The passing probabilities of zones of geometrical complexity and their effect in surface rupture length
A set of scripts to characterize different zones of geometrical complexity along surface ruptures and passing probabilities as a function of geometry.
Propagating earthquakes must overcome geometrical complexity on fault networks to grow into large, surface rupturing events. We map step-overs, bends, gaps, splays, and strands of length scales ~100-500 meters from the surface ruptures of 31 strike-slip earthquakes, recording whether ruptures propagated past the feature. We find that step-overs and bends can arrest rupture and develop a statistical model for passing probability as a function of geometry for each group. Step-overs wider than 1.2 km, single bends larger than 32°, and double bends larger than 38° are breached by rupture half of the time. ~20% of the ruptures terminate on straight segments. We examine how the distribution of geometrical complexity influences surface rupture length, inferring an exponential relationship between rupture length and event probability. Our findings support that geometrical complexity limits the size of large events and provide insights into the competition between energy supply and dissipation during rupture propagation.
This repository contains the scripts required to reproduce the results in Rodriguez Padilla et al. 202X, and to measure the geometry and passing probabilities of different types of geometrical complexity. The data required to run the scripts is provided in the following Zenodo repository and listed below:
- Data repository directories and content
- geometrical_complexity_shapefiles_v1 Shapefiles of the zones of geometrical complexity mapped for each event.
- Regional_maps Shapefiles with the regional fault maps for each event.
- FDHI_data Event information and displacement data for each event stored in excel file. Data from the FDHI database appendix.
- event_rupture_shp
primary rupture shapefiles for each event in the FDHI database. Kmz files used to generate the shapefiles sourced from the FDHI database appendix. Shapefiles generated using the kzm2shp script.
- event_rupture_shp primary rupture shapefiles for each event in the FDHI database. Kmz files used to generate the shapefiles sourced from the FDHI database appendix. Shapefiles generated using the kzm2shp script.
Refer to the readme file in the Zenodo repository linked above for additional details.
The subset of the scripts that measure the geometry of different features from shapefiles are in Matlab. The specific dependencies are listed on top of each script. Some of the scripts rely on functions downloable from Mathworks, and are provided as part of this repository in the source_code directory. The scripts for estimating passing probabilities and event likelihood are available as Python Jupyter Notebooks, with the functions stored in the .utils file in the directory. These scripts have been tested on MacOS.
To make a Python environment to run the Notebooks:
conda env create -f EQ_gates.yml
conda activate EQ_gates
Running the scripts: measuring fault geometry, estimating passing probabilities, and estimating event likelihood
-
To measure the geometry of features in a shapefile
- Run the "measure_geometry.m" Matlab script. This will output a csv file with the characterized features (geometries.csv).
- To measure the spacing between zones of geometrical complexity along a rupture, run the "gatespacing.m" script. This script produces a pdf output fitting log-normal, Weibull, and exponential CDFs to the ECDF of the feature spacings (supplemental Figure S6).
-
To estimate passing probabilities and event likelihood
- Run the "analysis_rupture_probabilities.ipynb" Jupyter Notebook. Requires the csv containing the measured geometries generated from the Matlab code in the previous step. This file ("geometries.csv") is also provided as part of this repository for users lacking access to Matlab or interested in accessing the geometry measurements directly without downloading the shapefiles and running the Matlab code.
- This script estimates passing probability as a function of geometry using logistic models.
- This script also estimates the event likelihood based on the probabilities.
- All figures in the main body of the manuscript except for Figure 1 and Figure 4b can be reproduced by running this code. All supplemental figures except figure S6 can be reproduced using this code too.
-
To reproduce the rupture maps in the appendix with the earthquake gates plotted over them
- Run the "map_maker.ipynb" script (requires the rupture maps and regional fault maps to be in shapefile format, provided in the Zenodo data repository).
Code link: https://github.com/absrp/passing_probabilities_EQgates_strikeslip
Data link: https://zenodo.org/records/12696715
Manuscript link: (open access) https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2024GL109957
Please report questions, suggestions, and issues:
Email: [email protected], [email protected]