I have a 10m USGS DEM for my entire map extent (~100sq miles)

Inside of that, I have specific regions of higher quality GPS point data. (From a rover using a OPUS corrected RTK base station)

How do I seamlessly merge my data/data products with the "background" USGS data? If I simply merge/replace the USGS data with mine, the edges don't align and my entire region looks "sunk in" or "raised up" in comparison to the USGS "background"

The gaps that are being filled in by the USGS DEM are up to a full square mile, so I can't just interpolate the blank area from the edges of my survey.

My first though is to "feather" the USGS DEM. Obviously at the edge where the two meet, my data is more accurate, so feathering the USGS data to meet smoothly with my survey seems like it should work, but I am unsure of the technical way to approach this or the accuracy of this method.

The data products that are being produced from this are hillshades, viewsheds, streamflow, contours, and more - and none of them seem to work well with a sharp boundary transition.

1 Answer 1

  • Create outline of new dem
  • convert to polyline Create transition zone buffer around it. to polyline
  • Interpolate shape using old dem
  • Interpolate shape using new dem
  • Triangulate using both. Map calculator will do the rest


to simulate distortions I took this DEM: ORIGINAL

clipped and flipped part in the middle and merged original with flipped: DISTORTED


arcpy.gp.RasterCalculator_sa("""Con( ~IsNull("flipped"),1)""","C:/FELIX_DATA/SCRARCH/one")
arcpy.Buffer_analysis("boundary","C:/FELIX_DATA/SCRARCH/buffer.shp","20 Meters","FULL")

arcpy.CreateTin_3d("C:/FELIX_DATA/SCRARCH/tin","#","inside Shape.Z Hard_Line <None>;outside Shape.Z Hard_Line <None>;buffer <None> Soft_Clip <None>","DELAUNAY")
arcpy.TinRaster_3d("TIN","C:/FELIX_DATA/SCRARCH/transition","FLOAT","LINEAR","CELLSIZE 1","1")




  • @Luke see little:) more detail added
    – FelixIP
    Commented Oct 8, 2016 at 22:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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