I have two files both projected in Sphere_Miller_Cylindrical with a central meridian set at 100 and both have GCS_Sphere_ARC_INFO. One comes from a PGDB and the other is a shapefile. I believe their origins is somehow responsible for this slight misalignment. Zoomed into the African southwest coast, the coincident boundaries are off by 1.5 inches, and on the Indian southwest coast they are off by 6.5 inches. Elsewhere, such as the American west coast, they align perfectly.
My questions are: Is there a way to align these troubled spots perfectly? How? How do I know which output is the correct one? What is the evidence? Basically, what is going wrong?
My guess is it has to do with their different origins (.mdb and .shp) and/or the slightly different extents.
I realize for a map of such a scale that this ever so slight misalignment is redundant, but I need to know what is happening.
The central meridian is +100. I've examined some other spots at the boundaries of the two also do not line up at western-southwestern sides of places like Japan (2in), New Zealand (19in), Burma (4in). But places like Alaska and South America align perfectly. I cannot find any pattern to the misalignments other than they tend to be on the western-southwestern sides of features.
That's...interesting. The data appears to line up better at the edges (North/South America), not so good in the middle east and west portions (Africa, Japan), horrible middle/east south (New Zealand). If it was a sphere/ellipsoid issue, I would expect it to be symmetrical. I almost feel like one is using a different projection, but I've no idea what it could be. Because the edges still work, mostly likely a cylindrical projection like Gall's or Plate Carree.