[ncl-talk] Too many missing values in 'trend_manken' precipitation output..

[Rashed Mahmood]

Hi Qudsia,
If you set zero as missing value, then *it is obvious that you would get
missing values at grid points that has 0 Sen's slope* (i.e. NO TREND).

However, I do not think that you should explicitly set 0 as missing value.
If you do not want to show 0 values in the map, you can choose colormaps
that contain white in between. e.g.:
https://www.ncl.ucar.edu/Document/Graphics/ColorTables/ViBlGrWhYeOrRe.shtml

You may need to play with setting resources, e.g:
   res at cnLevelSelectionMode  = "ExplicitLevels"
   res at cnLevels                     =   set your contour levels here.


If you still have problems then please provide full script along with shape
files.

Cheers,
Rashed











On Mon, Oct 21, 2019 at 10:00 AM qudsia zafar <missquaddus at gmail.com> wrote:

> Hi Rashed,
>
> As i am masking the data over the map, i am setting 'zero' as a missing
> value. This shows me that there are also zeroes in Sen's estimate output
> (which i have already set to missing with  attribute _FillValue = 0).
>
> Perhaps you could also check with you by putting zero as a missing value
> (as is also visible from your output).
>
> Attached is what i get from CRU when i don't put zero as a missing value
> (after masking)
>
> best,
>
> qudsia
>
>
>
> On Mon, Oct 21, 2019 at 12:34 PM Rashed Mahmood <rashidcomsis at gmail.com>
> wrote:
>
>> Hi Qudsia,
>> Please see the attached plot, I did not get any missing values for the
>> specified data range. I do not have your shape file. The script is also
>> attached.
>>
>> What version of NCL you are using?
>> you can check ncl version using command:
>>
>> ncl -V
>>
>> I am using 6.6.2, which is latest. It is always better to use latest
>> version of NCL or any other software, for that matter.
>>
>> Cheers,
>> Rashed
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> On Mon, Oct 21, 2019 at 8:17 AM qudsia zafar <missquaddus at gmail.com>
>> wrote:
>>
>>> Hi Rashed,
>>>
>>> Thank you for your quick response. As you suggested i have downloaded
>>> the latest CRU (1901-2018) data file and have tried the script you
>>> attached. However, still there are missing data in the south east and south
>>> west of the map (kindly see attached).
>>>
>>> More over i have also tried the trend test with CORDEX data, which gives
>>> me the same issue of missing trend data when analyzing precipitation data
>>> (kindly see attached outputs from two CORDEX experiments.
>>>
>>> The issue of missing data does not arise in case of temperature data.
>>>
>>> Any further help would be appreciated.
>>>
>>> best,
>>>
>>> qudsia
>>>
>>> On Fri, Oct 18, 2019 at 5:51 PM Rashed Mahmood <rashidcomsis at gmail.com>
>>> wrote:
>>>
>>>> Hi Qudsia,
>>>> The data file that you are using has issues and this data is no longer
>>>> distributed by the owners, here is the page that mentions this:
>>>> https://crudata.uea.ac.uk/cru/data/hrg/
>>>> Scroll down until you find CRU TS v. 3.24.
>>>>
>>>> The latest data is CRU TS v.4.03, you can download it here:
>>>>
>>>> https://crudata.uea.ac.uk/cru/data/hrg/cru_ts_4.03/cruts.1905011326.v4.03/pre/
>>>> Select the file "cru_ts4.03.1901.2018.pre.dat.nc.gz
>>>> <https://crudata.uea.ac.uk/cru/data/hrg/cru_ts_4.03/cruts.1905011326.v4.03/pre/cru_ts4.03.1901.2018.pre.dat.nc.gz>
>>>> " and download it and unzip it.
>>>>
>>>> Attached is a script that would do exactly the same as you were trying
>>>> to do (except i did not plot the results because you can do it yourself).
>>>>
>>>> It is better to define variables for time and lats/longs as you can see
>>>> in the attached script. This would save you time once you want to change to
>>>> different time period, for example.
>>>>
>>>> Let me know if you still find issues.
>>>>
>>>> Good Luck.
>>>> Rashed
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Fri, Oct 18, 2019 at 3:27 AM qudsia zafar via ncl-talk <
>>>> ncl-talk at ucar.edu> wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I am trying to plot spatial sen's estimate (1975-2005) from CRU
>>>>> precipitation data using Mann-Kendall trend function:
>>>>>
>>>>>
>>>>> f0 = addfile(dir+ cru_ts3.24.1901.2015.pre.dat.nc_1,"r")
>>>>>
>>>>> precip1 = doubletofloat(f0->pre(888:1259,{23:38},{60:78}))
>>>>> precip_trend = trend_manken(precip1, False, 0)
>>>>>
>>>>> print(precip_trend(1,:,:))     ; print 'trend' values
>>>>>
>>>>> ...
>>>>> ...
>>>>> ...
>>>>>
>>>>> map = gsn_csm_contour_map(wks, precip_trend(1,:,:),res)
>>>>>
>>>>> However i get too many missing values in the precipitation trend
>>>>> output (kindly see attached). i would highly appreciate if the problem is
>>>>> figured out.
>>>>>
>>>>> best,
>>>>>
>>>>> qudsia
>>>>>
>>>>> _______________________________________________
>>>>> 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/20191021/e0e14c99/attachment.html>