Re: rcm2points

The inverse distance squared is very sensitive to distance :-)
I some cases, it would 'almost' a nearest neighbor method.

If the variable being interpolated is smoothly varying [eg: temperature]
it is likely that you will see little difference between the two methods.

A 'random' variable like precipitation where you might have 0,0,0,100 or
100,0,0,0 could surround a target point could give very different
results depending upon distance.

On 8/4/11 9:17 PM, Yang Yang wrote:
> Hi:
>
> I have a simulated daily rainfall accumulation by a regional model. I want to interpolate the rainfall on 15 points. I used two methods:
>
> 1. Using rcm2points by setting opt=1
> 2. Using 4 grid point values around each station and the inverse distances to the point as weights to do the interpolation.
>
> The results showed pronounced difference between them. I attached the diagrams for each method in which the color shading for the 2D rainfall was plotted. Regarding the consistency between the station values and the color shading, method 2 seems better than rcm2points. For rcm2points method, large inconsistency can be found in the region with the maximum.
>
> My question is how many grid points are used in rcm2points for the interpolation of each point? Any limitation for using rcm2points? Any information is much appreciated.
> Thanks!
>
> Ed
>
>
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Fri Aug 5 13:55:01 2011