Re: RCM to RCM

Hi Mary:
 
Thanks very much for your help!
 
Our IT has upgrade our NCL to the latest version.
 
Cheers!
Ed
 

On Saturday, November 9, 2013 7:23 AM, Mary Haley <haley@ucar.edu> wrote:
Hi Ed,

Using the ESMF library will require you to upgrade to NCL V6.1.0 or later (6.1.2 is recommended). This is because there are built-in functions required (like "sparse_matrix_mult"), and not just NCL functions.

It is possible for you to install your own version of NCL in your own local directory, if your IT team doesn't have time to do this. It's not too difficult, as I can probably provide you with a precompiled binary for your system.

If you want to try this route, then I need to know the following information:

  uname -a
  gcc --version

In addition, if you have a Linux system, I need to know what flavor (RedHat, Debian, etc). If you have a Mac, then which Mac OS  (10.7, 10.8, etc).

Thanks,

--Mary

On Nov 7, 2013, at 2:33 PM, Yang Yang <yang123yang@yahoo.com> wrote:

> Hi Marry:
> 
> Thanks very much for your help!
> 
> The ncl I am using is v.6.0. I couldn't find esmf library. I am not sure if esmf was not installed by our IT or this version does not have esmf.
> 
> I am wondering if you can email me ESMF_regridding.ncl for temporary use. Thanks!
> 
> Regards
> Ed
>
>
>
> On Friday, November 8, 2013 4:23 AM, Mary Haley <haley@ucar.edu> wrote:
> Dear Ed,
>
> You may want to consider using the ESMF suite of regridding routines.
>
> See example ESMF_regrid_16.ncl at:
>
> http://www.ncl.ucar.edu/Applications/ESMF.shtml#ex16
>
> which regrids from a high-res curvilinear grid to a low-res curvilinear grid.
>
> You can also see some ESMF regridding templates at:
>
> http://www.ncl.ucar.edu/Applications/Templates/
>
> The one you want is "ESMF_curv_to_curv.ncl.
>
> --Mary
>
>
>
> On Nov 6, 2013, at 9:07 PM, Yang Yang <yang123yang@yahoo.com> wrote:
>
>> Hi There:
>> 
>> I would like to do interpolation from a low-resolution RCM to a high-resolution RCM. I checked NCL and couldn't find a function like rcm2rcm. I had to use rcm2points with the following codes:
>>    do i = 1, 1000
>>    do j= 1, 1000
>>      pp(i,j) = rcm2points (tlatb,tlonb,mslpb,tlat(i,j),tlon(i,j),0)
>>      uu(i,j) = rcm2points (tlatb,tlonb,u10mb,tlat(i,j),tlon(i,j),0)
>>      vv(i,j) = rcm2points (tlatb,tlonb,v10mb,tlat(i,j),tlon(i,j),0)
>>    end do
>>    end do
>> However, the codes takes a lot of CPU time. Very slow. Can someone give me some suggestions how to optimize the codes? 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
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Nov 11 12:55:37 2013