A problem - vector velocity representation

• To: vis5d-list@xxxxxxxxxxxxx, visad-list@xxxxxxxxxxxxx
• Subject: A problem - vector velocity representation
• From: Sergei Maurits <maurits@xxxxxxxx>
• Date: Tue, 14 Jul 1998 13:39:24 -0800

Hi Vis5D users,

I'd like to report what appears to be a bug in the version 
4.3 and ask if someone had similar experience.

We need to visualize the polar model scalar and vector 
output. The model domain has the Pole in the center.
For reference, the model has the regular Cartesian grid, 
projected from the geographic coordinates using azimuthal
equidistant projection (AEP). This Cartesian grid is a 
square with the Pole at its very center. 

The most close to the AEP projection available in Vis5D is 
the  stereographic projection mode (=3), which we use to 
display  the data with underlying geographic map.  

Now the problem. Vis5D can't properly dysplay the velocity 
vectors if we use zonal/meridional components of the velocity 
for U and V, as it's supposed to be with the Vis5D stereogra-
phic projection. I beleive, this is a bug, since the scalar 
values of the velocity components are not corresponding to 
the resulting vector reperesentation.

A simple fix is to use the Cartesian velocity components
(computed in our case for the central Pole location, co-
inciding with the Cartesian coordinateds origin), instead 
of the zonal/meridional components pair for U.V. With the
signs adjustement to the Vis5D conventions about the axis 
directions, the vector pattern is OK including the cor-
respondence to the scalar velocity components. 

This fix indicates that the Vis5D code misses an adjustement 
for the high-latitude convergence of the meridians in the 
spherical geometry, hence making the vector representation 
obviously wrong when the projection center is somewhere at the 
high latitudes. Actually, I think it's somewhat wrong everywere, 
it's just less noteceable than the longtitudinal span is much
less than global. 

I wonder if someone can point out, where in the Vis5D code 
the vector directions are computed for different projections. 
For stereographic projection it should be a longtitudinal 
correction for the current meridional direction, also depending
on the geographical location of the projector center. Also, I 
wonder if this is fixed in v.5.0.

Thank you,
Sergei



-- 
Sergei Maurits

HPC Visualization Specialist
Arctic Region Supercomputing Center
University of Alaska Fairbanks

P.O. Box 756020                    tel: (907) 474 5591         
910 Yukon Dr., suite 108           fax: (907) 474 5494
Fairbanks, AK 99755-6020           e-mail: maurits@xxxxxxxx
WWW: http://www.arsc.edu           http://dac3.gi.alaska.edu/~sergei