When I buffer a point (red dot on image) by 3 miles, it is 3 miles on all sides. When I buffer a line (blue line on image) by 3 miles, it is more like 2.67 miles on the sides parallel to the line. I tried various projections, but I basically get the same result.
Any ideas on why this happens or workarounds?
enter image description here
-
1What coordinate system is your point/line data in? If it is lat/long this may be the cause, try projecting your data into a coordinate system in meters then do the buffer?Hornbydd– Hornbydd2015年04月29日 21:21:35 +00:00Commented Apr 29, 2015 at 21:21
-
1Which projections did you try; were they all geographic? How did you create the buffers (which tool, settings)? Any geoprocessing environment settings? It's possible to create planar or geodesic buffers depending on your methods. See the help resources.arcgis.com/en/help/main/10.1/index.html#//… along with the How It Works link at the top of that page.Chris W– Chris W2015年04月29日 21:22:02 +00:00Commented Apr 29, 2015 at 21:22
-
1Which ArcGIS version?mkennedy– mkennedy2015年04月29日 21:29:05 +00:00Commented Apr 29, 2015 at 21:29
-
3At 10, buffering lines is still Cartesian only. I would have expected state plane to be closer to "truth". Possibly do a custom azimuthal equidistant, centered at the data. Or try Buffer Wizard, and set it for feature-optimized.mkennedy– mkennedy2015年04月29日 22:32:07 +00:00Commented Apr 29, 2015 at 22:32
-
1Just to put closure on this, I made the line into state plane. Made the map projection the same state plane and then buffered. 3 miles all around!! Thank you all!Tx_Dan– Tx_Dan2015年04月30日 13:20:10 +00:00Commented Apr 30, 2015 at 13:20
1 Answer 1
When buffering (among other distance related operations) in a geographic coordinate system, your result may appear or be distorted in 2D because geographic is on a sphere/ellipsoid, not flat. Geographic uses angular degrees for measurement, and a degree of longitude can vary significantly in terms of distance depending on latitude (which is relatively consistent). So what appears as a circle near the equator would be a very squashed ellipse near the poles - see Tissot's indicatrix.
If you are in a projected coordinate system (like a State Plane zone), that spherical/3D surface has been flattened out, with some distortion, to a 2D plane. Therefore when you create your buffer, the results should be much more consistent in terms of distance - though possibly inaccurate over large distances. The Esri article Understanding Geodesic Buffering give some detailed explanation.
It sounds like you may have been running into an issue with either the original data coordinate system being geographic, on-the-fly reprojection not being set up correctly, or the buffer using the original CRS (as geographic) and not the dataframe (as projected). Note also that if two CRS or projections have different datums, you also need to specify a correct transformation for the dataframe to use (there are other questions here on that subject).
Related:
Explore related questions
See similar questions with these tags.