Layer does not match imagery by 400 meters..

FrancoisGuegan

Hello all,

I have been looking for answers, from coordinate systems issues to projection issues but have not found a solution to this sticky one yet. As you can see below, my river SHP does not match the river on the satellite imagery... only by... 400 meters. Coordinate systems and projections are the same (I projected the river shapefile (was WGS84UTMzone47N) to the projection of the imagery. Would anyone have a clue what the problem is?

Thanks!

GGN406

NeilAyres

This sort of shift is caused by an incorrect datum (change in the underlying ellipsoidal model).

I took your mainrivers shapefile, I see it is of the Mekong river in Cambodia.

Yes, it's coordinate system is defined as UTM 47N and based on the WGS84 datum. However, if you assume that WGS84 is not the original datum for this data (depends on the source of these river polys, old carto sheets perhaps).

I looked up at ASPRS Clifford Mugnier Grids & Datums page and also at EPSG.org. There are several suggestions as to the original classical datums used by the French authorities in colonial days. One of which is India1960. (Don't know why they would have used this).

So, I copied the shapefile, gave it a new name, and then redefined the coord sys, changing the GCS to India 1960.

So the blue is the original, the green te new one, and I have applied one of the available transforms between India1960 & WGS84.

The base map in National Geographic (good enough in this case).

You can see that the green one is much closer to the basemap image. The difference here between old and new is ~500m.

The fit is not exact (I wouldn't expect it to be), because we do not know on which datum this data was originally based.

The only thing we do know for sure is that it definitely wasn't WGS84.