2

I have a raster image of water depth, float point values 1-15, in a very small area some erroneous values are in the raster with values of 100 - 375. I'm looking for a repeatable work flow in ArcMap 10.5, Spatial Analyst level. I've used one work flow to set the values to nodata, one work flow to set to a fixed float point value, but what I'd really like is to create polygon or raster mask, and then have the grids of the original raster, within the mask area, to be replaced with nearest neighbor values from the original raster outside the mask area.

2

The tool you are looking for is nibble. Nibble does exactly what you are asking for. Your work flow would be something along the lines of using your polygons to create a mask raster as the second input. Other related tools to consider (but probably not what you want for this use-case) are shrink and expand.

  • I tried Nibble and did not get the desired results, that why I came here. I not performing the workflow correctly in someway – L. Hogan Mar 6 at 18:43
  • Nibble is the correct tool. Re-read the documentation carefully and I think you'll find the solution. – MappaGnosis Mar 6 at 19:02
  • nibble is for integers documentation says help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#/Nibble/… – L. Hogan Mar 6 at 20:45
  • You can make an integer raster based on your original multiplied by 1000 (or whatever decimal precision you want), perform the nibble operation then divide by 1000 - it's a common solution where you are forced to work with integers. – MappaGnosis Mar 6 at 20:53
  • I was looking for a more concise workflow - maybe with Con or Raster Calculator, that could somehow avoid the conversion. The raster calculator examples I found online did not work for me, Con(IsNull(Raster), FocalStatistics(Raster, NbrCircle(10, "CELL"), "MEAN"), Raster) – L. Hogan Mar 6 at 21:12

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.