Skip to content

Latest commit

 

History

History
70 lines (52 loc) · 4.38 KB

README.md

File metadata and controls

70 lines (52 loc) · 4.38 KB
Raw

thermal-structure

This repository contains code used to investigate how glacier surges alter glacier thermal structure. In pursuit of our research objectives, we've employed a thermomechanically coupled numerical ice-flow model, Elmer/Ice in a 2D ($x-z$) configuration. We conduct a suite of experiments ranging from initialization through periodic perturbations to the basal slip coefficient, mean to mimic observed surges. This repository contains code used pre-process the necessary input data, execute and post-process the model runs, and do some plotting and analysis.

B_0.00031623_SP_2_QP_28_2--3ka_bitrate_3000_001.mp4

Project layout:

The repository structure is as follows:

├── Makefile    # Global makefile used to compile code throughout the repository
├── README.md  
├── bin         # Folder to hold the compiled code (mainly Elmer user functions)
├── config      # Repository wide configuration files 
├── expr        # Various experiments we conducted using `Elmer/Ice`
├── include     # External code used in our analysis
├── input_data  # Surface mass balance and topography data need to run `Elmer` simulations
├── notebooks   # Collection of jupyter notebooks doing data pre-processing and analysis
├── src         # source code folder, containing `Elmer` boundary condition code and more
└── study

Getting started

To get started first clone the repository by running,

$ git clone [email protected]:andrewdnolan/thermal-structure.git

Then, navigate into the top directory (i.e. cd thermal-strucure) and compile/install the code by running

$ make

which will execute the global Makefile that compiles external FORTRAN code from the include folder, compiles the Elmer user functions (USFs) from the src/elmer_src folder, and builds the legacy NetCDF post-processing FORTRAN program contained in the src/elmer2nc folder.

Finally, you'll also need to run

pip install --editable src/thermal/

which will create a locally editable copy of the python post-processing library thermal. thermal is built around xarray and dask to enable efficient, out of memory, and parallel postprocessing of the terabytes of data produced in the various experiments.

Installing Elmer/Ice

If working from a Linux machine, follow the compilation instructions from the Elmer/Ice documentation. Compiling Elmer/Ice on non-Linux machines (e.g. OSX and Windows) is notoriously challenging. To circumvent this problem, I have written a Docker container. Instructions on how to install the Docker can be found in the associated readme. All within code/files in this repository assume the Elmer executables (e.g. ElmerSolver, ElmerGird) are available.

Note on reproducibility

Given the computational cost of thermomechanically coupled numerical modeling, most of the model execution has been done on high performance computing (HPC) resources (e.g. Compute Canada's cedar cluster). Therefore, many of the files various expr directories are not meant to be run on a local laptop/workstation. That being said, command line scripts to run one-off simulations, which still could take < 24 hours, are available in each of the expr directories.