Zonal Statistics results are not equal

[PamFroemke]

I???m trying to get a value for mean water runoff by wilderness area. I am using the Zonal Statistics as Table tool with wilderness areas as the input raster and water runoff as the input value raster.
Problem:
Zonal stats produces higher values for total runoff and area when combining wilderness with owners and regions compared to wilderness with just owners.
Procedure:
1. Set environment - projection, extent, cell size (100 meters), and 100-meter snap raster.
2. I start with two versions of wilderness by federal agency polygon shapefiles: one divided by Forest Service regions and one that isn???t.
3. Then I use the Feature to Raster tool for converting the two wilderness shapefiles to rasters.
4. Project and resample the runoff grid from 5 km to 100 meters so that it matches the wilderness rasters.
Zonal Statistics as Table Results:
1. Total wilderness area is greater for wilderness by agency and region compared to wilderness by agency.
2. Total runoff is greater for each wilderness when split by agency and region compared to wilderness by agency.
Does anyone have any idea why I am getting these different results?
Thanks,
Pam

[curtvprice]

Does anyone have any idea why I am getting these different results?

Pam, just an approach to try:

I would compare the two input zone rasters (wilderness, wilderness x regions) to make sure the cell sizes, extents, cell counts match up the way you think they are matching up BEFORE you run the zonal statistics.

Then before I ran the zonal stats I'd make sure the GP environment is set the same for both runs.

[PamFroemke]

Curt,
I compared my two input datasets and found that the extents were slightly different. They were also different than the extent I defined before I ran the model. Is there another setting that over-rides the model's environment settings?

Thanks for the help,
Pam