Re: bugs in NCL 5.2.0 for contour map plot (guangshan chen)

Hi Guangshan,

Just FYI, I have been looking at this problem, and I agree with Mary
that NCL 5.2.0 is behaving correctly compared with the earlier
version. There was a fix to the transformation routines between 5.1.1
and 5.2.0.
  -dave

On May 27, 2010, at 8:31 AM, Guangshan Chen wrote:

> Hi Mary,
>
> Thanks.
>
> Guangshan
>
> On May 26, 2010, at 6:15 PM, Mary Haley wrote:
>
>> Hi Guangshan,
>>
>> I was able to reproduce the difference you mentioned. However, it
>> seems to me that V5.2.0 is behaving more correctly, as the plot is
>> following the area of the lat/lon locations. I will double-check
>> with Dave Brown on this, but p<contour_plot_test_v511.png>lease see
>> the two attached PNG<contour_plot_test_v521.png>files (run with
>> V5.1.1 and the latest 5.2.1 internal version). The white space is
>> caused by the fact that there are no lat/lon values outside that
>> area.
>>
>> --Mary
>>
>> On May 26, 2010, at 4:12 PM, guangshan chen wrote:
>>
>>> Hi Youzuo,
>>>
>>> The data used and the script used are exactly same. Also the
>>> computer is same too. The difference is the version of NCL.
>>>
>>> The top figure in each plot is same. It means that
>>> "gsn_contour_map_ce" are same to plot two dimension data (lat,
>>> lon) in the two version.
>>>
>>> The middle and the low figures in each plot are different. It
>>> means that "gsn_contour_map_ce" are different to handle one
>>> dimension data using "res@sfXArray" and "res@sfYArray".
>>>
>>> Guangshan
>>>
>>>
>>> On May 26, 2010, at 4:45 PM, Li, Yongzuo wrote:
>>>
>>>> Guangshan,
>>>>
>>>> I have looked at your 52 1/2 plots.
>>>>
>>>> It seems that something is wrong with data interpolation.
>>>> The figures at the top for both 5.1 and 5.2 are good.
>>>>
>>>> Are you using exactly the same ncl script for 5.1 and 5.2?
>>>>
>>>> Yongzuo
>>>> ________________________________________
>>>> From: guangshan chen [gchen9@gmail.com]
>>>> Sent: Wednesday, May 26, 2010 3:29 PM
>>>> To: Li, Yongzuo
>>>> Subject: Re: bugs in NCL 5.2.0 for contour map plot
>>>> (guangshan chen)
>>>>
>>>> Yongzuo,
>>>>
>>>> Here are them.
>>>>
>>>> Guangshan
>>>>
>>>>
>>>> On May 26, 2010, at 2:53 PM, Li, Yongzuo wrote:
>>>>
>>>>> Guangshan,
>>>>>
>>>>> Put it in png format.
>>>>>
>>>>> Yongzuo
>>>>>
>>>>> ________________________________________
>>>>> From: guangshan chen [gchen9@gmail.com]
>>>>> Sent: Wednesday, May 26, 2010 2:45 PM
>>>>> To: Li, Yongzuo
>>>>> Cc: ncl-talk@ucar.edu
>>>>> Subject: Re: bugs in NCL 5.2.0 for contour map plot
>>>>> (guangshan chen)
>>>>>
>>>>> YongZuo,
>>>>>
>>>>> I don't have http website to put. Lower left corner is fine. I
>>>>> tried to attached the figures in the last email. But the size is
>>>>> bigger. I will try attached again.
>>>>>
>>>>> There are two files:
>>>>> contour_plot_test_v52data_v51plot.ps ------- plotted by V5.1.1
>>>>> contour_plot_test_v52data_v52plot.ps -------- plotted by V5.2.0
>>>>>
>>>>> You can use my data and script to plot the figures using two
>>>>> versions too.
>>>>>
>>>
>>> _______________________________________________
>>> 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 Thu May 27 11:22:25 2010