Re: [ferret_users] Trouble Displaying Curvilinear Data

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi Jonn,
Bill's right, but the data's much wierder than that.

A good way to start diagnosing anything with curvilienar data is to look 
at the longitude and latitude fields separately, and the data without 
trying to use the curvilinear coordinates.  The STAT command works well 
to get at what's in the coordinates and data fields.

  set data "http://205.156.4.56/cgi-bin/nph-dods/st_johns/efdcgrid01.nc";
  shade lon
  shade lat
  shade temp[k=1]

It looks like the data has a flag of 0 for all of the fields, except 
where there are grid locations where the idealized river is. (plot 
attached below, with first just SHADE LON and then with some levels 
specified so we can see what the values are)

The zero's are what is throwning off the curvilinear plotting algorithm. 
I'm not sure how we could get the method we have of visualizing 
curvilinear data to work at all with this sort of grid.  We can't for 
instance just set the missing-value flags to zero, because the 
curvilinear method requires the longitude and latitude data to have no 
missing data.

Ferret would expect that the longitude and latitude fields be completely 
specified, and the data fields to have a flag where the data does not 
exist.  I can't think of a way around this right off (but it's sure nice 
to have the data available via DODS so we can check out what's happening!)

Here is the script I used to get the plot below:


  yes? set data 
"http://205.156.4.56/cgi-bin/nph-dods/st_johns/efdcgrid01.nc";
  yes? cancel mode logo

  yes? set view ul
  yes? shade lon
  yes? set view ur
  yes? shade/lev=(-82,-81,0.1) lon

  yes? set view ll
  yes? shade lat
  yes? set view lr
  yes? shade/lev=(29,31,0.2) lat


John Schattel wrote:

> Hi All,
>
> I am trying to display some curvilinear model data and I am getting 
> some unexpected results.  You can see the results of a 
> shade/title="Surface Currents" u,lon,lat command in the attached gif 
> file (st_johns_river.gif).  The data should have been for a stretch of 
> the Eastern coast of Florida ( 
> http://www.csc.noaa.gov/csp/images/fl_region.gif ).  Does anyone have 
> a suggestion on what might be going wrong or how to diagnose this 
> problem.  If you would like to see the netCDF data, it is available 
> via the following DODS URL or I could e-mail it to you offline.
>
> http://205.156.4.56/cgi-bin/nph-dods/st_johns/efdcgrid01.nc.info
>
> In advance, thanks for any help you might provide.
>
> John