4. Thermo-compositional convection

In this demo, we will mimic a dense dyke injected into an ice shell, its collapse and entrainment by thermal convection.

Thermo-compositional convection in an ice shell.

Download

Here you can download the bash script, the modified material properties file, the parameter file, and the matplotlib script. Detailed explanation is given below.

Attention

In this simulation, modification of the m_material_properties.py is needed.

Note

Copy the bash script, parameter file, and plotting script from /demos/demo4/ into the main directory.

Material properties file modification

We add the dependence of ice density on the composition, enhancing the density where the dyke material is present.

\[\rho(T, C) = \rho_i(T)\cdot (C_{\rm ice} + 1.004\cdot C_{\rm salt})\]

def rho(Temp, composition, xm):
        # .
        # .
        # .

        # --- Demo 4 ---
        return ((1.0-xm)*rho_s*mm/VV + xm*rho_m)*(composition[0] + 1.004*composition[1])

Parameter file modification

First, we redefine the directory name:

# --- Name of the directory with results ---
name = "demo4_convection_comp"

Then, we modify the interface() function as a gaussian peak in the middle of the ice shell to represent the dyke. Afterwards, in the first position of materials, we enter the whole domain (the ice), and in the second position the dyke material defined as the material below the interface.

def interface(x):
    return 75e3*exp(-(x - 100e3)**2/1e8)

# --- Leave empty for a single material ---
materials = [["rectangle",  0, length, 0, height], ["interface",  "below"]]

To resolve the dyke well, we refine the lower half of the domain (same as in the original example) and the region that contains the dyke.

refinement = [0, length, 0, height/2, 80e3, 120e3, height/2 + 1e3, 80e3]