[gradsusr] Change of month

Jennifer Adams jma at cola.iges.org
Tue Sep 29 10:06:20 EDT 2015


When you have monthly data, GrADS is happiest when the hour and day in the TDEF statement are "00z01". Even if your data variable is an accumulation over the entire month, and those variables usually take the time stamp of the end of the accumulation period, everything in GrADS works better (for monthly data) if you use the first of the month in the date stamp instead. If your data is in grib format, use -b with gribmap. 


> Here's the problem.  Your first file TDEF is:
> 
> tdef 1 linear 18Z30sep2015 1mo

Yes, this is a problem. It should be “TDEF 1 linear 00z01sep2015 1mo"

> 
> and the second file TDEF is:
> 
> tdef 1 linear 00Z01oct2015 1mo
This is okay. 

> 
> So, when you open the first file, the first time value will be set to 1800 on Sept. 30, 2015.  The increment is define as one month (1mo), so t=2 would presumably be 1800 Oct. 30, 2015, t=3 same for Nov., and so on.  The combined file has:
> 
> tdef 30 linear 06Z29sep2015 6hr

There are two problems with this entry: the start time and the increment. This is monthly data, correct? 
Try “TDEF 30 linear 00z01sep2015 1mo” instead for the case where you are templating files together. But be careful about your file naming convention and your DSET entry — make sure you are templating over month, not forecast hour. If you need help figuring out templating, please have a look at http://iges.org/grads/gadoc/templates.html


Also, when diagnosing problems, please be sure to run gribmap in the verbose mode (-v) and make sure that all records in your data file are matching. If not, there will be some informative messages about why the record didn’t match. If you’re still having problems, please send the gribmap output along with your descriptor files. 

—Jennifer


> 
> Here, the initial time is 0600 on Sept. 29, 2015 and the increment is 6 hours, so the next time step should be 1200 on Sept. 29.  I suspect that your files either have too many, or too few, time steps.  Maybe the files contain forecast information that overlaps?  It looks like the names of the files are f066 then f072, so maybe 6 time-steps in each, which means (complete speculation here) that they have 6 time-steps in each for a total of 36 hours?  If true, the first file starts at 0600 Sept 29, the next file should start 1200 Sept 30?
> 
> Jim
> 
> On 9/29/15 2:32 AM, Miquel Bernis wrote:
>> Thanks for your instructions Jeff and James, I think I understood what you said, and I tried that way. It seems to be everything right, but when I open de .ctl and try to display a var, I get the error:
>> 
>> Cannot contour grid - all undefined values 
>> 
>> These are commands that I execute to create de template, and after that de .ctl header that results. I don't know what might the mistake.
>> 
>> Sorry for being so hopeless on that.
>> Miquel
>> 
>> >>>
>> 
>> #el convertim a .ctl
>> perl g2ctl.pl GFS025_f%f3.grb >GFS025.ctl
>> 
>> #creem arxiu .idf
>> gribmap -i GFS025.ctl
>> 
>> >>>
>> 
>> dset ^GFS025_f%f3.grb
>> index ^GFS025_f006.grb.idx
>> undef 9.999E+20
>> title GFS025_f006.grb
>> * produced by g2ctl v0.1.1
>> * command line options: GFS025_f%f3.grb
>> * griddef=1:0:(53 x 41):grid_template=0:winds(N/S): lat-lon grid:(53 x 41) units 1e-06 input WE:SN output WE:SN res 48 lat 35.000000 to 45.000000 by 0.250000 lon 355.000000 to 8.000000 by 0.250000 #points=2173:winds(N/S)
>> 
>> dtype grib2
>> options template
>> ydef 41 linear 35.000000 0.25
>> xdef 53 linear -5 0.250000
>> tdef 30 linear 06Z29sep2015 6hr
>> * PROFILE hPa
>> zdef 3 levels 100000 85000 50000
>> options pascals
>> vars 27
>> 
>> 
>> 
>> 2015-09-29 1:23 GMT+02:00 Jeff Duda <jeffduda319 at gmail.com>:
>> As James mentioned and I suspected, doing this one file at a time is harder than using templates. What I would do is download all the GFS data files, then use one single control file to open all of them. Use templating, as described here. Then adjust the TDEF line of the control file accordingly. The last thing to do is to run gribmap on that control file.
>> 
>> Jeff
>> 
>> On Mon, Sep 28, 2015 at 5:48 PM, Miquel Bernis <bernisfeina at gmail.com> wrote:
>> Thanks again Jeff and James. 
>> 
>> I've pasted here a pair of my .ctl file example headers. After that, you have the commands I use for downloading the GFS 0.25 data, and for the conversion to .ctl. As you can see I use the perl g2ctl. 
>> 
>> I really appreciate your help. 
>> Miquel  
>> 
>> >>>>
>> 
>> dset ^GFS025_f066.grb
>> index ^GFS025_f066.grb.idx
>> undef 9.999E+20
>> title GFS025_f066.grb
>> * produced by g2ctl v0.1.1
>> * command line options: GFS025_f066.grb
>> * griddef=1:0:(53 x 41):grid_template=0:winds(N/S): lat-lon grid:(53 x 41) units 1e-06 input WE:SN output WE:SN res 48 lat 35.000000 to 45.000000 by 0.250000 lon 355.000000 to 8.000000 by 0.250000 #points=2173:winds(N/S)
>> 
>> dtype grib2
>> ydef 41 linear 35.000000 0.25
>> xdef 53 linear -5 0.250000
>> tdef 1 linear 18Z30sep2015 1mo
>> * PROFILE hPa
>> zdef 3 levels 100000 85000 50000
>> options pascals
>> vars 27
>> ....
>> 
>> >>>>
>> dset ^GFS025_f072.grb
>> index ^GFS025_f072.grb.idx
>> undef 9.999E+20
>> title GFS025_f072.grb
>> * produced by g2ctl v0.1.1
>> * command line options: GFS025_f072.grb
>> * griddef=1:0:(53 x 41):grid_template=0:winds(N/S): lat-lon grid:(53 x 41) units 1e-06 input WE:SN output WE:SN res 48 lat 35.000000 to 45.000000 by 0.250000 lon 355.000000 to 8.000000 by 0.250000 #points=2173:winds(N/S)
>> 
>> dtype grib2
>> ydef 41 linear 35.000000 0.25
>> xdef 53 linear -5 0.250000
>> tdef 1 linear 00Z01oct2015 1mo
>> * PROFILE hPa
>> zdef 3 levels 100000 85000 50000
>> options pascals
>> vars 27
>> ....
>> 
>> >>>>
>> 
>> URL="http://nomads.ncep.noaa.gov/cgi-bin/filter_gfs_0p25.pl?file=gfs.t${fhr}z.pgrb2.0p25.f${hr}&lev_10_m_above_ground=on&lev_2_m_above_ground=on&lev_500_mb=on&lev_850_mb=on&lev_1000_mb=on&lev_entire_atmosphere=on&lev_high_cloud_layer=on&lev_low_cloud_layer=on&lev_middle_cloud_layer=on&lev_surface=on&var_APCP=on&var_CAPE=on&var_CFRZR=on&var_CICEP=on&var_CRAIN=on&var_CSNOW=on&var_DPT=on&var_GUST=on&var_HGT=on&var_LFTX=on&var_PRATE=on&var_RH=on&var_TCDC=on&var_TMAX=on&var_TMIN=on&var_TMP=on&var_UGRD=on&var_VGRD=on&subregion=&leftlon=-5&rightlon=8&toplat=45&bottomlat=35&dir=%2Fgfs.${date}${fhr}"
>> 
>> # download file
>> curl "$URL" -o GFS025_f"$hr".grb
>> sleep 1
>> 
>> #convert to .ctl
>> perl g2ctl.pl GFS025_f"$hr".grb > GFS025_f"$hr".ctl
>> 
>> #create .idf file
>> gribmap -i GFS025_f"$hr".ctl
>> 
>> 
>> 2015-09-28 21:22 GMT+02:00 James T. Potemra <jimp at hawaii.edu>:
>> Miquel,
>> 
>> We'd need to see you control files, but I suspect the TDEF line is causing problems.  I believe this gets set with your first open file.  For example, if you open a file and "set t 1", time will get set to some date as defined in that control file.  If the subsequent files don't have this date defined you'll get the grid coordinate error.  You could issue a "set dfile" command to re-set the coordinates.
>> 
>> Jim
>> 
>> 
>> On 9/28/15 3:12 PM, Miquel Bernis wrote:
>>> Thanks a lot for your answer Jeff. I'm not sure how can I give more clues about the error. There's no problem when I work with the files separately, only when I operate between files in a different month. I realized that there's something wrong about the time vars. When I open consecutive files 24z, 30z, 36z, 42z... the first one has the right time value, but the next ones shows a change of month, instead of just a 6h advance:
>>> 
>>> ga-> set t 1
>>> Time values set: 2015:9:30:6 2015:9:30:6 
>>> ga-> set t 2
>>> Time values set: 2015:10:30:6 2015:10:30:6 
>>> ga-> set t 3
>>> Time values set: 2015:11:30:6 2015:11:30:6 
>>> ga-> set t 4
>>> Time values set: 2015:12:30:6 2015:12:30:6 
>>> ga-> 
>>> 
>>> If open the files separately there's no time problem. Kind of weird.
>>> 
>>> Anyway, you suggested:
>>> 
>>> "I would just use templates and do the sum over different time index values"
>>> 
>>> I'm sorry, I don't understand that. Maybe you can tell the same in another way or with more details. It would be helpful.
>>> 
>>> Thanks a lot again.
>>> Miquel
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> , 
>>> 
>>> 2015-09-28 17:51 GMT+02:00 Jeff Duda <jeffduda319 at gmail.com>:
>>> Perhaps some more information on the specifics of the problem would help us to troubleshoot. However, if you're summing precip totals from a single GFS run, I would just use                                                   templates and do the sum over different time index values.
>>> 
>>> Jeff Duda
>>> 
>>> On Mon, Sep 28, 2015 at 4:19 AM, Miquel Bernis <bernisfeina at gmail.com> wrote:
>>> Hi everyone,
>>> 
>>> I'm stucked in a problem that might be simple to fix, maybe any of you could help me. The point is that I can't operate with files when there's change of month. I'm working with GFS 0.25, and when I try to sum the rain of the last file of month with the rain of the first file of the moth ahead, it returns and error like this:
>>> 
>>> Invalid grid coordinates
>>> World coordinates convert to non-integer  grid coordinates
>>> Variable = apcpsfc  Dimension = 3 
>>> 
>>> Has any of you a solution for that? 
>>> Thank you very much.
>>> I'll really appreciate your help.
>>> 
>>> Miquel
>>> 
>>> _______________________________________________
>>> gradsusr mailing list
>>> gradsusr at gradsusr.org
>>> http://gradsusr.org/mailman/listinfo/gradsusr
>>> 
>>> 
>>> 
>>> 
>>> -- 
>>> Jeff Duda
>>> Graduate research assistant
>>> University of Oklahoma School of Meteorology
>>> Center for Analysis and Prediction of Storms
>>> 
>>> _______________________________________________
>>> gradsusr mailing list
>>> gradsusr at gradsusr.org
>>> http://gradsusr.org/mailman/listinfo/gradsusr
>>> 
>>> 
>>> 
>>> 
>>> _______________________________________________
>>> gradsusr mailing list
>>> gradsusr at gradsusr.org
>>> http://gradsusr.org/mailman/listinfo/gradsusr
>> 
>> 
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
>> 
>> 
>> 
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
>> 
>> 
>> 
>> 
>> -- 
>> Jeff Duda
>> Graduate research assistant
>> University of Oklahoma School of Meteorology
>> Center for Analysis and Prediction of Storms
>> 
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
>> 
>> 
>> 
>> 
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
> 
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr

--
Jennifer M. Adams
Center for Ocean-Land-Atmosphere Studies (COLA)
111 Research Hall, Mail Stop 2B3
George Mason University
4400 University Drive
Fairfax, VA 22030 





-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20150929/94396245/attachment-0001.html 


More information about the gradsusr mailing list