Dumb question: should datum be included in a longitude/latiude coordiate?
Posted 09 October 2012 - 03:02 PM
Posted 10 October 2012 - 12:56 AM
Posted 10 October 2012 - 01:10 PM
Most current national datums (geographic coordinate reference systems) are based on the ITRF so will give similar results. Older datums are not "earth-centered" and can show much larger differences. I often explain this by identifying the older datums as 'local' datums. They're tied to the earth's surface in the area of interest, rather than centered at, or near to, the earth's center of mass.
Peter H. Dana made a nice graphic showing the differences for a bunch of different datums.
Posted 10 October 2012 - 09:13 PM
In practical terms it rarely makes a big difference. I routinely have to guess at a datum when a student I'm working with is unsure of the actual data collection method, or if the data spans a time period that covers say both NAD27 & NAD83 in a single table. The results for anything but the most precise work is usually just fine.
GIS Reference and Instruction Specialist, Stanford Geospatial Center.
Posted 11 October 2012 - 01:55 PM
So I'm having this discussion with friends. When we give the latitude and longitude of a location, should we also include the datum? I didn't think so because the latitude and longitude are only angles to the equator and prime meridian and should have nothing to do with the shape of the earth. The datum is not needed unless maps are involved. Theoretically, you can measure the latitude and longitude of the location using the stars and a sextant. Am I correct or did I miss something?
You did miss someting . Latitude and longitude are indeed angles with respect to the Equator and Prime Meridian, but where is the Prime Meridian? Even assuming you mean the Greenwich Meridian (rather than the Paris Meridian, or the Washingon Meridian, ...), that has moved more than once, and now with WGS84 is not defined as running through any single point on earth - instead it is an avarage of measurements made at various observation points round the world, to even out continental drift. It's currently about 100m away from the old observatory through which the meridian was originally defined.
Indeed the latitude and longitude of a fixed physical object will change over time, due to continental drift, and in some parts of the world (for example Australia), that drift is several metres per century. So, to really define a position with lat/long you need a datum and an epoch (time), sometimes combined like ITRF89.
Does that help you when you reconvene your argument with your friends?
ESRI Europe (firstname.lastname@example.org)
Posted 23 October 2012 - 06:11 AM
When we returned to Australia I plotted 200,000 locations into Google Earth and all seemed to closely correlate with the regions visited, using church spires as pseudo-geodetics. However, the Prime Meridian itself baffled me. I had this discrepancy between recorded latitude and Google Earth marked as a projection error on the Google tile, being familiar with the formula they use to project WGS84 onto software tiles. However, you have now provided the answer. It is analagous to UTC, where the average of multiple atomic clocks is used to define the UTC standard.
Cheers, Bill, Canberra, Australia
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users