Compare diurnal cycles

[Mike Bosilovich]

Hi Oliver,

It should be possible, but you didn't say what format the data is in.
If it is flat binary with a ctl file, make a new ctl with your center
point time instead of the report time (in other words, in the TDEF
line, change 00Z to 01:30Z, and in the other 00Z to 00:30Z). It all
should work, unless you have the hour templated in the file name.

If that is the case, or if you have grib or a self defined format, it
may not be possible to center the average directly. I haven't tried
changing the time in an DDF file, so I'm not sure that that would
work. One way would be to write a binary with lats4d (lats4d -i infile
-o outfile -format stream), then create the ctl with the center time
point.

The lats4d option may not work for single point time series in a self
defined format. It typically hasn't liked -lat and -lon being points.

I suggest creating a new ctl so that in the future, after you haven't
used the data in a while, it will be apparent in your scripts which
uses the centered average, versus the reported time.

Good Luck,

Mike

On Mon, Jul 28, 2008 at 10:56 AM, Oliver Fuhrer
<Oliver.Fuhrer at meteoswiss.ch> wrote:
>
> Hi all!
>
> I have two variables in Grads containing the following diurnal cycles...
>
> flux1 (defined at 00Z, 03Z, 06Z, 09Z, 12Z, 15Z, 18Z, 21Z)
> flux2 (defined at 00Z, 01Z, 02Z, ..., 23Z)
>
> f1 at 00Z is the mean flux from 00Z to 03Z.
> f2 at 00Z is the mena flux from 00Z to 01Z.
>
> Now when I use Grads to plot the two on the same line chart, they are
> phase-shifted in time, since the flux1 at 00Z should be plotted at
> 01:30Z and the flux2 at 00Z should be plotted at 00:30Z, but they are
> not.
>
> How can I get Grads to time shift the two variables?
>
> Thanks for you help!
> Oliver
>
>