[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20010216: gdlist question

>From: address@hidden (Chris Hennon)
>Organization: UCAR/Unidata
>Keywords: 200102161840.f1GIeUL27371

>Steve -
>
>Perhaps you can clear up a bit of confusion I'm having with gdlist.  I
>have a NCEP Reanalysis gempak formatted file (global domain), and I'm
>trying to list out the data for the domain 0;-100;50;-1.5  :
>
>Creating process: gplt for queue 2226552
> GDATTIM   Grid date/time                    980601/0000
> GLEVEL    Grid level                        500
> GVCORD    Grid vertical coordinate          pres
> GFUNC     Scalar grid                       tmpk
> GDFILE    Grid file                         tempk_199806.gem
> GAREA     Graphics area                     0;-100;50;-2.5
> PROJ      Map projection/angles/margins|dr  CED
> SCALE     Scalar scale / vector scale       0
> OUTPUT    Output device/filename            t
>
>According to my hand calculations, it should give me back rows 57 thru 37 
>(21 rows), since the dataset has 2.5 degree resolution.  Instead it
>returns rows 58 thru 36 (23 rows).  

If you use GFUNC=mul(latr,rtd) it shows that it is giving you rows for 
52.5 to -2.5 degrees latitude.  The $GEMPAK/source/gemlib/dg/dgarea.f 
subroutine is adding 1 "EXTEND" distance to all sides for the subset
area. Without doing this, you would not be able to compute values
if you gfunc() involved a dirivative adjacent to a subset boundary (that is,
when the GAREA is set, the computations only occur on the subset grid,
so the subset grid must include the adjacent rows in order to compute
dirivaties in functions).

In this case, you are using GFUNC=tmpk which doesn't involve a gradient,
but if you used gfunc = avor(wnd) or grad(tmpk), etc you would need the
extra row  in order to use a centered finite difference formula instead
of a left or right finite difference. This is more important in gdcntr etc
where your contours go up to the boundary.


>Furthermore, the data values in row 58
>do not match the data values in row 58 when I use GAREA = dset.  What
>exactly is gdlist doing here?  Thanks.


They appear the same too me.... however, you may be looking in the wrong 
location.

When you use your garea, you are gettings COLUMNS 104 to 144. When you use 
GAREA=dset, you are getting COLUMNS 1 to 144. So the value that is at 
ROW 58 column 104 in the output with your subset garea is the first value
listed. But, when garea=dset, it is at the end of the 13th line of
output for ROW 58.

Let me know if your output does not match this!


>
>Chris
>
>================================================
>| Chris Hennon        Ohio State University   |
>| Tropical Meteorology      address@hidden   |
>|                                              |
>| Dept of Geography   Office: 1155 Derby Hall  |
>| 1036 Derby Hall     Phone : (614) 292-2704   |
>| Columbus, OH 43210  Fax   : (614) 292-6213   |
>================================================
>


Steve Chiswell