I am interpolating air pollution values from 6 data locations using IDW in ArcGIS. I know that more the number of points, more reliable the results are. But, I am not sure if a minimal number of data points is necessary.
-
It should be easy enough to test?– John PowellCommented Dec 21, 2015 at 17:56
-
Welcome. Please take the tour here: gis.stackexchange.com/tour. As John mentioned, trying it out with a few different values should help you decide what minimum you should use.– MaryBethCommented Dec 21, 2015 at 18:01
Add a comment
|
1 Answer
As a method, no.
IDW tries to interpolate (=guess) the grid value based on observations.
The idea is; the closer the better. So IDW tries to look around to find any observations, applies a weighting factor (the inverse distance, hence the name), averages and derives the answer.
How long or short is the minimum distance? That's defined by the user. It's one of the 'cons' of IDW.
Now, If in the searching distance, IDW cannot find any observation, well it cannot derive any value, so the Answer would be null.
In the case, it finds only one observation, it has nothing to average it with, so the grid will take the value cell.
So as a method for a valid value it only requires at least one observation to have an answer besides null.
It's up to the user (you) to decide if one observation is representative for the problem at hand, or not. (otherwise we all would be doing air pollution modeling ;P )
