EQdyna.2Dcycle

2D finite-element code for physics-based multicycle earthquake dynamics on geometrically complex fault systems.

Install

./install.sh             # autodetects OS, installs python deps, builds binary

Or explicitly:

./install.sh -e macos    # macOS  (with python deps)
./install.sh -e ubuntu   # Linux  (with python deps)
./install.sh -m ubuntu   # build only, deps assumed present

Sets EQDYNA2DCYCLEROOT, adds bin/ and scripts/ to PATH, and builds bin/run_eqdyna2d_<VERSION>.

Dependencies: gfortran, Python 3 with numpy, matplotlib, xarray.

Run

One-step demo (paper.saf.A — Liu et al. 2022 Model A):

bash example_workflow.sh

Manual flow (any compset):

create.newcase --work_dir work/my_case --compset paper.saf.A
cd work/my_case
python3 case.setup
bash run.sh

Available compsets: create.newcase --list.

Restart from a previous run

Every cycle the binary refreshes a binaryop file holding the full restart state. To extend a finished run from cycle N+1 → M:

cd work/my_case
# 1. Bump icstart to last_cycle + 1, raise icend.
sed -i 's/par\.icstart, par\.icend = .*/par.icstart, par.icend = N+1, M/' user_defined_params.py
# 2. Regenerate FE_Global.txt + run.sh from the new params.
python3 case.setup
# 3. Launch — binary loads binaryop and resumes at cycle N+1.
bash run.sh

Notes:

• For C_mesh=3 (.gmsh.lite) cases binaryop already lives in the case dir. For C_mesh=2 (paper.saf.A) it gets moved to aRawSimuData/ after a finished run — copy it back to the case dir before step 3.
• The new outputs land as totalop.txt<icstart> / cyclelog.txt<icstart> / interval.txt<icstart> (separate from the original *.txt1), and run.sh's rm -f totalop.txt* only wipes the case dir, so aRawSimuData/ keeps the previous output safe.
• Post-processing scripts (plotRuptureDynamics, analyze_catalog.py, plot_event_slips_overtime_fig4.py) auto-discover all totalop.txt* files via saf_result_utils.discover_cycle_tags and stitch them into one continuous catalog.

Outputs

After a run finishes, run.sh moves raw outputs into aRawSimuData/:

File	Content
totalop.txt<N>	per-cycle stack: shear, normal, slip, slip-rate, rupture-time per fault node, for cycles starting at icstart=N
cyclelog.txt<N>	one line per cycle: cycle_id nucleation_node
interval.txt<N>	one value per cycle: interseismic duration (yr)
binaryop	restart state for resuming with icstart > 1

Plus the input/mesh files used by the run (vert.txt, fac.txt, nsmp.txt, nsmpGeoPhys.txt, meshGeneralInfo.txt, Rate_direction.txt, FE_*.txt).

Plots go to aPlots/.

Post-processing

All scripts live in scripts/ (and are on PATH after install). Run from inside a case dir.

Script	What it does
plotRuptureDynamics	Per-cycle 4-panel plot (shear, normal, slip, rupture-time vs along-strike). Defaults: only saves cycles with M ≥ MIN_PLOT_MAGNITUDE (env var, default 6.5). Set FORCE_REPLOT=1 to redo existing plots.
CATALOG=1 plotRuptureDynamics	Catalog mode: no figures, ~10× faster. Writes aPlots/catalog.csv with columns: eqId, magnitude, moment_Nm, nuc_x_km, nuc_y_km, nuc_ft, rup_dur_s, peak_slip_m.
analyze_catalog.py [case_dir] [--mc M] [--mmax M]	Reads aPlots/catalog.csv. Outputs b-value (LSQ on [Mc, Mmax] window), magnitude-frequency distribution, magnitude-vs-cycle scatter, nucleation-along-strike-vs-cycle scatter. Saves aPlots/catalog_analysis.png. For characteristic-fault MFDs use --mmax 7.0 to exclude the bump.
plot_event_slips_overtime_fig4.py [case_dir]	Paper Figure-4 style slip-distribution stacks (slip vs along-strike, vertically offset by event time). --duration = window in kyr (default 3); --threshold = min event slip (m).
monitor_runs.sh [seconds]	Background poller (default 600s). Tails progress for all paper.saf.A.* and saflite cases under work/, re-runs Figure 4 + rupture-dynamics plots each tick.
compare_cycle_over_strike.py	Overlay a chosen cycle's stress/slip/rupture-time curves from multiple cases (e.g. saf.gmsh.lite vs paper.saf.A) for direct comparison.

Quick example: catalog + analyze on a finished case:

cd work/my_case
CATALOG=1 plotRuptureDynamics      # writes aPlots/catalog.csv
analyze_catalog.py . --mmax 7.0    # writes aPlots/catalog_analysis.png

Authors

• Dunyu Liu — Institute for Geophysics, Jackson School of Geosciences, The University of Texas at Austin — dliu@ig.utexas.edu
• Benchun Duan — Center for Tectonophysics, Department of Geology and Geophysics, Texas A&M University — bduan@tamu.edu

Contributors

• Claude (Anthropic Claude Code) — development assistance, refactoring, testing, documentation

Citations

If you use EQdyna.2Dcycle in your research, please cite:

• Duan, B., & Oglesby, D. D. (2006). Heterogeneous fault stresses from previous earthquakes and the effect on dynamics of parallel strike-slip faults. Journal of Geophysical Research, 111(B5), B05309. https://doi.org/10.1029/2005JB004138
• Liu, D., Duan, B., Scharer, K., & Yule, D. (2022). Observation-constrained multicycle dynamic models of the southern San Andreas and the northern San Jacinto faults: Addressing complexity in paleoearthquake extent and recurrence with realistic 2D fault geometry. Journal of Geophysical Research: Solid Earth, 127(2), e2021JB023420. https://doi.org/10.1029/2021JB023420

License

MIT — see LICENSE.