Re: white line at longitude 0

Chao,

I had a discussion with Adam Phillips about this, and he suggested
trying the lonFlip function:

http://www.ncl.ucar.edu/Document/Functions/Contributed/lonFlip.shtml

You need to load the "contributed.ncl" script to use this function:

load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/contributed.ncl"

If this doesn't fix things, then, if you are running version 4.3.1 of
NCL, you can try a new resource that I created, gsnDebugWriteFileName.

Set this resource to some string for just *one* of your plots:

   res = True
   ...
   res_at_gsnDebugWriteFileName = "test"
   ...
   plot= gsn_csm_contour_map(wks,...,res)

   delete(res_at_gsnDebugWriteFileName)
   ...

What will happen is that the contour data (and any coordinate
arrays) will be written to a netcdf file called "test.nc",
and an NCL script with all the resources you are setting
will be created and called "test.ncl".

You can then send me "test.ncl" and "test.nc", and I can run your
script here with the same data and plot resources that you are using.

--Mary

On Mon, 24 Sep 2007 cluo_at_uci.edu wrote:

> Hi,
>
> I just try anthoer bw plot. There is same thing with vertical white line
> at center (a gap is not plotted at the Greenwich Meridian), with the
> .hluresfile in my home directory and put res_at_gsnAddCyclic function. I
> don't know what caused this problem. This problem is nothing to do with
> NCL?
> Please see my another plot:
> http://dust.ess.uci.edu/chaoluo/fig7_bw.pdf
>
> Thanks,
> Chao
>
>>
>>
>> cluo_at_uci.edu wrote:
>>> Hi all,
>>>
>>> I have a troble to make the black and white plot. There is always a
>>> white
>>> line at longitude 0, from latitude -90 to +90. I have no idea how to do
>>> it.
>>> Here's URLs of my plot and ncl script.
>>> http://dust.ess.uci.edu/chaoluo/fig10_bw.pdf
>>> http://dust.ess.uci.edu/chaoluo/test_bw.ncl
>>>
>>> Thanks in advance for any help or suggestions!
>>
>> [1] Are you talking about the horizontal "white line"
>> at the top and bottom of the plot? If so, then
>> I speculate your grid does not extend to the poles.
>> eg: a gaussian grid, or one that spans (say)
>> -89 to +89.
>> In these cases there are no data at the poles.
>>
>> [2] I think it has been mentioned before, please put the
>> following .hluresfile in your home directory.
>> Amongst other things it makes helvetica the default
>> font.
>>
>> http://www.ncl.ucar.edu/Document/Graphics/hlures.shtml
>>
>> [2] Why are you doing
>>
>> do ilev=0,27
>> do ilat=0,93
>> do ilon=0,191
>> deltp(ilev,ilat,ilon)=(hyblev(ilev+1)-hyblev(ilev))*sufp(ilat,ilon)
>> end do
>> end do
>> end do
>>
>> There is a function for this:
>>
>>
>> http://www.ncl.ucar.edu/Document/Functions/Built-in/dpres_hybrid_ccm.shtml
>>
>> It is fast and you do *not* need to preallocate the array "deltp"
>>
>> [3] General rule for programming .... do not hardwire sizes.
>> In your do loop:
>>
>> do ilev=0,klvl-1
>> do ilat=0,nlat-1
>> do ilon=0,mlon-1
>>
>> [4] Use NCL's array syntax it is much faster and cleaner
>>
>> replace
>>
>> do ilev=0,27
>> do ilat=0,93
>> do ilon=0,191
>> bcwet_tot(ilat,ilon)=bcwet_tot(ilat,ilon)+bcwet(ilev,ilat,ilon)
>> end do
>> end do
>> end do
>>
>> with
>> do ilev=0,klvl-1
>> bcwet_tot = bcwet_tot + bcwet(ilev,:,:)
>> end do
>>
>>
>>
>>>
>>> Chao
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> ncl-talk_at_ucar.edu
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk_at_ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Sep 24 2007 - 13:49:43 MDT