wrong coordinates when plotting from an .nc file

[Costas Douvis]

In my case the correct map (with the correct coodrinates) can be plotted by gradsc using a .ctl file with the line:

pdef 98 78 lcc 44.00 15.00 49.00 39.00 55.00 31.00 15.00 60000. 60000.

But in that case the data file must be the usual fortran unformatted and direct output file. I know that for sure because the usual output of the model I use (RegCM3) is like that. This time I have problems because I used the PostProcess routines to interpolate the CRU database to the same grid and unfortunately the output is an .nc file.

If I could somehow change the format of the file from .nc to the usual direct and unformatted fortran file I would be OK for now. Do you know how I could do that?

Brian, in case you want to use this method be careful. If I remember correctly there is a number of ways to describe your grid and "lcc" (in the pdef line) is only one of them

Kostas

>Date:    Mon, 7 Jan 2008 05:58:07 -0800
>From:    brian vant-hull <brianvanthull at YAHOO.COM>
>Subject: Re: wrong coordinates when plotting from an .nc file
>
>Hi Costas;
>I hope to do something similar in the near future, but
>haven't started work on it yet.   What I want to do is
>feed in  the data plus somewhat irregular (but
>gridlike) lat and lon, and get out the data in a
>projection that GrADS will plot.
>
>It looks like the projections described in the user
>manual *might* do this, but I haven't figured out how
>to feed in the original unprojected lat lon data.  I
>fear the lat lon grids are expected to be already
>mercator projected.  If so I'll have to use IDL or
>MatLab to warp my grid to the projection first, or
>write a rebinning program and perhaps lose some
>resolution.
>
>Anyone else out there: do we need to warp our data to
>a given projection outside of GrADS, or will GrADS do
>this for us?  I see no provision for feeding arbitrary
>lat/lon grids into grads.
>
>(unless plotting as station data - though it doesn't
>look like this will provide contoured output.)
>
>
>--- Costas Douvis  <cdouvis at YAHOO.COM> wrote:
>
>> Thank you for your response Brian
>>
>> My grid has an interval of 50km. So when I use the
>> usual "rectilinear" map projection (where the
>> coordinates are latitude and longitude and a
>> "square" of 10x10 degrees does appear as a square)
>> my domain should look stretched towards the pole,
>> narrow towards the equator and curved. But it
>> doesn't: it appears rectangular
>>
>> I'm running SUSE 10.0 - 32bit
>> and Grads version 1.8SL11
>>
>> Please let me know if this information is still not
>> enough
>>
>> Kostas
>>
>>> Date:    Fri, 4 Jan 2008 07:18:23 -0800
>>> From:    brian vant-hull
>>>
>>> Subject: Re: wrong coordinates when plotting from an
>>> .nc  file
>>>
>>> You'll need to be more precise...for example if the
>>> coordinates are in equal increments of lat and lon
>>> then in an equal distance mapping they'd get
>>> narrower
>>> towards the poles, but the typical lat-lon plot
>>> would
>>> stretch it.  You must have something different, but
>>> hard to tell what it is.
>>>
>>> I may not be able to answer the question, but for
>>> someone else to answer it they'll need to know alot
>>> more details!
>>>
>>> -Brian.
>>>
>>> > --- Costas Douvis  wrote:
>>> >
>>> > Hello and happy new year to everyone
>>> >
>>> > I have problem plotting the data from an .nc file.
>>> I
>>> > believe that the values of the field are  correct
>>> but
>>> > the coordinates are "stretched" so that they fit
>>> in
>>> > a rectangle. The real grid comes from an RCM model
>>> > and is getting narrower towards the equator.
>>> >
>>> > Thanks in advance for any help
>>> > Kostas Douvis







---------------------------------
Be a better friend, newshound, and know-it-all with Yahoo! Mobile.  Try it now.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20080108/11870f5f/attachment.html