Just an observation... It looks like 273 is being subtracted from the value on
the good plot to give you the value on the bad plot. It looks like an incorrect
attempt to change the plotted temps from Celsius to Kelvin.
Quoting James Murakami <tenki@xxxxxxxxxxxxxxxxxxxx>:
You're right. Sort of...
If you plot values first and contour second, everything goes right (see
attached gif image). If you contour first and plot values second, you
get weird numbers ploted (see second attached gif). The contour scheme
remains fine. Even if you invoke clear and reset, the plotted numbers
are incorrect (they change with each new attempt). So, yes, there seems
to be a bug in the software.
(For reference, I just used current NAM212 data set and used default
settings for displaying 500mb temperature(deg C). I'm running version
5.8.2a on a Solaris8 machine from source.)