Re: Time and sigma coordinates

william weibel (
Wed, 26 Jan 94 17:03:36 PST

Eric Pepke writes:
> Now, I have another question.  Does anybody do what they tell me are called
> "sigma coordinates" in netCDF, and if so, how?  These are coordinates that are
> just like rectilinear coordinates, except that the top of the grid is at a
> fixed level above sea level and the bottom follows the terrain.  The levels
> between the flat top and the deformed bottom are deformed gradually so that for
> every x and y, the z values are evenly spaced.  As you go up, the levels get
> progressively less and less deformed until they hit the top.
> Eric Pepke                                     INTERNET:

I have been trying to get some people here to implement something for awhile.
But, I might as well push a rock uphill.  Regardless, here are some tips.

Most modellers transform their results onto constant pressure surfaces before
they plot, even if the results are stored in sigma coordinates.  So, you will
need to store enough information in your NetCDFs for the analysis 
programs can do the transforms.  You will need to store the surface pressure
to be able to convert to pressure coordinates, and you will need a topography
field, in order to caluculate z(sigma).

You should also have an attribute declaring that you are using a
special coordinate system.  I think conventions for that have been
discussed here in the past.  Please write me if you want to discuss this more,
or if you get better advice from someone else.

  William Weibel			|
  Department of Atmospheric Sciences	|
  UCLA					|  Tel. (310)206-4441
  Los Angeles, CA  90024-1565  		|  Fax  (310)206-5219
  U.S.A.				|