[ncl-talk] pattern_cor significance test

Dennis Shea shea at ucar.edu
Wed Mar 15 15:26:58 MDT 2017


Some wild guesses:

Generally, in geophysics, adjacent grid locations are not independent.

Levezey & Chen: 1983: Statistical Field Significance and its Determination
by Monte Carlo Methods

  http://dx.doi.org/10.1175/1520-0493(1983)111<0046:SFSAID>2.0.CO;2
<http://dx.doi.org/10.1175/1520-0493%281983%29111%3C0046:SFSAID%3E2.0.CO;2>

Perhaps a bootstrap (Monte Carlo) approach? as advocated in the paper.
Perhaps, the following might be helpful.

  http://www.ncl.ucar.edu/Applications/bootstrap.shtml

-------

Perhaps:

      http://www.ncl.ucar.edu/Document/Functions/Built-in/rtest.shtml

An issue would be the appropriate value of 'Nr'. Each spatial data set may
ve different number of independent  values. If so, use the lower number.

----

Look on the WWW /// maybe there is a more knowledgeable answer.

On Tue, Mar 14, 2017 at 9:17 PM, Sitan Heyl <heylsitan at gmail.com> wrote:

> Hello NCL users,
>     Does anyone have an idea how to make a significant test for the
> correlation coefficient calculated from the function "pattern_cor"?
>
>      Thanks in advance!
>
> Best,
> Yongli
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20170315/f54619a7/attachment.html 


More information about the ncl-talk mailing list