What you have to do is: write down your coordinates (in WGS 84) on paper and
    then type them in in 'add to waypoint list' and see that (down below in the
    add-box) the mapdatum is what you want. Ozi will think you are entering
    degrees in that mapdatum.

    Piet
     
    Stichting ST, Mar 9, 2005
    #16
  17. Thomas Magma

    jeff feehan Guest

    Stichting ST wrote:
    > On Wed, 09 Mar 2005 08:30:03 -0500, jeff feehan <> wrote:
    >
    >
    >>Stichting ST wrote:
    >>
    >>>On Tue, 01 Mar 2005 09:23:41 +1100, Graham W <>
    >>>wrote:
    >>>
    >>>
    >>>
    >>>>Thomas Magma wrote:
    >>>>
    >>>>
    >>>>>Does anyone know what the worst case position error would be when flipping
    >>>>>between two different datums? I heard it could be up to several kilometres
    >>>>>in certain regions around the world. Does anyone know of any reference or
    >>>>>website that might have this information?
    >>>>
    >>>>I assume you mean the largest distance between identical coordinates in
    >>>>different datums. I believe the older Japanese datum differs from WGS84
    >>>>by up to 2 km.
    >>>
    >>>
    >>>All datums made after say 1950 do not differ very much between each other.
    >>>Some tenth of meters.
    >>>Older mapdatums say Swiss CH1903 differs about 200 meter and yes indeed old
    >>>Tokio is very far off with about 500 meters.
    >>>If you have a program like Ozi Explorer you can experiment with many different
    >>>mapdatums. Recalibarte and see where your waypoint moves to.

    >>
    >>i have done this, and my waypoints don't move at all, no matter which datum
    >>i select. i figured something was wrong, or that maybe once a datum was selected,
    >>it can't be changed without starting from scratch.
    >>
    >>jeff

    >
    >
    > Maybe I was not clear in my saying.
    > In Ozi you should not do a 'display different mapdatum' but do
    > (menu:file)'check and recalibrate map.'
    > Give the map itself a new datum.


    that is what i was doing.



    > Things maybe further complicated by the fact that waypoints imported from
    > files are normally considered WGS84 and wil be converted to the new mapdatum
    > on import.


    that could explain my observation.

    > What you have to do is: write down your coordinates (in WGS 84) on paper and
    > then type them in in 'add to waypoint list' and see that (down below in the
    > add-box) the mapdatum is what you want. Ozi will think you are entering
    > degrees in that mapdatum.


    thanks. i'll try it.

    jeff
    >
    > Piet
    >
     
    jeff feehan, Mar 9, 2005
    #17
    1. Advertisements

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Mike
    Replies:
    6
    Views:
    369
  2. Chirisophus

    Datum error in lat/long & UTM ??

    Chirisophus, Sep 15, 2003, in forum: General GPS
    Replies:
    15
    Views:
    652
    Una Smith
    Oct 1, 2003
  3. Bad Idea

    Worst GPSR prices

    Bad Idea, Jul 24, 2004, in forum: General GPS
    Replies:
    4
    Views:
    333
    GSV Three Minds in a Can
    Jul 27, 2004
  4. Chirisophus

    Datum error in lat/long & UTM ??

    Chirisophus, Sep 15, 2003, in forum: Magellan
    Replies:
    15
    Views:
    1,009
    Una Smith
    Oct 1, 2003
  5. Chirisophus

    Datum error in lat/long & UTM ??

    Chirisophus, Sep 15, 2003, in forum: Global Navigation Satellite Systems
    Replies:
    15
    Views:
    695
    Una Smith
    Oct 1, 2003
Loading...

Share This Page