Choosing the right method

677
2
05-25-2011 09:03 AM
AshleyMott
New Contributor III
Hi,
I am attempting to create an interpolated surface of noise with the Geostatistical Analyst extension. I am not an expert, so please excuse my ignorance here.

I have 27 points of decibel measurements over approximately 475 square miles. One of the points towards the edge of the study area would be considered an outlier (it is about 25 db more than the second highest measurement).

After reviewing the Classification trees of the interpolation methods offered in Geostatistical Analyst, I think I should use either IDW or GPI. I am reluctant, however, to rely on the GPI, because of my outlier on the edge of my study area.

I am concerned about my sample size of 27. I have read that having less than 30 points can hurt reults in these types of spatial operations, which is why I choose methods that don't rely on spatial autocrrelation. Is there anything else I should be aware of?

I am also concerned about any behind the scenes math that the models may use, because decibels have a logarithmic relationship, one cannot simply arithmetically add, subtract, or average decibel levels. For example if you have two 60 dB noise sources occurring simultaneously they combine to equal 63 dB.

Help and advice is most appreciated!
0 Kudos
2 Replies
EricKrause
Esri Regular Contributor
All interpolation methods assume there is spatial autocorrelation.  Some methods explicitly model it (kriging, for example), and others just make assumptions about it (IDW, for example).

With only 27 points (and a possible outlier), I wouldn't try to use anything other than IDW.  If you're concerned about the logarithmic scale, you can make a new field and calculate the antilog, then take the logarithm of the results.

I wouldn't attempt to do kriging with so few points and an outlier.
0 Kudos
AshleyMott
New Contributor III
Thank you very much Eric. I was feeling a bit unsure of myself, so I am glad that you clarified my spatial autocorrelation worry. Luckily, this is for a school project (it is more about GIS and not so much about saying something about noise), so the result doesn't have to live in the real world and now, thanks to you, I can explain my choices.
0 Kudos