Modelling glaciovolcanic caves and chimneys - MSc Thesis (Simon Fraser University)

Tryggvi Unnsteinsson

This folder contains the code we used to investigate the formation and melt-through of glaciovolcanic voids in chapter 3 of my MSc thesis.

define_variables.sh - A file that takes in the user inputs of the desired physical parameters, imports them to the appropriate files and preps the working directory for simulation.

run_iterative.sh - The file that runs the simulation and iterates between different parameter combinations. It is created from the file run_iterative_fillin.sh which is edited by define_variables.sh to contain the appropriate variables.

elmer_job.sh - Equivalent as previous file but is used to run the code on Compute Canada's supercomputer clusters. (This file will be edited to not contain account details before publishing.)

create_geometry.py - Creates the mesh for each parameter combination and void geometry. The mesh is created using GMSH and exported in .msh format.

diangostic.sif - The diagnostic solver input file for Elmer/Ice. The file is exported into each parameter folder with the appropriate parameters changed.

prognostic.sif - The prognostic solver input file for Elmer/Ice. The file is exported into each parameter folder with the appropriate parameters changed.

remesh_deformed.py - A file that takes in the prognostic .vtu result, remeshes the void and top surface and returns a new mesh in .msh format.

Execution

To use the code, navigate to a working directory where the results are to be saved and download the "main" folder to said directory.

To setup for a simulation for desired physical parameters (glacier thickness, bed slope, total heat flux and mode of heat transfer) update the define_variables.sh bash file to include those parameters (see Physical parameters). Each variable has to have at least one option specified. Once the variables have been defined run the command:

bash main/define_variables.sh

This should create parameter folders with names of the format 'H{}_I{}_Q{}_M{}', e.g. for 100 m thickness, 5° bed slope, 2500 kW total heat flux and mode of heat flux nr. 3 the associated folder name is 'H100_I5_Q2500_M3'. A .txt file containing all the different parameter folder names is also created. The command should also create two separate files within the main directory, elmer_job.sh and run_iterative.sh, which are used to run the code on either a computer cluster or your personal computer, respectively.

Then to run the simulation over the chosen parameters simply run the command:

bash main/run_iterative.sh 

Simulations with each parameter composition should then run in serial and populate the parameter folders with the .vtu result files from each iteration which can be viewed in Paraview.

Physical parameters

The user has four variables to choose from:

1. Glacier thickness, H [m].

   • The thicknesses used in our simulations are 50, 100, 150 and 200 meters.

2. Bed inclination, I [°].

   • The bed slopes used in our simulations are 0°, 5°, 10° and 15°.

3. Total heat flux, Q [kW].

   • The heat fluxes used in our simulations range from 0 to 3.5 kW.

4. The mode of heat transfer, M.

   1. Uniform melt over the entire surface.
   2. Melt is a function of radial distance from vertical.
   3. Melt is a function of radial distance from the centre line of the void.
   4. Melt is a function of radial distance from the geothermal point source.

   Modes (ii) and (iii) compute a cylindroid surface with some minimal radial distance (R_min) that the total heat flux is applied to. Specific heat fluxes decrease radially, q_M = f(R_min/R), from the surface of that cylindroid. If a point is within the cylindroid the specific heat flux is the same as at the surface. The melt rate (dr/dt) is computed as: dr/dt = q / (rho_i * L_f). The melt amount is hence: dr = q / (rho_i * L_f) * dt.

Additional variables which the user must define are:

• Geometry (geo), as either "cave" or "chimney".
• Initial geometry radius (R).
• Plume radius (r)
• Timestep of the prognostic iterations in hours (dt).
• Total number of iterations (N).
• Total runtime for each simulation in hours (hrs).
• Maximum number of simultaneous simulations if run on a cluster (max_number_simulations)