Problem in sst plot

Jennifer Adams jma at COLA.IGES.ORG
Fri May 15 14:52:41 EDT 2009


Wacky!
My earlier sample was grib, but I get the same good results with the  
grib2 version of this data.... on my mac. When I tried it on my 64-bit  
linux box, I got an error I've never seen before:

ga-> d sst
  jpc_decode return = 0
Constant field.  Value = 271.33
ga->

That "jpc_decode" message is in the grib2 library in the  
dec_jpeg2000.c souce file. I have no idea what it means. Or where it  
got the 271.33 number. I will have to poke at it further. Could be a  
classic case of operator error. In the meanwhile ... Can you try using  
the grib1 version of this data instead? Go back to the original FTP  
site that you started with:  ftp://polar.ncep.noaa.gov/pub/sst/ophi/

Jennifer
p.s.  Why do these puzzles show up on Friday afternoons?



On May 15, 2009, at 2:27 PM, Fil wrote:

> Hi Jennifer,
>
> With 0.083333 in the ctl it does shows "LON set to 0 360". Maybe  
> it's something I'm doing wrong but even then it won't draw lon < 0º.  
> I'm sending the files I'm using for testing. The grib can be  
> downloaded from:
>
> http://www.ftp.ncep.noaa.gov/data/nccf/com/gfs/prod/sst.20090514/rtgssthr_grb_0.083.grib2
>
> Thanks.
>
> 2009/5/15 Jennifer Adams <jma at cola.iges.org>
> Fil,
> I have downloaded a data sample and done some testing. When you open  
> your descriptor file with GrADS, look at the info written to the  
> command window. Does it say "LON set to 0.042 359.944" or "LON set  
> to 0 360"?
>
> The answer will depend on whether you have enough significant digits  
> in the grid increment value in the XDEF and YDEF statements in your  
> descriptor file. When I use "0.083333" instead of "0.083", then  
> GrADS calculates that the file does indeed 'wrap' the globe and 'set  
> lon -180 180' works as it should.
>
> Jennifer
>
>
> GrADS automatically issues a 'set lon 0 360' command when it opens a  
> file that wraps the globe. It
>
>
> On May 15, 2009, at 1:31 PM, Fil wrote:
>
>> You are right, I swapped the results. It seems that for some reason  
>> grads won't draw longitude < 0º with 'set lon' starting with a  
>> negative number. This won't be a problem (drawing first lon > 0º  
>> and then lon < 0º) except for the robinson projection which has to  
>> start with -180.
>>
>> Adding more digits unfortunately didn't work.
>>
>> Thank you very much for your help.
>>
>> 2009/5/14 Diane Stokes <Diane.Stokes at noaa.gov>
>> Sorry that didn't fix all cases for you.  It worked for me running  
>> grads
>> v1.9b4 and v2.0.a3 on IBM AIX.
>>
>> I'm confused by your result with:
>>  'set lon -180 180'
>> you only see -180º to 0º.  And with:
>>  'set lon 180 540'
>> you only see 0º to 180º.  Do you have those results swapped in the  
>> text
>> of your msg?
>>
>> If you don't get advice from anyone else, I could only suggest you  
>> add a
>> couple more significant digits (3's in this case) to your increment  
>> values.
>>
>>   Good luck.
>>      Diane
>>
>>
>> Fil wrote:
>> Hi, Diane
>>
>> Thanks a lot for your help, your solution works perfectly for polar
>> stereographic projections. But on other projections it still doesn't
>> show anything from -180º to 0º longitude. If I try to plot just for
>> North America, I get an "Entire Grid Undefined" message.
>>
>> With 'set lon -180 180' it only draws from -180º to 0º. With lon  
>> 180 540
>> only draws from 0º to 180º. With lon -360 0 it won't draw anything.  
>> And
>> with 0 360 it will draw the entire globe, but with the problem (for  
>> me)
>> of getting centered in lon 180º which cuts Europe in half.
>>
>> Thanks.
>>
>> 2009/5/13 Diane Stokes <Diane.Stokes at noaa.gov
>> <mailto:Diane.Stokes at noaa.gov>>
>>
>>
>>    Hi, Filipe.
>>
>>    I recently ran into the same problem with a different dataset on  
>> the
>>    same grid.
>>
>>    The spatial resolution the data you are viewing is 1/12 of a  
>> degree.
>>    The increment in the xdef and ydef statements of your control  
>> file is
>>    0.083.  (That's how the increment is stored in the grib file, so  
>> it's
>>    not a result of anything you did wrong).
>>
>>    That increment value is not sufficient to wrap the data completely
>>    around the globe.  You have 4320 points in your xdef:
>>     4321 x 0.083 = 358.643
>>
>>    If you zoom in on the north pole, you'll see the data displayed  
>> does not
>>    extend far enough north either.
>>
>>    If you add a few significant digits to the increment values in  
>> your xdef
>>    and ydef statements, eg:
>>
>>     xdef 4320 linear 0.042000 0.083333
>>     ydef 2160 linear -89.958000 0.083333
>>
>>    you'll get a more accurate mapping of the data.  The values  
>> above will
>>    also bring the left and right edge of the grid sufficiently  
>> close that
>>    GrADS will know that the data does wrap the globe.  You'd then  
>> be able
>>    to display the field when setting longitude values less than 0 or
>>    greater than 360.
>>
>>      Diane
>>
>>
>>
>>    Filipe wrote:
>>
>>        Hello all,
>>
>>        I'm trying to plot sst data from
>>        ftp://polar.ncep.noaa.gov/pub/sst/ophi/ ,
>>        but for some reason I get a white strip at longitude 0º when
>>        using nps
>>        projection, as you can see from this pic:
>>
>>        http://img8.imageshack.us/img8/2815/userlist2.png
>>
>>        When using any other projection, grads only draws from lon >  
>> 0º,
>>        as you can
>>        see from these pics:
>>
>>        http://img8.imageshack.us/img8/5487/userlist1.png
>>        http://img8.imageshack.us/img8/7201/userlist3.png
>>
>>        I don't know what I'm doing wrong, hope someone can help me.  
>> I'm
>>        using
>>        latest grib2ctl.pl, wgrib v1.8.0.12x and grads 2.05a (I also
>>        tested v1.9 and
>>        opengrads 2.05). I used the script http://pastebin.com/f3dc65253
>>
>>        Best regards.
>>
>>
>>
>>
>> --
>> Fil - MeteoPT.com
>>
>
> --
> Jennifer M. Adams
> IGES/COLA
> 4041 Powder Mill Road, Suite 302
> Calverton, MD 20705
> jma at cola.iges.org
>
>
>
>
> <grads.gs><sst.ctl><sst.png>

--
Jennifer M. Adams
IGES/COLA
4041 Powder Mill Road, Suite 302
Calverton, MD 20705
jma at cola.iges.org



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20090515/ebb704dc/attachment.html 


More information about the gradsusr mailing list