Specifying Z levels

[Ricardo Hallak]

Dear Andrea,

if I understood your problem correctly, you can try this:

if your vertical coordinate z is dependent on (x,y), there is no way to plot
exactily your data in GrADS. You can do it only by approximation on GrADS or
try use NCAR Graphics. On the other hand, if your vertical coordinate depends
only on z, you can write a .ctl file where de ZDEF VALUES are explicit, as the
example bellow:

ZDEF   40  LEVELS 50.00    100.00   250.00   450.00   750.00   1250.00
2000.00   3750.00   4250.00   4750.00   5250.00   5750.00   6250.00   6750.00
  7250.00   7750.00   8250.00   8750.00   9250.00   9750.00  10250.00
10750.00  11250.00  11750.00  12250.00  12750.00  13250.00  13750.00  14250.00
 14750.00  15250.00  15750.00  16250.00  16750.00  17250.00  17750.00
18250.00  18750.00  19250.00  19500.00

Here, the zlevs are in meters from the ground. There is no necessity to have a
logical interval of values. GrADS doesn't care.

Remember that if your vertical coordinate is dependent on (x,y), this solution
will work, but with same (I don know how many) variable error by approximation.

Another cumbersome solution is to do an interpolation to constant z levels,
before writting the data to bin. Thus, GrADS will "become happy" to see all
your data on perfect cartesian coordinates.

Regards,

Ricardo

On Fri, 25 Feb 2005 18:47:36 +0100, Andrea Storto wrote
> I'll try to be clearer:
>
> In my model I've no grib output but binary files (not standard
> format), so I convert in ascci format and then re-convert in binary
> for GrADS. My output is in altimetric vertical coordinate and I'm
> interested in high resolution visualization near surface (so to
> interpolate to pressure levels or to get pressure variable and then
> write is not the best!), since my model is a Limited Area model with
> a very high resolution (subchilometric horizontal).
>
> So I can't use specified pressure levels for real vertical scale
> visualization.
> Moreover, I can fix the initial vertical spacing, and the
> stretching ratio of my grid:
>
> DZ(i+1)/DZ(i)=K (constant, for example 1.2)
>
> where DZ(i+1)=Z(i+1)-Z(i);
>
> if you develope you have something like
>
> z(n)=(K^n)*DZ0+SUM(i varying from 0 to n-1)(over Zi)
>
> where DZ0 is the initial vertical spacing (fixed)
>
> and this is not a loghartimic function since it depends on
> specific vertical level (SUM term).
> So I can't use set logz.
>
> I surely can find some solutions (using pressure levels or logz),
> but it could not be precise, and I'm also curious to understand
> wheter a way to have specifed vertical
> altimetric levels exists in GrADS.
>
> Thank you
>
> Andrea Storto
>
> ----- Original Message -----
> From: "Rodrigo Sanchez" <rodrigo at dgf.uchile.cl>
> To: <andrea.storto at TISCALI.IT>
> Sent: Wednesday, February 23, 2005 10:22 PM
> Subject: Re: Specifying Z levels
>
> > I'm a little bit confused ...
> > Do you have a working ctl file ?
> > or, Do you want to have one ?
> >
> > see, "set zlog"
> > ---
> > set zlog on|off
> > Sets log scaling of the Z dimension on or off. Sticks until reset.
> > ---
> >
> >
> > > Hi,
> > > I'll try your suggestion to know what I can get,
> > > even if I don't understand how it could help me,
> > > if I've only U,V,W variables (Components of wind)
> > > and I want to plot, for instance, a W contour in vertical plane
> > > (x-z) keeping real proportions of z levels?!
> > > Think of a stretched grid: the first vertical spacing (at ground level)
> is
> > > about 30 m, while the last one (a little above tropopause) is
> > > about 300 meters.
> > >
> > > Thanks
> > >
> > > Andrea
> > >
> >
> >


Hallak

Laboratório de Estudos de Tempestades Severas (LETS)
Departamento de Ciências Atmosféricas (DCA)
Instituto de Astronomia, Geofísica e Ciências Atmosféricas (IAG)
Universidade de São Paulo (USP)
Rua do Matão, 1226 - Cidade Universitária -São Paulo SP -05508-900
Fone.: +55 (0-XX-11) 3091-4735
Fax: +55 (0-XX-11) 3091-4714