[CIG-LONG] Lateral hydrostatic variation

[Walter Landry]

Taichi SATO <taichix at ori.u-tokyo.ac.jp> wrote:
> Hello,
> 
> I would like to specify HydrostaticTerm but I have some difficulties.
> 
> 1. How to apply TemperatureProfile(H0>0, described in P96) for
> HydrostaticTerm?
> I think HydrostaticTerm does not involve lateral temperature variation,
> although we can specify this variation using TemperatureProfile in
> StandardConditionFunction. Please see Gale manual P91 and P96.
> 
> 2. Is it possible to specify materialBoundary as  a function such as
> linear or exponetial?
> I think that constant depth material boundary does not fit for mid-ocean
> ridge modeling, because the lithosphere/asthenosphere boundary is not
> constant depth.

The idea of HydrostaticTerm is that you subtract off a non-dynamic
density profile.  With lateral temperature or density variations and a
finite viscosity, you will get dynamics.  So you have to come up with
a suitable non-dynamic density and temperature profile.  In the first
case, I would try to use

  T_0=TemperatureProfileTop
  linearCoefficient=(T_m - T_0)/H_m
  exponentialCoefficient1=TemperatureExponentialCoefficient1
  exponentialCoefficient2=TemperatureExponentialCoefficient2

For the value of materialBoundary, you will have to experiment with it
to get a reasonable value.

One thing to remember is that, in the limit of infinite resolution,
the answer does not depend on how you set up HydrostaticTerm.  You
could even omit in entirely.  You will get better answers with lower
resolution if you make a good choice.  With a bad choice, the answer
will be obviously inaccurate.

Cheers,
Walter Landry
walter at geodynamics.org