[ncl-talk] climo_3.ncl + loop + conventional subscripts
Dennis Shea
shea at ucar.edu
Mon Nov 4 07:43:16 MST 2019
A quick overview of the script. It will 'work '
However, I would suggest you use
res at cnLevelSelectionMode = "ExplicitLevels"
and
res at cnFillMode = "RasterFill"
===
Also, your script will need to be changed to accommodate the precipitation
variable plot values
There are many examples of how to do this.
eg: *http://www.ncl.ucar.edu/Applications/HiResPrc.shtml*
<http://www.ncl.ucar.edu/Applications/HiResPrc.shtml>
On Sat, Nov 2, 2019 at 11:26 PM zoe jacobs <zoejacobs1990 at gmail.com> wrote:
> Dear Dennis,
> Many thanks for your help. I used the attached script. I just want to
> check with you to make sure that the script is correct. Would you please
> check it?
> Best wishes,
>
> On Mon, Oct 28, 2019 at 12:31 AM Dennis Shea <shea at ucar.edu> wrote:
>
>> The following is untested. However, it gives you the idea.
>> It is your responsibility to look up documentation and look at the
>> printed output.
>>
>>
>>
>> fils = systemfunc ("ls precip.*.nc") ; file paths
>> a = addfiles (fils, "r")
>> printVarSummary (a)
>>
>> ;************************************************
>> ; Read the file
>> ;************************************************
>>
>> ymdStrt = 19790101 ; climatology start
>> year/month/day
>> ymdLast = 20151231 ; " last
>> year/month/day
>> time = a[:]->time
>> yyyymmdd= cd_calendar(time, -2) ; note -2
>> print(yyyymmdd)
>> ntStrt = ind(yyyymmdd.eq.ymdStrt) ; start time index
>> ntLast = ind(yyyymmdd.eq.ymdLast) ; last time index
>> print("ntStrt="+ntStrt+" ntLast="+ntLast)
>> prc = a[:]->precip(ntStrt:ntLast,:,:) ; read precip
>> from all files
>>
>> printVarSummary (prc)
>> printMinMax (prc,0)
>> print("=====")
>> ; you decide what you need
>> prcMonth = *calculate_monthly_values(*
>> <http://www.ncl.ucar.edu/Document/Functions/Contributed/calculate_monthly_values.shtml>prc,
>> "avg", 0, False) ; monthly mean
>> ;;prcMonth = *calculate_monthly_values(*
>> <http://www.ncl.ucar.edu/Document/Functions/Contributed/calculate_monthly_values.shtml>prc,
>> "sum", 0, False) ; monthly total
>>
>> printVarSummary (prcMonth)
>> printMinMax (prcMonth,0)
>> print("=====")
>> opt = True ; examine data distribute
>> [information only]
>> opt at PrintStat = True
>> stat_prc = stat_dispersion(prcMonth, opt )
>>
>> prcClm = *clmMonTLL*
>> <http://www.ncl.ucar.edu/Document/Functions/Contributed/clmMonTLL.shtml>(prcMonth)
>> ; Compute monthly climatology
>> printVarSummary (prcClm)
>> printMinMax (prcClm,0)
>>
>> ====
>> Add the plot code.
>> Look carefully at: http://www.ncl.ucar.edu/Applications/HiResPrc.shtml
>>
>> EG: The following may be inappropriate for your dataset but this is the
>> idea.
>>
>> res at cnLevelSelectionMode = "ExplicitLevels"
>> res at cnLevels = (/ 0.01, 0.02, 0.04, 0.08, 0.16, \
>> 0.32, 0.64, 0.96/)
>> res at cnFillColors = (/"white","cyan", "green","yellow",\
>> "darkorange","red","magenta","purple",\
>> "black"/)
>>
>>
>> On Sun, Oct 27, 2019 at 2:48 PM Dennis Shea <shea at ucar.edu> wrote:
>>
>>>
>>> [0]
>>> As noted on ncl-talk many time the first rule of data processing is
>>> *Look at your data!"
>>> The temperature data were *monthly means* [K].
>>> The precipitation data are* daily* totals [mm/day]. How do I know? I
>>> looked at the README file at the URL. You should do that.
>>>
>>> [1]
>>> a =* addfiles*
>>> <http://www.ncl.ucar.edu/Document/Functions/Built-in/addfiles.shtml>
>>> (fils, "r")
>>>
>>> You have the following which is appropriate for an *addfile*
>>> <http://www.ncl.ucar.edu/Document/Functions/Built-in/addfile.shtml>
>>> reference
>>> time = a->time
>>>
>>> However, you are using *addfiles*
>>> <http://www.ncl.ucar.edu/Document/Functions/Built-in/addfiles.shtml> It
>>> should be:
>>> time = a*[:]*->time ; *[:]* syntax is appropriate for an
>>> *addfiles* reference
>>>
>>> [2]
>>> Did you read the documentation for *cd_calendar *
>>> <http://www.ncl.ucar.edu/Document/Functions/Built-in/cd_calendar.shtml>as
>>> I suggested? The whole point of using *cd_calendar* is to return
>>> 'human-readable' start and end times from which the actual start/stop time
>>> index values could be determined.. Your script uses the actual time values
>>> [ 692496 , 1016808 ] mixed with human-readable yyyymm. Further, you used
>>>
>>> yyyymm = cd_calendar(time, *-1*) ; -1 returns YYYYMM
>>>
>>> The precipitation files contain [as noted] daily values. You must use a
>>> human readable time like YYYYMMDD [eg: 19790101]. The us
>>>
>>> yyyymmdd = cd_calendar(time, *-2*) ; -2 returns YYYYMMDD
>>>
>>> This converts the actual time values [eg: 692496] to
>>> 'human-readable' yyyymmdd [year=>yyyy, month=>mm, day=dd>*
>>> 692496=>197901*01] values.
>>>
>>>
>>> ===
>>> Also, your script will need to be changed to accommodate the
>>> precipitation variable plot values
>>> There are many examples of how to do this.
>>>
>>> eg: *http://www.ncl.ucar.edu/Applications/HiResPrc.shtml*
>>> <http://www.ncl.ucar.edu/Applications/HiResPrc.shtml>
>>> ===
>>> In a previous email, you mentioned that you converted degrees Kelvin to
>>> C.
>>> It is best to self-document the variable with its new units.
>>>
>>> tmp = tmp-273.15
>>> tmp at units = "degC" ; this will be automatically plotted
>>> ===
>>> Good Luck
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Sat, Oct 26, 2019 at 8:19 AM zoe jacobs <zoejacobs1990 at gmail.com>
>>> wrote:
>>>
>>>> Dear all,
>>>> I would like to have climatology precipitation like temperature
>>>> climatology discussed through previous emails. However, there are
>>>> precipitation files for each year from 1979- 2015 (
>>>> ftp://ftp.cdc.noaa.gov/Datasets/cpc_global_precip/). In the first
>>>> step, I tried to merge all those files as one, and then use Dennis script
>>>> to plot a map. I faced some errors and need your advice please. the script
>>>> I am running has been attached.Many thanks in advance,
>>>> Best regardsn Sat, Oct 26, 2019 at 3:20 PM zoe jacobs <
>>>> zoejacobs1990 at gmail.com> wrote:
>>>>
>>>>> I just noticed that I could use res at gsnRightString = " deg C" .
>>>>> Once again, many thanks for your guide.
>>>>> Best wishes,
>>>>>
>>>>>
>>>>> On Fri, Oct 25, 2019 at 11:07 PM Dennis Shea <shea at ucar.edu> wrote:
>>>>>
>>>>>> I forgot to mention that plots can be done in 'portrait' [default] or
>>>>>> 'landscape' mode.
>>>>>>
>>>>>> You should experiment to see which is appropriate for your needs.
>>>>>>
>>>>>> resP at gsnPaperOrientation = "landscape" ; "portrait" is default
>>>>>>
>>>>>> do nmo=0,11 ; loop over the months
>>>>>> res at gsnLeftString = months(nmo)
>>>>>> plot(nmo) = gsn_csm_contour_map(wks,tmpClm(nmo,:,:), res) ;
>>>>>> create plot
>>>>>> end do
>>>>>>
>>>>>> if (resP at gsnPaperOrientation .eq."landscape") then
>>>>>>
>>>>>> gsn_panel(wks,plot,(/3,4/),resP) ; 3 rows
>>>>>> x 4 columns
>>>>>> else
>>>>>> gsn_panel(wks,plot,(/4,3/),resP) ; portrait
>>>>>> ; 4 rows x 3 columns
>>>>>> end if
>>>>>>
>>>>>> On Fri, Oct 25, 2019 at 12:15 PM Dennis Shea <shea at ucar.edu> wrote:
>>>>>>
>>>>>>> Hello,
>>>>>>>
>>>>>>> We do not usually do this but this is offline from ncl-talk.
>>>>>>>
>>>>>>> [1]
>>>>>>> Conventional subscripts start at 1 for Fortran and Matlab. Hence,
>>>>>>> for 12 months [nmos=12]
>>>>>>> Fortran: do nmo=1,nmos
>>>>>>> Matlab: for 1:nmos or 1:1:nmos [I am not a Matlab user so
>>>>>>> this is a guess.]
>>>>>>> NCL 'conventional' subscripts start at 0 [like C/C++/IDL/Python]
>>>>>>> NCL: do nmo=0,ntim-1
>>>>>>>
>>>>>>> [2] Dave illustrated 'coordinate subscripting' which uses the *{*...
>>>>>>> *}* syntax.
>>>>>>>
>>>>>>> The script I am attaching uses conventional subscripting. Just like
>>>>>>> Matlab/Fortran except it uses a range fro 0 to 11 [12 elements].
>>>>>>>
>>>>>>> [3] Please carefully read the documentation for *clmMonTLL*
>>>>>>> <http://www.ncl.ucar.edu/Document/Functions/Contributed/clmMonTLL.shtml>,
>>>>>>> *cd_calendar*
>>>>>>> <http://www.ncl.ucar.edu/Document/Functions/Built-in/cd_calendar.shtml>
>>>>>>> ,* ind*
>>>>>>> <http://www.ncl.ucar.edu/Document/Functions/Built-in/ind.shtml>
>>>>>>>
>>>>>>> [4] Look at the output from '*printVarSummary*
>>>>>>> <http://www.ncl.ucar.edu/Document/Functions/Built-in/printVarSummary.shtml>'.
>>>>>>> USE *printVarSummary *frequently.
>>>>>>> Note the dimensions and added attributes:
>>>>>>> time_op_ncl : Climatology: 29 years
>>>>>>> info : function clmMonTLL: contributed.ncl
>>>>>>>
>>>>>>> [5] NCL offer a large number of color tables [palettes]:
>>>>>>>
>>>>>>> *http://www.ncl.ucar.edu/Document/Graphics/color_table_gallery.shtml*
>>>>>>> <http://www.ncl.ucar.edu/Document/Graphics/color_table_gallery.shtml>
>>>>>>> Using a color map that emphasizes features can be very useful.
>>>>>>>
>>>>>>> [6] You must invest the time to learn any new language.
>>>>>>> Karin Meier-Fliesher and Michael Bottinger [DKRZ] wrote a wonderful
>>>>>>> tutorial.
>>>>>>>
>>>>>>> *http://www.ncl.ucar.edu/Document/Manuals/NCL_User_Guide/*
>>>>>>> <http://www.ncl.ucar.edu/Document/Manuals/NCL_User_Guide/>
>>>>>>>
>>>>>>> I suggest you read it.
>>>>>>>
>>>>>>> Good luck
>>>>>>> ==============
>>>>>>> *%>* ncl zoe_jacobs.hgcn_cams.ncl
>>>>>>>
>>>>>>> will produce a png file.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On Fri, Oct 25, 2019 at 11:16 AM Dave Allured - NOAA Affiliate via
>>>>>>> ncl-talk <ncl-talk at ucar.edu> wrote:
>>>>>>>
>>>>>>>> 1. In your example, please notice that the month loop has a stride
>>>>>>>> of 3. Therefore it is making plots for only four months: January, April,
>>>>>>>> July, and October.
>>>>>>>>
>>>>>>>> If you want all 12 months, remove the stride. Then make sure the
>>>>>>>> array "plot" is dimensioned 12. I think you can interchangeably use either
>>>>>>>> a 1-D or 2-D plot array (i.e. 3*4), your choice.
>>>>>>>>
>>>>>>>> The secondary index "i" is used to write the four plots into
>>>>>>>> positions 0, 1, 2, 3 in the "plot" graphics array. If you are making 12
>>>>>>>> plots, you do not really need to use a secondary index.
>>>>>>>>
>>>>>>>> Panel plotting uses a graphics array containing multiple plots.
>>>>>>>> Please study basic examples and documentation for making panel plots.
>>>>>>>>
>>>>>>>> 2. That data file has full coordinates. Therefore you can use
>>>>>>>> either conventional or coordinate subscripting, or mixed, your choice.
>>>>>>>>
>>>>>>>> I suggest using one of the date functions with coordinate
>>>>>>>> subscripting, to index the time subset that you want. Something like this:
>>>>>>>>
>>>>>>>> time_units = f->air&time at units
>>>>>>>> time1 = cd_inv_calendar (year1, 1, 1, 0, 0, 0, time_units, 0)
>>>>>>>> time2 = cd_inv_calendar (year2, 12, 31, 23, 0, 0, time_units, 0)
>>>>>>>> air_subset = f->air({time1:time2},:,:)
>>>>>>>>
>>>>>>>> Use printVarSummary to ensure that the subset has the dimension
>>>>>>>> sizes that you expect. Then you can proceed to compute the climatology.
>>>>>>>>
>>>>>>>>
>>>>>>>> On Fri, Oct 25, 2019 at 10:14 AM zoe jacobs via ncl-talk <
>>>>>>>> ncl-talk at ucar.edu> wrote:
>>>>>>>>
>>>>>>>>> Hi all,
>>>>>>>>> regarding climo_3.ncl (
>>>>>>>>> https://www.ncl.ucar.edu/Applications/Scripts/climo_3.ncl) ,
>>>>>>>>> I have 2 questions:
>>>>>>>>> 1. I need to plot all months on one panel (say 3*4), and cannot
>>>>>>>>> understand the logic behind the below loop, which used in the climo_3.ncl
>>>>>>>>> script :
>>>>>>>>>
>>>>>>>>> i = -1 ; Climatologies
>>>>>>>>> do nmo=0,11,3 ; loop over the months
>>>>>>>>> i = i+1
>>>>>>>>> res at gsnCenterString = months(nmo)+":"+time(0)/100 +"-"+ time(ntim-1)/100
>>>>>>>>> plot(i) = gsn_csm_contour_map(wks,prcClm(nmo,:,:), res) ; create plot
>>>>>>>>> end do
>>>>>>>>>
>>>>>>>>> How does it work??/
>>>>>>>>>
>>>>>>>>> 2. I would like to show climotology temperature from 1987- 2015. Data which I am using is air.mon.mean.nc ( https://www.esrl.noaa.gov/psd/data/gridded/data.ghcncams.html) and it is using conventional subscripts. I am not familiar with conventional subscripts. So how can I convert coordinate subscripting to conventional subscripts?!!
>>>>>>>>>
>>>>>>>>> Please kindly advice me .
>>>>>>>>>
>>>>>>>>> Many thanks in advance,
>>>>>>>>>
>>>>>>>>> Best regards,
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>> 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/20191104/5b50865e/attachment.html>
More information about the ncl-talk
mailing list