Zonal Statistics as Table

[Anonymous User]

Original User: Shiko

Hi folks,
So this time I am once again, using Model Builder to run the "Zonal Statistics as Table" tool to create database (.dbf) files.
See model below:
[ATTACH=CONFIG]27583[/ATTACH]

After the model runs, I can access the .dbf files and pull them into excel for further analysis.  For each time it runs with the different tiff layers, the format is as follows:
[ATTACH=CONFIG]27587[/ATTACH]

The issue is that each time I run the model with the different raster (.tif) images, the AREA for the same UNIQ_FIDs is different.
From what I understand about this tool, is that it performs the statistic specified (in this case = MEAN) of all the raster  pixels that it counts within that zone. So I can understand why the COUNT would be different but I don't know why the AREA would change from dbf to dbf?!! Isn't the zone (in this case= polygon referenced by UNIQ_FID) remaining the same? :confused:

Maybe I have the whole Zonal Analysis idea wrong in my mind. Could anyone help me make sense of why the area would be changing.

[TOMKONG]

Hi folks,

The issue is that each time I run the model with the different raster (.tif) images, the AREA for the same UNIQ_FIDs is different.
From what I understand about this tool, is that it performs the statistic specified (in this case = MEAN) of all the raster  pixels that it counts within that zone. So I can understand why the COUNT would be different but I don't know why the AREA would change from dbf to dbf?!! Isn't the zone (in this case= polygon referenced by UNIQ_FID) remaining the same? :confused:

Maybe I have the whole Zonal Analysis idea wrong in my mind. Could anyone help me make sense of why the area would be changing.

1.Are your zonal layer and all raster dataset in same projection?
2.Are all raster dataset in same cellsize?
3.Is the tiff converted to GRID (ESRI raster format)?

[Anonymous User]

Original User: Shiko

Yes, yes, and yes!
I run the model two separate times with the different GRID files, with the same cell size, and coordinate system and the AREA values are still different.
Any other ideas?

[TOMKONG]

Yes, yes, and yes!
I run the model two separate times with the different GRID files, with the same cell size, and coordinate system and the AREA values are still different.
Any other ideas?

It should be "NoData" cell issue. Your different GRID dataset have different "NoData" cell coverage, so the raster area with one statistic value should be different from another.

[Anonymous User]

Original User: Shiko

It should be "NoData" cell issue. Your different GRID dataset have different "NoData" cell coverage, so the raster area with one statistic value should be different from another.

Oh okay! I figured that might be it since that's all that was left as an uncommon factor!
So does that mean that the AREAis not that of the zone? and rather, is that of the cumulative pixel area?

[TOMKONG · Accepted solution]

Oh okay! I figured that might be it since that's all that was left as an uncommon factor!
So does that mean that the AREAis not that of the zone? and rather, is that of the cumulative pixel area?

Yes that the count number is the pixcel number with one value you selected to do statistics, so the Area value is besed on these pixcel number and cellsize.