Re: Regridding

[Marcelo Barreiro <marcelo@xxxxxxxxxxxxx>]

Mark,

Thanks for your answer, but my English prooved to be very bad. 

I just wanted to do the opposite: I am trying to map a 1-D variable (415
points) in the "good" points of a 2-D grid (33*18=594 total grid
points) wich has 179 "bad" points (over the continents).
Any suggestions to this?

Marcelo


P.D. I tried to use the function "unravel" typing

yes? plot unravel(sst)

and I got this error message:

 **ERROR: command syntax: unravel(sst)
          unravel before ( is illegal



-----------------------------------------------------------


On Thu, 10 Feb 2000, Mark A. Verschell wrote:

> Marcelo,
> 
> 	Let me restate what I think you are trying to do, and then give you a
> (poor) solution.
> 
> 	You are trying to map the good data values (SST) of a 2-dimensional
> grid (33*18=594 total grid points) with 179 grid points being "bad" (over
> continents) to a 1-dimensional grid that contains only the 415 remaining good
> points.
> 
> If this is so, I only have a partial solution for you. You can use the ferret
> function UNRAVEL to turn this into a 1-D grid :
> 
> yes? plot unravel(sst)
> 
> The problem is to skip over the bad-data points. I don't know of anyway to do
> this in ferret directly. If you have version 5.0 you could probably do this in
> an external function. I don't have this, and so the only thing I can suggest is
> a multi-step process:
> 
> write out the data:
> list/nohead/form=(f10.4)/file=tmp1.dat unravel(missing(sst,-9999))
> 
> Filter out the bad data values:
> spawn grep -v tmp1.dat > tmp2.dat
> 
> Read in the compressed 1-d sst vector:
> file/var=sst tmp2.dat
> 
> Of course, the smart people at PMEL probably have a way to filter out the bad
> data values inside ferret, and so my reply to you is likely unnecessary.
> 
> Mark
> 
> On Feb 9,  9:13pm, Marcelo Barreiro wrote:
> > Subject: Regridding
> >
> > Hi all,
> >
> > I am trying to regridd a variable var_a with the grid of a variable
> > var_g.
> > -var_g has 594 (33*18*1*1) points of which 179 are flagged as bad (var_g
> > is SST and the bad data are over the continents). Thus we are left with
> > 415 good points.
> > -var_a is a row vector with 415 points.
> >
> > I tried
> >
> > ? file/var=ac/grid=var_g[d=1]/columns=415 var_a.dat
> >
> > but it did not work: it did not consider the region where var_g has bad
> > data (it did not take into account the continents).
> > Is there any way to regridd var_a such that its values are distributed
> > only in the region where var_g has good values? How can I tell ferret to
> > add bad data to var_a in the same region as var_g?
> >
> > Thanks!
> >
> > Marcelo
> >
> > Texas A&M University
> > Dept. of Oceanography.
> >
> >-- End of excerpt from Marcelo Barreiro
> 
> 
> 
> -- 
>  Mark Verschell                     NASA/Goddard Space Flight Center
>  verschell@neptune.gsfc.nasa.gov    Code 970/Lab. for Hydrospheric Proc.
>  301-614-5669  Fax: 301-614-5666    Greenbelt, MD 20771
> 
> 
> 		     Better Living Through Denial
>                ** USQC Certified: 100% Microsoft Free **
>