Hi Yilmaz :
On 3/11/2011 8:55 AM, Yilmaz Arslanoglu wrote:
Hi;
I was wondering the following points regarding dimensionless vertical
coordinates:
1) Assume that we have a variable U (let's say ocean current eastward)
and is defined on a dimensionless vertical coordinate (for example
ocean-s coordinate).
Do the variables indicated in "formula_terms" of this vertical
coordinate need to share exactly the same longitude and latitude
coordinates (possibly auxiliary) with
the main variable U, or can they be defined over their own
longitude and latitude coordinates independently?
I will give my opinion as someone who has written software to deal with
this: 1) the variables in the formula must be on the same coordinates as
the vertical coordinate, and 2) a vertical coordinate can only be used
by data variables on the same horiz coordinate.
this is a real problem with staggered grids, which i assume is your use
case. One needs an interpolation or regridding scheme to solve it. So
gridspec is one solution, though heavyweight. Rich SIgnell proposed
another "simple regridding" proposal which we should revisit.
2) Just given the "positive" attribute and the information whether the
sigma coordinate is monotonically increasing or decreasing, is it
possible to
conclude that the array of z values z(sigma, t, x, y) are always
increasing or decreasing? That is, is it possible to make the
following conclusion
for any dimensionless vertical coordinate ? ( sigma here is just
an example )
I was actually wondering what will the "positive" attribute
exactly mean for a sigma coordinate.
My understanding is that positive, which means "Direction of increasing
vertical coordinate value" refers to the actual sigma coordinate value,
not any coordinates derived from it. So i think your table is correct,
since sigma = 1 is at the surface and =0 is at the top of the
atmosphere. Thus larger values of sigma go "down".
Sigma Positive Z
--------- ------------ --------------
increasing down increasing
increasing up decreasing
decreasing down decreasing
decreasing up increasing
Best regards,
Yilmaz
_______________________________________________
CF-metadata mailing list
[email protected]
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
_______________________________________________
CF-metadata mailing list
[email protected]
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
