Re: Drawing streamlines freezes NCL script

From: Mary Haley <haley_at_nyahnyahspammersnyahnyah>
Date: Mon, 20 Apr 2009 13:13:09 -0600 (MDT)

My apologies Samuel, I had you reverse the latitude values, but not
the actual values themselves, so the latitude values are not correctly
representing your data.

There's something odd about this data that we need to look further
into. I tried to reverse the latitude dimension of both uu and vv, and
it hung again. If I only reverse the latitude of one of these arryas,
then it's okay.

For example, instead of this:

   uu&g0_lat_0 = uu&g0_lat_0(::-1)

I tried:

   uu = uu(::-1,::) ; Reverse the latitude dimension
   vv = vv(::-1,::)

which hangs. If I comment one of the two above lines, it doesn't hang,
but I'm not sure the streamlines are correct. I'm going to hand this
off to Dave Brown to see if he can find the problem.

--Mary

On Mon, 20 Apr 2009, Mary Haley wrote:

>
> Samuel,
>
> I think the problem is that you have 1D lat/lon coordinate arrays, but
> the latitude values are monotonically decreasing instead of
> increasing.
>
> I'm surprised ncl didn't issue a warning (I've CC'ed Dave Brown on
> this).
>
> Meanwhile, right before the lines:
>
> lat_0 = uu&g0_lat_0
> lon_1 = uu&g0_lon_1
>
> add the following line to reverse the latitude values:
>
> uu&g0_lat_0 = uu&g0_lat_0(::-1)
>
> Dave, let me know if this shouldn't be an issue, and I'll provide
> you with a sample script offline.
>
> --Mary
>
> On Mon, 20 Apr 2009, Leung Samuel wrote:
>
>> forgot to attach the ncl script
>>
>> On Mon, Apr 20, 2009 at 11:18 AM, Leung Samuel <shwl0608_at_gmail.com> wrote:
>>
>>> Hi Mary,
>>>
>>> I have checked the code in office. I had the problem the ncl hangs even if
>>> I draw the streamline in standalone plot (the attached ncl script is what I
>>> used). And I noticed the location where it hangs is not at the line
>>> gsn_csm_streamline but the next line which is draw(plot).
>>>
>>> By the way, how can I check if the settings here and your side are the
>>> same?
>>>
>>> Samuel
>>>
>>>
>>> On Sat, Apr 18, 2009 at 12:12 PM, Samuel Leung <shwl0608_at_gmail.com> wrote:
>>>
>>>> Hi Mary,
>>>>
>>>> Thanks for your effort put on my scripts. It's almost the same as what I'm
>>>> working on. For ISBL_INDEX, simply use either 6,8,10 or 12, or directly run
>>>> the spt in the pack. (FYI, there are totally 4 kinds of images I need to
>>>> generate, 500, 700, 850 and 925hPa. The ISBL_INDEX is used for the GRIB to
>>>> identify the isobaric levels)
>>>>
>>>> Do you mean the images (both ps and png) can be generated successfully in
>>>> your environment? My situation is, the test.ps is getting bigger and
>>>> bigger (as if an infinite loop) while the ncl script is hanging. What are
>>>> the filesizes of the 4 test.ps postscripts after they are generated?
>>>>
>>>> I gotta verify if all the setting is correct after I am back to the office
>>>> next Monday.
>>>>
>>>> Thanks again. This problem really bothers me a lot.
>>>>
>>>> Samuel
>>>>
>>>>
>>>> On 18 Apr 2009, at 3:39, Mary Haley <haley_at_ucar.edu> wrote:
>>>>
>>>> Samuel,
>>>>>
>>>>> I've been unable to get the code you sent me to hang, but I'm not
>>>>> entirely sure I'm running it under the same conditions.
>>>>>
>>>>> With the code you sent, does it hang as is, or do I need to change
>>>>> something?
>>>>>
>>>>> --Mary
>>>>>
>>>>> On Apr 17, 2009, at 3:58 AM, Leung Samuel wrote:
>>>>>
>>>>> Hi Dave,
>>>>>>
>>>>>>> In your code excerpt I can't tell whether these are set for res, but
>>>>>>> they are definitely missing for resS.
>>>>>>
>>>>>> For the res and resS I tried both gsnFrame and gsnDraw but the problem
>>>>>> still exists.
>>>>>>
>>>>>>> Please insert print statements between the last 5 lines of the program
>>>>>> to find out exactly where NCL is hanging.
>>>>>>
>>>>>> Yes and I noticed that the NCL is hanging at gsn_csm_streamline
>>>>>>
>>>>>>> Also try to make the streamline plot directly without using overlay.
>>>>>>
>>>>>> I did. I commented the base plot to solely draw the streamline but the
>>>>>> problem persists.
>>>>>>
>>>>>>> Also, what array sizes are are you using?
>>>>>>
>>>>>> The arrays of uu and vv are 285 x 389.
>>>>>>
>>>>>> Regards,
>>>>>> Samuel
>>>>>> _______________________________________________
>>>>>> 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 Apr 20 2009 - 13:13:09 MDT

This archive was generated by hypermail 2.2.0 : Fri Apr 24 2009 - 14:12:40 MDT