When running a point-in-polygon operation, is it valid to run that with polygon features in straight up WGS84 and a point in lat/lng, or should both be projected into a non degree based coordinate system?

  • 2
    WGS84 is fine for this operation... projections would be more useful for getting area/lengths, etc. – DPSSpatial Nov 20 '14 at 17:57
  • Are you using ArcGIS for Desktop to do this? – PolyGeo Nov 20 '14 at 20:11

It is entirely valid, but it may not produce the results you expect in any Cartesian coordinate space. So it isn't a problem with degrees or WGS84, but rather the mapping from the ellipsoid to a plane.

If the polygon is very large, the mapping of the curved surface on the ellipsoid to the plane can produce some artifacts where a point looks like it would be just inside (or outside) the polygon, but isn't quite. That can also be an issue if some of the polygons cross the anti-meridian (the -180/180 part). If those might be an issue for your data, and you can handle the performance impact, consider something like a PostGIS geography approach.

  • 1
    Mainly depends on how the software connects two points in a geographic CRS. Are the points connected with a geodesic or normal section or a 'Cartesian' line. Ex. between two points on the same latitude line, the Cartesian line is the latitude line. The geodesic/normal section might pass above or below the latitude line. – mkennedy Nov 20 '14 at 21:40
  • 3
    It's not the size of the polygon so much as the size of the gaps between vertices in the polygon. If the intervertex distances are set to approximate a great-circle route, then even a Cartesian point-in-poly will be "accurate" in a GCS (within the definition of "accurate" established for the polygon). – Vince Nov 21 '14 at 2:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.