trouble with using template with forecast files

[Jennifer Adams]

On Oct 16, 2008, at 1:29 PM, Roz MacCracken wrote:

> Hi Jennifer,
>
> Thanks.  That worked for the one day...but, I have a dataset from
> Oct 1
> - 16 (actually, it's bigger than that...).
>
> I tried doing:
> TDEF 105 linear 00z01oct2008 6hr               (for   105 = 7 files
> per
> day * 15 days)
If your data is 6hourly, then your have four files per day times 15
days, plus the 00hr time step =  TDEF 61

> and that didn't work..."Undefined grid".
what did the gribmap -v output look like?

> What kind of TDEF do I use for a long time series of data?

Same as for a long forecast data set. There is only one time dimension.
Are you trying to make a long time series out of forecasts
initialized at different times?


> also, in my grads script, how do I indicate that I want to use
> forecast
> times?  (i.e., "set time blah?")
Just use 'set t 1' or 'set t 2' or 'set time 18z17oct2008'

We are working on a way to use forecast offset times in expression
handling, like this:
'display var(ftime=24hr)'  or 'display var(ft=2)'
but that is still in the design phase.

Jennifer


>
> Roz
>
> Jennifer Adams wrote:
>> Hi, Roz --
>> Your templated descriptor file should have the following entries:
>>
>> DSET ^%iy4%im2%id2%ih2.AWIPSF%f2.tm00.sg
>> OPTIONS template
>> TDEF 7 linear 00z14oct2008 6hr
>>
>> I feel it's worth the effort to make your metadata as accurate as
>> possible. Use an accurate number of time steps -- why say you have
>> 999
>> time steps when in reality you only have 6? Actually, the correct way
>> to set up this forecast data set is to put 7 in the TDEF entry
>> because
>> the 00hr time step, although not present in your list of data files,
>> is essential for resolving the %iy, %im, %id, and %ih substitution
>> strings properly in your file names. Plus it just makes sense to have
>> the first time step of your forecast data set be the initialization
>> time, even if you don't have any data. GrADS will just set the first
>> time step to missing -- you'll need to 'set t 2' before being able to
>> display a variable. Don't forget you must also rerun gribmap if you
>> change any of the grid definition or variable entries in your
>> descriptor file. Use the -v option with gribmap and check to make
>> sure
>> everything is matching properly for each data file.
>>
>> --Jennifer
>>
>>
>>
>>
>> On Oct 16, 2008, at 12:08 PM, Roz MacCracken wrote:
>>
>>> Hi All,
>>>
>>> I am trying to create a templated control file for a couple of
>>> forecast
>>> files.
>>>
>>> I have files with the name:
>>> 2008101400.AWIPSF06.tm00.sg - 2008101400.AWIPSF36.tm00.sg (there are
>>> 6 files
>>> total).  I have created control files called:
>>> eta6.awipsf06.ctl - eta6.awipsf36.ctl  (mainly because I can't
>>> get the
>>> %f2 option to work in my grads scripts.)
>>>
>>> Then, I created an index file called eta6.awipsf06.idx, using:
>>> gribmap -i eta6.awipsf06.ctl,
>>> which was created without errors.
>>>
>>> I can display data using the respective control file, pointing
>>> at the corresponding index file.  (this is without templating).  So,
>>> the control files do display data without templating.
>>>
>>> I have changed the dset line in the control file to:
>>> dset ^%y4%m2%d2%h2.AWIPSF06.tm00.sg, changed tdef to 999,
>>> and have added the "options template" to the control file.
>>>
>>> I have 2 questions..fairly easy, but, I can't get this to work...
>>>
>>> First, when I change the control file to the dset, and "options
>>> template"
>>> above, and "set t 1", I get the error, "Undefined Grid", with no
>>> data.
>>> Is there something that I've missed doing in my control file so that
>>> I can display data using a template?
>>>
>>> Should I have changed the dset to %y4%m2%d2%h2.AWIPSF%
>>> f2.tm00.sg?  And,
>>> if this is a better way, how do you set the date within a
>>> grads script to get it to display the forecasted time?
>>> (i.e., "set time 12Z14Oct2008(f=12)"...that may not be right...)
>>>
>>> Thanks.
>>>
>>> Roz MacCracken
>>>
>>> Sample of part of the control file:
>>>
>>> dset /ETA6/%y4%m2/%y4%m2%d2%h2.AWIPSF06.tm00.sg
>>> index /ETA6/ctl/maps/test.AWIPSF06.tm00.sg.idx
>>> undef 9.999E+20
>>> title 2008100100.AWIPSF06.tm00.sg
>>> *  produced by grib2ctl v0.9.12.5p41
>>> dtype grib 212
>>> options template
>>> pdef 185 129 lcc 35.0 -95.0 105 49 25 25 -95 40635 40635
>>> xdef 181 linear -140 0.5
>>> ydef  91 linear  15 0.5
>>> tdef 100 linear 00Z01oct2008 1dy
>>> zdef 1 linear 1 1
>>> vars 31
>>> ACPCPsfc  0 63,1,0  ** surface Convective precipitation [kg/m^2]
>>> ALBDOsfc  0 84,1,0  ** surface Albedo [%]
>>> APCPsfc  0 61,1,0  ** surface Total precipitation [kg/m^2]
>>> BGRUNsfc  0 234,1,0  ** surface Baseflow-groundwater runoff [kg/m^2]
>>> CAPEsfc  0 157,1,0  ** surface Convective Avail. Pot. Energy [J/kg]
>>> CFRZRsfc  0 141,1,0  ** surface Categorical freezing rain
>>> [yes=1;no=0]
>>> CICEPsfc  0 142,1,0  ** surface Categorical ice pellets [yes=1;no=0]
>>> CNWATsfc  0 223,1,0  ** surface Plant canopy surface water [kg/m^2]
>>> CRAINsfc  0 140,1,0  ** surface Categorical rain [yes=1;no=0]
>>> etc....
>>
>> --
>> Jennifer M. Adams
>> IGES/COLA
>> 4041 Powder Mill Road, Suite 302
>> Calverton, MD 20705
>> jma at cola.iges.org <mailto:jma at cola.iges.org>
>>
>>
>>

--
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/20081016/cdc249a1/attachment.html