weighted mean of a raster?

[AutumnSartain]

I have a raster dataset of water depths and am trying to find the mean depth within a polygon. I know to use the Zonal Statistics, and that works. But, my polygon is a blob shape resulting in some cells being completely inside, while others are just barely inside.

My question is whether the Zonal Statistics takes this into account. Is the mean it gives me weighted based on how much of certain cells are within the polygon? Or is every cell that overlaps with the polygon (even if just a small corner) weighed the same in the calculation?

Thanks!

[EricRice]

Greetings,

Please review this other thread where I explain how a cell is determined to be within the mask or not.  The same applies to the Zonal tools.  Think of the zones as masks... Your zones are also converted to rasters in the background; most spatial analyst tools will let you input features, but we internally convert to raster.

http://forums.arcgis.com/threads/40549-Criteria-for-raster-cell-extraction-by-mask?p=137795#post1377...

Regards,
Eric

[AutumnSartain]

Thanks Eric, that is great information and very helpful. 

I noticed in the other thread you mentioned that she should set her snap raster to the original one she was using. I just used default environment settings. Should I set the snap raster in the environment settings to match the original raster? I also have a clipped version that I was using because it loads faster.

Thanks again!

[EricRice]

Hi Autumn,

You're welcome!  I personally set the snap raster environment on pretty much everything I do.  It's there for you to control, and ensure pixel alignment.  For even more control, convert your zones to raster in advance as discussed in the documentation for the zonal tools. 

Eric