What I find a bit disconcerting with GPS measurements is the question of datums. Different measurements may be relative to different dataums. For instance on Google Maps, the Greenwich meridian and 0 degrees longitude do not coincide. Has any one found cases of confusion and inaccuaracy in practice, such as not correctly locating a grave? Peter

"Peter J Seymour" <moz@pjsey.demon.co.uk> wrote in message news:fk04pr$4n9$1$8300dec7@news.demon.co.uk... > What I find a bit disconcerting with GPS measurements is the question of > datums. Different measurements may be relative to different dataums. For > instance on Google Maps, the Greenwich meridian and 0 degrees longitude > do not coincide. Has any one found cases of confusion and inaccuaracy in > practice, such as not correctly locating a grave? Welcome to the wonderful world of geodesy. Different datums do not exist just for fun. They came about because the original geodetic surveys that established the shape of the earth were not connected to each other. Thus, each country (continent) would use which ever spheroid (datum) best fit their local situation. Only when the shape of earth could be measured using satellites were we able to develop spheroidal models that provided a best fit for the entire earth, rather than just for a particular region. ITMT, all those legal property surveys were done and maps published using the existing local datums. So, countries couldn't just toss out the old datum when a better one came along. This means that the latitude and longitude of an object measured in the local datum (e.g. Tokyo, Gauss) can be as much as SEVERAL HUNDRED METERS removed from the same position described in a geocentric datum such as WGS84. For this reason surveyor will ALWAYS include information about the datum used on any survey he performs. ALL accurate maps (topographic, navigation, etc) have information about the datums and projections used. There is no way around this until the entire world adopts a single datum and converts all the old maps, plats, etc to that datum. That's the bad news. The good news is that the North American Datum (NAD27) that has been in use for US maps since 1927 is pretty close to the WGS84 (World Geodetic System 1984) datum used by GPS at least over most of North America. The best policy for recording the position of anything located by GPS is to always also record the datum used. You also need to understand that nothing is ever positioned exactly, never, ever. There is always an error, although in some cases the error can be made very small. When using GPS you should also record the error. (e.g. +-25m). -- Donald R. Newcomb DRNewcomb (at) attglobal (dot) net

Wasn't it Peter J Seymour who wrote: >What I find a bit disconcerting with GPS measurements is the question >of datums. Different measurements may be relative to different dataums. >For instance on Google Maps, the Greenwich meridian and 0 degrees >longitude do not coincide. Has any one found cases of confusion and >inaccuaracy in practice, such as not correctly locating a grave? You're very unlikely to find anything other than the WGS 84 datum being used in any recent software. The GPS satellites themselves use WGS 84, and so do things like Google Earth and Google Maps. (Although Google did use the ancient Tokyo Datum for locations in Japan when they first started those services). [There have been slight improvements in things like the WGS 84 ellipsoid and its Earth Gravity Model since WGS 84 came out, but they only make a difference of a few millimetres to locations on the Earth's surface. They refinements are mainly useful for extremely accurate positioning of satellites.] Before 1960, it wasn't particularly easy to obtain lat/lng coordinates accurate to within 100 feet, so moving the 0 degree line away from the historical Prime Meridian with WGS 60 didn't cause many problems. The differences between WGS 60, WGS 66, WGS 72 and WGS 84 changed locations by only a few feet in the worst case, so finding a grave that was marked using one of those datums would be affected more by the GPS error than by the difference in datum. Before 1 May 2000, the non-military GPS signals contained a random error of up to 10 metres anyway, so that would swamp any error due to a datum difference from records made before May 2000. -- Mike Williams Gentleman of Leisure