[CIG-LONG] thermal diffusivity scaling and specifying correct temperature profile in hydrostatic term

[Walter Landry]

<Guillaume.Duclaux at csiro.au> wrote:
> Hi Karen,
> 
> 
> On 05/08/2010, at 11:45 PM, Karen Paczkowski wrote:
> 
> Hi,
> 
> I had a few questions about adding a thermal field to my subduction
> zone model.
> 
> 1)  On page 16 in the new manual (18 in the old manual) it says that
> the thermal diffusivity needs to be scaled like the velocity and the
> viscosity.  So for a thermal diffusivity of 1E-6 m^2/s and a viscosity
> of 1E21, this should scale the thermal diffusivity to something like
> 1E15 when viscosity is rescaled to 1.  The example dike xml file has
> the thermal diffusivity rescaled like this, but the example subduction
> zone program does not have this rescaling.  I was wondering if the
> rescaling should or shouldn't be done?
> 
> 
> You're right, thermal diffusivity has to be scaled. I guess the
> diffusivity in the subduction example has been missed.  If you're
> using any radiogenic heat production, it also needs to be scaled.

Exactly.  The subduction example is not documented because I did
not fully work it out.

> 2)  My simulation currently goes from the earth's surface to a depth
> of 440km.  I have a thermal field given by TemperatureProfile such
> that the temperature goes from 273 K at the top of the simulation to
> 1623K at a depth of 100km.  The temperature is then constant below
> 100km depth.  I can specify all of the terms for the temperature
> profile (upper density, upper alpha, T_0, height, linearcoefficient,
> etc) but there is nowhere to specify at what depth the temperature
> becomes a constant.  I think my hydrostatic term thinks that my linear
> temperature profile extends down through the whole simulation box.  I
> believe that this is causing large errors in my flow field.  I would
> expect a downward flow driven by the subduction slab (the only forcing
> in the system), but instead I get an upward flow with a magnitude over
> 2 orders of magnitude larger than the slab velocity.  How can I
> correctly tell my hydrostatic term what my temperature profile is?

There is a parameter "TemperatureProfileMax" that sets a maximum
temperature.

> I have attached an input file to the email.  It is currently written
> for Gale 1.4.  I was also wondering where I can find the instructions
> on how to upgrade my xml input files to be run with the new version of
> Gale?

There is a file UPGRADE which should describe everything you need to
do.  I have used it myself several times.

Cheers,
Walter Landry
walter at geodynamics.org