Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› epsl-figures-and-tables
epsl-figures-and-tables
GitHub专为EPSL论文设计，指导构建核心图表与表格。涵盖触发场景、精简原则、上下文展示及不确定性可视化规范。针对不同子领域（如年代学、地震学）提供具体图表标准，确保跨学科可读性并符合期刊格式要求。
Trigger Scenarios
决定哪些图表放入正文，哪些移至补充材料
起草位置图、剖面图、同位素图或相图
审稿人指出图表缺乏上下文或不确定性信息
准备上传至Editorial Manager的图表文件
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill epsl-figures-and-tables -g -y
SKILL.md
Frontmatter
{
"name": "epsl-figures-and-tables",
"description": "Use when building figures and tables for an Earth and Planetary Science Letters (EPSL) manuscript. A Letter lives or dies by a few exhibits that a geochemist, seismologist, and geodynamicist can all read — maps with tectonic context, concordia\/isochron plots with full error ellipses, model sections with resolution shown. It designs exhibits; it does not run the analysis."
}
Figures & Tables (epsl-figures-and-tables)
In a letters-format paper the figures are the argument: with a main text near 6,500 words (re-check the cap on the official site), EPSL papers typically carry a small number of dense, layered exhibits plus a supplement. The venue-specific pressures are cross-disciplinary legibility — your map will be read by someone who has never worked in that orogen — and uncertainty made visible, because the sub-fields EPSL serves treat an error-free plot as a red flag.
When to trigger
- Deciding which few exhibits carry the Letter and what moves to the supplement
- Drafting location maps, cross-sections, isotope plots, tomography slices, or phase diagrams
- A reviewer called an exhibit unreadable, missing context, or missing uncertainties
- Preparing figure files for Editorial Manager upload
Principles
- Ration the slots. Every main-text figure must advance the argument; reference-quality material (all spectra, all sections, parameter sweeps) belongs in the supplementary material.
- Context panel first. Earth-science readers orient spatially: give a location/tectonic-setting inset with coordinates, scale, and the geodetic datum or projection named; planetary figures give the body, region, and data source (mission/instrument).
- Uncertainty is drawn, not asserted. Error ellipses on concordia, error bars sized to the stated σ-level, credible envelopes on age models, damping/resolution shading on inversions.
- Sub-field conventions, executed exactly. Concordia and isochron plots labeled with MSWD and n; stratigraphic columns with height and stage boundaries; seismic sections with depth scales and vertical exaggeration stated; phase diagrams with the calibration source.
- Color that survives review. Perceptually uniform colormaps for continuous fields (avoid rainbow/jet — the geophysics community actively campaigns against it), colorblind-safe categorical palettes, grayscale-legible line work; vector formats for line art.
- Tables carry data, figures carry claims. Full analytical tables (every analysis, every correction) go to the supplement and repository; the main text carries at most a compact summary table.
Exhibit choices by sub-field
| Result type | The load-bearing exhibit | Non-negotiable elements |
|---|---|---|
| Geochronology | concordia / weighted-mean plot | ellipses at stated σ, MSWD, n-of-N, excluded grains visible |
| Isotope tracing | covariation (x–y) plot with mixing/evolution curves | endmember values sourced, 2σ bars, sample symbols by group |
| Seismic structure | depth slice + cross-section | colorbar units, resolution overlay, no rainbow |
| Geodynamic model | time-series snapshots + regime diagram | parameter values, dimensionless numbers, benchmark noted |
| Paleo-record | proxy vs age with age model envelope | tie points marked, stage boundaries labeled |
| Planetary | mapped observable + interior model | mission/instrument, datum, uncertainty band |
Worked micro-example (illustrative — one figure carrying a slab-fluid argument)
The subduction-water Letter needs geochemistry and seismology in one exhibit (illustrative):
- Panel a: arc transect map with sample sites, trench, slab-depth contours, and the tomographic section line — the seismologist's context and the geochemist's sampling frame in one panel.
- Panel b: δ11B vs distance-to-trench with 2σ bars and model curves for shallow vs deep dehydration — the discriminating plot.
- Panel c: velocity cross-section (perceptually uniform colormap) with the low-velocity conduit and a resolution-test inset showing the feature is not smearing.
One three-panel figure replaces three single-discipline figures and makes the joint argument inspectable at a glance — the letters-format ideal.
Referee-pushback patterns and the venue-specific fix
- "Ellipses/error bars missing or undefined." → Draw them at a stated σ-level on every isotope and age plot; define once in the caption.
- "I can't locate the samples." → Add the context inset with coordinates, scale, projection.
- "The colormap manufactures structure." → Replace jet/rainbow with a perceptually uniform map and re-examine which "anomalies" survive.
Anti-patterns
- A main text bloated with every supporting plot while the one discriminating figure is missing
- Maps without scale, coordinates, or projection; sections without depth units or exaggeration
- Age plots without MSWD, n, or the excluded analyses
- Rainbow colormaps on tomography or gridded fields
- Full 400-row analytical tables typeset in the main text instead of deposited
- Captions that name the panel but not the σ-level, model, or data source
Output format
【Main exhibits】list + the claim each carries
【Context】location/setting panel with datum + scale? [Y/N]
【Uncertainty drawn】ellipses/bars/envelopes at stated σ? [Y/N]
【Conventions】sub-field norms met (MSWD, resolution overlay, ...)? [Y/N]
【Color/format】perceptually uniform + colorblind-safe + vector? [Y/N]
【Main vs supplement】split decided
【Next】epsl-reporting-and-reproducibility
Supplementary resources
../../resources/external_tools.md— GMT/PyGMT, GPlates, ObsPy, IsoplotR plotting../../resources/worked-examples/01-introduction.md— how exhibits pair with a Letter's prose
Version History
- 9f86f09 Current 2026-07-19 15:17
