2

I am trying to make an IDW-interpolation in arcgis 10.1.
In my sample file i have points with the same x,y and z values. According the arcgis-resource-site, the points are considered duplicates and have no effect on the output.

However, i want these points to have an effect:
the more often a certain value has been observed in a location, the more important that z-value should be.

Is there a way to circumvent this ignoring of duplicate sample points, or to use the number of duplicate points as a 'weight-value' to indicate relative importance of sample points?

  • I think you are looking for Kernel Density function which takes into account number of observations in the same location however it is hard to interpret it in an interpolated fashion. In your case you can pre-process your data by adding up coincident point values into the same location. Given the logic behind IDW, the result that you receive will be the same even if you shift your points slightly (even further apart depending on your neighbourhood defition). – fatih_dur Jun 23 '15 at 11:31
  • I'm not an expert, but I don't believe the IDW method allows for that simply because it's the distance between point values that matters, not that certain points can be considered more correct/absolute than others. To do that you'd need to use a different interpolation method like Spline, which forces the surface to pass through the sample points, or Kriging, which is even more complicated to properly use (many discussions on that here). – Chris W Jun 28 '15 at 6:16
1

You can shift the duplicate points slightly (for example, using a Field Calculator to generate small random values in the attribute table, then adding these to the coordinate values generated by Calculate Geometry and creating a new point layer).

This way, ArcGIS should not detect and ignore the points as duplicates, and the nature of IDW assures the point cluster will then manifest its value in the resulting surface more visibly (as the surface values are calculated as sums over all input points).

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.