Data¶

Four datasets sit behind the dashboards: one produced by simulation, three compiled from field observations of real earthquakes.

Unfamiliar terms?

Every abbreviation on this page — DEM, FDHI, SURE, DZW, FZW — is defined in plain language in the glossary.

Simulation outputs (DEM)¶

The paper reports 3,434 distinct element method (DEM) experiments — a technique that models sediment as many interacting particles, so faults and collapses emerge rather than being prescribed. Each experiment is one combination of geological and fault conditions. Within an experiment the modelled fault slips progressively, and the ground surface is measured at every 0.05 m of slip — 346,834 model stages in total.¹ So a row is a stage, not an experiment: the fixed parameters repeat down the rows of one experiment while slip advances.

The experiments sweep sediment depth, density, homogeneous and heterogeneous sediment strengths, fault dip, and the thickness of unruptured sediment above the fault tip. From each stage a computer-vision model measures four surface characteristics: scarp height, uplift, deformation zone width and scarp dip.¹ The simulations are two-dimensional; a set of 3D cases exists in the wider project but is not part of these dashboards.⁴

Provenance: the model dataset is published open-access on DesignSafe — DOIs 10.17603/ds2-gfsj-pp60 and 10.17603/ds2-xpq0-gw80.¹

This is the cloud in Model vs reality, the whole subject of Response curves, the histograms in Distributions, the context distribution in Per-event boxplots, and the population the slip regression fits are computed over.

FDHI flatfile¶

The Fault Displacement Hazards Initiative measurement flatfile — a published compilation of surface-rupture measurements across many earthquakes, with per-measurement location, displacement components and event metadata. It is the field dataset the paper compares its models against.¹

Provenance: UCLA Dataverse, DOI 10.25346/S6/Y4F9LJ.

The project cleans this flatfile in-pipeline rather than relying on a pre-filtered extract. That yields two tables: a 4,121-row measurement population across 25 events, which backs the per-event boxplots and supplies the FDHI reference lines on the distributions histograms, and a much smaller scatter-overlay subset that backs the model-vs-reality view.²

SURE database¶

SURE — "a worldwide and unified database of surface ruptures … for fault displacement hazard analyses", version 2.0.⁵ A public compilation of surface-rupture observations across many historical earthquakes. Roughly 1,400 measurement records covering identifiers, location, strike-slip / fault-normal / vertical displacement components and their uncertainties, scarp height, and event metadata.³

It supplies the fault-normal-component and scarp-height panels in the per-event boxplots, and its measurements stand as reference lines on the distributions histograms. Note that the release carries no event magnitude; the magnitudes shown on those panels come from a lookup curated inside this project, with every value sourced from the SURE 2.0 data descriptor (Nurminen et al. 2022).³

Kern County (1952)¶

A hand-compiled merge of three sources of surface-rupture measurements from the 1952 M 7.36 Kern County earthquake on the White Wolf fault in California: the classic Buwalda & St. Amand (1955) field survey, the Kern entries from the FDHI flatfile, and the event's SDC — surface deformation characteristics, the authors' umbrella term for the measured surface quantities: scarp height, deformation zone width and scarp dip.³

Kern is the project's worked example for inverting the model — placing a measured vertical displacement on a fitted relationship to infer the slip that produced it.³ That inversion is the slip regression dashboard.

How the data reaches the dashboards¶

The pipeline is linear. Raw CSVs land in a local, untracked data/raw/ directory and are read by typed loaders; the FDHI flatfile is cleaned in-process into the analysis tables. Each table is written as Parquet in a directory-per-table layout, and a DuckDB file of view definitions is generated over that Parquet — this is where cross-source normalisation and the event-magnitude lookup happen, and it is the local SQL layer for desktop analysis.⁴ Because Tableau Public cannot connect to either DuckDB or a cloud warehouse, a final step exports each view to CSV, and those CSVs are what the published dashboards on this site are built from.²

Every step is reproducible from the repository — one command builds the tables, views and schemas, a second exports the CSVs — and the pipeline fails fast if any raw input is missing rather than producing a partial set of artifacts.

Most exports are simply a table made readable. Four are different. Three carry results the pipeline computed, so that the slip regression dashboard displays numbers the project's tests pin rather than recomputing them in the browser; the fourth is assembled from the three field datasets for the distributions reference lines, its populations likewise pinned by tests.

Export	What it holds
`dem_regression.csv`	One row per fault dip: the fitted slope, intercept and r² of vertical displacement against slip.
`dem_regression_lines.csv`	Two endpoints per dip, so the fitted line can be drawn as a line rather than re-fitted.
`kern_inferred_slip.csv`	Each Kern County vertical displacement back-projected through every dip's fit.
`historic_events.csv`	One row per field measurement — the FDHI flatfile, SURE and Kern County unioned, each row keeping whichever of deformation-zone width or scarp height it carries — drawn as the reference lines on the distributions histograms.

Raw data is not redistributed here

The raw inputs are not committed to the repository — they come from the sources cited above, or from the project owner. The repository documents the expected filenames and their provenance.

How to cite this data¶

These dashboards are a convenience layer. If you use the data in your own work, please cite the underlying publications and archives rather than this site — the dashboards only repackage them, and the papers are the citable record. Any DOI issued for this site or its source code identifies the software that builds these views; it does not replace the citations below.

Please cite the study itself as:

Chiama, K., Bednarz, W., Moss, R., Plesch, A., and Shaw, J. H. (2025). "Quantifying relationships between fault parameters and rupture characteristics associated with thrust and reverse fault earthquakes." Earthquake Spectra, 41(5), 3977–4014. DOI: 10.1177/87552930251346434

The 2D DEM experiments behind these dashboards, archived open-access on DesignSafe-CI:

Chiama, K., W. Bednarz, R. Moss, A. Plesch, and J. Shaw (2024a). "Homogeneous 2D DEM Experiments," in Influence of sediment depth, sediment strength, fault dip, and slip on fault scarp morphology, DesignSafe-CI. DOI: 10.17603/ds2-xpq0-gw80
Chiama, K., W. Bednarz, R. Moss, A. Plesch, and J. Shaw (2024b). "Heterogeneous 2D DEM Experiments," in Influence of sediment depth, sediment strength, fault dip, and slip on fault scarp morphology, DesignSafe-CI. DOI: 10.17603/ds2-gfsj-pp60

The wider project's 3D DEM models (not part of these dashboards) are described in

Chiama, K., Plesch, A., and Shaw, J. H. (2025). "Along-Strike Variability of Surface Deformation on Thrust and Reverse Fault Ruptures: Insights from 3D Distinct Element Method Models." Seismological Research Letters 96(6), 3473–3489. DOI: 10.1785/0220250173

and deposited on DesignSafe-CI:

Chiama, K., Plesch, A., and Shaw, J. H. (2025a). "Case 3 — Variable Fault Gouge in 3D DEM Models." DOI: 10.17603/ds2-8kb3-5g63
Chiama, K., Plesch, A., and Shaw, J. H. (2025b). "Case 2 — Variable Fault Dip in 3D DEM Models." DOI: 10.17603/ds2-zt8x-6e73
Chiama, K., Plesch, A., and Shaw, J. H. (2025c). "Case 1 — Cylindrical 3D DEM Models." DOI: 10.17603/ds2-xgqp-ay07

The field compilations carry their own citations: the FDHI flatfile is UCLA Dataverse DOI 10.25346/S6/Y4F9LJ, and SURE is Baize et al. (2019).⁵

Where to go next¶

Glossary — what DZW, FZW, vertical separation and the sediment settings actually mean.
Model vs reality — these datasets on one plot.
The paper — which figure each dashboard replaces.

Please cite as:

Chiama, K., Bednarz, W., Moss, R., Plesch, A., and Shaw, J. H. (2025). "Quantifying relationships between fault parameters and rupture characteristics associated with thrust and reverse fault earthquakes." Earthquake Spectra, 41(5), 3977–4014. DOI: 10.1177/87552930251346434

If you use the underlying data itself, cite the archives it comes from as well — the DEM experiments are deposited open-access on DesignSafe-CI, and the field compilations carry their own citations. The full list, with DOIs, is under How to cite this data.

Any DOI issued for this site or its source code identifies the software and the website, and does not replace the citations above.

Related work, for the wider project's 3D models — not a source for anything shown here:

Chiama, K., Plesch, A., and Shaw, J. H. (2025). "Along-Strike Variability of Surface Deformation on Thrust and Reverse Fault Ruptures: Insights from 3D Distinct Element Method Models." Seismological Research Letters 96(6), 3473–3489. DOI: 10.1785/0220250173

Chiama et al. (2025), abstract — see The paper. ↩↩↩↩
notes/dashboard-3-build-spec.md in the source repository. ↩↩
docs/datasets.md in the source repository — reference notes on each input dataset and how the legacy analyses used it. ↩↩↩↩
notes/Roadmap.md in the source repository. ↩↩
Baize, S., Nurminen, F., Sarmiento, A., et al. (2019). "A worldwide and unified database of surface ruptures (SURE) for fault displacement hazard analyses." Seismological Research Letters 91: 499–520 — the reference Chiama et al. (2025) cites for this dataset. ↩↩