Here's my stab at a summary. Please correct me if you can (there are some trained surveyors here). Originally, NAD83 was directly related to a fixed version of the GRS80 ellipsoid. GRS80 has since been adjusted to depend on the ITRF, and is now dynamic. NAD83 (CSRS and NSRS) are related to this new dynamic GRS80 ellipsoid. (This also affects WGS84.) Normally, for GIS purposes, it isn't necessary to worry about these differences. For synchronic data (positions measured or made correct at the same time), it is correct just to use GRS80 as the ellipsoid, since this is same reference frame at the same time. However, if there is an explicit need to convert between data measured or made accurate under the old, fixed versions of NAD83/GRS80, and data relative to the new dynamic versions (and if differences in the order of several centimetres are significant for our purpose), then we need to use a transformation. The official 7-parameter Helmert transformation model for NAD83 (for both fixed and arbitrary timeframes) is given in Realization and Unification of NAD83 in Canada and the U.S. via the ITRF, available from the US NOAA here. So yes, I think we can do this. Because I'm not on very firm ground I want to sit on this for a moment--and give others a chance to chime in with observations and corrections--before writing up some XML for the transformation(s). Then we would need to test, with known good data measured in NAD83 (original) and NAD83 (CSRS or NSRS) measured at a known time.
|