[gradsusr] wrfout projection in grads

[LI Qi]

Hi Jeff,


Thanks for your clear and inspirational reply! It really helps!


Besides, I found a page explaining how the re-projected data display in grads in details, paste it here as a supplementary. Somebody else might find it helpful too.
http://www.iges.org/grads/gadoc/pdef.html


Best,
Qi


Message: 2
Date: Tue, 2 Jun 2015 20:10:06 -0500
From: Jeff Duda <jeffduda319 at gmail.com>
Subject: Re: [gradsusr] wrfout projection in grads
To: GrADS Users Forum <gradsusr at gradsusr.org>
Message-ID:
<CAAig09D-sZFuzcVVib-773HM_6mpg+MrLt0qkgPD8uuWMGNt-Q at mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

On Tue, Jun 2, 2015 at 7:14 PM, LI Qi <liqi123sh at qq.com> wrote:

> Dear grads community,
>
> I don't really understand how the wrfout projection works in grads.
> Below is a descriptor file from ARWpost as an example.
> The model domain is using lambert conformal projection with grid spacing
> of 30 km.
> title  OUTPUT FROM WRF V3.7 MODEL
> pdef  231 171 lcc  35.178  110.000  116.000   86.000  60.00000  30.00000
>   110.00000  30000.000  30000.000
> xdef  783 linear   57.03648   0.13513513
> ydef  392 linear    7.01026   0.13513513
> zdef   19 levels
>
> My question is:
> [1] Do "xdef" and "ydef" mean the data have been interpolated with
> resolution of 0.13513513 deg? But the original resolution is
> approximately 0.27 deg.
>
Yes. XDEF and YDEF can be whatever value you want. XDEF and YDEF control
the resolution of the plotted data, but do not change the resolution of the
raw model output. Think of it as a sampling rate. You can sample a signal
more frequently to get smoother looking data, but the data themselves are
not changed.

> [2] How are the xsize of 783 and ysize of 392 derived?
>
In ARWpost, the code finds the maximum and minimum latitude and longitude
value in your model grid and divide the distance by some arbitrary fraction
(hard-coded within the program)

> [3] If I would like to remap observations onto the wrf model output to
> facilitate comparisons, should I remap the obs onto the original
> 0.27deg-resolution grid, or the interpolated 0.135deg-res grid? Is the
> latter feasible?
>
That's your call, but you'll get "the right" kind of comparison if you
stick to the native model grid.

You can set the namelist parameter to grads_low_res to .true. to tell the
program to ignore these arbitrary settings and just stick to your raw model
grid for the control file XDEF and YDEF entries. If you want to use
ARWpost, I recommend doing that.

ARWpost is somewhat outdated software, however. I would recommend using the
Unified Post Processor software. There are no weird issues about different
grid settings when you use it.

>
> Your reply would be much appreciated!
>
> Best,
> Qi
>
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20150604/c708d4d4/attachment-0001.html