Trying to find best interpolation method for small dataset

Discussion created by neildgarr on May 5, 2013
Latest reply on May 6, 2013 by neildgarr
I've got a set of 14 points scattered across the Gulf of Mexico and I'm trying to find the best interpolation method for it.  I've gathered archived buoy data from 2011 that contains hourly or twice hourly observations of wave height (with data gaps in some buoys).  I then categorized each observation as suitable (made that ob = 1) or unsuitable (= 0), added all of the observations for each buoy, and divided by the total observations for each buoy (ranged from 4k to 16k).  So, each buoy now has a value equal to the percentage of total obs taken in 2011 that were suitable.  These values range from 48 to 87.


Unfortunately there are only 14 buoys in the entire Gulf of Mexico that measure wave height.  So there are few points and they are scattered.  The data is also far from normally distributed.  I've read in several places where I should stay clear of Kriging b/c there are so few points and several transformations are unsuccessful at normalizing the distribution.  I've also tried Kernel Smoothing with not much success.  The best Root Mean Squared error I've been able to get is about 9.  Is there an interpolation method in ArcMap 10.1 that would be suitable for this sort of dataset?  Any help would be greatly appreciated.