I'm performing a fairly routine task, which is to add height information (from a 50m DEM) to survey polygons to work out mean height, as well as max and min.

I've noticed that zonal statistics (Spatial Analyst) is very slow, and depending on the raster analysis settings, can miss out polygons due to raster sampling size. For example, in the image below, the red polygon is missed out of zonal statistics because it's too thin to adequately intersect enough (or indeed any) raster cells.

missing polygons

Interpolate shape -> Add surface information, by contrast, seems much much faster. But I don't understand what's being interpolated here for a polygon feature. Is it just the vertices, or is the interior of the polygon also being compared with my underlying DEM? At what resolution?

Which method would provide more accurate results? And, are there any other solutions. I often find that terrain operations work equally well with rasters containing non-height information, and vice versa.

1 Answer 1


Interpolate shape is used to create a 3D image of a 2D shape containing Z values.


To capture values for your red shape resample the raster to a finer scale. You can use either the bilinear or majority interpolation.


Furthermore, if you have having difficulty running tools while working on a network you can transfer to the local hard drive to hopefully speed up the process. Setting a computational mask is also a way to speed up raster calculations.

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.