Cut Fill Error Margin

[KaraLamantia]

I'm using the Cut and Fill tool in the 3D analyst toolbox to find the difference in volume between two sets of raster files.  

I'm wondering if there is some sort of error margin that is possible each time I run the tool? For instance, does it round values so there might be an error of 0.5 meters in each pixel?

Very familiar with the tool and the 3D toolbox, but I have yet to find anything beyond the basic description of how the tool works

[DanPatterson_Retired]

Not sure I follow, but since you are working with rasters and raster representation of geometry, then there is an inherent approximation.  What sort of cell size are you working with?

[KaraLamantia]

The entire area is a few square kilometers. I was assuming there is some sort of approximation, but I was hoping to quantify that for research publishing purposes

[DanPatterson_Retired]

Kara...  I meant the size of each raster cell.  If the cell size is 1 meter (aka cell width), then area is 1m^2 and volume 1m^3.  If you are delineating an area, then the boundary 'replication' is going to better the smaller the cell size, hence area and volume would be better 'estimates'.  If people don't specify the cell size when using any tool in ArcToolbox's Spatial Analyst extension, then the cell size is determined as 1/250th the width/height.  So for a 1 km boundary edge, that would be 4 metres in length, 16m^2 in area etc.  Is that good enough? If not, you use a finer cell size and weigh the benefits of that versus data processing issues.  Of course, the totally leaves out how well any process is an actual representation of what it is purporting to measure.

[KaraLamantia]

Dan,

My cell size is about 5m, the raster themselves are derived from a series of mass point measurements across the surface of a glacier. We're looking at volume loss over time.

[DanPatterson_Retired]

Kara, if you are looking for some volume change +/- some margin of error, then the error would lie in the accuracy of those measures from the mass points.  The precision (fineness) of the raster is simply how it represents those measures. 

[KaraLamantia]

Okay, that is what I originally thought, but I just wanted to make sure I wasn't missing something. Thank you for your help!

[DanPatterson_Retired]

No problem Kara