[aspect-devel] Question regarding heat fluxes

[Ian Rose]

Hi all,

I have been doing some tests with aspect involving heat fluxes out of the
top and bottom boundaries, and am getting some behavior I don't
understand.

When the convection reaches steady state, the heat fluxes out of the top
and bottom should be equal and opposite, otherwise, of course, it wouldn't
be steady.  However, if I look at a simple convection model in with 2d
shells, I find that the flux out the top is considerably lower than the
flux out of the bottom at steady state.  I am attaching a prm file and a
plot of heat fluxes as demonstration.  The qualitative change at ~10 Gyr is
due to a transition from organized to chaotic convection.  The basic heat
flux story is unchanged, though.

Looking at the code for heat flux statistics, nothing seems obviously
wrong.  I should note that doing the same test for a box geometry where the
top and bottom have the same area produces the expected results.

Am I missing something about how this should work?  Any insights as to what
is going on would be appreciated.
Best,
Ian

PS All calculations are done with r1353
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_flux.prm
Type: application/octet-stream
Size: 2737 bytes
Desc: not available
Url : http://geodynamics.org/pipermail/aspect-devel/attachments/20121115/d61471d4/attachment-0001.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: flux.ps
Type: application/postscript
Size: 54871 bytes
Desc: not available
Url : http://geodynamics.org/pipermail/aspect-devel/attachments/20121115/d61471d4/attachment-0001.ps