# Cell size on cost path calculation

3385
2
Jump to solution
04-17-2015 04:01 AM New Contributor II

Hi All,

I have a basic question about cell size (i.e. in meters of my calculation; 30 meters compare with 300 meters) on cost path calculations.

Is it true, if I have smaller cell size of cost surface will have bigger cost path values?

In my understand the smaller cell size will have more accumulated cells of cost distance, and vice versa.

I still don't understand this paragraph at :

When the input source data is a feature class, the source locations are converted internally to a raster before performing the analysis. The resolution of the raster can be controlled with the Output cell size parameter or the Cell Size environment. By default, the resolution will be determined by the shorter of the width or height of the extent of input feature, in the input spatial reference, divided by 250.

Thank you,

[bayu]

Tags (3)
1 Solution

Accepted Solutions by MVP Esteemed Contributor

If no cell size is given the larger of the extent width or height is divided by 250 to give the cell size.  When you are doing cost analysis, you are still working with a cell size.  which means that you have to know the cost per cell area.  You simply can't take the number of cells as the total 'cost' in some economic sense etc, it is a relative measure.  I suggest you experiment

• generate a cost surface using a large cell area say 100 m^2 (ie 10m cell size, then run cost path
• resample the cost surface so that the cell size is 1 m (ie 1m^2 area), and rerun the cost path.
• compare

since you have resampled the cost surface and not generated a new cost surface which may differ in shape you can now see whether the paths differ when going from one location to another.  Produce these examples, then we can talk some more

2 Replies by MVP Esteemed Contributor

If no cell size is given the larger of the extent width or height is divided by 250 to give the cell size.  When you are doing cost analysis, you are still working with a cell size.  which means that you have to know the cost per cell area.  You simply can't take the number of cells as the total 'cost' in some economic sense etc, it is a relative measure.  I suggest you experiment

• generate a cost surface using a large cell area say 100 m^2 (ie 10m cell size, then run cost path
• resample the cost surface so that the cell size is 1 m (ie 1m^2 area), and rerun the cost path.
• compare

since you have resampled the cost surface and not generated a new cost surface which may differ in shape you can now see whether the paths differ when going from one location to another.  Produce these examples, then we can talk some more New Contributor II

In my case:

- I know the cell size is 30m X 30m (900 sqm)

- I know/define the cost per cell size, I stated it as 'normalization cost' which have values between 1 to 21.54. This values is defined as "cost multiplier" of my infrastructure development. e.g. total cost (\$) = norm. cost * cost per km.

OK, I will try to reproduce with different cell size (cell area) and update here.

Thanks Dan. 