[ncl-talk] ESMF in NCL

Jeff Auger jauger83 at gmail.com
Mon May 23 08:21:50 MDT 2016


Hi Mary,
That ended up being the issue, just a large domain. The script works great
now that we aren't running the entire CONUS and just chose the region of
interest.
Thanks for the ideas!
Cheers,
Jeff

On Mon, May 23, 2016 at 10:15 AM, Mary Haley <haley at ucar.edu> wrote:

> Jeff,
>
> Do you still need help with this?
>
> You're right, this is quite a large domain. It helps if you just do the
> regridding over the area of interest. For example, if you have regional
> data, then set the SrcURCorner/SrcLLCorner/DstURCorner/DstLLCorner to an
> area just large enough to contain the domain of interest. This could result
> in a noticeable improvement in time.
>
> --Mary
>
>
> On Fri, May 20, 2016 at 4:40 PM, Jeff Auger <jauger83 at gmail.com> wrote:
>
>> Mary,
>> No problem. I tried a smaller domain and it works now. Apparently the
>> domain was just too big for the memory. I tried sending the files but it
>> was too big and my response was kicked by admin there.
>> Thanks for taking a look at it.
>> Cheers,
>> Jeff
>>
>> On Fri, May 20, 2016 at 6:02 PM, Mary Haley <haley at ucar.edu> wrote:
>>
>>> Jeff,
>>>
>>> I won't get to this before I leave for the day, but to save time, I
>>> would think that inside the "do m=0,11" loop, you could set the last
>>> argument of resize_grid to True for the first iteration, and then to False
>>> for the other 11 iterations.
>>>
>>> This will probably not fix the original problem, but I wanted to point
>>> it out for future use.
>>>
>>> --Mary
>>>
>>>
>>> On Fri, May 20, 2016 at 3:06 PM, Jeff Auger <jauger83 at gmail.com> wrote:
>>>
>>>> ​Hello Mary,
>>>>
>>>> I attached a compressed folder including the ncl script and input file.
>>>> The script is set up to produce the weights file first then once produced,
>>>> change the "True" to "False" on line 206 and rerun. The goal is to regrid
>>>> CCSM4 to PRISM resolution, 800m. Quite a jump, but not sure that would be
>>>> the issue... I had to close the domain around the CONUS to cut down on
>>>> computation time and file size. Again, not sure if that would be a problem
>>>> as well. But perhaps useful information as you look it over.
>>>>
>>>> Thanks,
>>>> Jeff​
>>>>
>>>> On Fri, May 20, 2016 at 4:51 PM, Mary Haley <haley at ucar.edu> wrote:
>>>>
>>>>> Hi Jeff,
>>>>>
>>>>> This is indeed a bit obscure. This error is not coming from within NCL
>>>>> itself, but from the ESMF_RegridWeightGen application that NCL calls.
>>>>>
>>>>> I'm afraid there's probably something fundamentally wrong with the
>>>>> lat/lon grid that you are either regridding from, or regridding to.
>>>>>
>>>>> We will need see your script, and know the details of the type of
>>>>> lat/lon grids you are working with as a start.  If this is not a big file,
>>>>> then it would be even better if you could provide the full script and data
>>>>> so we can run it here.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> --Mary
>>>>>
>>>>>
>>>>> On Fri, May 20, 2016 at 1:40 PM, Jeff Auger <jauger83 at gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> I have run into a problem trying to produce a weights file.
>>>>>>
>>>>>> ESMF_regrid_gen_weights: 'ESMF_RegridWeightGen' was not successful.
>>>>>>
>>>>>> The PET file reads:
>>>>>> 20160520 150737.592 INFO             PET0 Running with ESMF Version
>>>>>> 6.3.0rp1
>>>>>> 20160520 151940.816 ERROR            PET0 ESMF_IOScrip.F90:1322
>>>>>> ESMF_OutputScripWeightFile netCDF Status Return Error
>>>>>> 20160520 151940.817 ERROR            PET0
>>>>>> ESMF_RegridWeightGen.F90:1395 ESMF_RegridWeightGenFile Failure - Internal
>>>>>> subroutine call returned Error
>>>>>>
>>>>>> I have looked at the ESMF source code but could not find what the
>>>>>> error means within the code.
>>>>>> Does anyone know how to find/fix the error here?
>>>>>>
>>>>>> Thanks,
>>>>>> Jeff
>>>>>>
>>>>>> _______________________________________________
>>>>>> 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/20160523/f391d745/attachment.html 


More information about the ncl-talk mailing list