Double precision & adaptable z dimension

[Jennifer Adams]

On Jun 29, 2006, at 5:14 PM, Jim Benedict wrote:

> I have two questions:
>
> 1.  Can GrADS handle double precision numbers (8 byte)?  I have only
> had success with integers and floats (reals) as mentioned in the
> GrADS documentation... just wanted to confirm that 8-byte doubles
> cannot be used.
That's right. Internally, everything is carried around as a float.
There is no way to read a flat binary file that contains doubles. If
it's netcdf, the doubles will be converted to floats, with a likely
loss of precision.

> 2.  I write two variables to the same file.  The first variable
> (rec=1) lists pressure values, the second (rec=2) heating rates.  I'd
> like to be able to plot the heating rate vertical profile using the
> listed pressure values as the z dimension.  Can I do this without
> listing every pressure value in the control file?  The pressure
> values are not uniformly spaced (can't use LINEAR for ZDEF) and are
> subject to change (can't use LEVELS for ZDEF).  Also, I have 199
> vertical levels, so writing the list out in the control file would be
> difficult.

The only way you can have an adaptable Z dimension is with station
data. As long as you are writing out a data file, you might want to
consider changing the format from the gridded data model to the station
data model.

Jennifer

--
Jennifer Miletta Adams
Center for Ocean-Land-Atmosphere Studies (COLA)
4041 Powder Mill Road, Suite 302
Calverton, MD 20705 USA
jma at cola.iges.org