[Thread Prev][Thread Next][Index]

Re: Contours on a Gaussian grid

> Hi Michel,
> I assume that your data is in an ASCII file.
> What I don't know from your message is whether you have the latitude
> locations in that file or whether you know them analytically, or
> whatever ...
> If I'm understanding the problem correctly, the solution will be is to
> make sure that the grid upon which your variable is defined has the
> correct (unequally spaced) Y axis by using "DEFINE AXIS/FROM_DATA
> ...".
> Lets start at the solution at the point that you have the latitude
> locations (by whatever means) in a Ferret variable called "lats". Then
> simply use:
> -------
> 	DEFINE AXIS/X=lo:hi:delta xlongs
> 	DEFINE AXIS/Y/FROM_DATA/NAME=ylats lats	  ! using the values of "lats"
> 	DEFINE GRID/X=xlongs/Y=ylats gg
> 	FILE/var=myvar/grid=gg mydata.dat
> 	SET VARIABLE/TITLE="my title"/UNITS="some units" myvar

It works great, thanks !  But what does it do exactly ?  Does it only
stretch the plot without regridding ?  If it is so, that is great.
Regridding might give strange results anyway because I have some points
set to 1.0E35 to indicate invalid values (as it is suggested somewhere
in the documentation).  At least I get strange results if I try to use
FILL instead of SHADE to plot the data (I mean with or without an
irregular axis).  It seems that FILL regrids the whole domain to higher
resolution but it does so also with invalid values, wich appear to take
more space in the plot afterwards.  Since these invalid values happen to
be mostly over land, the continents, which show as a white area, get
suddenly bigger and deformed...

Michel Beland                 beland@CERCA.UMontreal.CA
professionnel de recherche    tel: (514)369-5223  fax: (514)369-3880
CERCA (CEntre de Recherche en Calcul Applique)
5160, boul. Decarie, bureau 400(423), Montreal (Quebec), Canada, H3X 2H9 

[Thread Prev][Thread Next][Index]

Dept of Commerce / NOAA / OAR / ERL / PMEL / TMAP

Contact Us | Privacy Policy | Disclaimer | Accessibility Statement