Re: uv2dvF and uv2dv_cfd

Both functions *require* that the data be ordered South-to-North.
If they are N->S then the signs will be opposite of what you expect.

http://www.ncl.ucar.edu/Document/Functions/Built-in/uv2dv_cfd.shtml
http://www.ncl.ucar.edu/Document/Functions/Built-in/uv2dvF.shtml

If the data are ordered North->South

    u = f->U ; (time,lev,lat,lon)
    u = u(:,:,::-1,:)
    printVarSummary(u) ; look at the coordinates

The spherepack software is available at:
     http://www.cisl.ucar.edu/softlib/SPHERE.html

Good luck

On 11/15/10 8:04 PM, 武伟 wrote:
> Hi, developers
>
> I have computed divergence of a vector (U*Q,V*Q) , with uv2dvF and
> uv2dv_cfd, seperately.
> Now the question is that even though their global patterns are
> similar, the results in some sub-regions are so different that uv2dvF
> result is *contradictory* to conventional knowledge (with opposite
> sign). (p.s. I do believe that spheric harmonic result will be better
> than that of difference)
> So would you please offer the source code of this function, or its basic
> maths libraries? Or could you please recommend some papers about it?
>
> Thanks in Advance!
>
> --
> Will Wu (wewe870521 at gmail.com <http://gmail.com>)
> Center for Monsoon Study and Research,
> Institute of Atmospheric Physics,
> Chinese Academy of Science,
> Beijing, P.R.China
>
>
>
> _______________________________________________
> 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 Mon Nov 15 20:30:43 2010