I have a DEM raster with some small and large areas of NoData. I also have a raster defining zones (the zones consists of a rectangular grid).

I want to create a new raster marking the zones where there exists any nodata values in the DEM raster.

My initial approach was to use the Zonal Statistics tool, using any statistics type (for example Maximum) and uncheck the "Ignore NoData" option. According to what I can read in the documentation, this should give me an output raster with NoData in the entire zones where the input raster contained any NoData pixels. (I could then perform some trivial operations to give me the desired final result).

However, this doesn't work as expected. The output raster from the Zonal Statistics tool have some holes, but there are several places where there are definitely NoData in the input raster and data in the output.

So, are there any other possibilities of achieving what I want? Or have I misunderstood how I should use the Zonal statistics tool?

  • I finally found out that the real problem was that I used Zonal Statistics from ArcToolbox. When I tried the MapAlgebra function ZonalSum instead, everything worked as expected. The reason probably being that I didn't realize that ArcToolbox don't care about SpatialAnalyst Options set in the toolbar. – matli Aug 31 '11 at 7:40

Perform a zonal summary of the NoData indicator grid, IsNull([my DEM]). This indicator has 1's where NoData exist and 0's elsewhere. Thus, a sum, maximum, range, standard deviation, or average that exceeds zero will all characterize zones containing at least one NoData value. The sum actually tells you how many NoData cells are in each zone.


You could reclassify (link) the NoData to a value, and then manipulate the two files using the additional information.

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.