brown_sdc_2025_scripts

Repository for scripts to run the test casese and plot the results for Brown 2025 paper: Fast-wave slow-wave spectral deferred correction methods applied to the compressible Euler equations

The test case scripts in this repository use the Gusto model, which can be accessed here: [https://github.com/firedrakeproject/gusto].

The plotting scripts in this repository use the tomplot plotting library, which can be accessed here: [https://github.com/tommbendall/tomplot]

How to get results & plots:

---

1. Make sure you have a firedrake with gusto installed, and tomplot installed. The results in the paper were ran from gusto at commit a5e045a6e16ac318eb00db9bf291595df449aacb.

---

2. Run the following commands from the test_cases directory:

For Figure 1 simply run python williamson1_convergence.py, this should generate all the data for the self convergence test

For Figure 2 run:

1. python gravity_wave.py to generate the example solution in the plot
2. run mpiexec -n N python 1 gw_convergence_true.py, mpiexec -n N python 3 gw_convergence_true.py and mpiexec -n N python 5 gw_convergence_true.py to generate the order 1, 3 and 5 reference solutions. Note this takes some time. N = 20 would be a reasonable choice here
3. run mpiexec -n N python gw_convergence_o1.py, mpiexec -n N python gw_convergence_o3.py and mpiexec -n N python gw_convergence_o5.py to generate the solutions for the convergence test. N = 3 would be a reasonable choice here

For Figure 3 run:

1. mpiexec -n N python moist_bf.py to generate the solution with the LU and FE Qdelta matrices
2. mpiexec -n N python moist_bf_parallel.py to generate the solution with the MIN-SR-FLEX and MIN-SR-NS Qdelta matrices N = 5 would be a reasonable choice here

For Figure 4 run mpiexec -n N python dry_baroclinic_channel.py. N = 30 to N = 60 would be reasonable choices here

---

3. Run all plotting scripts from the plotting_scripts directory. They are named based on which figure in the paper they produce.

---