# Modelling the thermal convection of the ice shell of Ganymede using ASPECT

Good morning all.
I would like to ask you if I could obtain a cylindrical gplate 2d model just for the ice shell of Ganymede, I need to simulate the thermal convection in this layer but I am not really sure if I could adapt the same technique (ASPECT mantle convection) just for the superficial ice shell.
I have studied the input for this cookbook and I should replace the values of the subsection spherical shell (inner and outer radius) by the top and bottom thickness of the ice shell (70 Km) but I am not really sure if I could do it.

please, help me with these doubts.

Regards,
Katherine

Katherine,
replacing the geometry is not going to be very difficult. If you look at, for example, the file `cookbooks/shell_simple_3d.prm`, you will see how you define the radii of the lower and upper boundaries of the shell you want to simulate.

As for the boundary conditions: Can you elaborate what you mean by “cylindrical gplate 2d model”? Are you thinking of a 2d cross section of Ganymede? And are there gplates files for Ganymede, or how does gplates enter into the picture here?

Separately, I don’t think that we have ice rheologies already in ASPECT. Have you given some thoughts on how you will model the rheology of the icy layer on Ganymede?

Best
Wolfgang

thanks for the clarification Dr. Bangerth.
My idea is to observe the changes of the heating in the latitud using the tidal heating equation (I think that is possible to do a discretization as an external scrip and after to add it in the principal input). I am not going to consider the rheology in my thermo elastic model. My doubt was oriented to the thickness because I tried to change the values but the simulation failed; so, I am going to read again which could be the mistakes that don’t allow me to obtain the model with the depth I need.
I asked for the gplate 2d becuase is the perfect template that I need to use for observing the changes in the latitude and equator.

sorry, I mean to do the discretization of the tidal equation as an external plugin (actually I hope to reach it), the reference that I have comes from Gerya but he uses finite differences so I need to figure out how to adapt it reading the algorithms from deal library

Hi Dr. Bangerth,

I checked the cookbook/shell_simple_3d.prm but it has the same values
of the the 2d model:

subsection Geometry model
set Model name = spherical shell

subsection Spherical shell
end
end

I tried to run my simulation with these parameters:

subsection Geometry model
set Model name = spherical shell

subsection Spherical shell
end
end

and the input doesn’t run, that’s why I wonder if I can apply the
gplate 2d template in my model but with a different thickness.

kind regards.
katherine villavicencio

Wolfgang Bangerth via Computational Infrastructure for Geodynamics
geodynamics@discoursemail.com escribió:

I changed the values again and it’s running now. I have to wait for
visualizing the new model.
Thanks again.

Katherine

Wolfgang Bangerth via Computational Infrastructure for Geodynamics
geodynamics@discoursemail.com escribió:

Katherine,
let us know what you find! For the future: When you report that “it didn’t work”, or “I get an error”, try to be specific: Show us the error message you get so that we have an idea what may be the reason why it didn’t work.
Best
W.

Dear Dr. Bangerth,

I could not run the simulation with the parameters I need and I
obtained this message. Please, help me to find what I did wrong.

Regards,
Katherine Villavicencio
Wolfgang Bangerth via Computational Infrastructure for Geodynamics
geodynamics@discoursemail.com escribió:

Katherine – there can be many reasons for why the solver can’t solve the problem. In your case, I believe that you have a very thin shell. Try and see whether things work with a much thicker shell – say one in which the ratio of inner to outer ratio is more earth-like, since that is what ASPECT was primarily developed for. It could also be that the rheology you use is too extreme – in that case experiment with an isoviscous rheology for the moment. It could also be the use of free surfaces, complicated boundary conditions, or a few other things.

The point I’m trying to make is that we generally start with a “simple” model and make sure that that runs. Then we gradually incorporate changes that make things more realistic. If at some point the solver breaks, at least you know which change it was that broke things.

Can you try and isolate which part of the model it is that causes the problem?

Best
Wolfgang

Hi Dr. Bangerth,

I have an idea on how to solve this problem, I am going to try and let
you know.
I watched the tutorials on Youtube, in the tutorial 5 part 2 Dr.
Dannberg menctions that the gravity model should be changed as the
geometry and this is exactly the values I need to replace; but now I
am confused with this becuase that means I could change the values of
the radius because is part of the geometry, but the program doesn’t
solve it.
One reference that I have read is in Zhang & O’Neill 2015, related to
Geodynamic Evolution on Mars; it’s quite similar the thing I need to
simulate but the difference is that they considered the mantle
convection (in this case the radio is the half of Earth) so, I think
that is possible to change the radius data, but I don’t know why I am
having problems.
I am reading again the manual and trying figure out which are the
mistakes I am doing.