Re: 19990519: NetCDF query

>To: address@hidden
>From: Roger Proctor <address@hidden>
>Subject: query
>Organization: .
>Keywords: 199905192208.QAA23759

Hi Roger,

> i'm a new user of netcdf.
> if i have two scalar components of wind, say uwind & vwind, how do
> i combine them into a vector variable so that i can plot the vectors
> in a package like ferret?

Good question.  We just use separate variables u_wind and v_wind, but
then applications must "know" that these represent components of a
single vector variable.  This can be done in other ways by convention,
depending on what conventions you choose to use.  I'm not sure what
conventions Ferret uses for such variables, and in a quick look at the
documentation, I was unable to figure it out.  The Ferret
documentation states that data must conform to the COARDS conventions,
described at

  http://ferret.wrc.noaa.gov/noaa_coop/coop_cdf_profile.html

but I couldn't find this subject addressed in the COARDS conventions.

In the "GDT conventions document" at

  http://www-pcmdi.llnl.gov/drach/GDT_convention.html

they say

  ... Components of vector or tensor quantities could be contained in
  a single data variable by giving the variable a dimension over
  components. While there exist advantages for manipulating such a
  variable in memory, we see no strong advantage in introducing this
  complexity into the netCDF description, and do not recommend it.

Descriptions of other conventions are available from 

   http://www.unidata.ucar.edu/packages/netcdf/conventions.html

You might try contacting Steve Hankin <address@hidden> to get a
definitive answer about how to make Ferret understand vector
variables.  I'd be interested in the answer.

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden                     http://www.unidata.ucar.edu

---------------------------------------
 Date: Tue, 25 May 1999 19:09:13 +0100
 From: roger proctor <address@hidden>
 Organization: Natural Environment Research Council
 X-Mailer: Mozilla 3.01SGoldC-SGI (X11; I; IRIX 6.3 IP32)
 MIME-Version: 1.0
 To: address@hidden
 Subject: NetCDF query [Fwd: Re: ferret]

 Russ

 Attached is the answer to the question about forming vector quantities.

 I'm quite happy with this.

 Thanks for your idea.

 Cheers

 Roger
 -- 
 CCMS - Proudman Oceanographic Laboratory, 
 Bidston Observatory, Birkenhead, Merseyside L43 7RA, U.K. 
 Tel: +44-151-653-8633  Fax: +44-151-653-6269
 Email: address@hidden    URL: http://www.pol.ac.uk

 --------------4531158753D5
 Content-Type: message/rfc822
 Content-Transfer-Encoding: 7bit
 Content-Disposition: inline

 Return-Path: <address@hidden>
 Received: from zeta.pmel.noaa.gov by unixa.nerc-bidston.ac.uk 
(8.8.7/NERC-1.5(Solaris 2.x, SMTP))
         id AAA20535; Fri, 21 May 1999 00:12:41 +0100 (BST)
 Received: from pmel.noaa.gov (outlaw.pmel.noaa.gov)
  by ZETA.PMEL.NOAA.GOV (PMDF V5.1-12 #27385)
  with ESMTP id <address@hidden> for
  address@hidden; Thu, 20 May 1999 16:12:37 PDT
 Date: Thu, 20 May 1999 16:16:14 -0700
 From: Steve Hankin <address@hidden>
 Subject: Re: ferret
 To: Roger Proctor <address@hidden>
 Message-id: <address@hidden>
 Organization: NOAA/Pacific Marine Environmental Laboratory
 MIME-version: 1.0
 X-Mailer: Mozilla 4.5 [en] (WinNT; I)
 Content-type: text/plain; charset=us-ascii
 Content-transfer-encoding: 7BIT
 X-Accept-Language: en
 References: <address@hidden>

 The two variables "uwnd" and "vwnd" are entirely independent. The
 command
         VECTOR uwnd, vwnd
 is the point at which they are associated into a "vector".

 In the GUI identifies likely variable pairs based on clues in the
 variable names. It is only a figment of the user interface. If you use
 the "SCRIPTS ->COMMAND LINE" menu (something close to those names),
 you'll see that you could make vector pairs out of anything.

 So in your netCDF file just write the variables
         temp
         salt
         u
         v
         w

 Ferret will handle the rest. Easy as pie (we hope :-} ).

         - steve



 Roger Proctor wrote:
 > 
 > Hello Steve
 > I'm a new user of ferret. I have looked at the examples and particularly
 > the navy winds example which has three data variables
 > uwind, vwind and uwind,vwind
 > thus allowing the wind vector to be drawn. I want to do the same
 > but don't understand how to create the variable (uwind,vwind) in a netcdf
 > file. I can see from the documentation how to do this within ferret
 > by loading uwind and vwind as scalars and creating the vector.
 > However I would like to create the vector outside of ferret in the
 > netcdf file. doing a nvdump on the navy winds file doesn't provide me
 > with any clues. Can you help me? (oops, nvdump = ncdump).
 > 
 > What i really want to do is output my model data directly in netcdf format,
 > i.e. creating a file with 3d variables salinity, temperature and currents.
 > Surely someone must have done this!
 > 
 > best wishes
 > 
 > Roger Proctor
 > 
 > ps what i should have added is that i am a new user of netcdf too! I
 > contacted the netcdf help & they are not sure how to do it!!
 > Centre for Coastal and Marine Sciences, Proudman Oceanographic Laboratory,
 > Bidston Observatory, Birkenhead, Merseyside L43 7RA, U.K.
 > Tel: +44-151-653-8633  Fax: +44-151-653-6269
 > URL: http://www.pol.ac.uk====================Internet: address@hidden

 -- 

                 |  NOAA/PMEL               |  ph. (206) 526-6080  
 Steve Hankin    |  7600 Sand Point Way NE  |  FAX (206) 526-6744
                 |  Seattle, WA 98115-0070  |  address@hidden