[aspect-devel] Question on layered and temperature-dependent viscosities

[Wolfgang Bangerth]

> So, for the setting I showed you, does it means that if a hot anomaly was ~10 
> times less viscous than its surroundings on layer 1 (with an initial ambient 
> log visc = 22), then if that hot mass goes through the interface, its 
> viscosity will down-jump two orders of magnitude following the new environment 
> (where initially log visc = 20 ) , and it will continue being something like 
> ~10 times less viscous than the environment (if thermal diffusion didn't 
> occur), so it will be approximately log visc=18 ?

Yes. That's because it does not matter where the material came from. It just 
matters that at a particular point the temperature of the plume is higher than 
the are to its left and right, and it will have a viscosity lower than the 
ares to the left and right.

In your setup, there is no history. Forget about where the material came from. 
All you can ask is what the temperature is at a given point, not how it 
happened that it is hot or cold here.

Best
  W.

-- 
------------------------------------------------------------------------
Wolfgang Bangerth          email:                 bangerth at colostate.edu
                            www: http://www.math.colostate.edu/~bangerth/