Re: multi netcdf

[verschel@xxxxxxxxxxxxxxxxxxxxx (Mark A. Verschell)]

Joerg,

This should work pretty easy (try having different length timesteps some time!)
By the way, using seconds could get you in trouble due to very large numbers.
Since you are using 21600 as your time step, may I suggest switching to hours,
then your delta-T will be 6, a more manageable number.

In your example, your files should of course be named something like:

psib_1985.pbp2.nc.001 - January
psib_1985.pbp2.nc.002 - February
         :
psib_1985.pbp2.nc.012 - December

And then the step files will all have to be seperately entered (since you will
have a different number of timesteps per file. I will assume that you are
making a climatological year (year 1) for this example. (What I did to get
S_START and S_END was to create an axis/grid in ferret and then show grid/l,
i.e., define axis/t="1-jan-0001":"31-dec-0001:18"/np=1460/unit=seconds myt):

 $STEPFILE_RECORD
   S_FILENAME      = 'psib_1985.pbp2.nc.001',
   S_AUX_SET_NUM   = 0,
   S_START         = 31622400.,
   S_END           = 34279200.,
   S_DELTA         = 21600.,
   S_NUM_OF_FILES  =   1,
 $END
 $STEPFILE_RECORD
   S_FILENAME      = 'psib_1985.pbp2.nc.002',
   S_AUX_SET_NUM   = 0,
   S_START         = 34300800.,
   S_END           = 36698400.,
   S_DELTA         = 21600.,
   S_NUM_OF_FILES  =   1,
 $END

          :

 $STEPFILE_RECORD
   S_FILENAME      = 'psib_1985.pbp2.nc.012',
   S_AUX_SET_NUM   = 0,
   S_START         = 60480000.,
   S_END           = 63136800.,
   S_DELTA         = 21600.,
   S_NUM_OF_FILES  =   1,
 $END


See what I mean about the large number? If you used hours instead (or even
days) the S_START and S_END would then be: 8784,9522 (or 366,396.75).

Mark Verschell

On Dec 1,  1:09pm, =?iso-8859-1?Q?J=F6rg?= Kaduk wrote:
> Subject: multi netcdf
>
> [ text/plain
>   Encoded with "quoted-printable" ]:
> Hi there,
>
> I have a question about constructing a multi-netcdf file:
>
> I have twelve monthly files with 4 data pre day in them, so
> there are 124 data per variable in the Jan file, 112 in the
> Feb file, 124 in the March file, and so on.
> My time unit is the second. So the data are every 21600 seconds.
> How does the descriptor file have to look like for ferret
> to read the data ok, i.e. what should there be in the
> stepfile record? I've tried a couple of things, but somehow
> I do not get it yet.
> Any hints greatly appreciated,
> thanks!
> Joerg
>
>
>   $STEPFILE_RECORD
>    s_filename           = 'psib_1985.pbp2.nc.009',
>    S_AUX_SET_NUM        = ,
>    S_START              = ,
>    S_END                = ,
>    S_DELTA              = ,
>    S_NUM_OF_FILES       = 1,
>    S_REGVARFLAG         = ' ',
>-- End of excerpt from =?iso-8859-1?Q?J=F6rg?= Kaduk




-- 
 Mark Verschell                     NASA/Goddard Space Flight Center
 verschell@neptune.gsfc.nasa.gov    Code 970/Lab. for Hydrospheric Proc.
 301-614-5669  Fax: 301-614-5666    Greenbelt, MD 20771


		     Better Living Through Denial
               ** USQC Certified: 100% Microsoft Free **